Datasets:

keerthanapacific

/

sample3

like 0

545971

From theory to practice: improving the impact of health services research

Background While significant strides have been made in health research, the incorporation of research evidence into healthcare decision-making has been marginal. The purpose of this paper is to provide an overview of how the utility of health services research can be improved through the use of theory. Integrating theory into health services research can improve research methodology and encourage stronger collaboration with decision-makers. Discussion Recognizing the importance of theory calls for new expectations in the practice of health services research. These include: the formation of interdisciplinary research teams; broadening the training for those who will practice health services research; and supportive organizational conditions that promote collaboration between researchers and decision makers. Further, funding bodies can provide a significant role in guiding and supporting the use of theory in the practice of health services research. Summary Institutions and researchers should incorporate the use of theory if health services research is to fulfill its potential for improving the delivery of health care.

Background While significant strides have been made in medical research over the past several decades, many research results considered important by researchers and expert committees are not being used by health care practitioners. While the value of health services research must be judged by its validity, its utility cannot be taken for granted. There has been an assumption that when research information is available it will be accessed, appraised and then applied [ 1 ]. However, knowledge of a research-based recommendation is by itself insufficient to ensure its adoption. While the value of research evidence as a basis for decision making in health care is well established, the incorporation of such evidence into decision-making remains inconsistent [ 2 ]. The gap between research evidence and its' incorporation into practice has led to an increase in research in how to bring new knowledge to bear on everyday health care. Factors influencing the adoption of research evidence have been studied extensively [ 3 - 5 ]. Personal attributes, time, organizational boundaries, geography and educational background all contribute to decision-makers' responses to research evidence [ 6 , 7 ]. An area that has received less attention is the incorporation of theory in health services research. The authors of this paper propose a need for a stronger theoretical base in health services research wherein health services research would be more informative and influential, facilitating the adoption of research results into practise. Integrating theory into health services research is an important first step. In this paper we first describe the importance of theory followed by how theory driven research changes the manner researchers interact with decision-makers. We conclude on how theory driven research may influence the training, practice and the funding of health services research. Discussion The importance of theory In recent years a number of researchers have advocated a greater role for the use of theory in strengthening the practice of research [ 8 - 12 ]. However, health services research has continued to focus primarily on evaluating outcomes with less attention to the mechanisms by which these outcomes are produced [ 10 , 13 , 14 ]. The emphasis on method at the expense of theory has led to several criticisms. Chen and Rossi [ 15 ] argue that an atheoretical approach to research is characterized by adherence to a step-by-step cookbook method for doing outcomes studies. In this situation they contend that research is reduced to a set of predetermined steps that are mechanically applied to various interventions without concern for the theoretical implications of intervention content, setting, participants or implementing organizations. The atheoretical approach tends to result in a simple input/output, or black box type of study [ 13 ]. Such simple evaluations may provide a gross assessment of whether or not an intervention works under one set of conditions but fail to identify the reasons why. As such, the conclusions are often less than satisfying to consumers of research results and not easily transferable to different settings. Theory provides a systematic view of a phenomena by specifying the relations among variables and propositions with the purpose to explain or predict phenomena that occurs in the world [ 16 , 17 ]. In health services research theory can provide a framework to understand the relationship between program inputs (resources), program activities (how the program is implemented) and their outputs or outcomes [ 11 , 13 ]. In addition to identifying the mechanisms by which programs are effective, theory may consider program implementation and contextual factors. While it is important to know the extent to which an intervention attains intended outcomes, it is also essential to know what occurred in the implementation of the intervention. Variation in the implementation of the intervention may be due to differences among program providers, target population characteristics, and differences among sites on how the intervention is delivered. Theory also offers the opportunity to specify the contextual conditions that will influence the effectiveness of an intervention. Attitudinal factors at the provider level as well as structural, cultural factors at the organizational level have been under appreciated in exploring variations in health care outcomes [ 9 , 18 , 19 ]. Understanding the influence that contextual factors have on program implementation and outcomes facilitates successful application of the intervention in alternate settings, therein, addressing the generalizabilty of an intervention. Theory offers many advantages to the health services researcher. Theory helps to identify the appropriate study question and target group; clarify methods and measurement issues; provide more detailed and informative descriptions on characteristics of the intervention and supportive implementation conditions; uncover unintended effects; assist in analysis and interpretation of results; and, the successful application of an intervention to different settings [ 11 , 12 ]. Theory-driven studies are addressing the challenge of both decision-makers and funding agencies to move beyond simplistic explanations of significance in health services research. Decision-makers are seeking explanations about how an intervention works and whether it will work in a fashion similar to the intervention that was evaluated when applied to a different environment [ 10 , 12 , 20 ]. Despite these potential benefits, there are a number of reasons offered as to why there has been a failure to integrate theory into research. Ironically, clinical randomized control trials have discouraged the use of theory in health services research. Given the genesis of clinical trial methodology, this may derive, in part, from the very origins of epidemiology, whereby John Snow allegedly ended an epidemic of cholera by removing the handle from the Broad Street water pump, even though he had no concept of what actually caused cholera. By ignoring the need for theory, Snow was able to overcome the fact that the theories he would have needed had not yet been elucidated. Similarly, we know that lung cancer incidence can be reduced by elimination of cigarette smoking, even though we do not know exactly how cigarette smoke causes lung cancer. Experimental trials often determine intervention effects without considering how the component features of an intervention work together to bring about study outcomes [ 13 , 15 , 21 ]. The more complex the intervention, the more difficult it is to know what the treatment entailed. There is a growing recognition for the need to establish the theoretical bases of interventions. The United Kingdom Medical Research Council recently proposed a framework for the development and evaluation of randomized control trials for complex interventions where theory is viewed as valuable in assisting hypothesis development and steering decisions on strategic design issues [ 22 ]. Adopting a theory-driven approach in health services research is not without its challenges. Given the typical training of researchers and the uni-disciplinary nature of the practice the first challenge is the capacity of researchers to engage in theory driven research. Second, a theory driven approach requires organizational conditions that support researchers and decision makers collaborating in the development and testing of theory. Finally, theory development and testing is cumulative in nature, encouraging researchers to pursue a programmatic approach in research. This approach has implications on how funding agencies support health services research. Despite the potential challenges, a theory based approach offers promise for a greater understanding on what happens when interventions work to address social/health problems. The importance of collaboration with decision-makers Collaborative research partnerships between academic researchers and decision-makers describe a relationship and process between individuals from different backgrounds, who together, develop an integrative cooperative approach to resolve a research problem [ 23 ]. It has been identified as a significant strategy that holds multiple benefits [ 23 - 27 ]. Collaborative practice has also been identified as a key strategy in facilitating a theory driven approach. Weiss [ 28 ] recommends that the first criterion in selecting a theory to guide the evaluation of a program is to draw the theory out from those associated from the program, including designers of the program, program personnel and relevant clinical staff. The argument is that few programs are theory driven. Rather, they are typically the product of the experience and values of those who are associated with the program. In recent years a number of techniques have been developed for this purpose. Strategies range from unstructured interviews, to highly structured iterative interactions between program personnel and researchers [ 29 - 31 ]. Perspectives of service providers can be rounded out by a review of the research literature. In fact, a number of researchers suggest a combination of these two approaches [ 10 , 32 ]. Viewing program stakeholders as a key source in developing theory in health services research demands stronger collaboration between researchers and program decision makers [ 9 ]. In this fashion, collaborative practice becomes a methodological strategy in health services research. Lomas [ 7 ] has stressed that a first step in encouraging meaningful partnerships between researchers and decision-makers is to view linkage and exchange between the two as a process not as a discrete event. Establishing and maintaining ongoing links offers a more comprehensive understanding between the two groups. Researchers uncover the desired program outcomes, the causal change of the program intervention and develop a better understanding of the contextual factors that influence the variation on intervention implementation and outcomes. Similarly, decision-makers will develop a deeper understanding of the research process and thus can influence the development of feasible and sustainable interventions for practice settings. The role and impact of the researcher and the research process in practice settings have received greater attention in other fields such as program evaluation, nursing, anthropology and community psychology. For example, core principles of community psychology practice include: a) consistency of goals and values between the researcher and the setting, and b) the notion that interventions should have the potential for being "institutionalized" or systematically established within the setting in such a way that strengthens the natural resources of the setting [ 25 - 27 ]. Rather than reinventing the wheel, health services research could benefit from theoretical frameworks developed within these disciplines. Implications for the practice of health services research Recognizing the importance of theory calls for new expectations in the practice of health services research. There are a number of challenges that must be met in order for these perspectives to gain acceptance in the health services research community. Evolving perspectives on the practice of health services research require recognition that few disciplines are able to span the breadth of responsibilities associated with the research process. To date there has been a tendency for health services research to be practiced as a uni-discipline where clinical disciplines tend to practice separately from the social science disciplines. A priority is to encourage the formation of research teams that are inter-disciplinary. Pursuing this agenda will promote the formation of research teams that may include: business, anthropology, sociology, psychology, education, engineering, nursing and medicine. Combined disciplinary skills would, in a complementary fashion, address the breadth of skills required in a more complex research environment that includes the development and testing of theory. A second point concerns broadening the training for those who will practice health services research. By and large, academic training has focused on methodological issues. While a focus on research methods has made an important contribution to the practice of health services research, relying on research methods as a core curriculum has led to limitations in the training of health services researchers such as inadequate attention to the value of theory driven research. As health services research expands its methodological repertoire beyond the classical randomized control trial, researchers face increased ambiguity in attributing the source of intervention impact. It is in this circumstance that theory can guide health services researchers in understanding the causal linkages within an intervention. Further, students are educated in separate departments with little planned, formal activity across disciplines, which discourages co-operative approaches to research and service [ 33 - 35 ]. Education programs are not generally structured to facilitate the importance of inter-disciplinary strategies. Identifying the processes associated with creating effective linkages between researchers and decision-makers are also not typically part of training. Rethinking the current assumptions and practices regarding the training of health services researchers will enable trainees in health services research to be better prepared for their evolving responsibilities. Collaboration between researchers and decision-makers are contingent upon supportive organizational conditions for both partners. Researchers have, and most likely will continue to operate from university-based settings where incentives for promotion and tenure can act as barriers to changes in the practice of health services research [ 7 , 36 ]. Most academic institutions award tenure and promote faculty based upon the frequency and quality of publications and on obtaining peer review funding [ 36 ]. The time involved in collaborating with decision-makers, joint planning and implementing research often represents activities that are not recognized by tenure promotion committees. As well, these activities may slow the production of research results and the generation of publications. Recognizing these factors requires academic centres to generate new criteria for evaluating contributions to knowledge and practice. Decision-making organizations also play a significant role in ensuring the success of collaborative relationships. The clearest indication of institutional support for research is to provide the time and resources for decision-makers to participate in collaboration activities with researchers. Funding bodies have the potential to play a significant role in guiding and integrating these considerations into health services research. Research sponsors can develop evaluation criteria that encourage the application of theory. As an example, the Agency for Healthcare Research and Quality (AHRQ) in the USA funded an initiative (Translating Research into Practice) to identify sustainable and reproducible strategies that will: 1) accelerate the impact of health services research on direct patient care; and 2) improve the outcomes, quality, effectiveness, efficiency, and/or cost effectiveness of care through partnerships between health care organizations and researchers [ 37 ]. Further, research sponsors are beginning to move away from supporting single shot studies that are conducted in relative independence from one another. This focus on supporting programmatic research should be encouraged. Programmatic research offers a cumulative environment that allows researchers the opportunity to develop and test the application of theory. In a similar fashion, collaborative practice is also best practiced in a programmatic environment. Developing and maintaining linkages with decision-makers is predicated on developing and maintaining long-term relations. Embedded within these linkages are fundamental professional and personal attributes that include; credibility, familiarity, mutual understanding and trust [ 38 , 39 ]. Summary This paper has examined the importance of theory in health services research. We have argued that by strengthening the role of theory encourages collaborative practice between researchers and decision-makers. It has been noted that a theory driven approach in health services research is not without its challenges. However, given the modest advances towards incorporating research evidence into healthcare decisions, a theory driven approach is well worth the effort. The implication of this approach for health services research is that it has impact on the training and practice of health services research. Institutions and researchers should consider this emerging model of practice if health services research is to fulfill its potential for improving the delivery of care. Competing interests The authors declare that there are no competing interests. Disclaimer: The opinions expressed are the authors' and do not necessarily represent official policy of AHRQ or the Department of Health and Human Services Authors' contributions KB drafted the manuscript, edited and revised the contents, EO edited and revised the manuscript, KB, EO, MC, RS, DS all contributed to the conceptual development, editing and review of the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545971.xml

549185

Global gene expression in neuroendocrine tumors from patients with the MEN1 syndrome

Background Multiple Endocrine Neoplasia type 1 (MEN1, OMIM 131100) is an autosomal dominant disorder characterized by endocrine tumors of the parathyroids, pancreatic islets and pituitary. The disease is caused by the functional loss of the tumor suppressor protein menin, coded by the MEN1 gene. The protein sequence has no significant homology to known consensus motifs. In vitro studies have shown menin binding to JunD, Pem, Smad3, NF-kappaB, nm23H1, and RPA2 proteins. However, none of these binding studies have led to a convincing theory of how loss-of-menin leads to neoplasia. Results Global gene expression studies on eight neuroendocrine tumors from MEN1 patients and 4 normal islet controls was performed utilizing Affymetrix U95Av2 chips. Overall hierarchical clustering placed all tumors in one group separate from the group of normal islets. Within the group of tumors, those of the same type were mostly clustered together. The clustering analysis also revealed 19 apoptosis-related genes that were under-expressed in the group of tumors. There were 193 genes that were increased/decreased by at least 2-fold in the tumors relative to the normal islets and that had a t-test significance value of p < = 0.005. Forty-five of these genes were increased and 148 were decreased in the tumors relative to the controls. One hundred and four of the genes could be classified as being involved in cell growth, cell death, or signal transduction. The results from 11 genes were selected for validation by quantitative RT-PCR. The average correlation coefficient was 0.655 (range 0.235–0.964). Conclusion This is the first analysis of global gene expression in MEN1-associated neuroendocrine tumors. Many genes were identified which were differentially expressed in neuroendocrine tumors arising in patients with the MEN1 syndrome, as compared with normal human islet cells. The expression of a group of apoptosis-related genes was significantly suppressed, suggesting that these genes may play crucial roles in tumorigenesis in this syndrome. We identified a number of genes which are attractive candidates for further investigation into the mechanisms by which menin loss causes tumors in pancreatic islets. Of particular interest are: FGF9 which may stimulate the growth of prostate cancer, brain cancer and endometrium; and IER3 (IEX-1), PHLDA2 (TSS3), IAPP (amylin), and SST, all of which may play roles in apoptosis.

Background Multiple Endocrine Neoplasia type 1 (MEN1, OMIM 131100) is an autosomal dominant disorder characterized by endocrine tumors of parathyroid, pancreatic islets and pituitary [ 1 ]. The prevalence of MEN1 is estimated to be 2–10 per 100,000 [ 2 ]. Based on loss of heterozygosity in tumors and Knudson's "two-hit" hypothesis, the MEN1 gene was classified as a tumor suppressor [ 2 , 3 ] and the gene was isolated in 1997 by positional cloning [ 4 ]. The MEN1 gene spans 9 kb of the genome, is comprised of 10 exons, and codes for a 610 amino acid protein termed menin [ 4 ]. More than 300 independent germline and somatic mutations have been identified [ 5 ]. Recently, five new germline mutations which affect splicing of pre-mRNA transcribed from MEN1 gene were identified in our laboratory [ 6 ]. The nature of all the disease-inducing mutations points to a loss of function of menin, which is characteristic of a tumor suppressor. Database analysis of menin protein sequence reveals no significant homology to known consensus protein motifs. Menin is widely expressed in both endocrine and non-endocrine tissues [ 4 ]. Menin is primarily localized in the nucleus and contains two nuclear localization signal sequences near the carboxyl terminus of the protein [ 7 ]. Studies on the function of menin have not yielded a clear picture as to the role of menin as a tumor suppressor; however, the results of these studies suggest some interesting possibilities. Two groups [ 8 , 9 ], based on yeast two-hybrid screening of a human adult brain library, reported that menin interacts with JunD (a member of the AP-1 transcription factor family) and represses JunD mediated transcription. Recently, Agarawal et al [ 10 ] reported that when JunD loses its association with menin it becomes a growth promoter rather than a growth suppressor. Other reports suggest some relevance of the menin-JunD interaction. JunD null male mice exhibit impaired spermatogenesis [ 11 ]. In postnatal mouse, Men1 was found to be expressed in testis (spermatogonia) at high levels [ 12 ]. Lemmens et al [ 13 ] by screening a 12.5 dpc mouse embryo library with menin, identified a homeobox-containing mouse protein, Pem. Interestingly, both menin and Pem showed a very similar pattern of expression, especially in testis and Sertoli cells. These findings along with the fact that some MEN1 patients have idiopathic oligospermia and non-motility of spermatozoa [ 14 ] suggest that menin-JunD and menin-PEM interactions may play a vital role in spermatogenesis. Kaji et al [ 15 ] observed that menin interacts with Smad3 and inactivation of the former blocks transforming growth factor beta (TGF-β) signaling in pituitary tumor derived cell lines. Recently, two more menin interacting proteins, NF-kappa B [ 16 ] and a putative tumor metastasis suppressor nm23 [ 17 ] have been identified. Interactions among AP-1 family members, Smad proteins and NF-kappa B have been documented [ 18 - 21 ] and such cross talk among signaling pathways is not uncommon. Despite the above studies, a clear consensus of the molecular mechanisms leading to neoplasia, following the loss of menin, has not emerged. Very little is known about the gene expression changes in human neuroendocrine tumors following the loss of menin. Global gene expression analyses, using cDNA microarrays, have been used to classify other human tumors into clinically distinct categories [ 22 - 26 ]. Wu [ 27 ] has discussed the mathematical and statistical considerations for the use of DNA microarrays to identify genes of specific interest, and Harkin [ 28 ] has used expression profiling to identify downstream transcriptional targets of the BRCA1 tumor suppressor gene. Our objective was to identify genes that might be directly or indirectly over or under-expressed as a consequence of loss of menin expression. Results Patients and Controls Eight neuroendocrine tumors from six MEN1 patients were included in this study. The patient ages were 19, 22, 42, 51, 57, and 57 years at the time of surgery (Table 1 ). One was female, and five were male. Two of the patients had clinical and laboratory findings consistent with insulinoma. Three tumors were analyzed from one of these patients. One patient had findings consistent with VIP-oma (vasoactive intestinal polypeptide secreting tumor). Two patients, with no specific symptoms, had non-functioning or pancreatic polypeptide secreting tumors. One patient had symptoms of gastrinoma from a duodenal tumor (not used for this analysis). A pancreatic tumor from this patient, found incidentally, was used in this study. Pathological examination of tumors from the 6 patients resulted in the classification of 3 insulinomas, 3 neuroendocrine tumors, 1 VIP-oma and 1 glucagonoma. The ages of the individuals donating normal pancreatic islets were 42, 52(2), and 56 years. Two were female, and two were male. Table 1 Characteristics of patients and normal subjects. Pt.# T # Age Sex Clinical LN Mets T Vol. (ml) Menin Defect [6] 1 1 19 F Insulinoma 0/1 8.28 Large Deletion, exon 1 & 2 2 2 42 M Neuroendocrine Tumor 0/14 18.75 Nonesense Mutation, exon 7 6 6 60 M VIP-oma 1/16 288 8 bp Deletion, exon 5 7 7 51 M Neuroendocrine Tumor 2/30 3.75 2 bp Deletion, exon 2 8 8–10 22 M Insulinoma 2/8 6.9 2 bp Deletion, exon 2 11 11 57 M Gastrinoma 1/1 0.5 4 bp Deletion, exon 3 N1 N1 52 M Normal NA NA NA N2 N2 56 F Normal NA NA NA N3 N3 52 F Normal NA NA NA N4 N4 42 M Normal NA NA NA Quality of Hybridization The RNA isolated from 8 tumor specimens (6 patients) and 4 normal islet preparations was of acceptable quality for hybridization, as determined by preliminary small hybridizations on test chips. The dChip computer program returned data concerning the percent of genes judged to be present, and the percent of single and array outlier events (Table 2 ). The expression data from one normal islet preparation had 5.94% array outliers, which prompted dChip to issue a warning (a warning indicates more than 5% array outliers detected). However, since we had only four normal specimens, we elected to include all four in our analysis. The average level of gene expression was computed for each gene (Figure 1 ). The average gene expression level for all genes followed an exponentially decreasing pattern; the greatest number of genes had expression values less than 100, and only a few genes had expression levels greater than 4000. Table 2 Overall statistics on the quality of each the processed GeneChips. One chip was used for each tumor/normal specimen. The "Median Intensity" refers to the overall brightness of the fluorescence of the genes. The "Present Call" refers to whether the gene was "present" or "absent". Chip Name Median Intensity Present Call (%) Array outlier % Single outlier % Warning T1 170 49.4 1.12 0.11 T2 107 46.2 1.54 0.15 T6 160 51.4 1.16 0.12 T7 132 47.7 0.50 0.08 T8 158 51.0 0.59 0.10 T9 114 48.9 0.66 0.10 T10 158 50.6 0.42 0.07 T11 121 46.1 3.34 0.30 N1 142 48.4 2.65 0.26 N2 179 49.7 2.72 0.24 N3 75 48.3 3.38 0.31 N4 73 33.2 9.50 0.63 * Figure 1 Histogram showing the frequency of genes being expressed at levels between 50 and 7875 (arbitrary expression units). Overall Consistency of Gene Expression Average expression and standard deviation was computed for each gene in both the group of 4 normal islets, and the group of 8 islet tumors and expressed as the coefficient of variation (CV). Genes with average expression levels less than 50 were excluded from this analysis. Figure 2 shows that the average (11,416 genes and expressed sequences) CV in the group of 8 tumors was 30%. There was a linear regression of CV values as the average minimum expression level of the genes increased. Genes with an average minimum expression level of 7000 or more had an average CV level of 12.7%. The analysis of genes expressed in the normal islets gave similar results. However, when the tumors were combined with the normals, the CV was higher than either group alone. This was caused by the true differences in gene expression levels between the tumors and the normals. Figure 2 Coefficient of variation (CV) of genes being expressed at levels between 50 and 6000. For each gene expressed at an average level of 50 or above, the CV was computed for the group of 8 tumors, for the group of 4 normals, and for the group of all 12 tumors and normals. As the lower limit of expression was increased, the number of genes represented in the CV decreased: there were 12,000 genes with expression levels of 50 or more, but only a few genes with expression levels of 6,500 or more. Clustering The experimental groups were clustered (figure 3 ) using a hierarchical clustering procedure [ 29 , 30 ]. This cluster was based on the inclusion of all genes which had 33% to 67% of "present" calls made by the GeneChip software. The assignment of tumor type was made on the basis of principal hormone messenger RNA levels that were consistent with the clinical and biochemical findings (Table 3 ). The principal bifurcation in the clustering occurred between the group that included the normal specimens and the three tumors with a predominance of insulin expression, on one hand, and the other tumor types on the other. The four normal islet preparations clustered together, separate from the tumors. Among the normal islets, the females clustered separately from the males. Among the tumors, all 3 insulinomas clustered together, separate from the VIP-oma, the glucagonoma and the PP-omas (pancreatic polypeptide producing tumors). It is also interesting that all the specimens clustered in a pattern of increasing malignancy going from normal at the bottom of the cluster to most malignant at the top. Table 3 Gene expression levels of islet hormone mRNAs in tumors and normals. VIP: Vasoactive intestinal polypeptide; PP: Pancreatic polypeptide. T1 T2 T6 T7 T8 T9 T10 T11 N1 N2 N3 N4 pre-Gastrin 864 530 678 392 600 383 209 395 1036 775 28 1192 Insulin 9990 13 179 401 10195 240 8971 1831 10010 9752 9580 8158 Glucagon 10 6482 2783 1198 10 8370 10 10 9037 8425 9043 7800 VIP 351 278 10243 374 334 276 362 202 806 436 334 389 PP 246 7257 577 5845 70 1805 211 8895 1897 7605 3598 1177 Figure 3 Clustering of tumors and normals according to overall gene expression patterns. The predominant type of hormone expression (Table 3) is noted for each tumor/normal specimen. The genes were also clustered by the dChip software. A group of apoptosis-related genes was identified whose expression was significantly correlated with the Tumor/Normal assignment of the data. Twenty-four apoptosis-related genes represented by 26 different Affymetrix probes were identified in the overall hierarchical clustering. Nineteen of these genes were more highly expressed in the normal islets than in the islet tumors (Figure 4 ). Eighteen of the nineteen under expressed genes in the set of tumors had t-test p values (tumor vs. normal) <= 0.037. All five of the apoptosis-related genes, that were more highly expressed in the tumors, had t-test p values >0.05 Figure 4 Clustering of apoptosis-related genes in tumors (T) and normals (N). Pink indicates strong, white indicates moderate, and blue indicates weak expression. Evaluation of Student's t-test Since the Student's t-test was designed to compare only one parameter in two populations, the simultaneous measurement of multiple genes might lead to an excessive number of false positives. In order to empirically determine the potential false positive rate, we started with 923 genes which had a p value <=.05 and repeatedly scrambled the individual tests into groups 4 and 8 and then performed new t-tests. The average number of genes having a p value = < .05 in 20 such scrambles was 51 (5.5% of 923 genes). This was only slightly more than the 46 genes expected (0.05 × 923). We therefore concluded that there was little chance of excess false positives in repeatedly using the Student's t-test. Hormone Expression Profiles In order to obtain a better picture of the nature of the tumors and normal islets in this study, the expression levels of the principal hormone RNA of pancreatic islets was examined (Table 3 ). Tumors 1, 8, and 10 had high levels of insulin expression and came from patients with the clinical diagnosis of insulinoma. Tumor 6 had high levels of VIP and came from a patient with the clinical syndrome of VIP-oma. Tumors 2 and 7 had high levels of pancreatic polypeptide, and came from patients with only a diagnosis of neuroendocrine tumor. Tumor 9, which came from a patient with a clinical diagnosis of insulinoma had a high level of glucagon expression; the clinical diagnosis was apparently due to the other tumor (#8) which did have a high level of insulin expression. One other apparent discrepancy between the clinical diagnosis and hormone expression profile occurred with tumor 11, which had high a level of glucagon expression. This patient had an additional duodenal tumor that was responsible for the gastrin secretion and the clinical diagnosis. All the normal islet preparations had high levels of insulin and glucagon expression, as expected. Comparison of tumor and normal gene expression The reporting of differentially expressed genes was restricted to those in which the absolute ratio of Tumor to Normal was greater than or equal to 2, and which had a Student's t-test p value of less than or equal to .005. There were 193 genes that met the criteria. Expressed sequences with no known protein product were not included. There were 45 genes that were increased in the tumors relative to the normals, and 148 genes that were decreased. The fold-change in expression values ranged from +179 to -449. Genes were assigned to functional categories based on the Gene Ontology Consortium assignments . There were 16 genes related to cell growth, 13 genes related to signal transduction, and 16 genes related to other functions which were increased in the group of tumors relative to the group of normal islets (Table 4 ). There were 44 genes related to cell growth, 10 related to cell death, 10 related to embryogenesis, 5 related to nucleic acid binding, 21 related to cell signaling, and 58 related to other functions in the group of genes which were decreased in the islet tumors relative to the controls (Tables 5 , 6 , 7 , 8 ). Table 4 Genes significantly increased in tumors. GeneBank Accession Gene Symbol Normal Mean Tumor Mean Fold Change P value Cell Growth/Cycle X16323 hepatocyte growth factor HGF 11 116 10.77 0.003305 AB017642 oxidative-stress responsive 1 OSR1 58 428 7.41 0.000819 AL078641 phorbolin-like protein APOBEC3G 15 92 6.21 0.000158 L17128 gamma-glutamyl carboxylase GGCX 64 346 5.37 0.000018 D21089 xeroderma pigmentosum, complementation group C XPC 292 1278 4.38 0.000284 AL050223 vesicle-associated membrane protein 2 VAMP2 360 1533 4.26 0.002196 D38145 prostaglandin I2 synthase PTGIS 29 121 4.09 0.000448 AF092563 structural maintenance of chromosomes 2-like 1 SMC2L1 58 185 3.21 0.002352 AF006087 actin related protein 2/3 complex, subunit 4 ARPC4 292 865 2.96 0.000565 AC004537 inhibitor of growth family, member 3 ING3 46 114 2.47 0.003976 AF013168 tuberous sclerosis 1 TSC1 35 86 2.45 0.001232 AJ236876 ADP-ribosyltransferase polymerase)-like 2 ADPRTL2 32 76 2.34 0.003874 Cell Death/Apoptosis D38435 postmeiotic segregation increased 2-like PMS2L1 74 193 2.6 0.002976 M61906 phosphoinositide-3-kinase, regulatory subunit PIK3R1 43 104 2.4 0.004387 Signal Transduction U26710 Cas-Br-M ectropic retroviral transforming sequence b CBLB 21 177 8.4 0.000082 AB010414 guanine nucleotide binding protein, gamma 7 GNG7 59 334 5.68 0.003835 U59913 mothers against decapentaplegic homolog 5 MADH5 14 73 5.22 0.004731 AB004922 Homo sapiens gene for Smad 3 MADH3 93 443 4.76 0.001024 L11672 zinc finger protein 91 ZNF91 428 2007 4.69 0.000376 D14838 fibroblast growth factor 9 FGF9 27 108 3.97 0.000752 W27899 member RAS oncogene family RAB6B 68 232 3.43 0.00501 U48251 protein kinase C binding protein 1 PRKCBP1 40 127 3.18 0.001999 U90268 cerebral cavernous malformations 1 CCM1 53 151 2.87 0.004392 AL050275 cysteine rich with EGF-like domains CRELD1 195 543 2.79 0.000828 AB014600 SIN3 homolog B, transcriptional regulator SIN3B 177 425 2.39 0.001924 M27691 cAMP responsive element binding protein 1 CREB1 107 229 2.15 0.003559 U85245 phosphatidylinositol-4-phosphate 5-kinase, type II, beta PIP5K2B 244 518 2.12 0.000441 W25793 ring finger protein 3 RNF3 163 326 2 0.004947 Nucleic Acid Binding D50912 RNA binding motif protein 10 RBM10 96 443 4.6 0.001925 U41315 makorin, ring finger protein, 4 MKRN4 404 808 2 0.000262 Ligand Binding X67155 kinesin-like 5 KIF23 64 368 5.76 0.001584 AB028985 ATP-binding cassette, sub-family A, member 2 ABC1 65 262 4.04 0.001234 Z48482 matrix metalloproteinase 15 MMP15 139 495 3.56 0.003946 Enzyme X13794 lactate dehydrogenase B LDHB 396 1606 4.05 0.000845 X15334 creatine kinase, brain CKB 939 2083 2.22 0.002008 X60708 dipeptidylpeptidase IV DPP4 133 291 2.19 0.000697 AC004381 SA homolog SAH 283 599 2.11 0.000168 AF000416 exostoses-like 2 EXTL2 134 271 2.02 0.001314 Embryogenesis U48437 amyloid beta precursor-like protein 1 APLP1 851 2433 2.86 0.001043 U66406 ephrin-B3 EFNB3 168 438 2.6 0.00309 D50840 UDP-glucose ceramide glucosyltransferase UGCG 85 211 2.5 0.002554 Other/Unknown L48215 hemoglobin, beta HBB 12 2099 178.78 0.001299 J00153 hemoglobin, alpha 1 HBA1 15 1249 82.25 0.001889 U30521 P311 protein C5orf13 157 453 2.88 0.001431 AB011169 similar to S. cerevisiae SSM4 TEB4 140 300 2.15 0.00154 AL031432 GCIP-interacting protein P29 99 198 2 0.002036 Table 5 Genes significantly decreased in tumors. GeneBank Accession Gene Description Symbol Normal Mean Tumor Mean Fold Change P value Cell Growth/Division D17291 regenerating protein I beta REG1B 6286 13 -499.46 0.000095 X67318 carboxypeptidase A1 CPA1 3928 121 -32.57 0.003205 AI763065 regenerating islet-derived 1 alpha REG1A 5641 334 -16.88 0.000001 D29990 solute carrier family 7, member 2 SLC7A2 2988 445 -6.72 0.002204 AB017430 kinesin-like 4 KIFF22 1223 316 -3.87 0.000177 Z25884 chloride channel 1 CLCN1 2511 655 -3.84 0.00013 X81438 amphiphysin AMPH 2686 752 -3.57 0.000002 L03785 myosin, light polypeptide 5 MYL5 207 59 -3.51 0.000233 W28062 guanine nucleotide-exch. Prot. 2 ARFGEF2 66 19 -3.46 0.003602 X52486 uracil-DNA glycosylase 2 UNG2 2555 756 -3.38 0.000514 M81933 cell division cycle 25A CDC25A 312 96 -3.25 0.000005 M69136 chymase 1 CMA1 360 115 -3.13 0.004413 U90543 butyrophilin BTN2A1 685 226 -3.04 0.000023 X69086 utrophin UTRN 1325 457 -2.90 0.000011 AF039241 histone deacetylase 5 HDAC5 1124 393 -2.86 0.000319 U49392 allograft inflammatory factor 1 AIF1 165 58 -2.82 0.000105 U81992 pleiomorphic adenoma gene-like 1 PLAGL1 330 118 -2.80 0.004717 L26336 heat shock 70kD protein 2 HSPA2 90 32 -2.79 0.000689 F27891 cytochrome c oxidase subunit VIa COX6A2 872 313 -2.79 0.000342 D87673 heat shock transcription factor 4 HSF4 1964 721 -2.73 0.000453 X97795 RAD54-like RAD54L 392 144 -2.72 0.001345 X92689 UDP-N-acetyl-alpha-D-galactosamine GALNT3 80 32 -2.50 0.000243 Y08683 carnitine palmitoyltransferase I CPT1B 1038 420 -2.47 0.000573 U40622 X-ray repair complementing defective repair 4 XRCC4 177 72 -2.45 0.000678 U64315 excision repair, complementation group 4 ERCC4 2122 868 -2.44 0.000045 AB020337 beta 1,3-galactosyltransferase B3GALT5 1489 635 -2.34 0.002613 U40152 origin recognition complex ORC1L 3671 1702 -2.16 0.001425 M10943 metallothionein 1F MT1F 5691 2653 -2.14 0.001707 X79882 major vault protein MVP 758 376 -2.02 0.001719 AF035960 transglutaminase 5 TGM5 3097 1542 -2.01 0.002951 Cell Death/Apoptosis S81914 immediate early response 3 IER3 2209 480 -4.60 0.000307 D80007 programmed cell death 11 PDCD11 457 129 -3.55 0.002358 AF013956 chromobox homolog 4 CBX4 1599 492 -3.25 0.00034 U33284 protein tyrosine kinase 2 beta PTK2B 693 237 -2.93 0.000763 U90919 likely partner of ARF1 APA1 2687 1021 -2.63 0.000015 X57110 Cas-Br-M retroviral transforming CBL 1889 784 -2.41 0.000033 AL050161 pro-oncosis receptor PORIMIN 1178 497 -2.37 0.00031 U40380 presenilin 1 PSEN1 1301 569 -2.29 0.00012 D83699 harakiri, BCL2 interacting protein HRK 768 338 -2.27 0.001321 U07563 v-abl viral oncogene homolog 1 ABL1 1415 631 -2.24 0.000248 M95712 v-raf oncogene homolog B1 BRAF 338 157 -2.16 0.004207 M16441 lymphotoxin alpha LTA 2106 985 -2.14 0.000239 AF035444 pleckstrin homology-like domain, family A, member 2 PHLDA2 334 166 -2.01 0.001759 Table 6 Genes significantly decreased in tumors (continued). GeneBank Accession Gene Description Symbol Normal Mean Tumor Mean Fold Change P value Signal Transduction J00306 somatostatin SST 7701 284 -27.09 0 AI636761 somatostatin SST 7224 598 -12.09 0.000001 AB011143 GRB2-associated binding protein 2 GAB2 2237 402 -5.57 0.001816 M93056 serine (or cysteine) proteinase inhibitor SERPINB1 505 105 -4.80 0.004637 X68830 islet amyloid polypeptide IAPP 2231 477 -4.68 0.001221 AB029014 RAB6 interacting protein 1 RAB6IP1 824 181 -4.56 0.000155 AI198311 neuropeptide Y NPY 610 154 -3.96 0.004817 M28210 member RAS oncogene family RAB3A 2566 672 -3.82 0.000048 J04040 glucagon GCG 8620 2351 -3.67 0.000396 AF030335 purinergic receptor P2Y P2RY11 2314 680 -3.40 0.000058 M29335 major histocompatibility complex HLA-DOA 906 268 -3.39 0.00159 L38517 Indian hedgehog homolog IHH 3013 897 -3.36 0.000055 U95367 gamma-aminobutyric acid A receptor, pi GABRP 668 202 -3.30 0.000837 W28558 pleiotropic regulator 1 PLRG1 704 216 -3.26 0.000068 L08485 gamma-aminobutyric acid A receptor, alpha 5 GABRA5 342 107 -3.20 0.000336 AF004231 leukocyte immunoglobulin-like receptor LILRB2 93 30 -3.08 0.001105 AF055033 insulin-like growth factor binding protein 5 IGFBP5 126 43 -2.96 0.000257 AJ010119 ribosomal protein S6 kinase RPS6KA4 1532 522 -2.94 0.000201 U46194 Human renal cell carcinoma antigen RAGE 2057 754 -2.73 0.000324 L13858 son of sevenless homolog 2 SOS2 964 354 -2.72 0.000268 Z29572 tumor necrosis factor receptor superfamily TNFRSF17 184 68 -2.69 0.000178 U01134 fms-related tyrosine kinase 1 FLT1 910 379 -2.40 0.003257 D78156 RAS p21 protein activator 2 RASA2 327 144 -2.26 0.002332 U77783 glutamate receptor GRIN2D 518 240 -2.15 0.001379 D49394 5-hydroxytryptamine receptor 3A HTR3A 197 98 -2.02 0.002493 Nucleic Acid Binding Z30425 nuclear receptor subfamily 1, group I, member 3 NR1I3 1008 356 -2.83 0.000329 U18760 nuclear factor I/X NFIX 5796 2216 -2.62 0.000711 AI223140 purine-rich element binding protein A PURA 1137 506 -2.25 0.002448 AF015950 telomerase reverse transcriptase TERT 561 255 -2.20 0.002839 U40462 zinc finger protein, subfamily 1A, 1 ZNFN1A1 662 308 -2.15 0.001171 Z93930 X-box binding protein 1 XBP1 2223 1061 -2.09 0.000277 AB019410 PET112-like PET112A 1422 707 -2.01 0.001309 Ligand Binding X00129 retinol binding protein 4, plasma RBP4 1517 68 -22.27 0.004809 AJ223317 sarcosine dehydrogenase SARDH 3844 1069 -3.60 0.000085 AB017494 LCAT-like lysophospholipase LYPLA3 906 326 -2.78 0.001131 U78735 ATP-binding cassette, sub-family A, member 3 ABCA3 1914 706 -2.71 0.000288 AF026488 spectrin, beta, non-erythrocytic 2 SPTBN2 1604 671 -2.39 0.00005 U83659 ATP-binding cassette, sub-family C, member 3 ABCC3 1287 551 -2.34 0.00244 R93527 metallothionein 1H MT1H 5093 2196 -2.32 0.002937 AA586894 S100 calcium binding protein A7 S100A7 507 221 -2.29 0.000537 U91329 kinesin family member 1C KIF1C 2981 1484 -2.01 0.000518 Table 7 Genes significantly decreased in tumors (continued). GeneChip Accession Gene Description Symbol Normal Mean Tumor Mean Fold Change P value Enzyme M81057 carboxypeptidase B1 CPB1 4534 79 -57.09 0.001106 X71345 protease, serine, 4 PRSS3 3859 76 -51.11 0.004102 X01683 serine (or cysteine) proteinase inhibitor, clade A SERPINA1 2550 74 -34.64 0.004833 M24400 chymotrypsinogen B1 CTRB1 5158 207 -24.95 0.001744 M18700 elastase 3A, pancreatic ELA3A 7058 384 -18.37 0.000009 U66061 protease, serine, 1 PRSS1 7291 645 -11.31 0.000047 L22524 matrix metalloproteinase 7 MMP7 595 54 -11.03 0.002591 AI655458 5-oxoprolinase (ATP-hydrolysing) OPLAH 446 99 -4.52 0.004072 H94881 FXYD domain-containing ion transport regulator 2 FXYD2 3116 708 -4.40 0.000539 AL021026 flavin containing monooxygenase 2 FMO2 905 215 -4.21 0.000804 AC005525 plasminogen activator, urokinase receptor PLAUR 1779 566 -3.14 0.000031 U40370 phosphodiesterase 1A, calmodulin-dependent PDE1A 268 89 -3.03 0.004023 R90942 sialyltransferase 7D SIAT7D 3148 1052 -2.99 0.002319 M84472 hydroxysteroid (17-beta) dehydrogenase 1 HSD17B1 1196 440 -2.72 0.000991 X55988 ribonuclease, RNase A family, 2 RNASE2 480 203 -2.36 0.001314 AB003151 carbonyl reductase 1 CBR1 4538 1945 -2.33 0.000511 X08020 glutathione S-transferase M1 GSTM1 2766 1376 -2.01 0.000519 Embryogenesis U15979 delta-like homolog SIGLEC5 3384 402 -8.41 0.002927 M60094 H1 histone family, member T HIST1H1T 976 230 -4.23 0.001639 U50330 bone morphogenetic protein 1 BMP1 3298 973 -3.39 0.001637 M74297 homeo box A4 HOXA4 501 176 -2.85 0.000477 AJ011785 sine oculis homeobox homolog 6 SIX6 530 190 -2.79 0.000286 U66198 fibroblast growth factor 13 FGF13 191 73 -2.61 0.001068 D31897 double C2-like domains, alpha DOC2A 1151 451 -2.55 0.000068 U12472 glutathione S-transferase pi GSTP1 3122 1524 -2.05 0.000237 Transcription AL049228 pleckstrin homology domain interacting protein PHIP 257 33 -7.69 0.000782 M27878 zinc finger protein 84 ZNF84 54 15 -3.64 0.001108 U77629 achaete-scute complex-like 2 ASCL2 438 184 -2.38 0.000058 D50495 transcription elongation factor A, 2 TCEA2 1330 595 -2.23 0.000019 U49857 transcriptional activator of the c-fos promoter CROC4 542 259 -2.09 0.003894 Table 8 Genes significantly decreased in tumors (continued). GeneBank Accession Gene Description Symbol Normal Mean Tumor Mean Fold Change P value Other/Undefined X72475 immunoglobulin kappa constant IGKC 1409 276 -5.11 0.000111 D17570 zona pellucida binding protein ZPBP 355 71 -5.02 0.001107 M90657 transmembrane 4 superfamily member 1 TM4SF1 592 141 -4.20 0.004537 AF063308 mitotic spindle coiled-coil related protein SPAG5 2015 502 -4.01 0.000588 U66059 T cell receptor beta locus TRB@ 3022 779 -3.88 0.000266 AL022165 carbohydrate sulfotransferase 7 CHST7 359 94 -3.82 0.001738 U10694 melanoma antigen, family A, 9 MAGEA9 1039 272 -3.82 0.000067 M73255 vascular cell adhesion molecule 1 VCAM1 80 22 -3.66 0.004179 U47926 leprecan-like 2 protein LEPREL2 1003 319 -3.15 0.00013 L05424 CD44 antigen CD44 1439 471 -3.05 0.001361 AI445461 transmembrane 4 superfamily member 1 TM4SF1 463 161 -2.88 0.002911 AF010310 proline oxidase homolog PRODH 1194 421 -2.84 0.000005 AF000991 testis-specific transcript, Y-linked 2 TTTY2 700 254 -2.76 0.000542 X57522 transporter 1, ATP-binding cassette, sub-family B TAP1 781 287 -2.72 0.000971 AA314825 trefoil factor 1 TFF1 1657 616 -2.69 0.000011 AB020880 squamous cell carcinoma antigen SART3 3228 1224 -2.64 0.000135 AF040707 homologous to yeast nitrogen permease NPR2L 1131 437 -2.59 0.001537 U47292 trefoil factor 2 TFF2 359 141 -2.54 0.000684 X69398 CD47 antigen CD47 350 144 -2.42 0.000853 U27331 fucosyltransferase 6 FUT6 1105 473 -2.34 0.000872 AI827730 cyclin M2 CNNM2 5863 2535 -2.31 0.000484 U05255 glycophorin B GYPB 1606 717 -2.24 0.00013 M34428 pvt-1 oncogene homolog, MYC activator PVT1 1231 550 -2.24 0.004423 U86759 netrin 2-like NTN2L 2039 937 -2.18 0.000204 D90278 CEA-related cell adhesion molecule 3 CEACAM3 4388 2024 -2.17 0.000902 L40400 ZAP3 protein ZAP3 1549 719 -2.15 0.000776 U48224 beaded filament structural protein 2, phakinin BFSP2 568 271 -2.10 0.000166 AI138834 deltex homolog 2 DTX2 311 148 -2.10 0.000687 M13755 interferon-stimulated protein, 15 kDa G1P2 1507 741 -2.03 0.001157 X52228 mucin 1, transmembrane MUC1 1523 756 -2.02 0.001707 Validation of GeneChip Data with Quantitative RT-PCR In order to evaluate how accurately the GeneChip data was representing actual gene expression levels, eleven genes were tested with quantitative RT-PCR (Q-PCR). The results are shown in Table 9 . The correlation coefficients ranged from 0.964 to 0.235 with an average of 0.655. The lower correlation coefficients were associated with genes with larger numbers of exons. There was some association of low correlation with low average numerical expression values. The lowest correlations were associated with very faint image intensity of the involved genes in the dChip visual representation. The correlation coefficients of 4 genes, identified as apoptosis-related, was examined in detail (Figure 5 ). IER3, IAPP, SST, and PHLDA2 all had good correlation between GeneChip and Q-PCR results. FGF9, a potential growth stimulating gene was also examined (Figure 6 ). Again, there was overall good correlation between the individual GeneChip and Q-PCR results. Table 9 Correlation of GeneChip expression with quantitative RT-PCR. Gene Symbol Correlation Probe Set Exons Gene Size (bp) Fold Change (T/N) P value GeneChip T vs. N IER3 0.964 1237_at 1 1236 -4.6 0.0000 SST 0.925 37782_at 2 351 -12 0.0000 PHLDA2 0.909 40237_at 2 913 -2.01 0.0003 REG1B 0.875 35981_at 6 773 -499 0.0000 IAPP 0.823 37871_at 3 1462 -4.68 0.0033 REG1A 0.814 38646_s_at 6 808 -16.9 0.0000 FGF9 0.74 1616_at 3 1420 3.97 0.0031 CBLB 0.327 514_at 21 3923 3.01 0.0009 XPC 0.318 1873_at 16 3658 4.38 0.0018 HRK 0.273 34011_at 2 716 -2.27 0.0011 PTK2B 0.235 2009_at 38 4715 -2.94 0.0019 Average 0.655 Figure 5 The expression levels of 4 apoptosis-related genes are shown by GeneChip and quantitative RT-PCR: a) IER3; b) IAPP; c) SST; d) PHLDA2. Normals (N) and tumors (T) are shown. Solid bars represent GeneChip and open bars represent Q-PCR results. Figure 6 FGF9 expression levels in tumors (T) and normals (N) by GeneChip and quantitative RT-PCR. Solid bars represent GeneChip and open bars represent Q-PCR results. Discussion Whether there were degradative processes acting on the tissues prior to or during or after the extraction of the RNA can be guessed by the quality of the RNA. Each RNA specimen in this study was tested on an Affymetrix test chip, and each was found to be acceptable. Additional quality assessment was made by the dChip software. Only one specimen, a normal control, had Array Outliers greater than 5%, suggesting that it was subnormal (Table 2 ). However, since the percent outliers was only 5.94, the chip was included in the analysis. Although, only solid tumor was utilized, there were undoubtedly a small percentage of blood, blood vessel, and connective tissue elements intermixed with the tumor tissue. Rarely, there might be a small amount of exocrine tissue. In the case of the normal islets used as controls, microscopic examination showed that greater than 90% of the tissue was islet. Any contaminants would probably have the effect of reducing the discriminant power to differentiate tumor from normal. Thus, t-test p values and fold changes would tend to under-represented and some, otherwise significant, genes might be missed. The actual data, represented by the hierarchical specimen clustering (Figure 3 ), showed strong differential gene expression relating to group identity as would be expected if the overall gene expression levels were accurate. All the normals clustered together, separate from all the tumors. Within the normals, the two male specimens clustered in one group, and the two female in another. All the normal islet preparations, which are composed predominantly of beta cells, clustered closer to the insulinoma tumors than to the other neuroendocrine tumor types. The gene clustering results revealed 19 apoptosis-related genes whose expression was suppressed in the islet tumors relative to the normals. This suggests that apoptosis may play a significant role in the development of these tumors. One might have expected more variation in the gene expression levels in the tumors than in the normal islets, since tumors are often heterogonous. However the data on the average CV of the genes in the normal and tumor groups suggested that there was no more variation in the tumors (average CV of 30%) than in the normals (average CV of 31%). The low CV in the tumors may relate to the single mode of tumor formation (induction by the loss of the menin tumor suppressor). However, there was increased variation noted when the tumors and normals were combined (Figure 2 ). This was probably the result of the differences in expression between the tumors and the normals. Of particular interest was the high proportion (3/8) of tumors expressing principally PP hormonal RNA. This was entirely consistent with pathological studies showing the preponderance of PP containing tumors in the pancreas of MEN1 patients [ 31 ]. The fact that the clinical classification of two patients (9 and 11) was different than indicated by the hormone expression profile of the tumor analyzed was a consequence of the facts that those patients had multiple tumors secreting multiple hormones but only insulin and gastrin and sometime PP over secretion are likely to result in a clinical diagnosis. The use of the Students t-test for comparison of multiple genes might be questioned because the test was designed for comparison of only two groups. In this study, we confirmed that comparison of 923 genes would not generate an excess number of false positive results. Nevertheless, in the group of 193 genes finally selected at a p < = .005, we can expect that 1 of those genes is a false positive. This study suggests that the overall effect of loss of function of menin is the suppression of gene expression. Nevertheless, there were 86 genes that were over-expressed in the tumors relative to the normals. Although we associate tumorigenesis with increased rates of growth, only two of eleven Cell Cycle and Cell Proliferation genes were increased in the tumors. Since tumor growth may also be significantly affected by rates of cell death, it is perhaps significant that there were no Cell Death genes significantly increased in the tumors relative to the controls. The correlation of GeneChip results with quantitative real-time PCR (Q-PCR, Table 9 ) was relatively good. However, there were some genes that correlated poorly (correlation coefficient less than 0.6). Interestingly, most of the genes with poor correlation coefficients had a large number of exons, whereas those with high correlation coefficients had a low number of exons. Since exhaustive testing of alternative primer pairs for Q-PCR was not made, it is possible that correlation coefficients of some genes could be improved by the use of other primers. Four studies of global gene expression in pancreatic islets have been published recently [ 32 - 35 ]. Cardozo et al [ 32 ] have used microarrays to look for NF-kB dependent genes in primary cultures of rat pancreatic islets. Shalev et al [ 33 ] have measured global gene expression in purified human islets in tissue culture under high and low glucose concentrations. They noted that the TGFβ superfamily member PDF was down regulated 10-fold in the presence of glucose, whereas other TGFβ superfamily members were up regulated. In the current study, none of the TGFβ superfamily members were significantly different between tumor and normal. Scearce et al [ 34 ] have used a pancreas-specific micro-chip, the PanChip to analyze gene expression patterns in E14 to adult mice. Only a few specific genes were noted in the paper, and none of them had human homologs of significance to the current study. Maitra et al [ 35 ] conducted a study which in many ways was similar to the current one. They compared gene expression, using the Affymetrix U133A chip, in a series of sporadic pancreatic endocrine tumors with isolated normal islets. There was no overlap in the genes they identified (having a three-fold or greater difference in expression) with the genes we identified (having a two-fold or greater difference in expression). This is quite surprising, but perhaps suggests that sporadically arising tumors may have a quite different pattern of gene expression than tumors arising as a result of menin loss or dysfunction. Another possible cause of the differences may be the different Affymetrix GeneChips used in the two studies. The question of which (if any) of the genes delineated in this study are a direct and necessary affect of loss-of-menin tumorigenesis cannot be determined by this study alone. Firstly, the activity of many genes are regulated both by their levels of expression and by post-translation modifications, such as phosphorylation. Secondly, the microchips used in this study represent only about 1/3 of the total number of human genes. Thirdly, it is not certain that the initiating gene changes caused by loss-of-menin are persistent in the tumors that develop. However, there were some genes, which because of their association with growth or apoptosis are of special interest. The general suppression of apoptosis related genes noted in this study (Figure 4 ) has been highlighted by the recent study of Schnepp et al , [ 36 ] who showed a loss of menin suppression of apoptosis in murine embryonic fibroblasts through a caspase-8 mechanism. Specific apoptosis-related genes which were suppressed in the tumors in the current study, and which were confirmed by Q-PCR include IER3, SST, PHLDA2, and IAPP. IER3 (IEX-1) is regulated by several transcription factors and may have positive or negative effects upon cell growth and apoptosis depending upon the cell-specific context [ 37 ]. Several studies have shown that it can be a promoter of apoptosis [ 38 - 40 ]. Somatostatin has shown a wide range of growth inhibitory activity in vitro and in vivo [ 41 - 57 ].PHLDA2 (TSSC3) is an imprinted gene homologous to the murineTDAG51 apoptosis-related gene [ 58 ], and may be involved in human brain tumors [ 59 ]. IAPP (amylin) is a gene which has contrasting activities and has been associated with experimental diabetes in rodents [ 60 ]. Amylin deposits were increased in islets of patients with gastrectomy-induced islet atrophy [ 61 ]. On the other hand, exposure of rat embryonic islets to amylin results in beta cell proliferation [ 62 ]. In contrast, amylin has been shown to induce apoptosis in rat and human insulinoma cells in vitro [ 63 , 64 ]. In contrast to the suppression of apoptosis-related genes, FGF9 (Figure 6 ), a growth promoting gene, was significantly increased in the neuroendocrine tumors. This protein has been reported to play roles in glial cell growth [ 65 ], chondrocyte growth [ 66 ], prostate growth [ 67 ], endometrial growth [ 68 ], and has been suggested to have a role in human oncogenesis [ 69 ]. A recent report by Busygina et al [ 70 ] suggested that loss of menin can lead to hypermutability in a Drosophila model for MEN1. The spectrum of mutation sensitivity suggested that there was a defect in nucleotide excision repair. Whether the defect was a direct or indirect effect of menin loss was not stated. In the current study, there was a 2.44-fold decrease, in the tumors, in the expression of ERCC4 (Table 5 ), a gene involved in nucleotide excision repair. In addition, XRCC4, a gene involved in double-strand break repair, was also decreased in the tumors in the current study. Conclusion This first study of global gene expression in neuroendocrine tumors arising in the pancreas of patients with the MEN1 syndrome has identified many genes that are differentially expressed, as compared with normal human islet cells. A number of these genes are strongly over/under expressed and are attractive candidates for further investigation into the mechanisms by which menin loss causes tumors in pancreatic islets. Of particular interest was a group of 24 apoptosis-related genes that were significantly differentially expressed (mostly underexpressed) in the group of neuroendocrine tumors. The underexpression of these apoptosis-related genes may be related to neoplastic development or progression in these MEN1-related neuroendocrine tumors. Methods Human Tissue Specimens Tumor specimens were obtained from patients with the MEN1 syndrome who had undergone surgery for islet-cell tumors of the pancreas. The specific germline mutations in the menin tumor suppressor gene were identified and previously reported [ 6 ] for each of the patients. Six of the patients had frame-shift mutations and one had a nonsense mutation. Informed consent was obtained in advance, and tumor tissues not needed for pathological analysis were snap frozen in liquid nitrogen, and kept frozen at -70° prior to RNA extraction. Normal pancreatic islets (which were originally intended for human transplatation studies, but were not used) were isolated from brain-dead donors by a collagenase procedure, as previously described [ 71 ], and were then frozen until used for extraction of RNA. Human Studies Committee approval from Washington University School of Medicine was obtained for this study. Isolation of RNA from Tissue Specimens Approximately 50 mg of tissue was removed from each frozen tumor specimen and homogenized with a mortar and pestle (Qiagen, Qiashredder Kit), and RNA was extracted using the Rneasy Mini Kit (Qiagen, Inc.), and quantified by UV absorbance. RNA was similarly isolated from the normal human islet preparations. GeneChip Hybridization and Analysis The RNA was submitted to the GeneChip facility of the Siteman Cancer Center at Washington University School of Medicine. There, biotin labeled cRNA was prepared and hybridized to U95Av2 microarray chips (Affymetrix). The fluorescence of individual spots was then measured and the data returned on compact discs. We analyzed the gene expression data and made comparisons between groups using the dChip computer program [ 30 ]. Following normalization (to equalize the overall intensity of each chip), the expression of each gene was determined by statistical modeling procedure in the dChip software. Each gene was represented by an array of 10 perfect match oligonucleotide spots and 10 mismatch oligonucleotide spots on the U95Av2 chip. The dChip program examines all the spots on all the chips involved in the study, and by a statistical procedure determines single and array outliers. These outliers can be considered as "bad" readings, and removed from further consideration. Quantitative RT-PCR The same preparations of total RNA that were used to probe the GeneChips were also used to prepare c-DNA for quantitative RT-PCR analysis of gene expression. C-DNA was first prepared using Superscript II reverse transcriptase (Invitrogen, Inc.). Primers for each gene were designed to produce products of 100 to 150 bp that spanned exon boundaries (when possible). The primer pairs are shown in table 10 . Table 10 Gene Forward Primer Reverse Primer CBLB cacgtctaaatctatagcagccagaac tgcactcccaagcctcttctc FGF9 cggcaccagaaattcacaca aaattgtctttgtcaactttggcttag HRK agctggttcccgttttcca cagtcccattctgtgtttctacgat IAPP ctgctttgtatccatgagggttt gaggtttgctgaaagccacttaa ER3 ccagcatctcaactccgtctgt caccctaaaggcgacttcaaga SST cccagactccgtcagtttctg tacttggccagttcctgcttc PHLDA2 tgcccattgcaaataaatcact ctgcccgcccattcct PTK2B gtgaggagtgcaagaggcagat gccagattggccagaacct REG1A cctcaagcacaggattccaga acatgtattttccagctgcctcta REG1B gggtccctggtctcctacaagt catttcttgaatcctgagcatgaa XPC gcccgcaagctggacat atcagtcacgggatgggagta The Sybr Green technique on an Applied Biosystems model GeneAmp 5700 instrument was utilized. Relative quantitation using a standard c-DNA preparation from an in vitro islet tumor cell line was utilized. Competing interests The author(s) declare that they have no competing interests. Authors' contributions All authors contributed equally to this manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549185.xml

517945

The effects of moderate alcohol supplementation on estrone sulfate and DHEAS in postmenopausal women in a controlled feeding study

Background We have demonstrated that moderate alcohol consumption (15 g/d, 30 g/d) for 8 weeks resulted in significantly increased levels of serum estrone sulfate and DHEAS in 51 postmenopausal women in a randomized, placebo-controlled trial. We now report on the relationships between serum estrone sulfate and dehydroepiandrosterone sulfate (DHEAS) levels after 4 weeks of moderate alcohol supplementation, and compare the results to the 8 weeks data to elucidate time-to-effect differences. Methods Postmenopausal women (n = 51) consumed 0 (placebo), 15 (1 drink), and 30 (2 drinks) g alcohol (ethanol)/ day for 8 weeks as part of a controlled diet in a randomized crossover design. Blood samples were drawn at baseline, at 4 weeks and at 8 weeks. Changes in estrone sulfate and DHEAS levels from placebo to 15 g and 30 g alcohol per day were estimated using linear mixed models. Results and Discussion At week 4, compared to the placebo, estrone sulfate increased an average 6.9% (P = 0.24) when the women consumed 15 g of alcohol per day, and 22.2% (P = 0.0006) when they consumed 30 g alcohol per day. DHEAS concentrations also increased significantly by an average of 8.0% (P < 0.0001) on 15 g of alcohol per day and 9.2% (P < 0.0001) when 30 g alcohol was consumed per day. Trend tests across doses for both estrone sulfate (P = 0.0006) and DHEAS (P < 0.0001) were significant. We found no significant differences between the absolute levels of serum estrone sulfate at week 4 versus week 8 (P = 0.32) across all doses. However, absolute DHEAS levels increased from week 4 to week 8 (P < 0.0001) at all three dose levels. Conclusions These data indicate that the hormonal effects due to moderate alcohol consumption are seen early, within 4 weeks of initiation of ingestion.

Background Epidemiological evidence consistently shows a positive association between alcohol, even low to moderate intake, and breast cancer risk [ 1 ]. However, during the past two decades, it has become evident that moderate drinking is associated with longer life [ 2 ], reduced rates of heart disease [ 3 ] and stroke [ 4 ]. What does this mean for women when the epidemiologic data show an exposure is associated with both benefits and harms? Recommendations regarding the use or avoidance of moderate alcohol, must take into consideration both its potential benefit on cardiovascular disease, as well as its potential risk for breast cancer. To understand the biologic parameters potentially influenced by alcohol, there is a need for well-controlled mechanistic studies of dose and duration of use on markers of risk in the causal pathway of breast cancer. A reanalysis of nine prospective studies shows that levels of endogenous sex hormones are strongly associated with breast cancer risk in postmenopausal women [ 5 ]. Clear evidence that moderate alcohol consumption increases levels of hormones associated with increased risk for breast cancer would provide support for a causal relationship. In a controlled study of acute alcohol ingestion, serum estrone levels were significantly elevated in postmenopausal women on hormone replacement therapy (HRT), but did not affect serum estrone levels in women not using HRT [ 6 ]. To better understand the effects of moderate long-term alcohol ingestion on sex hormones, we conducted a controlled feeding study in postmenopausal women not using HRT. As previously reported, there were significant elevations in both serum estrogen sulfate and dehydroepiandrosterone sulfate (DHEAS) after 8 weeks of supplementation [ 7 ]. Here, we evaluate the relationships between serum estrogen sulfate and DHEAS after 4 weeks of supplementation with moderate alcohol, and compare the results to the 8 week data to elucidate time-to-effect differences. Methods Details of the Women's Alcohol Study (WAS) was previously published [ 7 ]. Briefly, a total of 51 postmenopausal women not on HRT completed the study and are included in the analysis. The WAS utilized a crossover design. Each participant rotated through three 8-weeks controlled dietary periods during which she consumed a beverage daily that contained no alcohol (placebo), 15 g alcohol, or 30 g alcohol in random order. Each of the three dietary periods was preceded by a 2- to 5-week washout period during which the women consumed no alcohol. Alcohol was supplied as 95% ethanol (Everclear™ Pharmaco Products, Inc., Brookfield, CN) in orange juice (12 ounces). All meals were prepared at the Beltsville Human Nutrition Research Center and the participants ate breakfast and supper at the Center and had carryout lunches on weekdays. On weekends, food and beverages were packaged for consumption at home. The calorie level for each subject was adjusted to maintain constant body weight. Blood for hormone analyses was collected after an overnight fast at weeks 4 and 8. Serum was separated, and aliquots were frozen at -70°C. In the samples taken after 4 weeks of alcohol supplementation, we only measured serum estrone sulfate and DHEAS, as they were the only two hormones that were significantly elevated at 8 weeks [ 7 ]. Hormone concentrations were transformed using natural log. Changes in hormone concentrations from placebo to 15 g and 30 g of alcohol per day were estimated at 4 and 8 weeks using linear mixed models including a random intercept and alcohol levels as fixed effects treated as two indicator variables. Separate models used alcohol as a continuous variable to test for trend. Regressions to evaluate the effect of several baseline covariates on the precision of the parameter estimates for alcohol consumption included age, BMI, and years since menopause modeled as continuous fixed effects; assignment order, dietary period, hysterectomy, and race were modeled as indicator variables. Effect modification by race, assignment order, dietary period, age, BMI, and years since menopause were assessed by likelihood ratio tests of improvement in the model fit after addition of cross-product terms to models that included main effects for alcohol and the characteristic being evaluated. All tests of statistical significance were two-sided. Statistical analyses were performed using S-PLUS (S-PLUS version 6.1 for Windows. Seattle (WA): Insightful Corporation; 2002.) Results and Discussions Table 1 summarizes the mean hormone concentrations at four weeks and eight weeks in the participants when not consuming alcohol and the percent changes from no alcohol consumption when consuming 15 g or 30 g of alcohol per day, respectively. At 4 weeks of alcohol intake the inclusion of age, years since menopause, race, and baseline BMI in the models did not change the precision of parameter estimates for alcohol doses and thus results from simple models are presented. At week 4, compared to the no alcohol placebo, estrone sulfate increased an average 6.9% (P = 0.24) when women consumed 15 g of alcohol per day and 22.2% (P = 0.0006) when they consumed 30 g of alcohol per day. DHEAS concentrations also increased significantly by an average 8.0% (P < 0.0001) on 15 g of alcohol per day and 9.2% (P < 0.0001) when 30 g alcohol was consumed per day at week 4. Trend tests for both estrone sulfate (P = 0.0006) and DHEAS (P < 0.0001) were highly significant. Table 1 Geometric mean hormone levels (ng/dL) for participants on 0 g alcohol and % change (Δ) in hormone levels from 0 g to 15 g and 30 g alcohol per day at weeks 4 and 8 Hormone 0 g/d, mean (95% C.I.) 15 g/d, Δ (95% C.I.)* 30 g/d, Δ (95% C.I.)* P -trend† Estrone Sulfate (4 wk) 47.5 (39.2–57.4) 6.9% (-4.2%–19.3%) 22.2% (9.4%–36.5%) 0.0006 (8 wk) 47.4 (40.5–55.6) 7.5% (-0.3%–15.9%) 10.7% (2.7%–19.3%) 0.009 DHEAS (4 wk) 55.3 (47.0–65.1) 8.0% (4.5%–11.6%) 9.2% (5.6%–12.8%) <0.0001 (8 wk) 59.4 (50.5–70.0) 5.1% (1.4%–9.0%) 7.5% (3.7%–11.5%) 0.0001 * Estimates of percent change are from linear mixed models, including participant as a random effect and alcohol levels as fixed effects treated as two indicator variables. † P -trend values (two-sided) are from linear mixed models, including participant as a random effect and alcohol levels as a continuous fixed effect with values 0, 15, and 30. The effects of alcohol supplementation on serum estrone sulfate and DHEAS levels did not vary with age, BMI, race, or years since menopause. These hormone concentrations also did not differ among the three dietary periods, and the order of the assignment to the treatment regimens did not modify the associations with either of the two hormones. At 8 weeks of alcohol intake, the 15 g dose versus the placebo increased serum estrone sulfate and DHEAS by 7.5% and 5.1% respectively, whereas the 30 g dose compared to the placebo, increased levels of estrone sulfate and DHEAS by 10.7% and 7.5% respectively. When comparisons using linear mixed models are made, after controlling for alcohol intake across all doses, there was no statistically significant difference between the absolute levels of serum estrone sulfate at week 4 versus 8 (P = 0.32). However, controlling for alcohol intake absolute DHEAS levels increased between weeks 4 and 8 (P < 0.0001). The increase in DHEAS between weeks 4 and 8 did not occur in just the 15 or 30 g per day groups, however; the increase was similar in each of the three groups. In the 0 g per day group the geometric mean shifted from 55.3 ng/dL at 4 weeks to 59.4 ng/dL at 8 weeks. In the 15 g per day group the geometric mean shifted from 59.7 ng/dL to 62.5 ng/dL, and in the 30 g per day group the geometric mean shifted from 60.4 ng/dL to 63.9 ng/dL. The models did not suggest an interaction between measurement week and alcohol dose in serum estrone sulfate (P = 0.32) or DHEAS (P = 0.58). In postmenopausal women moderate alcohol consumption for four weeks resulted in statistically significant increased levels of serum estrone sulfate, the most abundant circulating estrogen, and DHEAS, the steroid hormone with the highest concentration in the blood. Compared to the placebo, serum estrone sulfate levels increased 6.9% (P = 0.24) and 22.2% (P = 0.0006) respectively among women who consumed 15 g and 30 g alcohol per day for four weeks. Serum DHEAS concentrations also increased by 8.0% (P < 0.0001) and 9.2% (P < 0.0001) respectively among women who consumed 15 g or 30 g alcohol per day for four weeks. At week 8, serum estrone sulfate levels increased 7.5% and 10.7% respectively among women who consumed 15 g and 30 g alcohol per day whereas DHEAS concentrations increased 5.1% and 7.5% respectively among women who consumed 15 g or 30 g alcohol per day compared to the placebo [ 7 ]. The increased levels of serum estrone sulfate concentrations at week 4 were essentially the same as those seen at week 8. The % change in serum estrone sulfate levels from week 4 to week 8 is -0.1%, 0.5%, and -8.9% respectively in the 0 g, 15 g, and 30 g alcohol doses. Estrone sulfate does not show a consistent change over time across all alcohol doses and there is no statistical evidence of an interaction between time and alcohol (P = 0.32). Although there appears to be a difference when looking at the point estimates in the 30 g alcohol dose, variance in estrone sulfate is so large at both 4 weeks and 8 weeks the difference between the two point estimates is not statistically significant. The large variance in estrogen sulfate at the 30 g dose will require more people to accurately judge this possible interaction. For DHEAS, the absolute levels were statistically (P < 0.0001) higher at week 8 compared to week 4 in all the alcohol groups. This increase in DHEAS between weeks 4 and 8 did not differ regardless of 0 g, 15 g, or 30 g per day dose (P = 0.58). Biological reasons may explain the increase in the 15 g and 30 g per day alcohol dose, but the reasons behind the increases in DHEAS for the 0 g dose need further research. We did not find any evidence for an effect modification between measurement week and alcohol supplementation in serum estrone sulfate (P = 0.32) or DHEAS (P = 0.58). These data indicate that the hormonal effects due to moderate consumption of alcohol equivalent to one or two drinks per day are seen early, within 4 weeks of initiation of ingestion. Importantly from a study design perspective, our study also demonstrates that it may be possible to utilize shorter study periods when assessing the effects of alcohol consumption on hormone levels, at least in postmenopausal women. To understand the earliest effects of moderate alcohol intake on hormone levels in postmenopausal women, future studies will have to be designed to assess serum levels earlier than 4 weeks or possibly later than 8 weeks. Postmenopausal women with elevated levels of serum estrone sulfate [ 8 , 9 ] and DHEAS [ 8 - 11 ] levels were reported to be at an increased risk of breast cancer in several prospective cohort studies. Results from our study showing statistically significant increased serum estrone sulfate and DHEAS concentrations after four weeks of supplementation with alcohol equivalent to one or two drinks per day, provide one possible mechanism by which moderate alcohol ingestion could increase breast cancer risk in postmenopausal women. Alcohol has many physiologic effects and could influence breast cancer risk through non-hormonal mechanisms as well. Experimental evidence suggests that alcohol interferes with folate absorption, transport, and metabolism, potentially limiting folate stores in the tissues and may interfere with DNA methylation [ 12 , 13 ]. Alcohol consumption and metabolism can result in increased production of several classes of DNA damaging molecules including reactive oxygen species [ 14 ] which can lead to increase DNA damage and the development of breast cancer [ 15 ]. Conclusions In conclusion, our results provide additional evidence for a mechanism by which moderate alcohol drinking could modify breast cancer risk, indicating that this effect occurs after a short time period (as early as four weeks), and thus provides further support for a causal association. Competing interests None declared.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC517945.xml

545965

Health care restructuring and family physician care for those who died of cancer

Background During the 1990s, health care restructuring in Nova Scotia resulted in downsized hospitals, reduced inpatient length of stay, capped physician incomes and restricted practice locations. Concurrently, the provincial homecare program was redeveloped and out-of-hospital cancer deaths increased from 20% (1992) to 30% (1998). These factors all pointed to a transfer of end-of-life inpatient hospital care to more community-based care. The purpose of this study was to describe the trends in the provision of Family Physician (FP) visits to advanced cancer patients in Nova Scotia (NS) during the years of health care restructuring. Methods Design Secondary multivariate analysis of linked population-based datafiles including the Queen Elizabeth II Health Sciences Centre Oncology Patient Information System (NS Cancer Registry, Vital Statistics), the NS Hospital Admissions/Separations file and the Medical Services Insurance Physician Services database. Setting Nova Scotia, an eastern Canadian province (population: 950,000). Subjects : All patients who died of lung, colorectal, breast or prostate cancer between April 1992 and March 1998 (N = 7,212). Outcome Measures Inpatient and ambulatory FP visits, ambulatory visits by location (office, home, long-term care facility, emergency department), time of day (regular hours, after hours), total length of inpatient hospital stay and number of hospital admissions during the last six months of life. Results In total, 139,641 visits were provided by family physicians: 15% of visits in the office, 10% in the home, 5% in the emergency department (ED), 5% in a long-term-care centre and 64% to hospital inpatients. There was no change in the rate of FP visits received for office, home and long-term care despite the fact that there were 13% fewer hospital admissions, and length of hospital stay declined by 21%. Age-sex adjusted estimates using negative binomial regression indicate a decline in hospital inpatient FP visits over time compared to 1992–93 levels (for 1997–98, adjusted RR = 0.88, 95%CI = 0.81–0.95) and an increase in FP ED visits (for 1997–98, adjusted RR = 1.18, 95%CI = 1.05–1.34). Conclusion Despite hospital downsizing and fewer deaths occurring in hospitals, FP ambulatory visits (except for ED visits) did not rise correspondingly. Although such restructuring resulted in more people dying out of hospital, it does not appear FPs responded by providing more medical care to them in the community.

Background Canadians have become increasingly vocal about the need for improved care for the dying. In response, the Special Senate Committee on Euthanasia and Assisted Suicide[ 1 ] declared that Canadian governments should make end-of-life care or palliative care a top priority in the restructuring of the health care system. As the leading cause of adult deaths, it is estimated that 67,400 Canadians died of cancer in 2003[ 2 ]. The aging of the Canadian population will result in increasing numbers of individuals presenting to the health care system with advanced chronic illnesses such as cancer[ 3 ]. In one Canadian province, Nova Scotia, the number of cancer deaths is projected to increase by 27% from 2,259 in 1999 to 2,870 by 2010[ 4 ]. Although cancer is the leading cause of death in those who receive care from comprehensive palliative care programs, access to such programs is hugely variable[ 5 , 6 ]. As these calls for better care for the dying go out, hospital roles in Canadian communities have been redefined. Hospital restructuring has transferred many aspects of inpatient care to community-based care, including the end-of-life or palliative care of those with cancer. End-of-life care is defined broadly as all care provided to dying persons. Multiple providers are involved in such care and include generalist providers such as family physicians, community nurses and other primary care providers, hospital-based providers, specialists, specialist palliative care providers, volunteers and family. Unless otherwise stated "end-of-life care" reflects any or all of these kinds of care as received by dying patients. When warranted, we will refer to "palliative care program" (PCP) as the comprehensive, organized and specialized program of care for the dying. Such PCPs are often hospital based and include an inpatient palliative care unit, consultant nurses and physicians and may provide consultations and / or care in the community. End-of-life care has been provided in the community in the past, however, this health system restructuring has forced an even greater emphasis on this location for care. Nova Scotia is a small, east coast Canadian province of just under one million people where the percentage of cancer deaths occurring outside of hospitals rose from 19.8% in 1992–93 to 30.2% in 1997–98, an increase of over 50%[ 7 ]. One direct result of this change, we feel, is the need for effective, available, continuous and increasingly complex care of the dying in the community. Options to provide this end-of-life care vary from the coordination and integration of existing community resources including family physicians to specialized palliative care program home support teams providing all the necessary visits. There are no free-standing, community-based hospices available to the dying in Nova Scotia. While health system restructuring has increased community-based care, restructuring has also affected the context in which family physicians provide this primary medical care. In Nova Scotia, substantial health system restructuring occurred during the study years and initially capped physician incomes, restricted practice locations, downsized hospitals, provided new hospital-in-the-home capacity, initiated redevelopment of a provincial home care program and introduced drug co-payments for seniors receiving government drug benefits. The purpose of this study was to describe the trends in the provision of family physician visits to those dying of the four cancers with the highest mortality in Nova Scotia and to whom, we believe, would be most frequently seen by a family physician, during the years concurrent with this health care restructuring. We hypothesized that, given our previous research showing the trends of advanced cancer patients spending more time out of hospital, and fewer dying in-hospital, family physician community-based services to them would increase. Methods Study subjects included all Nova Scotians who died due to lung, colorectal, breast or prostate cancer from April 1, 1992 to March 31, 1998 as indicated on the Vital Statistics death certificate (International Classification of Diseases. 9 th revision [ICD9-CM]). This population-based study involved the secondary data analysis of linked administrative health information. Individual level data were obtained from: (1) the Queen Elizabeth II Health Sciences Centre Oncology Patient Information System (OPIS) which encompasses the provincial Nova Scotia Cancer Registry (NSCR) and includes provincial Vital Statistics information, (2) the Nova Scotia Medical Services Insurance Physician Services (MSIPS), and (3) the provincial Hospital Admissions and Separates file (HAS). The MSIPS includes data pertaining to all visits provided by physicians in the province who are remunerated via fee-for-service schedules and for physicians who are paid alternatively that provide 'shadow' billing information. More than 96% of FPs in Nova Scotia were fee-for-service at the time of this study[ 8 ]. Health services were limited to the 'end-of-life (EOL)', defined in this study as the six months (180 days) prior to the date of death, or from the date of initial cancer diagnosis as recorded in the NSCR to death for persons living less than six months after diagnosis. Service fee codes (for fiscal years 1992–93 to 1995–96) and health service identification numbers (for fiscal years 1996–97 to 1997–98) were used to identify FP visits. Measures All visits provided by a family physician were counted for each patient during their end-of-life including those in the FP office, patient's home, long-term-care facility (LTC), emergency department (ED), and hospital inpatient settings. Ambulatory visits were defined as all visits, including those to the emergency department, but excluding those made to a patient during a hospital inpatient stay or outpatient procedures. Two 'time of care' categories were created: regular hours (8:01 am to 5 pm), and after hours care (5:01 pm to 8 am, weekends, holidays). Because a common code was not available to identify the time of visit to hospitalized patients, all hospital inpatient visits were considered as occurring during regular hours. In addition to the total length of inpatient hospital stay and the number of hospital admissions, we also examined the total number of specialty visits received by these patients. Demographic and clinical variables included sex, age, fiscal year of death, geographic region, time from diagnosis to death, and tumour site (lung, colorectal, breast or prostate). Analysis Initial analyses focused on frequency counts and descriptive measures (central tendency, dispersion) of all FP visits, hospital admissions, length of inpatient stay and specialty visits, overall and by fiscal year of death. Each count was expressed as the number of visits provided per 100 "end-of-life (EOL) person-days". Linear temporal trends were assessed using negative binomial regression with a logarithmic link function linking the dependent variable (for example, FP visits, hospital admissions, inpatient stay) to fiscal year. Fiscal year was included as a linear predictor adjusting for age and sex, and with log (EOL person-days) as an offset variable. Assessment of a nonlinear trend was made by including fiscal year as both a linear predictor and a qualitative predictor. The Type 3 analysis of year as a qualitative predictor in this model relates to the nonlinear component of trend. Adjusted regression coefficients were exponentiated and reported as rate ratios (RR) with associated 95% confidence intervals (CI). All analyses were conducted using SAS software[ 9 ]. Results In total, 7212 Nova Scotians were identified as having died due to lung, colorectal, breast or prostate cancer over the six-year study period. Males comprised a larger proportion of deaths, as did adults aged 65 years and older, those who died due to lung cancer and survivors of at least 150 days from date of initial cancer diagnosis (Table 1 ). The number of deaths across fiscal years remained relatively stable. Table 1 Characteristics of adults who died due to lung, colorectal, breast or prostate cancer in Nova Scotia between April 1, 1992 and March 31, 1998 Characteristic Number of deaths (%) Fiscal year of death 1992/93 1142 (15.8) 1993/94 1162 (16.1) 1994/95 1243 (17.2) 1995/96 1245 (17.3) 1996/97 1241 (17.2) 1997/98 1179 (16.4) Sex Female 3126 (43.3) Male 4086 (56.7) Age group (years) < 65 1740 (24.1) 65–74 2151 (29.8) 75–84 2247 (31.2) 85+ 1074 (14.9) Cancer cause of death Lung 3674 (50.9) Colorectal 1223 (17.0) Breast 1243 (17.2) Prostate 1072 (14.9) Survival time (days) <31 723 (10.0) 31–60 470 (6.5) 61–90 340 (4.7) 91–120 277 (3.8) 121–150 222 (3.1) >150 5180 (71.8) Region of death Halifax regional municipality 2293 (31.9) Cape Breton Island 1153 (16.0) All other regions of Nova Scotia 3751 (52.1) In total, 139,641 visits or a median of 13 visits per patient (mean 19.4; standard deviation [SD] 20.3), were provided by FPs to patients during their end-of-life with 94% of patients receiving at least one FP visit. Variability across fiscal years was minimal, ranging from 92.9% receiving at least one FP visit in 1993–94 to 95.3% in 1997–98. The majority of FP visits were provided to hospital inpatients (64%), followed by the office (15%), home (10%), the emergency department (5%) and long-term care (5%). Of visits provided to hospital inpatients, almost 71% were categorized as a 'subsequent hospital visit', which represents continuing in-hospital care. Other inpatient visits included initial hospital visits (4.7%), visits after four weeks (17.5%), supportive care visits (4.6%) and urgent or emergency care visits (2.5%). Temporal trends associated with ambulatory and inpatient FP visits per 100 EOL person-days are shown in Figure 1 along with the average number of days spent as hospital inpatient stay. Ambulatory visits by service location are illustrated in Figure 2 . After accounting for age, sex and survival time, the total number of FP visits were found to have decreased significantly over the time period (p < 0.0001 declining from 15.3 visits per 100 EOL person-days in 1992–93 to 11.8 visits per 100 EOL person-days in 1996–97 followed by a small increase to 13.6 visits per 100 EOL person-days in 1997–98. This nonlinear trend is primarily due to the decline and then rise in the number of inpatient visits made by FPs over time. A closer examination of these inpatient visits by category (e.g., initial hospital visits, subsequent visits) did not reveal any major shift in the distribution of inpatient visit types over time. In contrast, total ambulatory visits remained relatively stable over the six-year time period with no evident significant time trends. Stratification of ambulatory visits by location of visit indicate a significant linear trend in emergency department visits over time (p < 0.01). After accounting for age and sex, patients in 1997–98 made 18% more emergency department visits than patients in 1992–93 (adjusted RR 1.18; 95% CI 1.05, 1.34). No association was evident across time for visits provided in the office, at home or within a long term care facility. Figure 1 Family physician inpatient and ambulatory visits and length of hospital stay among advanced cancer patients over time Figure 2 Ambulatory family physician visits to cancer patients by location over time Examination of ambulatory FP visits provided during regular hours (8:01 am-5 pm) showed no significant change over time. However, ambulatory visits provided during 'after' hours (5:01 pm-8 am, weekends) were found to have increased significantly over the study period (p < 0.03). Compared to patients in 1992–93, patients in 1997–98 received 12% more ambulatory visits after hours (adjusted RR 1.12; 95% CI 1.01, 1.25). Although this increase represented a significant change over time, after hour visits were relatively few compared to the large number of other ambulatory visits and therefore exerted no impact on the overall temporal effect among all ambulatory visits. Significant temporal trends were evident with respect to both the total number of hospital admissions experienced by the patient and the total number of days they spent as a hospital inpatient. A decline in hospital admissions was seen over time, from 1.2 admissions in 1992–93 to 1.1 admissions per 100 EOL person-days in 1997–98. Compared to 1992–93, patients in all subsequent years experienced fewer hospital admissions after accounting for sex and age. By 1997–98 patients experienced 13% fewer hospital admissions than patients in 1992–93 (adjusted RR 0.87; 95%CI 0.82, 0.93). Over 85% of patients spent at least one day as a hospital inpatient. Patients spent on average a total of 22.7 days in hospital (SD 27.4; median 14 days; range 0–180 days) or 15.6 days per 100 EOL person-days. Total length of hospital inpatient stays declined from 18.6 days per 100 EOL person-days in 1992–93 to 14.8 in 1997–98. Results from the age and sex adjusted regression analysis indicate total length of hospital inpatient stays in 1997–98 were 21% shorter than experienced in 1992–93 (adjusted RR 0.79; 95%CI 0.71, 0.88). In total, 82,575 visits were made to a medical specialty during the end of life. The number of visits ranged from one to 169, with a median of 11 visits per patient (mean 11.4; SD 13.5). Age and sex adjusted regression analysis indicate visits to a medical specialty did not change significantly over time. Discussion Despite the move to a greater percentage of cancer patients dying out of hospital[ 7 ] and despite the findings of this study which show advanced cancer patients are spending more time out of hospital during the end-of-life, we have found no indication of increased family physician involvement in office, home or long-term care settings. There was an increase in FP visits in the emergency department. Many questions follow. By whom and how is the medical component of community-based end-of-life cancer care being provided? We have shown that the number of visits made to a specialist physician has not changed significantly over the same time period. Therefore, is care previously performed by family doctors now being offered by non-physicians? Might these patterns be influenced by changes in the provision of end-of-life care with an increased use of systemic therapies? Are patients receiving adequate and appropriate care at the end-of-life? When we compare these trends to concurrent trends in the province of Nova Scotia for all types of patients, a number of interesting points emerge. For all types of patients in the province from 1992–1999, the total number of office-based, home and long-term-care visits has declined slightly[ 8 ]. This is not true for the patients in our study. FPs may be continuing to see cancer patients in the office despite reduced office visits for other types of patients in an attempt to ensure comprehensive care for those with this serious illness. We expected to see an increase in home visits for those dying of cancer given the longer period people are spending out of hospital and the greater numbers dying at home. This was not the case. Home and long-term care visits among patients remained stable over time. This trend may have been facilitated by a revamped home care system in 1995 providing more nursing and assisted care visits in the home in the latter part of the decade perhaps reducing the need for family physician visiting. Hospital visits declined quite substantially in the early years of our study and then increased in the final year. Provincial information suggests this trend was true for non-cancer patients as well[ 8 ]. Since both length of stay and the number of hospital admissions have declined, it is possible that those admitted in 1997/1998 were those who had been cared for longer in the community but who had reached a critical point where they were sicker than in the previous years and required more visits during these shorter stays. In other words, if patients in hospital had greater severity of illness, greater medical visit intensity may have been required. This may account for the rise in hospital visits in the last study year. The increase in emergency department visits by family physicians parallels a rise during the study period among the general Nova Scotian population. (personal communication, M. Joyce, Department of Finance, Government of Nova Scotia). This greater emergency department utilization may be a reflection of reduced access to inpatient beds experienced by all patients, including those with advanced cancer. The lack of increased visits made by family physicians in end-of-life care in the community is concerning when one considers the evidence that their participation is associated with a greater likelihood of home death [ 10 - 12 ] and less emergency department use[ 13 ]. It is not, however, surprising given the overall decline in comprehensiveness of care by individual family physicians in Canada[ 14 ]. It is important to note that the advantage of using provincial administrative health databases is that we have information for the entire population regarding cancer mortality and physician and hospital utilization. However, there is no clinical information on severity of disease, which would provide a much better understanding of factors influencing health service utilization. In addition, we do not have concurrent home care utilization data or private long-term care facility data to factor into our modeling. Conclusions Despite health care restructuring of the 1990s which resulted in fewer days in-hospital for those dying with cancer in Nova Scotia, there was no concurrent increase in the family physician visits provided to the dying. This may represent a growing unmet need for community-based medical care. However, further research is needed to examine whether end-of-life-care needs are being met by other health providers such as specialized home palliative care nurses, or general home care nursing services. Competing interests The author(s) declare that they have no competing interests. Authors' contributions FB and BL participated in the conceptualisation and design of the project, the analysis and interpretation of the data, created the first draft of the article, and incorporated co-authors' comments into the final draft. GJ participated in the design of the project, the interpretation of data, and revising the manuscript. GF participated in the analysis and interpretation of data and the revising of drafts. All authors gave approval to the final version. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545965.xml

545795

Estimating genomic coexpression networks using first-order conditional independence

A computationally efficient statistical framework for estimating networks of coexpressed genes is presented that exploits first-order conditional independence relationships among gene expression measurements.

Background Analyses of functional genomic data such as gene-expression microarray measurements are subject to what has been called the 'curse of dimensionality'. That is, the number of variables of interest is very large (thousands to tens of thousands of genes), yet we have relatively few observations (typically tens to hundreds of samples) upon which to base our inferences and interpretations. Recognizing this, many investigators studying quantitative genomic data have focused on the use of either classical multivariate techniques for dimensionality reduction and ordination (for example, principal component analysis, singular value decomposition, metric scaling) or on various types of clustering techniques, such as hierarchical clustering [ 1 ], k -means clustering [ 2 ], self-organizing maps [ 3 ] and others. Clustering techniques in particular are based on the idea of assigning either variables (genes or proteins) or objects (such as sample units or treatments) to equivalence classes; the hope is that equivalence classes so generated will correspond to specific biological processes or functions. Clustering techniques have the advantage that they are readily computable and make few assumptions about the generative processes underlying the observed data. However, from a biological perspective, assigning genes or proteins to single clusters may have limitations in that a single gene can be expressed under the action of different transcriptional cascades and a single protein can participate in multiple pathways or processes. Commonly used clustering techniques tend to obscure such information, although approaches such as fuzzy clustering (for example, Höppner et al . [ 4 ]) can allow for multiple memberships. An alternate mode of representation that has been applied to the study of whole-genome datasets is network models. These are typically specified in terms of a graph, G = { V , E }, composed of vertices ( V ; the genes or proteins of interest) and edges ( E ; either undirected or directed, representing some measure of 'interaction' between the vertices). We use the terms 'graph' and 'network' interchangeably throughout this paper. The advantage of network models over common clustering techniques is that they can represent more complex types of relationships among the variables or objects of interest. For example, in distinction to standard hierarchical clustering, in a network model any given gene can have an arbitrary number of 'neighbors' (that is n -ary relationships) allowing for a reasonable description of more complex inter-relationships. While network models seem to be a natural representation tool for describing complex biological interactions, they have a number of disadvantages. Analytical frameworks for estimating networks tend to be complex, and the computation of such models can be quite hard (NP-hard in many cases [ 5 ]). Complex network models for very large datasets can be difficult to visualize; many graph layout problems are themselves NP-hard. Furthermore, because the topology of the networks can be quite complex, it is a challenge to extract or highlight the most 'interesting' features of such networks. Two major classes of network-estimation techniques have been applied to gene-expression data. The simpler approach is based on the notion of estimating a network of interactions by defining an association threshold for the variables of interest; pairwise interactions that rise above the threshold value are considered significant and are represented by edges in the graph, interactions below this threshold are ignored. Measures of association that have been used in this context include Pearson's product-moment correlation [ 6 ] and mutual information [ 7 ]. Whereas network estimation using this approach is computationally straightforward, an important weakness of simple pairwise threshold methods is that they fail to take into account additional information about patterns of interaction that are inherent in multivariate datasets. A more principled set of approaches for estimating co-regulatory networks from gene-expression data are graphical modeling methods, which include Bayesian networks and Gaussian graphical models [ 8 - 11 ]. The common representation that these techniques employ is a graph theoretical framework in which the vertices of the graph represent the set of variables of interest (either observed or latent), and the edges of the graph link pairs of variables that are not conditionally independent. The graphs in such models may be either undirected (Gaussian graphical models) or directed and acyclic (Bayesian networks). The appeal of graphical modeling techniques is that they represent a distribution of interest as the product of a set of simpler distributions taking into account conditional relationships. However, accurately estimating graphical models for genomic datasets is challenging, in terms of both computational complexity and the statistical problems associated with estimating high-order conditional interactions. We have developed an analytical framework, called a first-order conditional independence (FOCI) model, that strikes a balance between these two categories of network estimation. Like graphical modeling techniques, we exploit information about conditional independence relationships - hence our method takes into account higher-order multivariate interactions. Our method differs from standard graphical models because rather than trying to account for conditional interactions of all orders, as in Gaussian graphical models, we focus solely on first-order conditional independence relationships. One advantage of limiting our analysis to first-order conditional interactions is that in doing so we avoid some of the problems of power that we encounter if we try to estimate very high-order conditional interactions. Thus this approach, with the appropriate caveats, can be applied to datasets with moderate sample sizes. A second reason for restricting our attention to first-order conditional relationships is computational complexity. The running time required to calculate conditional correlations increases at least exponentially as the order of interactions increases. The running time for calculating first-order interactions is worst case O ( n 3 ). Therefore, the FOCI model is readily computable even for very large datasets. We demonstrate the biological utility of the FOCI network estimation framework by analyzing a genomic dataset representing microarray gene-expression measurements for approximately 5,000 yeast genes. The output of this analysis is a global network representation of coexpression patterns among genes. By comparing our network model with known metabolic pathways we show that many such pathways are well represented within our genomic network. We also describe an unsupervised algorithm for highlighting potentially interesting subgraphs of coexpression networks and we show that the majority of subgraphs extracted using this approach can be shown to correspond to known biological processes, molecular functions or gene families. Results We used the FOCI network model to estimate a coexpression network for 5,007 yeast open reading frames (ORFs). The data for this analysis are drawn from publicly available microarray measurements of gene expression under a variety of physiological conditions. The FOCI method assumes a linear model of association between variables and computes dependence and independence relationships for pairs of variables up to a first-order (that is, single) conditioning variable. More detailed descriptions of the data and the network estimation algorithm are provided in the Materials and methods section. On the basis of an edge-wise false-positive rate of 0.001 (see Materials and methods), the estimated network for the yeast expression data has 11,450 edges. It is possible for the FOCI network estimation procedure to yield disconnected subgraphs - that is, groups of genes that are related to each other but not connected to any other genes. However, the yeast coexpression network we estimated includes a single giant connected component (GCC, the largest subgraph such that there is a path between every pair of vertices) with 4,686 vertices and 11,416 edges. The next largest connected component includes only four vertices; thus the GCC represents the relationships among the majority of the genes in the genome. In Figure 1 we show a simplification of the FOCI network constructed by retaining the 4,000 strongest edges. We used this edge-thresholding procedure to provide a comprehensible two-dimensional visualization of the graph; all the results discussed below were derived from analyses of the entire GCC of the FOCI network. The mean, median and modal values for vertex degree in the GCC are 4.87, 4 and 2 respectively. That is, each gene shows significant expression relationships to approximately five other genes on average, and the most common form of relationship is to two other genes. Most genes have five or fewer neighbors, but there is a small number of genes (349) with more than 10 neighbors in the FOCI network; the maximum degree in the graph is 28 (Figure 2a ). Thus, approximately 7% of genes show significant expression relationships to a fairly large number of other genes. The connectivity of the FOCI network is not consistent with a power-law distribution (see Additional data file 1 for a log-log plot of this distribution). We estimated the distribution of path distances between pairs of genes (defined as the smallest number of graph edges separating the pair) by randomly choosing 1,000 source vertices in the GCC, and calculating the path distance from each source vertex to every other gene in the network (Figure 2b ). The mean path distance is 6.46 steps, and the median is 6.0 (mode = 7). The maximum path distance is 16 steps. Therefore, in the GCC of the FOCI network, random pairs of genes are typically separated by six or seven edges. Coherence of the FOCI network with known metabolic pathways To assess the biological relevance of our estimated coexpression network we compared the composition of 38 known metabolic pathways (Table 1 ) to our yeast coexpression FOCI network. In a biologically informative network, genes that are involved in the same pathway(s) should be represented as coherent pieces of the larger graph. That is, under the assumption that pathway interactions require co-regulation and coexpression, the genes in a given pathway should be relatively close to each other in the estimated global network. We used a pathway query approach to examine 38 metabolic pathways relative to our FOCI network. For each pathway, we computed a quantity called the 'coherence value' that measures how well the pathway is recovered in a given network model (see Materials and methods). Of the 38 pathways tested, 19 have coherence values that are significant when compared to the distribution of random pathways of the same size ( p < 0.05; see Materials and methods). Most of the pathways of carbohydrate and amino-acid metabolism that we examined are coherently represented in the FOCI network. Of each of the major categories of metabolic pathways listed in Table 1 , only lipid metabolism and metabolism of cofactors and vitamins are not well represented in the FOCI network. The five largest coherent pathways are glycolysis/gluconeogenesis, the TCA cycle, oxidative phosphorylation, purine metabolism and synthesis of N-glycans. Other pathways that are distinctive in our analysis include the glyoxylate cycle (6 of 12 genes in largest coherent subnetwork), valine, leucine, and isoleucine biosynthesis (10 of 15 genes), methionine metabolism (6 of 13 genes), phenylalanine, tyrosine, and tryptophan metabolism (two subnetworks each of 6 genes). Several coherent subsets of the FOCI network generated by these pathway queries are illustrated in the Additional data file 1. Combined analysis of core carbohydrate metabolism In addition to being consistent with individual pathways, a useful network model should capture interactions between pathways. To explore this issue we queried the FOCI network on combined pathways and again measured its coherence. We illustrate one such combined query based on four related pathways involved in carbohydrate metabolism: glycolysis/gluconeogenesis, pyruvate metabolism, the TCA cycle and the glyoxylate cycle. Figure 3 illustrates the largest subgraph extracted in this combined analysis. The combined query results in a subset of the FOCI network that is larger than the sum of the subgraphs estimated separately from individual pathways because it also admits non-query genes that are connected to multiple pathways. The nodes of the graph are colored according to their membership in each of the four pathways as defined by the Kyoto Encyclopedia of Genes and Genomes (KEGG). Many gene products are assigned to multiple pathways. This is particularly evident with respect to the glyoxylate cycle; the only genes uniquely assigned to this pathway are ICL1 (encoding an isocitrate lyase) and ICL2 (a 2-methylisocitrate lyase). In this combined pathway query the TCA cycle, glycolysis/gluconeogenesis, and glyxoylate cycle are each represented primarily by a single two-step connected subgraph (see Materials and methods). Pyruvate metabolism on the other hand, is represented by at least two distinct subgraphs, one including { PCK1 , DAL7 , MDH2 , MLS1 , ACS1 , ACH1 , LPD1 , MDH1 } and the other including { GLO1 , GLO2 , DLD1 , CYB2 }. This second set of genes encodes enzymes that participate in a branch of the pyruvate metabolism pathway that leads to the degradation of methylglyoxal (methylglyoxal → L-lactaldehyde → L-lactate → pyruvate and methylglyoxal → ( R )- S -lactoyl-glutathione → D-lactaldehyde → D-lactate → pyruvate) [ 12 , 13 ]. In the branch of methylglyoxal metabolism that involves S -lactoyl-glutathione, methyglyoxal is condensed with glutathione [ 12 ]. Interestingly, two neighboring non-query genes, GRX1 (a neighbor of GLO2 ) and TTR1 (neighbor of CYB2 ), encode proteins with glutathione transferase activity. The position of FBP1 in the combined query is also interesting. The product of FBP1 is fructose-1,6-bisphosphatase, an enzyme that catalyzes the conversion of beta-d-fructose 1,6-bisphosphate to beta-D-fructose 6-phosphate, a reaction associated with glycolysis. However, in our network it is most closely associated with genes assigned to pyruvate metabolism and the glyoxylate cycle. The neighbors of FBP1 in this query include ICL1 , MLS1 , SFC1 , PCK1 and IDP3 . With the exception of IDP3 , the promoters of all of these genes (including FBP1 ) have at least one upstream activation sequence that can be classified as a carbon source-response element (CSRE), and that responds to the transcriptional activator Cat8p [ 14 ]. This set of genes is expressed under non-fermentative growth conditions in the absence of glucose, conditions characteristic of the diauxic shift [ 15 ]. Considering other genes in the vicinity of FBP1 in the combined pathway query we find that ACS1 , IDP2 , SIP4 , MDH2 , ACH1 and YJL045w have all been shown to have either CSRE-like activation sequences and/or to be at least partially Cat8p dependent [ 14 ]. The association among these Cat8p-activated genes persists when we estimate the FOCI network without including the data of DeRisi et al . [ 15 ], suggesting that this set of interactions is not merely a consequence of the inclusion of data collected from cultures undergoing diauxic shift. The inclusion of a number of other genes in the carbohydrate metabolism subnetwork is consistent with independent evidence from the literature. For example, McCammon et al . [ 16 ] identified YER053c as among the set of genes whose expression levels changed in TCA cycle mutants. Although many of the associations among groups of genes revealed in these subgraphs can be interpreted either in terms of the query pathways used to construct them or with respect to related pathways, a number of association have no obvious biological interpretation. For example, the tail on the left of the graph in Figure 3 , composed of LSC1 , PTR2 , PAD1 , OPT2 , ARO10 and PSP1 has no clear known relationship. Locally distinct subgraphs The analysis of metabolic pathways described above provides a test of the extent to which known pathways are represented in the FOCI graph. That is, we assumed some prior knowledge about network structure of subsets of genes and asked whether our estimated network is coherent vis-à-vis this prior knowledge. Conversely, one might want to find interesting and distinct subgraphs within the FOCI network without the injection of any prior knowledge and ask whether such subgraphs correspond to particular biological processes or functions. To address this second issue we developed an algorithm to compute 'locally distinct subgraphs' of the yeast FOCI coexpression network as detailed in the Materials and methods section. Briefly, this is an unsupervised graph-search algorithm that defines 'interestingness' in terms of local edge topology and the distribution of local edge weights on the graph. The goal of this algorithm is to find connected subgraphs whose edge-weight distribution is distinct from that of the edges that surround the subgraph; thus, these locally distinct subgraphs can be thought of as those vertices and associated edges that 'stand out' from the background of the larger graph as a whole. We constrained the size of the subgraphs to be between seven and 150 genes, and used squared marginal correlation coefficients as the weighting function on the edges of the FOCI graph. We found 32 locally distinct subgraphs, containing a total of 830 genes (Table 2 ). Twenty-four out of the 32 subgraphs have consistent Gene Ontology (GO) annotation terms [ 17 ] with p -values less than 10 -5 (see Materials and methods). This indicates that most locally distinct subgraphs are highly enriched with respect to genes involved in particular biological processes or functions. Members of the 21 largest locally distinct subgraphs are highlighted in Figure 1 . The complete list of subgraphs and the genes assigned to them is given in Additional data file 2. The five largest locally distinct subgraphs have the following primary GO annotations: protein biosynthesis (subgraphs A and B); ribosome biogenesis and assembly (subgraph C); response to stress and carbohydrate metabolism (subgraph K); and sporulation (subgraph N). Several of these subgraphs show very high specificity for genes with particular GO annotations. For example, in subgraphs A and B approximately 97% (32 out of 33) and 95.5% (64 out of 67) of the genes are assigned the GO term 'protein biosynthesis'. Subgraph P is also relatively large and contains many genes with roles in DNA replication and repair. Similarly, 21 of the 34 annotated genes in Subgraph F have a role in protein catabolism. Three medium-sized subgraphs (S, T, U) are strongly associated with the mitotic cell cycle and cytokinesis. Other examples of subgraphs with very clear biological roles are subgraph R (histones) and subgraph Z (genes involved in conjugation and sexual reproduction). Subgraph X contains genes with roles in methionine metabolism or transport. Some locally distinct subgraphs can be further decomposed. For example, subgraph K contains at least two subgroups. One of these is composed primarily of genes encoding chaperone proteins: STI1 , SIS1 , HSC82 , HSP82 , AHA1 , SSA1 , SSA2 , SSA4 , KAR2 , YPR158w , YLR247c . The other group contains genes primarily involved in carbohydrate metabolism. These two subgroups are connected to each other exclusively through HSP42 and HSP104 . Three of the locally distinct subgraphs - Q, W and CC - are composed primarily of genes for which there are no GO biological process annotations. Interestingly, the majority of genes assigned to these three groups are found in subtelomeric regions. These three subgraphs are not themselves directly connected in the FOCI graph, so their regulation is not likely to be simply an instance of a regulation of subtelomeric silencing [ 18 ]. Subgraph Q includes 26 genes, five of which ( YRF1-2 , YRF1-3 , YRF1-4 , YRF1-5 , YRF1-6 ) correspond to ORFs encoding copies of Y'-helicase protein 1 [ 19 ]. Eight additional genes ( YBL113c , YEL077c , YHL050c , YIL177c , YJL225c , YLL066c , YLL067c , YPR204w ) assigned to this subgraph also encode helicases. This helicase subgraph is closely associated with subgraph P, which contains numerous genes involved in DNA replication and repair (see Figure 1 ). Subgraph W contains 10 genes, only one of which is assigned a GO process, function or component term. However, nine of the 10 genes in the subgraph ( PAU1 , PAU2 , PAU4 , PAU5 , PAU6 , YGR294w , YLR046c , YIR041w , YLL064c ) are members of the seripauperin gene family [ 20 ], which are primarily found subtelomerically and which encode cell-wall mannoproteins and may play a role in maintaining cell-wall integrity [ 18 ]. Another example of a subgraph corresponding to a multigene family is subgraph CC, which includes nine subtelomeric ORFs, six of which encode proteins of the COS family. Cos proteins are associated with the nuclear membrane and/or the endoplasmic reticulum and have been implicated in the unfolded protein response [ 21 ]. As a final example, we consider subgraph FF, which is composed of seven ORFs ( YAR010c , YBL005w-A , YJR026w , YJR028w , YML040w , YMR046c , YMR051c ) all of which are parts of Ty elements, encoding structural components of the retrotransposon machinery [ 22 , 23 ]. This set of genes nicely illustrates the fact that delineating locally distinct groups can lead to the discovery of many interesting interactions. There are only six edges among these seven genes in the estimated FOCI graph, and the marginal correlations among the correlation measures of these genes are relatively weak (mean r ~ 0.62). Despite this, the local distribution of edge weights in FOCI graph is such that this group is highlighted as a subgraph of interest. Locally strong subgraphs such as these can also be used as the starting point for further graph search procedures. For example, querying the FOCI network for immediate neighbors of the genes in subgraph FF yields three additional ORFs - YBL101w-A , YBR012w-B , and RAD10 . Both YBL101w-A and YBR012w-B are Ty elements, whereas RAD10 encodes an exonuclease with a role in recombination. Discussion Comparisons with other methods Comparing the performance of different methods for analyzing gene-expression data is a difficult task because there is currently no 'gold standard' to which an investigator can turn to judge the correctness of a particular result. This is further complicated by the fact that different methods employ distinct representations such as trees, graphs or partitions that cannot be simply compared. With these difficulties in mind, we contrast and compare our FOCI method to three popular approaches for gene expression analysis - hierarchical clustering [ 1 ], Bayesian network analysis [ 10 ] and relevance networks [ 7 , 24 , 25 ]. Like the FOCI networks described in this report, both Bayesian networks and relevance networks represent interactions in the form of network models, and can, in principle, capture complex patterns of interaction among variables in the analysis. Relevance networks also share the advantage with FOCI networks that, depending on the scoring function used, they can be estimated efficiently for very large datasets. Comparison with relevance networks Relevance networks are graphs defined by considering one or more scoring functions and a threshold level for every pair of variables of interest. Pairwise scores that rise above the threshold value are considered significant and are represented by edges in the graph; interactions below this threshold are discarded [ 25 ]. As applied to gene-expression microarray data, the scoring functions used most typically have been mutual information [ 7 ] or a measure based on a modified squared sample correlation coefficient [ 24 ]). We estimated a relevance network for the same 5007-gene dataset used to construct the FOCI network. The scoring function employed was with a threshold value of ± 0.5. The resulting relevance network has 13,049 edges and a GCC with 1,543 vertices and 12,907 edges. The next largest connected subgraph of the relevance network has seven vertices and seven edges. There are a very large number of connected subgraphs (3,341) that are composed of pairs or singletons of genes. To compare the performance of the relevance network with the FOCI network we used the pathway query approach described above to test the coherence of the 38 metabolic pathways described previously. Of the 38 metabolic pathways tested, nine have significant coherence values in the relevance network. These coherent pathways include: glycolysis/gluconeogenesis, the TCA cycle, oxidative phosphorylation, ATP synthesis, purine metabolism, pyrimidine metabolism, methionine metabolism, amino sugar metabolism, starch and sucrose metabolism. Two of these pathways - amino sugar metabolism and starch and sucrose metabolism - are not significantly coherent in the FOCI network. However, there are 12 metabolic pathways that are coherent in the FOCI network but not coherent in the relevance network. On balance, the FOCI network model provides a better estimator of known metabolic pathways than does the relevance network approach. Comparison with hierarchical clustering and Bayesian networks To provide a common basis for comparison with hierarchical clustering and Bayesian networks, we explored the dataset of Spellman et al . [ 26 ] which includes 800 yeast genes measured under six distinct experimental conditions (a total of 77 microarrays; this data is a subset of the larger analysis described in this paper). Spellman et al. [ 26 ] analyzed this dataset using hierarchical clustering. Friedman et al. [ 10 ] used their 'sparse candidate' algorithm to estimate a Bayesian network for the same data, treating the expression measurements as discrete values. For comparison with Bayesian network analysis we referenced the interactions highlighted in the paper by Friedman et al . and the website that accompanies their report [ 27 ]. For the purposes of the FOCI analysis we reduced the 800 gene dataset to 741 genes for which there were no more than 10 missing values. We conducted a FOCI analysis on these data using a partial correlation threshold of 0.33. The resulting FOCI network had 1599 edges and a GCC of 700 genes (the 41 other genes are represented by subgraphs of gene pairs or singletons). On the basis of hierarchical clustering analysis of the 800 cell-cycle-regulated genes, Spellman et al. [ 26 ] highlighted eight distinct coexpressed clusters of genes. They showed that most genes in the clusters they identified share common promoter elements, bolstering the case that these clusters indeed correspond to co-regulated sets of genes (see [ 26 ] for description and discussion of these clusters). Applying our algorithm for finding locally distinct subgraphs to the FOCI graph based on these same data (with size constraints min = 7, max = 75) we found 10 locally distinct subgraphs. Seven of these subgraphs correspond to major clusters in the hierarchical cluster analysis (the MCM cluster of Spellman et al. [ 26 ] is not a locally distinct subgraph). At this global level both FOCI analysis and hierarchical clustering give similar results. While the coarse global structure of the FOCI and hierarchical clustering are similar, at the intermediate and local levels the FOCI analysis reveals additional biologically meaningful interactions that are not represented in the clustering analysis. An example of interactions at an intermediate scale involves the clusters referred to as Y' and CLN2 in Spellman et al. [ 26 ] Genes of the CLN2 cluster are involved primarily in DNA replication. The Y' cluster contains genes known to have DNA helicase activity. The topology of the FOCI network indicates that these are relatively distinct subgraphs, but also highlights a number of weak-to-moderate statistical interactions between the Y' and CLN2 genes (and almost no interactions between the Y' genes and any other cluster). Thus the FOCI network estimate provides inference of more subtle functional relationships that cannot be obtained from the clustering family of methods. An example at a more local scale involves the MAT cluster of Spellman et al. [ 26 ] This cluster includes a core set of genes whose products are known to be involved in conjugation and sexual reproduction. In the FOCI network one of the locally distinct subgraphs is almost identical to the MAT cluster, and includes KAR4 , STE3 , LIF1 , FUS1 , SST2 , AGA1 , SAG1 , MFα2 and YKL177W ( MFα1 is not included in the FOCI analysis because there were more than 10 missing values). The FOCI analysis additionally shows that this set of genes is linked to another subgraphs that includes AGA2 , STE2 , MFA1 , MFA2 and GFA3 . This second set of genes are also involved in conjugation, sexual reproduction, and pheromone response. AGA1 and AGA2 form the bridge between these two subgraphs (the proteins encoded by these two genes, Aga1p and Aga2p, are subunits of the cell wall glycoprotein α -agglutinin [ 28 ]). These two sets of genes therefore form a continuous subnetwork in the FOCI analysis, whereas the same genes are dispersed among at least three subclusters in the hierarchical clustering. We interpret the difference as resulting from the fact that the FOCI network can include relatively weak interactions among variables, as long as the variables are not first order conditionally independent. For example, the marginal correlation between AGA1 and AGA2 is only 0.63, between AGA1 and GFA1 is 0.59, and between AGA2 and MFA1 only 0.61. Hierarchical clustering or other analyses based solely on marginal correlations will typically fail to highlight such relatively weak interactions among genes. Because hierarchical clustering constrains relationships to take the form of strict partitions or nested partitions, this type of analysis seems best suited to highlight the overall coarse structure of co-regulatory relationships. The FOCI method, because it admits a more complex set of topological relationships, is well suited to capturing both global and local structure of transcriptional interactions. Graphical models, like the FOCI method, exploit conditional independence relationships to derive a model that can be represented using a graph or network structure. Unlike the FOCI model, general graphical models represent a complete factorization of a multivariate distribution. In the case of Bayesian networks it is also possible to assign directionality to the edges of the network model. However, these advantages come at the cost of complexity - Bayesian networks are costly to compute - and generally this complexity scales exponentially with the number of vertices (genes). The estimation of a FOCI network is computationally much less complex than the estimation of a Bayesian network. Both methods allow for a richer set of potential interactions among genes than does hierarchical clustering. We therefore expect that both methods should be able to highlight biologically interesting interactions, at both local and global scales. Friedman et al. [ 10 ] analyzed the 800-gene dataset of Spellman et al . [ 26 ] and highlighted a number of relationships that are assigned high confidence in their analysis. Relationships that were recovered under both a multinomial and Gaussian model include STE2 - MFA2 , CTS1 - DSE2 ( YHR143w ), OLE1 - FAA4 , KIP3 - MSB1 , SHM2 - GCV2 , DIP5 - ARO9 and SRO4 - YOL007c . All of these relationships, with the exception of SRO5 - YOL007c , are present in the FOCI analysis of the same data. Comparisons of the local topology of each network, based on examining the edge relationships for a number of query genes, suggests that the FOCI and Bayesian networks are broadly similar. There are of course, examples of biologically interpretable interactions that are present in the FOCI analysis but not in the Bayesian network and vice versa. For example, using a multinomial model, Friedman et al. demonstrated an interaction between ASH1 and FAR1 , both of which are known to participate in the mating type switch in yeast. This relationship is absent in the FOCI network. Similarly, the relationship between AGA1 and AGA2 that is highlighted in the FOCI analysis does not appear in the multinomial Bayesian network analysis. Review of FOCI assumptions As with all analytical tools, careful consideration of the assumptions underlying the FOCI network method is necessary to understand the limits of the inferences one can draw. For example, our current framework limits consideration to linear relationships as measured by correlations and partial correlations. These assumptions may be relaxed, allowing for other types of distributions and relationships among variables (for example, monotone and curvilinear relationships), but there is an inevitable trade-off to be made in terms of computational complexity and statistical power. However, as seen in our analysis, many biologically interesting relationships among gene expression measures appear to be approximately linear. Biologically speaking, it is important to keep in mind that the graphs resulting from a FOCI analysis of gene-expression measurements should properly be considered coexpression or co-regulation networks and not genetic regulatory networks per se . While the clusters and patterns of coexpression summarized by the FOCI network may result from particular regulatory dynamics, no causal hypothesis of regulatory interaction is implied by the network. Conclusions Biology demands that the analytical tools we use for functional genomics should be able to capture and represent complex interactions; practical considerations stemming from the magnitude and scope of genomic data require the use of techniques that are computable and relatively efficient. The FOCI framework we have used for representing genomic coexpression patterns in terms of a weighted graph satisfies both these constraints. FOCI networks are readily computable, even for very large datasets. Comparisons with known metabolic pathways show that many key biological interactions are captured by FOCI networks, and the algorithm we provide for finding locally distinct subgraphs provides a mechanism for discovering novel associations based on local graph topology. The subgraphs and patterns of interactions that we are able to demonstrate based on such analyses are strongly consistent with known biological processes and functions, indicating that the FOCI network method is a powerful tool for summarizing biologically meaningful coexpression patterns. Furthermore, the kinds of interactions captured by network analysis are typically more natural than the clustering family of analyses where biased and unstable results can be forced by the algorithm. Secondary analysis based on the network properties also reveal additional subtle structure. For example, our procedure for finding locally distinct subgraphs reveals associated genes whose pairwise interactions may be globally weak but relatively strong compared to their local interactions. While the results reported here focus on the analysis of gene expression measurements, the FOCI approach can be applied to any type of quantitative data making it a generally suitable technique for exploratory analyses of functional genomic data. Materials and methods A statistical/geometrical model for estimating coexpression networks The approach we employ to estimate coexpression networks is based on a general statistical technique we have developed for representing the associations among a large number of variables in terms of a weighted, undirected graph. The technique is based on the consideration of so-called 'first-order' conditional independence relationships among variables, hence we call the graphs that result from such analyses first-order conditional independence, FOCI, networks. The network representation that results from a FOCI analysis also has a dual geometrical interpretation in terms of proximity relationships defined with respect to the geometry of correlations and partial correlations. We outline the statistical and geometrical motivations underlying our approach below. First-order conditional independence networks A FOCI network is a graph, G = { V , E }, where the vertex set, V , represents the variables of interest and the edge set, E , represents interactions among the variables. e ij is an edge in G , if and only if there is no other variable in the analysis, k ( k ≠ k ≠ k ) such that or , where is a modified partial correlation between i and j conditioned on k . takes values in the range -1 ≤ ≤ 1. is approximately zero when i and j are independent conditional on k . is positive when the marginal correlation, ρ ij , and the standard partial correlation, ρ ij . k , agree in sign, and is negative otherwise. Cases where the marginal and conditional correlations are of opposite sign are examples of 'Simpson's paradox', which usually indicates that there is a lurking or confounding effect of the conditioning variable (see [ 29 ] for a general discussion of such relationships). While true biological interactions may sometimes lead to inverted conditional associations, their interpretation can be complicated; therefore in the analysis presented above, we did not connect edges when the relationships became inverted. However, one can also keep such edges for subsequent analysis if there is reasonable functional justification. When such sign-reversed edges are ignored, we will call this the sign-restricted FOCI network. This definition means that variables i and j are connected in the FOCI network if there is no other variable in the analysis for which i and j are conditionally independent or which causes an association reversal. Because we restrict the conditioning set to single variables, these are so called 'first-order' conditional interactions (marginal correlations correspond to zero-order conditional interactions; partial correlations given two conditioning variables are second-order conditional interactions, etc). If i and j are conditionally independent given k we write this as ( i ⊥ j | k ). Using an information theoretic interpretation suggested by Lauritzen [ 9 ], the statement ( i ⊥ j | k ) implies that if we observe the variable k , there is no additional information about i that we gain by also observing j (and vice versa). Because the edges of the FOCI network indicate pairs of variables that are not conditionally independent, one can interpret the FOCI graph as a summary of all the pairwise interactions that can not be 'explained away' by any other single variable in the analysis. Unlike standard graphical models, a FOCI network does not represent a factorization of a multivariate distribution into the product of simpler distributions. However, below we show that a sign-restricted FOCI graph has a unique geometric interpretation in terms of proximity relationships in the multidimensional space that represents the correlations among variables. This geometric interpretation suggests that the FOCI model should be a generally useful approach for exploratory analyses of very high-dimensional datasets. Our FOCI approach is similar to a framework developed by de Campos and Huete [ 30 ] for estimating belief networks. These authors developed an algorithm based on the application of zero- and first-order conditional independence test to learn the 'prior skeleton' of a Bayesian network, followed by a refinement procedure that uses higher-order interactions sparingly. Geometrical model of first-order conditional independence Above we described the FOCI network model in statistical terms. Here we provide a geometrical interpretation of FOCI graphs. We show that a FOCI network is equivalent to a proximity graph of the variables of interest (genes in the current analysis). More specifically, we demonstrate that a sign-restricted FOCI network is a 'Gabriel graph' in the geometric space that represents the relationships among the variables. A Gabriel graph, introduced by Gabriel and Sokal [ 31 ], is a type of proximity graph. Let B ( x , r ) denote an open n -sphere centered at the point x with radius r , and let d ( p , q ) denote the Euclidean distance function. Given a set of points, P = { p 1 p 2 , ..., p n }, in an n -dimensional Euclidean space, ( p i , p j ) is an edge in the Gabriel graph if no other point, p k ( i ≠ k , j ≠ k ) in P falls within the diameter sphere defined by B (( p i = p j )/2, d ( p i , p j )/2). That is, p i and p j are connected in the Gabriel graph if no other point falls within the sphere that has the chord p i , p j as its diameter [ 32 ]. Geometry of marginal and partial correlations and conditional independence One can represent random variables as vectors in the space of the observations (often called object space or subject space [ 33 , 34 ]). In such a representation, a set of mean centered and standardized variables correspond to unit vectors whose heads lie on the surface of an n -dimensional hypersphere (where n is the number of observations). In this representation, the correlation between two random variables, x and y , is given by the cosine of the angle between their vectors. We will refer to this construction as the 'correlational hypersphere'. The partial correlation between x and y given z is equivalent to the cosine of the angle between the residual vectors obtained by projecting x and y onto z . The vectors x , y and z form the vertices, A, B, and C, of a spherical triangle on that hypersphere with associated angles γ , λ , and φ . Then, ρ xy . z = cos( φ ), ρ xz . y = cos( λ ), and ρ yz . x = cos( γ ) [ 35 ]. Given this geometric construction of partial correlations in terms of spherical triangles, conditional independence, defined as ρ xy . z = 0 for the multivariate normal, is obtained when cos( φ ) = 0 (that is, when the φ = π /2). The set of z vectors that satisfy this condition defines a circle (actually a hypersphere of dimension n - 1) on the hypersphere whose diameter is the spherical chord between x and y . If the projection of z onto the hypersphere lies outside of this circle then ρ xy . z is positive, inside the circle ρ xy . z is negative (with ρ xy . z = -1 along the chord between x and y ). The sign-restricted FOCI network construction corresponds to the graph obtained by connecting variables i and j only if no third variable falls within the diameter sphere defined by i and j on the correlational hypersphere, or by the diameter sphere defined by i and - j when r ij < 0 (allowing for deviations due to sampling). This is the same criteria of proximity that defines a Gabriel graph. A FOCI graph is therefore a summary of relative proximity relationships among the variables of interest, defined with respect to the geometry of correlations when restricted to the cases when the partial correlation signs are consistent with the marginal correlations. FOCI network algorithm A simple algorithm for estimating a network based on first-order conditional independence relationships is described below. The results of this algorithm can be represented as a graph where the vertices represent the variables of interest (genes) and the edges represent interactions among variables that show at least first-order conditional dependence. A library of functions for estimating FOCI networks, implemented in the Python programming language, is available from the authors on request. We use vanishing partial correlations [ 8 , 36 ] to test whether pairs of genes are conditionally independent given any other single variable in the analysis. Strictly speaking, if the data are not multivariate normal, then zero partial correlations need not imply conditional independence, but rather conditional uncorrelatedness [ 37 ]. However, regardless of distributional assumptions, zero partial correlations among variates are of interest as long as the relationship between the variables has a strong linear component [ 38 ]. FOCI algorithm 1. Estimate marginal associations. For a set of p variables, indexed by i and j , calculate the p × p correlation matrix, C , where C i , j = corr( i, j ) for all i , j ; i = 1... p , j = 1... p . 2. Construct saturated graph. Construct a p × p adjacency matrix, G . Let G i , j = 1 for all i , j . 3. Prune zero-order independent edges. For each pair of variables, ( i , j ), if C i , j < T crit (or some appropriately chosen function, f ( C i , j ) < T crit ), where T crit is a threshold value for determining marginal/conditional independence (see below), then set G i , j = 0. G defines a marginal independence graph. 4. Estimate first-order relationships. For each pair of variables ( i , j ) in G calculate , the minimum partial correlation between i and j , conditioned on each of the other variables in the analysis taken one at a time. for all k such that i ≠ k and j ≠ k and ( i , k ) and ( j , k ) are both edges in G . is the sample modified partial correlation coefficient as defined in equation (1). 5. Prune first-order independent edges. If < T crit (or f ( ) < T crit then set G i , j = 0. The resulting adjacency matrix G , can be represented as an undirected graph, with p vertices, whose edge set is defined by the non-zero elements in G . The edges of this graph can be represented as either unweighted (all edges having equal weight) or with weights defined by some function of corr( i, j ) or . If we assume multivariate normality we can use Fisher's z-transformation [ 39 ] to normalize the expected distribution of correlation/partial correlations and use standard tables of the normal distribution to define T crit for a given edge-wise false-positive rate. Alternatively, one can define T crit by other methods such as via permutation analysis to define a null distribution for . While the FOCI approach requires that one define a critical threshold for determining conditional independence, this threshold is in theory a function of the sample size and the null distribution of rather than the somewhat fuzzier distinction between 'strong' and 'weak' correlation that most pairwise network estimation approaches require. Estimating the yeast FOCI coexpression network We used the FOCI network algorithm to estimate a coexpression network for the budding yeast, Saccharomyces cerevisae . The data used in our analysis are drawn from publicly available microarray measurements of gene expression described in DeRisi et al. [ 15 ], Chu et al. [ 40 ] and Spellman et al. [ 26 ]. These data represent relative measurements of gene expression taken at different points in the cell cycle in yeast cultures synchronized using a variety of different mechanisms [ 26 ] or in the context of specific physiological process such as diauxic shift [ 15 ] or sporulation [ 40 ]. The data were log 2 -transformed, duplicate and missing data were removed and any ORFs listed as 'dubious' in the Saccharomyces Genome Database as of December 2003 were filtered out. The final dataset consisted of expression measurements for 5,007 ORFs represented by 87 microarrays (see Rifkin et al. [ 41 ] for a full description of the pretreatment of these data). The mean centered data were treated as continuous variables for the purposes of our analysis. Microarray measurements, especially spotted microarrays, are subject to a variety of systematic effects such as those due to dye biases and print-tip effects, and a number of methods have been devised to normalize and correct for such biases [ 42 , 43 ]. However, the data analyzed here include both spotted DNA microarray measurements and expression measurements based on Affymetrix arrays (experiments of Cho et al. [ 44 ] as reported by Spellman et al. [ 26 ]), making it difficult to apply a consistent correction. Another consideration is that the assemblage of experiments considered by Spellman et al. [ 26 ], have been frequently used to illustrate the utility of new analytical methods [ 7 , 10 , 45 ]. To facilitate comparison with previous reports we have chosen to analyze these data without any transformations other than the log-transformation and mean-centering described above. As noted above, zero partial correlations are exactly equivalent to conditional independence only for multivariate normal distributions. However, from the perspective of exploratory analyses, the more important assumption is that the relationships among the gene expression measures are predominantly linear. We tested each of these assumptions as follows. We used a Cramer-von Mises statistic [ 46 ] to test for the normality of each vector of gene expression measurements. Approximately 59% of the univariate distributions of the variables are consistent with normality ( p < 0.05). While a majority of the univariate distributions are approximately normal, a significant proportion of the trivariate distributions are clearly not multivariate normal. As a crude test of linearity for bivariate relationships we calculated linear regressions for 10,000 random pairs of gene expression measures (randomly choosing one of the pair as the dependent variables), and performed runs tests [ 47 ] for randomness of the signs of the residuals from each regression. Significant deviations from non-linearity in the bivariate relationships should manifest themselves as non-random runs of positive or negative residuals. For approximately 95% of the runs tests we can not reject the null hypothesis of randomness in the signs of the residuals ( p < 0.05). We therefore conclude that the assumption of quasi-linearity is valid for a large number of the pairwise relationships. Given these observations, in order to define an appropriate partial correlation threshold, T crit , for these data we considered both permutation tests and false-positive rates based on asymptotic expectations for the distribution of first-order partial correlations (see above). Permutation tests were carried out by independently randomizing the values for each gene expression variable such that each gene had the same mean and variance as its original observation vector, but both the marginal and partial correlations had an expected value of zero. We then sampled 1,000 such randomized variables and examined the distribution of for every pair of variables in this sample. For p ≤ 0.001 the permutation test indicates a value of T crit ~ 0.3. The asymptotic threshold for p ≤ 0.001 based on Fisher's z-transform is T crit ~ 0.3. We used the slightly more conservative value of T crit ~ 0.34. Metabolic pathways We used 38 metabolic pathways as documented in KEGG release 29.0, January 2004 [ 48 , 49 ] to test the biological relevance of the estimated yeast coexpression network. These pathways are listed in Table 1 . In our analysis we only considered metabolic pathways for which more than 10 pathway genes were represented in the gene expression dataset described above. The metabolic pathways we studied are not independent, as there are a number of genes whose products participate in two or more metabolic processes. However, for the purposes of the present analysis we have treated each pathway as independent. Testing the coherence of pathways using pathway queries We used the following method to compare our FOCI network to the metabolic pathways from KEGG. We say that a subset of vertices, H , is two-step connected in the graph G if no vertex in H is more than two edges away from at least one other element of H . Given a set of genes assigned to a pathway (the query genes), we computed the set of two-step connected subgraphs for the query genes in the GCC of our yeast coexpression network. This procedure yields one or more subgraphs that are composed of query (pathway) genes plus non-query genes that are connected to at least two pathway genes. We used two steps as a criterion for our pathway queries because our estimate of the distribution of path distances (Figure 2b ) indicated that more than 99% of gene pairs in our network are separated by a distance greater than two steps. Therefore, two-step connected subgraphs in our coexpression network represent sets of genes which are relatively close to each other with respect to the topology of the graph as a whole. Suppose we have a set of query genes from a known pathway denoted as P = {g 1 , g 2 ,... g k }. We construct the two-step connected graph of the elements of P from our FOCI estimated network denoted as F P ⊃ P . That is, F P is a subgraph from the FOCI network that contains elements of P and its neighbors according to the two-step connected criteria described above. F P may itself be composed of one or more connected components. We define F Pmax as the connected component of F P that has the greatest overlap with P . If the FOCI network was completely coherent with respect to P , than F P should constitute a single connected component (that is, F Pmax = F P ) whose vertex set completely overlaps P (that is, | F p ∩ P | = | P |). For cases in which the query pathway is less than perfectly represented in the estimated network we measure the degree of coherence as | F Pmax ∩ P | / | P |). We refer this ratio the 'coherence value' of the pathway P in the network of interest. However, we note that in a completely connected graph (that is, every vertex is connected to every other vertex), every possible pathway query would be maximally coherent but so would any random set of genes. It is therefore necessary to compare the coherence of a given pathway to the distribution of coherence values for random pathways composed of the same number of genes drawn from the same network. We estimated this distribution by using a randomization procedure in which we used 1,000 replicate random pathways to estimate the distribution of coherence values for pathways of different sizes. In Table 1 , pathways that are significantly more coherent than at least 95% of random pathways are marked with an asterisk. Locally distinct subgraphs of coexpression networks We describe an algorithm for extracting a set of 'locally distinct' subgraphs from an edge-weighted graph. We assume that the edge-weights of the graph are measures of the strength of association between the variables of the interest. We define a locally distinct subgraph as a subgraph in which all edges within the subgraph are stronger than edges that connect subgraph vertices to vertices not within the subgraph. Such subgraphs are 'locally distinct' because they are defined not by an absolute threshold on edge strengths, but rather by a consideration of the local topology of the graph and the distribution of edge weights. We describe an algorithm for finding locally distinct subgraphs below. An algorithm for finding locally distinct subgraphs Let G = { V , E } and w : E → R be an edge-weighted graph where w ( e ) is the edge weight function, and | V | = p and | E | = q . Define an ordering on E , O ( E ) = ( e 1 , e 2 ,..., e q ), such that w ( e i ) ≥ w ( e j ) for all i ≤ j (that is, order the edges from strongest to weakest). Let G ( τ ) = { V , E ( τ )} be a subgraph of G obtained by deleting all edges, e , such that w ( e ) < e τ . G ( τ ) an edge-level graph. Also let denote the k connected components of G ( τ ). Let Ω = C 1 ∪ C 2 ∪ … C n . Define L α , ζ = { l 1 , l 2 ,..., l m } where l i ⊆ Ω, l i ∩ l j = ( i ≠ j ) and α ≤| l i |≤ ζ . That is, L α , ζ is a collection of disjoint subgraphs of G , where every l i is a connected component of some G ( τ ) and the size of l i is between α and ζ . We call the elements of L α , ζ the α , ζ -constrained locally distinct subgraphs of G . We say L α , ζ is optimal if | l i ∪ l j … l m | is maximal and | L α , ζ | is minimal. Our goal is to find the optimal L α , ζ for the graph G given the constraints α and ζ . A simple algorithm for calculating the L α , ζ is as follows: 1. let L ← , i = 0 2. while i ≤ q : 3.  calculate G ( i ) and C i 4.  for in C i : 5.   if : 6.    for l in L : 7.     if : 8. L ← L - { l } 9. 10. i = i + 1 11. L α , ζ ← L The algorithm is straightforward. At each iteration, i , we calculate the connected components of the edge-level graph, G ( i ), and add those components which satisfy the size constraints to the candidate list L . Lines 6-8 of the algorithm serve to eliminate from L any non-maximal components. Biological significance of locally distinct subgraphs We applied the locally distinct subgraph algorithm to our yeast FOCI coexpression network. We used pairwise marginal correlations as the edge-weighting function, and set the size constraints as α = 7, ζ = 150. The subgraph search given these constraints yielded 32 locally distinct subgraphs (see Table 2 and Additional data file 2). For each locally distinct subgraph found we used the SGD Gene Ontology (GO) term finder of the Saccharomyces Genome Database [ 50 , 51 ] to search the set of genes in each subgraph for significant shared GO terms. We excluded from the term finder search any genes for which no biological process or molecular function term was assigned. Table 2 summarizes the primary GO terms assigned to each subgraph and the number of genes labeled with that GO term is shown in parentheses. The p -values in Table 2 indicate the frequency at which one would expect to find the same number of genes assigned to the given GO term in a random assemblage of the same size. Additional data files Additional data are available with the online version of this article. Additional data file 1 provides supplementary figures illustrating the connectivity distribution (on a log-log scale) of the estimated yeast FOCI network and additional examples of coherent subgraphs of the FOCI network generated by querying with known metabolic pathways. Additional data file 2 contains a table detailing each of the 32 locally distinct subgraphs generated from the yeast FOCI network via the unsupervised graph search algorithm described in the text. A listing is provided for each locally distinct subgraphs describing yeast ORFs assigned to that subgraph and the Yeast GO Slim annotations associated with each ORF. Supplementary Material Additional data file 1 Supplementary figures illustrating the connectivity distribution (on a log-log scale) of the estimated yeast FOCI network and additional examples of coherent subgraphs of the FOCI network generated by querying with known metabolic pathways Click here for additional data file Additional data file 2 A table detailing each of the 32 locally distinct subgraphs generated from the yeast FOCI network via the unsupervised graph search algorithm described in the text Click here for additional data file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545795.xml

548299

Extracellular degradation of lipoprotein lipase in rat adipose tissue

Background Recent studies in vivo indicate that short-term regulation of lipoprotein lipase (LPL) in rat adipose tissue is post-translational and occurs by a shift of the lipase protein towards an inactive form under the influence of another gene with short-lived message and product. It has not been possible to reproduce this process with isolated adipocytes suggesting that other cells are needed, and perhaps mediate the regulation. The objective of the present study was, therefore, to explore if explants of adipose tissue could be used for studies of the regulatory process. Results When explants of rat epididymal adipose tissue were incubated, LPL mass and activity decreased rapidly. Mass and activity within adipocytes remained constant for at least six hours, demonstrating that it was the extracellular portion of the enzyme that decreased. Adipocytes isolated from the explants after three or six hours of incubation retained their ability to secrete LPL to the medium. Addition of a cocktail of protease inhibitors to the incubation medium slowed down the decrease of LPL mass. Chloroquine was without effect, indicating that the degradation was not lysosomal. 125 I-labeled LPL added to the medium was degraded to acid soluble products, indicating that the degradation occurred extracellularly. Fragmentation of the labelled lipase occurred in conditioned medium and this process was virtually abolished by two MMP inhibitors. Conclusions The decrease of LPL mass and activity that occurs when explants of rat adipose tissue are incubated is due to proteolysis of extracellular LPL. The adipocytes continue to produce and secrete the enzyme. The effect of inhibitors indicates, but does not prove, that the degradation is mediated by MMPs. It appears that this process is accelerated in the tissue fragments compared to intact tissue.

Background Lipoprotein lipase (LPL) hydrolyzes triglycerides in very low-density lipoproteins and chylomicrons [ 1 ]. Tissue-specific regulation of LPL activity is a major mechanism to distribute lipids among tissues according to the physiological needs [ 2 ]. Current information indicates that in adipose tissue, the regulation is post-translational and occurs by a shift of the lipase protein towards an inactive form under the influence of another gene with short-lived message and product [ 3 ]. This information derives from in vivo experiments. To study the underlying mechanism an in vitro model is urgently needed. Experiments with isolated adipocytes do not seem to bring out the mechanism and the in vivo experiments indicate that it is the extracellular LPL that is the target for the regulation [ 4 ]. We have therefore explored the possibility to use tissue explants and report our experiences in this paper. The results support the view that it is the extracellular enzyme that is being regulated. Unfortunately the preparation of tissue explants seems to trigger a proteolytic response in the tissue. Results LPL activity and mass decreased when explants of rat adipose tissue were incubated In the first set of experiments we incubated explants of rat adipose tissue and followed LPL mass and activity (Figure 1 ). LPL mass decreased by more than 50% in three hours. The decrease then continued so that after six hours only around 20% of the original mass remained. With tissues from fed rats, in which most of the LPL protein is in the catalytically active form, the LPL activity decreased in parallel to LPL mass. In tissues from fasted rats, in which most of the LPL protein is in the catalytically inactive form, the decrease was less steep for LPL activity than for LPL mass. The inset in Figure 1 shows that there was a delay of about two hours before the decrease of LPL mass accelerated. There was no difference between tissue from fed and fasted rats. Figure 1 Changes in LPL mass and activity during incubation of adipose tissue explants. Epididymal adipose tissue was dissected from fed and 24 h-fasted rats, cut into small pieces and rinsed in cold PBS. About 100 mg tissue was then incubated at 37°C for the indicated times as described in the Methods section. LPL mass (triangles, panel A) and activity (circles, panel B) in tissue explants from fed (filled symbols, ● and ▼) or fasted (open symbols, ○ and ▽) rats. Values at the start of the incubations were for LPL mass: 17.4 ± 1.1 and 17.1 ± 2.5 ng/μg DNA and for LPL activity: 6.3 ± 1.8 and 3.5 ± 0.7 mU/μg DNA, in tissues from fed and fasted rats, respectively. The inset shows a separate experiment where incubations were stopped at shorter times. Only LPL mass was followed in that experiment. Some of the data points for fed and fasted rats fall on top of each other. Values are mean ± SEM for five parallel incubations. We tried variations in technique and several different incubation media in experiments such as those in Figure 1 . There was some variation in the absolute values but the results were in principle the same, a rapid decline of LPL mass and activity. The rate at which LPL mass decreased was similar with tissue explants from fed and fasted rats, and LPL activity roughly followed LPL mass in tissues from fed rats. One possible explanation could be that LPL was released from the tissue into the medium. The amounts of LPL mass or activity that appeared in the medium were, however, small (Table 1 ). To test whether lipase might have adsorbed to the plastic dishes these were rinsed out with warm SDS solution, but only small amounts of LPL protein were recovered (Table 1 ). Hence it is clear that there was a loss of LPL mass from the system. Table 1 Changes in LPL mass during incubation. Explants of rat adipose tissue were incubated for three hours as in Figure 1. The tissue explants and the medium were recovered and then the vessel was rinsed out with warm (~80°C) SDS solution. This was then suitably diluted with Triton X-100 to match the composition of the medium used for the ELISA. Mean ± SEM of five parallel incubations. LPL mass (ng/g tissue) Before incubation After incubation Adipose tissue 1882 ± 13 665 ± 11* Culture medium 26 ± 6 Washing solution 12 ± 3 Total 1882 ± 13 702 ± 8 * * P < 0.001 compared with before incubation Another possibility was that the cells lost their ability to produce LPL. Isolated adipocytes, incubated under the same conditions as the tissue explants, released LPL activity (Figure 2 ) and mass (data not shown) to the medium, while cellular activity (Figure 2 ) and mass (data not shown) increased slightly. Total LPL activity in the system increased by about 60% during four hours of incubation. When cycloheximide was added, the release of LPL to the medium was virtually abolished and cellular LPL activity decreased with time. Total LPL activity in the system decreased by almost 70%. This demonstrates that the cells depend on synthesis of new LPL protein to sustain LPL activity and secretion to the medium. Adipocytes isolated from tissue explants that had been incubated for three or six hours as in Figure 1 retained the ability to release LPL to the medium (not shown). Hence, the loss of LPL activity that occurred when tissue explants were incubated was not due to a loss of LPL production within adipocytes. In these experiments we also noted that the release of LPL from adipocytes was similar whether the cells were isolated from fed or fasted rats (not shown). Figure 2 LPL activity in cells and in medium during incubation of isolated adipocytes, and the effect of heparin. Adipocytes (from 180 – 200 g rats) were incubated under the same conditions as used for the tissue explants in Figure 1 without (filled symbols) or with (open symbols) 0.1 mg/ml cycloheximide. ●, ○ – adipocytes; ▼, ▽ – medium. Mean ± SEM for five wells at each time. Is the loss of LPL mass an intra- or extracellular event? Heparin release is often used to assess the LPL activity of tissues and releases mainly extracellular LPL [ 4 ]. Figure 3 shows that in fresh explants of adipose tissue a substantial fraction of tissue LPL could be released by heparin. More LPL was released from tissues of fed rats (Figure 3 ). When the tissue explants were incubated for three hours before the heparin challenge, much less LPL was released and after six hours virtually no LPL was released (Figure 3 ). This suggested that the decrease of LPL affected mainly the extracellular enzyme. To test this hypothesis we isolated adipocytes from tissue explants after three or six hours of incubation (Figure 4 ). The LPL activity and mass in the adipocytes was the same when the cells were isolated from tissue explants that had been incubated for three or six hours as when they were isolated from fresh tissue explants. Hence, it was extracellular LPL (calculated as the difference between tissue total and adipocytes) that accounted for the rapid decrease of tissue LPL during incubation. Figure 3 Changes in the amount of heparin-releasable LPL mass during the incubation. Conditions as in Figure 1 but at the designated times, the explants of adipose tissue were transferred to new medium containing heparin and incubated for a further 45 min. Values are means ± SEM for five parallel incubations. Black bars represent explants from fed rats; grey bars represent explants from fasted rats. Figure 4 LPL activity and mass within and outside the adipocytes. Conditions as in Figure 1 but adipose tissue explants from 180–200 g rats were used to get enough material to isolate adipocytes. At the end of the incubation some of the tissue explants were incubated with collagenase and adipocytes were isolated as described in the methods section. Filled symbols – fed rats; open symbols – fasted rats. Panel A shows LPL mass. ▲, △ – tissue, ▼, ▽ – adipocytes. Panel B shows LPL activity. ■, □ – tissue, ●, ○ – adipocytes. Values are means ± SEM for five parallel incubations. Some of the data points for fed and fasted rats fall on top of each other. Is the LPL protein degraded? These results indicated that the decline of LPL mass occurred through proteolytic cleavage of the extracellular enzyme. To test this hypothesis, we included a cocktail of protease inhibitors in the medium used for incubation of tissue explants. This slowed down the decrease of LPL mass (Figure 5A ). Chloroquine had no effect (Figure 5B ), indicating that the degradation did not occur in lysosomes. Figure 5 Effect of protease inhibitors on the decrease of LPL mass during incubation of tissue explants. Conditions as in Figure 1. A cocktail of protease inhibitors (panel A) or chloroquine (final concentration 150 mM, panel B) were included in the medium of some of the incubations. The tissue was from fed rats. Similar results (not shown) were obtained with tissue from fasted rats. ○ – without protease inhibitor, ● – with protease inhibitor. Mean ± SEM for five parallel incubations. To further study the proteolytic process, 125 I-LPL was added to the incubations. TCA soluble material appeared in the medium demonstrating that proteolytic degradation took place (Figure 6 ). We noted that some of the TCA precipitable material became associated with the tissue explants suggesting, binding and/or uptake of the lipase. To explore if the degradation required that the lipase was taken up into cells in the tissue we incubated the labelled lipase in conditioned medium. Analysis by SDS-PAGE showed that several fragments were formed (Figure 7 ). Addition of either of two non-specific MMP inhibitors, Captopril or GM6001, prevented the degradation almost completely. Figure 6 Fate of 125 I-LPL added to the incubation. Conditions as in Figure 1 but 125 I-LPL was added to the medium. ● – TCA precipitable radioactivity in medium, ▲ – TCA precipitable radioactivity in the tissue explants, ■ – TCA soluble radioactivity in medium. Mean ± SEM for five parallel incubations. Figure 7 Analysis by SDS-PAGE of the cleavage of 125 I-LPL in conditioned medium and the effect of an MMP inhibitor Explants of adipose tissue from fed rats were incubated as in Figure 1 for four hours and the medium was collected. 125 I-LPL was then incubated at 4°C (to minimize the risk of conformational changes) in this medium with or without the MMP inhibitor GM6001 (5 μg/ml). Lane 1 – fresh medium, lane 2 – conditioned medium, lane 3 – conditioned medium + inhibitor. The lower band in lane 1 is a proteolytic fragment of LPL that is always present in preparations of the enzyme from bovine milk [23]. Discussion The objective for this study was to find an in vitro system to study the mechanism for down-regulation of LPL activity in rat adipose tissue that occurs on food deprivation. Isolated adipocytes have been tried in several laboratories [ 5 - 9 ], but the differences with nutritional state are rather small. This is true whether one measures LPL within the cells or the rates at which the cells secrete LPL to the medium. These observations are repeated here. The lack of difference within adipocytes indicates that other cell types are needed and may in fact be responsible for the pronounced down-regulation of extracellular LPL activity that occurs on food deprivation [ 3 , 4 ]. We therefore tried to use tissue explants, which have proved valuable in other studies of adipose tissue [ 10 ]. Our results show that degradation of the enzyme was a major process when explants were incubated. The degradation occurred extracellularly; LPL mass and activity in the adipocytes did not change during incubation for up to six hours. Added 125 I-LPL was degraded and this was prevented by addition of MMP inhibitors to the medium. There must be damage of cells when the tissue is cut into small pieces, and there is probably some degree of hypoxia during incubation of the pieces. Cultured explants have, however, been widely and successfully used to explore various aspects of adipose tissue biology ([ 10 ] and references therein). In preliminary experiments we found that the explants retained their ability to take up glucose, to synthesize proteins and to secrete leptin. Adipocytes isolated from the explants after several hours of incubation had the same LPL activity as adipocytes from fresh tissue (Figure 4 ). This indicates that the cells produced LPL at an essentially unchanged rate, since the LPL activity in adipocytes decreased rapidly when protein synthesis was inhibited by cycloheximide (Figure 2 ). Hence, the rapid decrease of LPL when explants were incubated was not due to a general loss of functionality, but reflected specific processes leading to degradation of extracellular LPL. Our results are in line with observations made already in the sixties. There was evidence from chromatographic separations for at least two different forms of LPL in rat adipose tissue [ 11 ]. Cunnigham and Robinson found that incubation of fat pads from fed rats resulted in a rapid loss of LPL activity until a low activity, stable to prolonged incubation, was attained [ 12 ]. In contrast, the LPL activity of isolated fat cells was stable to prolonged incubation. The concept of stable and unstable forms of the lipase can now be interpreted as a reflection of extracellular lipase that is exposed to proteolysis, and intracellular lipase that is protected from the extracellular proteases. The degradation of LPL in conditioned medium was almost completely abolished by the two non-specific MMP inhibitors that we tested. We have not characterized the proteolytic activity further but note that adipose tissue produces at least two MMPs, 2 and 9 [ 13 ]. The loss of LPL mass during incubation of tissue explants was relatively slow during the first two hours and then accelerated. It is likely that the tissue trauma and/or the loss of blood circulation triggered an activation of the MMP system. It has been shown that primary culture of human adipose tissue explants dramatically alters adipocyte gene expression [ 14 ]. It is of interest to note that LPL activity does not decrease during perfusion of fat pads [ 15 ], whereas it does decrease when whole fat pads are cut out and incubated in vitro [ 12 ]. Two pathways for turnover of adipose tissue LPL have been demonstrated so far. One is dissociation of the lipase, perhaps after loss of catalytic activity, into the blood and degradation in the liver [ 16 ]. Release of LPL into blood from adipose tissue has been directly demonstrated by measurement of arterio-venous difference in man [ 17 ]. This pathway can, however, not operate in tissue explants. Another pathway, demonstrated with cultured fat cells, is endocytosis and degradation in lysosomes [ 18 ]. This pathway did not seem to contribute significantly in the present system since the rate at which LPL mass decreased was not affected by chloroquine. The present findings suggest a third pathway, extracellular proteolysis in the tissue. Conclusions The rapid decrease of LPL that occurs when adipose tissue explants are incubated engages only the extracellular enzyme. The adipocytes continue to produce and secrete the enzyme and intracellular LPL remains essentially constant for at least six hours. The decrease in extracellular LPL is due to proteolytic cleavage/degradation of both active and inactive forms of the enzyme. The effects of inhibitors indicate, but do not prove, that the degradation is mediated by MMPs. It appears that this process is accelerated in the tissue fragments compared to intact tissue. Methods Animals Male Sprague-Dawley rats were from Möllegaard Breeding Center (Ejby, Denmark). Unless otherwise stated, the rats were 23 days old and weighed around 60 g. After transport to Umeå they were allowed to acclimatize for seven to ten days by which time they had reached a weight of approximately 120 g. The rats were kept in a well ventilated, temperature (21°C) and humidity (40–45%) controlled room with free access to a standard laboratory chow (Laktamin AB, Stockholm) and tap water. The light in the room was on between 6 a.m. and 6 p.m. In experiments where the rats were to be fasted, food was withdrawn from the cages at 6 a.m. and a grid was placed at the bottom of the cages to prevent coprophagia. The adipose depot used in all experiments was the periepididymal one. The rats were killed by decapitation. Animal experiments were approved by the animal ethics committee in Umeå. Materials Cycloheximide, bovine serum albumin (BSA), the MMP inhibitors GM6001 (Galardin) and Captopril, chloroquine and collagenase were from SIGMA (St. Louis, MO, USA). Protease inhibitor cocktail tablets "Complete Mini" were from Roche Diagnostics, Mannheim, Germany. Heparin was from Lövens (Malmö, Sweden). Substrate for the LPL activity assay was 3 H-labelled triolein in Intralipid (10%) kindly prepared by Pharmacia-UpJohn (Stockholm, Sweden). Parker medium (Parker 199) was from SBL (Stockholm, Sweden). 125 I-LPL was prepared as before [ 19 ]. All other reagents were of the highest commercial grade possible. Assays LPL was extracted from tissues by homogenization in a Tris-HCL buffer (pH 8.2) containing detergents and protease inhibitors as described [ 20 ]. The homogenate was centrifuged for 15 min at 3000 rpm after which the intermediate phase (between the floating fat droplets and the pellet) was used for assay of LPL activity and mass. In most cases the extract was kept on ice and assayed within a few hours. Under these conditions LPL activity is stable. In some cases the extracts were frozen and kept at -70°C for later assay. LPL activity was measured as described previously [ 20 ]. Briefly, two μl of tissue homogenate (triplicate samples) was incubated for 60 min at 25°C with substrate in the presence of ten μl heat-inactivated serum from fasted rats (as source of apolipoprotein CII) and 6% BSA. The total volume was 200 μl. After termination of lipolysis by addition of organic solvents, the fatty acids were extracted and counted for radioactivity. One mU of lipase activity represents one nmol of fatty acids released per minute. LPL mass was measured with an ELISA as described [ 20 ]. The chicken antibodies used recognize both active and inactive forms of the lipase [ 21 ]. Briefly, three different dilutions of tissue homogenate were incubated in microtiter plate wells previously coated with affinity-purified chicken anti-LPL IgG. Detection was mediated via the 5D2 monoclonal antibody (a kind gift by Dr John Brunzell, University of Washington, Seattle) followed by a peroxidase conjugated anti-mouse IgG antibody. Absorbance at 490 nm was measured in a Spectramax microplate spectrophotometer (Molecular Devices, Sunnyvale, CA, USA). DNA content was assayed using Labarca's method [ 22 ]. In vitro incubation of adipose tissue Epididymal adipose tissue was dissected out from fed or 24 h fasted rats. The tissue was cut into small pieces (5 mg or less). A total of about maximal 100 mg tissue pieces were immediately put into culture plates. Each well contained 15 ml of Parker Medium 199 supplemented with 2% BSA, 0.5% casein hydrolysate, 10 mM glucose and adjusted to pH 7.4. Incubations were at 37°C and 5 % CO 2 : 95 % O 2 with continuous gentle shaking motion in a Cellstar Incubator (Queue Systems, Asheville, Canada). After incubation, the tissues were prepared for measurement of LPL activity and mass as described above. In some experiments samples of the medium were taken for assay of LPL activity and/or mass. In some experiments adipocytes were prepared after collagenase treatment of the tissue pieces as described [ 4 ]. To measure heparin releasable LPL (HR-LPL), adipose tissue explants were incubated with heparin (final concentration was 50 IU/ml) for 45 min at 37°C. Statistics Student's t-test was used for analysis of the data. Authors' contributions GW carried out the experiments and participated in their design and in writing of the manuscript, TO participated in the design of the study and drafted the manuscript. GO conceived of the study and coordinated the work. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548299.xml

434150

DNA Display III. Solid-Phase Organic Synthesis on Unprotected DNA

DNA-directed synthesis represents a powerful new tool for molecular discovery. Its ultimate utility, however, hinges upon the diversity of chemical reactions that can be executed in the presence of unprotected DNA. We present a solid-phase reaction format that makes possible the use of standard organic reaction conditions and common reagents to facilitate chemical transformations on unprotected DNA supports. We demonstrate the feasibility of this strategy by comprehensively adapting solid-phase 9-fluorenylmethyoxycarbonyl–based peptide synthesis to be DNA-compatible, and we describe a set of tools for the adaptation of other chemistries. Efficient peptide coupling to DNA was observed for all 33 amino acids tested, and polypeptides as long as 12 amino acids were synthesized on DNA supports. Beyond the direct implications for synthesis of peptide–DNA conjugates, the methods described offer a general strategy for organic synthesis on unprotected DNA. Their employment can facilitate the generation of chemically diverse DNA-encoded molecular populations amenable to in vitro evolution and genetic manipulation.

Introduction A number of strategies have been proposed recently to enable the in vitro selection and evolution of chemical libraries ( Harbury and Halpin 2000 ; Gartner and Liu 2001 ). These new approaches to molecular discovery rely upon DNA-directed synthesis, whereby a physical linkage is established between DNA “genes” and respective nonbiological synthetic “gene products.” Such encoded syntheses proceed through a repeated series of two associated steps: reading of sequence information in the DNA and execution of an appropriate chemical transformation. A fundamental obstacle in all cases is covalently constructing a synthetic entity in the presence of unprotected DNA. One approach takes advantage of hybridization to induce proximity between reactants covalently attached to oligonucleotides. “Reading” is accomplished by the hybridization of the reactant conjugates to a DNA template, whereas synthetic execution results from the reactants being positioned closely together. The strategy has been demonstrated for several types of chemistries ( Orgel 1995 ; Bruick et al. 1996 ; Xu et al. 2001 ; Gartner et al. 2002a ; Li et al. 2002 ). However, these proximity methods are necessarily limited to reaction conditions compatible with DNA hybridization, precluding a large number of potential chemical transformations. Moreover, only synthetic reagents that have been precoupled to DNA can be used. Rather than tailoring reactions to the narrow window of hybridization conditions, DNA reading and chemical transformation can be carried out in chronologically distinct steps ( Halpin and Harbury 2004b ). The DNA is first physically partitioned into subpools by hybridization, accomplishing the reading step. An appropriate reaction is then carried out on each physically separate subpool. As such, the chemical process can take place under DNA-denaturing conditions, permitting the use of organic solvents, high pH, and elevated temperature. Although DNA exhibits limited solubility in nonaqueous solvents, its immobilization on a solid phase can be exploited to access such environments. With a solid-phase chemistry approach, a large existing body of known chemical transformations becomes accessible to DNA-encoded synthesis. Solid-phase approaches also facilitate rapid, efficient, small-scale (nanomole) syntheses. Reagents can be used in vast excess, postreaction work-up involves only washing of the solid phase, and solution-phase manipulation steps that lead to material losses are avoided. However, the conventional attachment of the small molecule to a solid phase through an irreversibly cleavable linker is not adequate, because DNA must repeatedly come on and off the solid phase during reading steps. Here we report a detailed strategy for carrying out solid-phase organic chemistry on unprotected DNA that is suitable for encoded library synthesis by a partitioning approach. Furthermore, we demonstrate a number of tools for adapting new chemistries, and we use those tools to develop comprehensive methods for peptide synthesis on unprotected DNA. Results Solid-Phase Synthesis Resin We first had to choose a solid-phase material that exhibited several critical properties: reversible, efficient binding and release of unprotected DNA; robust solvent integrity; and resistance to chemical modification. The first requirement narrowed our focus to resins that noncovalently bind DNA. A number of resins were tested and excluded due to poor bind–release properties (diethylaminoethyl [DEAE] silica, Macro-Prep ceramic hydroxyapatite, and quaternary amine anion exchange resins). Others exhibited extensive compression in organic solvent (Sephacryl S-1000, macroporous methacrylate) or poor reswelling during organic to aqueous solvent transitions (Poros 50 HQ). Reverse-phase resins were excluded because they would presumably not retain DNA in many organic solvents. DEAE Sepharose, a tertiary amine anion exchange resin, exhibited excellent bind and release properties. By a high performance liquid chromatography (HPLC) assay, oligonucleotides were immobilized and eluted quantitatively in small volumes ( Figure 1 ). Single-stranded DNA (ssDNA) molecules as long as 340 bases were also bound and eluted with high efficiency (data not shown). No severe resin compression was observed using a number of solvents, including H 2 O, methanol (MeOH), dimethyl sulfoxide, N , N -dimethylformamide (DMF), ethyl acetate, and dichloromethane. Lastly, Sepharose has been used previously as a material for solid-phase synthesis ( Tegge and Frank 1997 ; Nakaie et al. 2003 ). All subsequent work was carried out with DEAE Sepharose. Figure 1 Peptide Coupling to DNA Supports (A) Fmoc-based peptide coupling reaction to an aminated 20-base oligonucleotide (NC20) where X represents a succinimidyl or EDC/HOAt-activated ester. (B) HPLC chromatograms of a nonaminated 10-base (10mer) and an aminated 20-base (NC20) oligonucleotide. HPLC traces show DEAE column load (solid black) and elute (broken black). DEAE column elutes after succinimidyl ester–mediated (solid red) or EDC/HOAt-mediated (broken red) Fmoc-Leu coupling and Fmoc deprotection are shown. The resulting amino acid–DNA conjugate is denoted (Leu-NC20). (C) Chemical transformations are carried out using small DEAE Sepharose columns and syringes (left). Washes are facilitated by a vacuum manifold with chemically resistant stopcocks (right). Peptide Chemistry We studied 9-fluorenylmethyoxycarbonyl (Fmoc)–based peptide synthesis because it has a well-established solid-phase precedent and offers a challenge in diverse chemical functionality ( Figure 1 A). As a DNA support, we chose a 20-base oligonucleotide modified with a 5′ primary amine (NC20; Figure 2 ) so that coupling reactions could be readily assayed by HPLC. NC20 and an unmodified 10-base oligonucleotide control (10mer) were bound to a DEAE Sepharose column, washed with DMF, and incubated with an Fmoc–amino acid succinimidyl ester solution in a closed system (see Figure 1 C). A solution of 20% piperidine in DMF was then used to remove the Fmoc group. The DNA was eluted and compared to starting material via an HPLC mobility assay ( Figure 1 B). Coupling and Fmoc-deprotection proceeded with high efficiency (95% or greater) in 30 min. Furthermore, the internal control oligonucleotide was unmodified and fully recovered, ruling out nonspecific DNA modification. Further experiments indicated that addition of the first amino acid proceeds more efficiently than subsequent additions. We therefore optimized coupling for the addition of a second amino acid to an oligonucleotide already acylated with leucine (Leu-NC20). Nearly quantitative coupling was observed for all amino acids, and all conjugates were verified by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) ( Table 1 ). However, the succinimidyl esters did not efficiently acylate proline residues. Figure 2 DNA Support Structure and Modified Amino Acids (A) Peptide synthesis is carried out on DNA modified with a 5′ C12 amino (NC20) or a 5′ PEG amino (NP20) linker. (B) Fluorescent lysine derivative (compound 1, Fmoc-Lys[coumarin]-OH) and BME/DBU labile protecting groups for lysine (compound 2, Fmoc-Lys[Ns]-OH), arginine (compound 3, Fmoc-Arg[Ns]-OH), and histidine (compound 4, Fmoc-His[CNP]-OH). Table 1 Amino Acid Coupling Efficiencies to Leu-NC20 Dde, 1-(4,4-dimethyl-2,6-dioxocyclohex-1-ylidene)ethyl; NorLeu, norleucine Footnotes denote deviations from standard coupling conditions: (a) 4 × 5 min, (b) 80% DMF in MeOH, (c) 40% DMF in MeOH, (d) 20% DMF in MeOH and 50 mM HOAt, (e) 20% DMF in MeOH, and (f) 100% DMF and 50 mM HOAt Efficiencies report the fraction of recovered DNA that was converted to product, and are indicated to the nearest 5%. Column headings “Succinimidyl Ester” and “In Situ Activation” indicate the method of activation. MALDI-MS denotes observed and calculated masses of reaction products. Reactions were performed with 100 pmol of Leu-NC20 starting material. Quantitative coupling is reported as ≥ 95%. Conjugates with N-terminal His(Trt) were not detectable by MALDI-MS, but acetylation of the N-terminus gave accurate mass measurements (data not shown) We next explored the possibility of in situ activation for peptide coupling. In a first pass, 1-(3-[dimethylamino]propyl)-3-ethylcarbodiimide hydrochloride (EDC) out-performed other activating reagents examined; other reagents gave poor coupling yields (benzotriazole-1-yl-oxy-tris-pyrrolidino-phosphonium hexafluorophosphate [PyBOP]), resulted in the formation of undesired side products (2-[1H-benzotriazole-1-yl]-1,1,3,3-tetramethyluronium tetrafluoroborate [TBTU]), or led to poor recovery of DNA (dicyclohexylcarbodiimide [DCC] and diisopropylcarbodiimide [DIC]). Thirty-minute EDC coupling reactions were typically less than 50% efficient without the addition of acylation catalysts such as N -hydroxysuccinimide, 1-hydroxy-7-azabenzotriazole (HOAt), or N -hydroxybenzotriazole ( Nozaki 1997 ). Of these, HOAt was superior, bringing coupling efficiencies above 90%. A number of reaction solvents were examined, including H 2 O, DMF, MeOH, isopropanol, dichloromethane, and mixtures thereof. In general, MeOH gave the best results, with DMF only slightly worse, followed by isopropanol then H 2 O. For most amino acids, exceptional coupling was achieved with 30-min coupling times using 50 mM Fmoc–amino acid-OH, 50 mM EDC, and 5 mM HOAt (see Figure 1 B; Table 1 , Figure S1 ). Importantly, amino acids activated in situ with EDC acylated proline efficiently. We observed equally efficient coupling using a more hydrophilic polyethylene glycol (PEG) linker (NP20; see Figure 2 ), which might be better suited for biological applications and in vitro selections ( Halpin and Harbury 2004b ). To examine whether the coupling conditions generalized to longer DNA fragments, we used an aminated 340-base ssDNA as the support. After coupling, the eluted amino acid–DNA conjugates were digested with nuclease P1, a 3′-to-5′ exonuclease that cleaves all but the 5′ phosphodiester bond of our ssDNA constructs. The 5′-terminal nucleotide, which maintains the linker and synthetic peptide product, was separated from the other nucleoside monophosphates by HPLC and verified by electrospray ionization mass spectrometry. Amino acid coupling to the 340-base ssDNA proceeded with efficiencies comparable to those observed with oligonucleotides (data not shown). In some cases, the sensitivity of this HPLC assay was increased using fluorescence detection. For these experiments, we synthesized a fluorescent lysine derivative, Fmoc-Lys(coumarin)-OH (see Figure 2 B, compound 1), that was incorporated as the C-terminal amino acid. The fluorescence signal allowed sensitive monitoring of subnanomolar quantities of product. By mass spectrometry and HPLC criteria, the peptide synthesis procedures did not damage oligonucleotides. As a more stringent test for possible DNA damage, we synthesized a pentapeptide on a 340-base ssDNA support and examined the ability of the conjugate to act as a template for primer extension. Polyacrylamide gel electrophoresis analysis of the radiolabeled extension products showed no truncated fragments ( Figure 3 ), providing further evidence that the synthetic procedures are DNA-compatible. Figure 3 Peptide–DNA Conjugate As Template for DNA Synthesis 5′ PEG amino-modified 340-base ssDNA was loaded onto two DEAE Sepharose columns. The pentapeptide [Leu]enkephalin was synthesized on one column using EDC/HOAt and Fmoc–amino acids. The DNA was eluted, desalted, and used as template for radiolabeled primer extension reactions. Denaturing polyacrylamide gel electrophoresis analysis of reaction products shows no difference between ssDNA (control) and [Leu]enkephalin–ssDNA ([Leu]enk) templates. Side Chain Protection Typically, acid-labile groups are used to protect reactive amino acid side chains in Fmoc peptide synthesis. Given the instability of DNA in strong acids, it was necessary to identify alternate protecting group strategies. In all cases, the protecting groups were required to be stable during peptide coupling procedures and removed under alternate DNA-compatible conditions. The carboxylic acid side chains of aspartic and glutamic acid are usually protected as esters during Fmoc peptide synthesis. Sterically bulky esters are required to suppress piperidine-induced imide formation, which leads to undesired side chain peptide bonds. We discovered that the conventional tert -butyl ( t Bu) esters, which are normally cleaved with trifluoroacetic acid, could be removed at pH 6.5 in an aqueous solution at 70 °C. This gentle condition offers a convenient approach for acid deprotection. To verify that the thermolytic tert -butyl ester deprotection did not proceed through intramolecular imide formation, we coupled either Fmoc-Asp( t Bu)-OH or Fmoc-Asp-O t Bu to Leu-NC20, followed by Fmoc-Phe-OH. The main and side chain isomers of these tripeptide–DNA conjugates were resolved by HPLC after removal of the t -butyl group. Interconversion of the isomers during deprotection was undetectable (less than 5%). When the experiment was repeated using a 10 mM NaOH solution for tert -butyl ester deprotection (where imide formation would be expected), interconversion of the side and main chain isomers was observed. These results indicate that the thermolytic deprotection maintains the regiochemistry of the initial peptide bonds. At this time, we have little other data that speak to the mechanism of deprotection. Protection of the primary amine side chain of lysine prevents the formation of branched peptides. 2-Nitrobenzenesulfonamide (nosylamide) protection was particularly attractive as a means for lysine protection because the protecting group is base-stable and removed under conditions known to be DNA-compatible ( Fukuyama et al. 1995 ). The lysine side chain was nosylated in good yield to produce Fmoc-Lys(Ns)-OH (see Figure 2 B, compound 2). The nosyl (Ns) group was removed quantitatively from Lys(Ns)-containing peptide–DNA conjugates by β-mercaptoethanol (BME)/1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in a DMF solution at 60 °C. Arginine does not absolutely require side chain protection ( Table 1 ). However, we observed a marked decrease in coupling efficiency as multiple unprotected arginine residues were added at adjacent positions in a synthetic peptide (first Arg approximately 95%; second Arg approximately 50%; third Arg less than 10%). These difficult transformations were accomplished in high yield by repeated high-temperature couplings (4 × 30 min, 37 °C). However, protection of the guanidino group of arginine would offer a more general solution. Given that the conventional arginine protecting groups are sulfonamides, we speculated that nosylamide protection could be applied. We synthesized Fmoc-Arg(Ns)-OH (see Figure 2 B, compound 3) from Boc-Arg-OH and found that it coupled well ( Table 1 ) and was quantitatively deprotected under the BME/DBU conditions. Side chain protection of histidine is essential to prevent acyl-transfer reactions and to suppress L,D-racemization. We found that the trityl (Trt) group of His(Trt) is rapidly removed under the thermolytic conditions used to deprotect tert -butyl esters. However, the trityl group is not ideal in aqueous conditions because of its hydrophobicity and extreme lability at high temperatures. To offer a more robust solution, we sought an H 2 O-compatible histidine protecting group. The 2,4-dinitrophenyl group, widely used in Boc peptide synthesis, is more hydrophilic than trityl and is removed under the nosyl-deprotection conditions. Unfortunately, 2,4-dinitrophenyl is not stable to the piperidine used for Fmoc removal ( Garay et al. 1997 ). After testing a number of nitro-phenyl derivatives, we found that the 4-cyano-2-nitro-phenyl (CNP) group exhibits the appropriate reactivity. We synthesized Fmoc-His(CNP)-OH (see Figure 2 B, compound 4) in one step from Fmoc-His-OH. The CNP group is stable to 20% piperidine in DMF and is deprotected fully with BME/DBU. Histidine racemization is a well-recognized concern in peptide synthesis. To assay the extent of racemization occurring during histidine coupling, we synthesized the oligonucleotide–dipeptide conjugate His(CNP)-Ala-NC20 using either L- or D-Fmoc-His(CNP)-OH. The diasteriomeric dipeptide–oligonucleotide products were resolvable by reverse-phase HPLC. Neither L-His nor D-His coupling resulted in detectable racemization. The experiment was repeated using L- and D-Fmoc-His(Trt)-OH with the same result. Fmoc-Cys(S t Bu)-OH was employed as the protected form of cysteine. The tert -butyl thioether coupled efficiently ( Table 1 ) and was deprotected under the same conditions used to deprotect lysine, arginine, and histidine (CNP). Although we were unable to recover free thiol-containing peptides from our HPLC system for unknown reasons, we could alkylate the deprotected thiol side chain and recover the thioether-containing peptide–oligonucleotide conjugates. Polypeptide Synthesis To characterize multistep syntheses ( Gartner et al. 2002b ), we prepared a number of peptide–DNA conjugates, ranging from two to 12 amino acids in length and varying in sequence and degree of side chain protection required ( Protocol S1 ). In all cases, absolute yields of conjugates with n amino acids exceeded (0.9) n ( Figure S2 ). The HPLC-purified conjugates were analyzed by Edman degradation. Peptide sequencing data were unambiguous and in full agreement with the intended synthetic peptide sequences ( Table 2 ), ruling out side chain modifications that could not be detected by MALDI-MS. Table 2 Sequenced Peptide–DNA Conjugates Peptides are written with the N-terminus to the left. N/A, not applicable Beyond the physical and chemical characterization of the peptide–oligonucleotide conjugates, it was important to examine their behavior in a biochemical setting. Thus, the [Leu]enkephalin pentapeptide was synthesized on an aminated 340-base ssDNA support, which was subsequently converted to duplex form. The conjugate exhibited a peptide-dependent electrophoretic mobility gel shift when incubated with the 3-E7 antibody ( Halpin and Harbury 2004b ), demonstrating its biochemical activity. Discussion DNA-directed synthesis requires “chemical translation” ( Gartner and Liu 2001 ; Halpin and Harbury 2004b ). Rather than simply acting as a tag to report a synthetic history ( Brenner and Lerner 1992 ), the DNA must actively determine the series of reactions that construct the molecule. No matter how this is achieved, it involves associated steps of DNA “reading” and synthetic execution. The synthetic steps must not damage the DNA and consequently compromise the reading process. Once each product is covalently attached to an amplifiable support that carries the information necessary for its synthesis, the product molecules are amenable to evolution and genetic manipulation. The proximity approach to chemical translation uses hybridization to induce proximity-driven chemical transformations. Because the DNA “reading” and chemical execution steps are simultaneous, reactions are necessarily performed in aqueous solutions with solute, pH, and temperature conditions that promote DNA–oligonucleotide hybridization. These conditions limit the generality, efficiency, and speed of possible organic transformations. The partitioning strategy separates the DNA reading and the chemical step, and also introduces a solid-phase format ( Halpin and Harbury 2004b ). The separation overcomes incompatibility between hybridization conditions and optimal reaction conditions. Synthetic transformations are carried out using standard solvents and elevated temperature. This advantage cannot be fully appreciated with solution-phase reaction formats, given the insolubility of DNA in organic solvents. Solid-phase formats, however, can take full advantage of flexibility in reaction conditions. Once DNA is bound to the solid phase, the chemical transformation, rather than DNA solubility, dictates solvent choice. We carry out transformations in H 2 O, MeOH, ethanol, isopropanol, DMF, dimethyl sulfoxide, dichloromethane, and ethyl acetate (data not shown). We also carry out reactions at high temperatures. For example, difficult peptide couplings are facilitated with elevated temperature, and the BME-mediated deprotection of lysine, arginine, histidine, and cysteine is carried out at 60 °C. We have recently used microwave-assisted methods ( Lew et al. 2002 ) to accelerate by 100-fold alkylation reactions on DNA-linked substrates (data not shown). Standard, commercially available reagents are used to perform chemistry, and we employ them in large excess (1,000- to 1,000,000-fold) over the DNA support, facilitating rapid reactions with high yields. The peptide coupling detailed here proceeds quantitatively in less than an hour, on par with the fastest standard solid-phase peptide synthesis coupling times. With the possible exception of tert -butyl esters, which are slowly removed during the 72 °C step of hybridization-mediated library splitting, the peptide chemistry and protecting groups presented here are suitable for use with DNA display ( Halpin and Harbury 2004b ). Another important aspect of the solid-phase method is reversible immobilization. This is an essential component of the DNA display chemical-translation cycle, where the DNA moves on and off a series of hybridization and chemistry columns ( Halpin and Harbury 2004a , 2004b ). In a simplistic view, we have essentially taken advantage of the polyanionic “handle” covalently attached to our synthetic substrates. The handle acts as a phase label ( Curran 1998 ) for solid-phase extractions with anion exchange resins. Previously, soluble polymer-supported synthesis has been accomplished with PEG and fluorinated hydrocarbon purification handles ( Han et al. 1995 ; Curran 1998 ). The polyanionic handle here uniquely accommodates both liquid- and solid-phase chemical steps. Our approach offers a general tool for derivatization of DNA. Unprotected peptide–DNA conjugates have been recognized as biochemically useful reagents for over 15 years ( Corey and Schultz 1987 ; Zuckermann and Schultz 1988 ; Allinquant et al. 1995 ; Tong et al. 1995 ; Troy et al. 1996 ). The methods described here are efficient, rapid, and inexpensive, and they utilize DNA of synthetic or enzymatic origin, offering advantages over previously reported techniques ( Robles et al. 1999 ; Stetsenko and Gait 2000 ; Debethune et al. 2002 ). Importantly, the protocols are not inherently limited to peptides. We have designed a set of tools for the adaptation of new chemistries. Reaction conditions are rapidly examined and optimized with oligonucleotides using HPLC mobility assays and MALDI-MS analysis. Nuclease P1 digestion facilitates the characterization of reactions on long DNA fragments and improves chromatographic and mass spectral resolution of synthetic products. MALDI-MS, primer extension analysis, and DNA sequencing reveal the presence of chemistry-induced DNA damage. We have used these tools to develop highly efficient protocols for solid-phase N -substituted polyglycine (“peptoid”) submonomer synthesis on unprotected DNA (data not shown). The chemistry used for peptoid synthesis is entirely different from peptide chemistry, illustrating the generality of the strategy. The potential for adapting other chemistries is essentially limitless. Wittig reactions, azide reductions, 1,3 dipolar cycloadditions, reductive aminations, Heck couplings, and a wide variety of other useful chemical transformations have been carried out in the presence of unprotected DNA without modification of DNA ( Bruick et al. 1996 ; Xu et al. 2001 ; Gartner et al. 2002a ; Li et al. 2002 ). Each could be used to synthesize and evolve interesting and potentially useful small molecule–DNA conjugate libraries. Our results demonstrate a robust method for solid-phase organic synthesis on unprotected DNA supports. Taken with the chemical-translation and DNA-manipulation strategies detailed elsewhere ( Halpin and Harbury 2004a , 2004b ), they facilitate a physical linkage between “genes” and synthetic “gene products” that is generalizable with respect to chemistry. The establishment of such a genetic underpinning to synthetic chemistry makes possible in vitro selection-based molecular discovery strategies for wholly abiotic molecular populations. Materials and Methods Materials. Fmoc amino acids were purchased from Novabiochem (La Jolla, California, United States), Chem-Impex International (Wood Dale, Illinois, United States), or Fluka (Basel, Switzerland). EDC was purchased from Omega Chemical (Levis, Quebec, Canada). N -hydroxysuccinimide was purchased from Chem-Impex. HOAt was purchased from Millipore (Billerica, Massachusetts, United States). Nuclease P1 (#27–0852-01), DEAE Sepharose Fast Flow (#17–0709-01), and Medium Grade G-25 Sephadex (#17–0033-01) were purchased from Pharmacia-LKB Technology (Uppsala, Sweden). Xba1 was purchased from New England Biolabs (Beverly, Massachusetts, United States). DEAE Sepharose columns were poured in Empty TWIST synthesis columns (#20–0030, Glen Research, Sterling, Virginia, United States). Kendall Monoject syringes (#1180100555, Kendall, Walpole, Massachusetts, United States) and a Promega (Madison, Wisconsin, United States) manifold with chemically resistant PFTE stopcocks (#121–0009, Argonaut, Foster City, California, United States) were used. All other chemical reagents or solvents were purchased from either Sigma-Aldrich (St. Louis, Missouri, United States) or Fisher Scientific International (Hampton, New Hampshire, United States). The internal control 10-base oligonucleotide had the sequence CGGACTAGAG. The reactive 20-base oligonucleotides had the sequence H 2 N-X-AGCAGGCGAATTCGTAAGCC, where X represents a C12 linker (NC20) or a longer PEG linker (NP20). NC20 was synthesized using the Glen Research 5′-Amino-Modifier C12 (#10–1922). NP20 was synthesized using the Glen Research Spacer Phosphoramidite 18 (#10–1918) followed by the 5′-Amino-Modifier 5 (# 10–1905). Reverse-phase HPLC assay. Coupling reactions were monitored by HPLC mobility shift using a C18 analytical column (Microsorb, Varian, Palo Alto, California, United States) and UV detection at 260 nm and 280 nm (Spectra FOCUS, Thermo Separation Products, San Jose, California, United States). Linear gradients from 0%–90% acetonitrile in 100 mM triethylammonium acetate (pH 5.5) were employed. Coupling efficiencies (recovered product DNA/ total recovered DNA) and yields (recovered product DNA/total starting material DNA) were determined by integration of elution peaks from the 260-nm channel. Chromophores added or removed during reactions cause changes in extinction coefficients less than the sensitivity (5%) of our HPLC assay (for NC20/NP20 ɛ 260nm ≈ 224.5 mM −1 cm −1 ) and were not considered in efficiency and yield determination. Reaction products were collected, concentrated to approximately 50 μM using centrifugal evaporation, and desalted over G-25 Sephadex. A mixture of 1 μl of desalted oligonucleotide and 1 μl of a freshly prepared saturated matrix solution was spotted on a matrix-assisted laser desorption/ionization target and allowed to air dry before mass spectrometry analysis. The matrix solution was made from 250 μl of H 2 O, 250 μl of acetonitrile, 25 mg of THAP, and 10 mg of ammonium tartrate. Peptide sequences (five or more amino acids) were verified by Edman degradation peptide sequencing. Nuclease P1 assay. DEAE elute buffer containing the peptide–DNA conjugates was neutralized and brought to 100 mM sodium acetate (pH 5.2) and 400 μM ZnSO 4 . Then 1 μg of nuclease P1 was added, and the mixture was incubated at 37 °C for 30 min. The entire reaction mixture was directly injected onto a C18 reverse-phase HPLC column and analyzed using linear gradients from 0%–90% acetonitrile in 10 mM ammonium acetate (pH 5.2). Yields were determined by integration of elution peaks from the 260-nm channel, using the P1 digestion product of unreacted starting material as a reference. Approximately 1 nmol of material was required for accurate UV detection. Products were collected, concentrated by centrifugal evaporation, and applied to a C18 SepPak cartridge (#WAT023590, Waters, Milford, Massachusetts, United States) in 25 mM triethylammonium acetate (pH 5.5). The cartridge was washed with 3 ml of 25 mM triethylammonium acetate (pH 5.5) and 1 ml of H 2 O. The products were eluted with 1 ml of 50/50 MeCN/H 2 O, concentrated to 100 μl by centrifugal evaporation, and analyzed by electrospray ionization mass spectrometry. For coumarin-labeled products, fluorescence was monitored (320 nm excitation/380 nm emission) with a scanning fluorescence detector (Thermo Separation Products, FL2000), and less than 50 pmol of material was necessary for accurate fluorescence detection. Primer extension assay. A 5′-aminated 340-base ssDNA support was generated as described ( Halpin and Harbury 2004a ). After loading the support onto DEAE Sepharose, the pentapeptide [Leu]enkephalin (Tyr-Gly-Gly-Phe-Leu) was synthesized using EDC chemistry. The resulting peptide–DNA conjugate was desalted over reverse-phase cartridge ( Halpin and Harbury 2004a ) and used as a template for primer extension ( Halpin and Harbury 2004b ). The radiolabeled duplex product was digested with Xba1, subjected to denaturing polyacrylamide gel electrophoresis, exposed to a phosphorimager cassette, and imaged on a Typhoon 8600 (dynamic range 10 5 /pixel). The intensity of full-length control and [Leu]enkephalin bands were similar to within 1% (S/N approximately 600). Upon peak integration along the entire lane, the full-length band represented a similar percentage of total intensity in the control (83%) and [Leu]enkephalin (81%) samples. The data suggest that, in the worst case, 3% of the DNA could have been modified during the course of peptide synthesis. Resin loading and eluting. Approximately 250 μl of DEAE Sepharose suspension was pipetted into an empty Glen Research column housing and washed with 20 ml of H 2 O followed by 12 ml of DEAE bind buffer (10 mM acetic acid and 0.005% Triton X-100) using a syringe or a syringe barrel, a male-male luer adapter, and a vacuum manifold (see Figure 1 ). The DNA was loaded onto the washed chemistry column in 1 ml of DEAE bind buffer at approximately 1 ml/min. The column was then washed with 3 ml of DEAE bind buffer, followed by 500 μl of H 2 O and 3–5 ml of the solvent required for the subsequent reaction. At least 50 nmol of oligonucleotide can be loaded onto one 250-μl DEAE Sepharose column. After the desired chemical transformations were carried out, the column was washed with 3–5 ml of the reaction solvent followed by 3–5 ml of DEAE bind buffer. The DNA was then eluted with 2 ml of DEAE elute buffer (1.5 M NaCl, 50 mM Tris-HCl [pH 8.0], and 0.005% Triton X-100) using a syringe. Long DNA molecules (340mers) were eluted with 4 ml of 1.5 M NaCl, 10 mM NaOH, and 0.005% Triton X-100 heated to 80 °C. Peptide coupling: succinimidyl esters. The following process was carried out twice. The column, with DNA bound, was washed with 3 ml of DMF. Using two syringes (see Figure 1 ), the column was incubated for 5 min at room temperature with a freshly prepared solution containing 225 μl of DMF, approximately 19 mg of Fmoc–amino acid-OSu, 67.5 μl of H 2 O, and 7.5 μl of diisopropylethylamine. Fmoc-Asn-OSu required four couplings rather than two to achieve quantitative yields. After the second amino acid incubation, the column was washed with 3 ml of DMF. Fmoc deprotection was carried out as follows: 3 ml of 20% piperidine in DMF was applied to a 3-ml syringe barrel attached to the top of the column. 1.5 ml was pushed through the column, followed by a 3-min incubation. An additional 1 ml was pushed through the column, followed by a 17-min incubation. The procedure was completed with a final 3-ml DMF wash. Peptide coupling: in situ activation. The column was washed with 500 μl of H 2 O and 3 ml of MeOH and then incubated for 30 min at room temperature with a freshly prepared 500-μl solution of 50 mM Fmoc–amino acid-OH, 50 mM EDC, and 5 mM HOAt in MeOH. These conditions were derived directly from Nozaki (1997) . The column was then washed with 3 ml of MeOH. Fmoc-Asn-OH and Fmoc-His(Trt)-OH required 50 mM HOAt for efficient coupling. The following amino acids couple optimally in DMF/MeOH mixtures: Fmoc-Arg(Ns)-OH, Fmoc-Asn-OH ,and Fmoc-Gln-OH (20% DMF); Fmoc-Phe-OH and Fmoc-Val-OH (40% DMF); Fmoc-Ala-OH (80% DMF); and Fmoc-His(Trt)-OH (100% DMF). These mixtures were determined primarily by solubility. Typical reactions were carried out with 100 pmol of aminated oligonucleotide (see Table 1 ). Peptide–DNA conjugates were synthesized on small (0.1–2 nmol) or preparative scales (greater than 10 nmol). For particularly difficult sequences or preparative-scale reactions, the coupling procedure was repeated multiple times to achieve high yields. In all cases, absolute yields for peptide–DNA products with n amino acids exceeded (0.9) n . Fmoc deprotection was carried out as described for succinimidyl ester coupling. See Supporting Information for a more detailed description of peptide coupling. Side chain deprotection. For Lys(Ns), Arg(Ns), Cys(S t Bu), and His(CNP), the column was washed with 3 ml of DMF and subsequently incubated for 30 min with 700 μl of DMF containing 500 mM BME and 250 mM DBU while submerged in a 60 °C H 2 O bath. The column was then washed with 3 ml of DMF and 12 ml of DEAE bind buffer. Lys(Ns) can also be deprotected quantitatively with a DMF solution containing 5% thiophenol and saturated K 2 CO 3 at 37 °C for 90 min. These conditions deprotect Arg(Ns) inefficiently, and have not been tested for Cys(S t Bu) or His(CNP). For Asp( t Bu), Glu( t Bu), and His(Trt), after HPLC purification, the tert -butyl ester and/or trityl containing oligonucleotide–peptide hybrid was incubated in a 20-mM MgCl 2 solution at 70 °C, yielding quantitative deprotection in 3 h (Asp, His) or 12 h (Glu). Deprotection can alternatively be carried out before HPLC purification: after eluting from the DEAE Sepharose column, NaOAc (pH 5.2) and MgCl 2 were added to final concentrations of 30 mM and 200 mM, respectively, and the solution was then incubated at 70 °C for the appropriate time. In contrast, acid deprotection on solid support was inefficient. Supporting Information Figure S1 MALDI-MS Analysis of Conjugates All reported conjugates were verified by MALDI-MS analysis. Example mass spectra of a conjugate before (A; Leu-NC20) and after (B; Arg-Leu-NC20) peptide coupling. Calculated masses are noted to the left of the mass peaks. (149 KB PDF). Click here for additional data file. Figure S2 Sequential Coupling Efficiencies and Yields of Peptide–DNA Synthesis (A) Reaction scheme for synthesis of GLFYG-NC20. Coupling efficiencies for individual steps are noted in black, and absolute yields from NC20 are noted in red. MALDI-MS results for all species are denoted under each species as “Observed (Calculated).” See Protocol S1 for precise coupling procedures. (B) HPLC analysis of sequential couplings during peptide synthesis monitored at 260 nm. Load and elutes from columns 1–5 are indicated. Sequential coupling efficiencies (black) were calculated by integration of recovered aminated DNA peaks. Absolute yields (red) were calculated by integration of intended product peak relative to load. A nonaminated 10-base oligonucleotide (10mer) was included as a control for nonspecific DNA loss and modification. Percent recovery of 10mer is noted in red. The HPLC analysis employed a 60-min gradient of 0%–45% MeCN in100 mM TEAA (pH 5.5). (368 KB PDF). Click here for additional data file. Protocol S1 Synthetic Methods (1.0 MB PDF). Click here for additional data file.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC434150.xml

549187

Correlation of breast cancer risk factors with HER-2/neu protein overexpression according to menopausal and estrogen receptor status

Background Several researchers have claimed that classification of tumours on the basis of HER-2/ neu overexpression or amplification may define a subset of breast cancer in which the net effect of a risk factor could be rather more obvious and its impact on breast cancer development more clear. We decided to investigate, in a group of patients from a geographical area with a low incidence of breast cancer, whether HER-2/ neu positive tumours are correlated with established or suspected risk factors for breast cancer and thus to identify distinct subgroups of high risk women. Methods This study analysed data from patients who attended the Breast Unit at the University Hospital of Heraklion, Crete, Greece between 1996 and 2002. 384 women with primary invasive breast cancer were compared with 566 screened women who were referred to the Unit and had not developed breast neoplasm by the time the data were analysed. Risk factor data were obtained from each subject by personal interviews using a structured questionnaire. The detection and scoring of the HER-2/ neu protein, estrogen and progesterone receptor expression were performed using immunochemistry. Odds ratios and 95% confidence intervals were determined by chi-square test and logistic regression analysis. Case-case odds ratios were calculated in order to measure the risk heterogeneity between HER-2/ neu + and HER-2/ neu -tumours. Separate analyses were performed for premenopausal and postmenopausal women and according to estrogen receptor status. Results In multivariate analysis without HER-2/ neu stratification, an increased breast cancer risk was associated with only four of the factors examined: use of oral contraceptives (OR = 4.40, 95%C.I: 1.46–13.28), use of HRT (OR = 7.34, 95%C.I: 2.03–26.53), an age at first full pregnancy more than 23 years (OR = 1.91, 95%C.I: 1.29–2.83) and body mass index more than 29 kg/m 2 (OR = 3.13, 95%C.I: 2.02–4.84). Additionally, a history of abortion or miscarriage (OR = 0.56, 95%C.I: 0.38–0.82) was correlated with a decreased risk of breast cancer. In the case to case comparison only BMI >29 kg/m 2 revealed a relative connection that was stronger with positive than with negative HER-2/ neu tumours (ratio of OR's = 2.23, 95%C.I: 1.20–4.15, p = 0.011). This may indicate evidence of heterogeneity of a rather significant degree for this factor. In the ER negative group an age at first full pregnancy >23 years and a BMI >29 kg/m 2 were associated with an increased risk in both HER-2/ neu groups, but the association was significantly stronger for the latter factor in the positive HER-2/ neu tumours (ratio of OR's = 2.46, 95%CI: 0.97–6.21). Conclusions Our study did not confirm that the established or putative hormonal breast cancer risk factors differ regarding their relations with HER-2/ neu + versus HER-2/ neu -breast tumours, with the exception of increased BMI. Further innovative studies with larger sample sizes are needed to examine how the status of these potentially modifiable breast cancer risk factors interacts with biological markers such as HER-2/ neu oncoprotein.

Background The HER-2/ neu oncogene, also known as c-erb-B2, c- neu or ERBB2, is located in chromosome 17q11.2-12, encoding an EGFR-family like glycoprotein [ 1 ]. Its amplification, which is strongly correlated with protein overexpression, occurs in about 15–43% of breast tumours [ 1 - 10 ]. The observation that morphologically similar neoplasmatic lesions of the breast can exhibit different biology has necessitated the identification of biological parameters that might improve risk assessment; the evaluation of HER-2/ neu expression is a typical example [ 11 ]. Indeed, several studies have demonstrated that HER-2/ neu amplification represents a prognostic and predictive marker; its expression is associated with early disease recurrence, relative resistance to chemo- and/or hormonotherapy and short survival [ 2 , 10 ]. In addition it has been shown that genetic alterations of the HER-2/ neu oncogene represent early events involved in breast carcinogenesis and tumour initiation, while their presence is observed in all stages of malignant development from in situ carcinomas to metastatic lesions [ 12 ]. As a result, some researchers have maintained that HER-2/ neu amplification and/or protein overexpression may also represent not only an important marker of prognosis but also a key indicator of the aetiological heterogeneity of breast carcinogenesis. [ 3 , 7 - 9 ]. On the other hand, the contribution of even well established breast cancer risk factors to the aetiology of carcinogenesis in the breast remains obscure, ill-defined and tenuous, mostly because of the existence of different pathways for the initiation and the evolution of a breast tumour [ 13 ]. In order to explain this incompatibility, several researchers have claimed that classification of tumours on the basis of HER-2/ neu overexpression or amplification may define a subset of breast cancer in which the net effect of a risk factor could be rather more obvious and its impact on breast cancer development more clear [ 3 , 7 , 8 ]. Thus, a close correlation of a risk factor with HER-2/ neu overexpression could indicate either that a HER-2/ neu alteration is the way that this risk factor evolves into the carcinogenesis or that there is a parallel interaction between them that leads to breast tumour initiation and development. Since the data in the literature supporting the above hypothesis are few and conflicting, we decided to investigate, in a group of patients from a geographical area with a low incidence of breast cancer, whether HER-2/ neu positive tumours are correlated with established or suspected risk factors for breast cancer and thus to identify distinct subgroups of high risk women. Methods This study analysed data derived from the database of the Breast Unit of the Department of Surgical Oncology at the University General Hospital of Heraklion on the island of Crete, Greece. The study considered all women who were consecutively diagnosed with primary, invasive breast cancer in our unit from 1996 to 2002. Subjects of other races, ethnicity, with residence outside Crete or diagnosed with DCIS or LCIS were excluded. Finally, 384 women, all originating from the island of Crete, were eligible for analysis. An age-stratified random sample of 566 women was used as a control group, derived from the Breast Unit's database of screened patients who had not developed breast cancer after a median follow up period of 40 months (range 12–92 months). Personal interviews were conducted with each woman during her first visit (both patients and controls) by a consultant or a senior resident. The interview followed a structured questionnaire, which did not change during the study period. Anthropometric measures were also made during the first visit. Women were classified as postmenopausal if their menstrual cycles had ended naturally at least 12 months before the interview or from surgery or radiation therapy at any age. Those who reported not having menstrual cycles for the last 10 months were considered as perimenopausal and were combined with premenopausal women for the purpose of our analysis. The following variables were analysed for all patients and controls: residence (rural/urban), age at interview (≤ 50 and >50 years), age at menarche (≤ 12 and >12 years old), age at first full birth (<23 vs. ≥ 23 years old), parity (nulliparous, 1 or 2, and >3), lactation (yes/no), use of medications to suppress lactation (yes/no), abortions and miscarriages (yes/no), age at menopause for postmenopausal women (≤ 50 and >50 years old), use of HRT for more than 2 months (yes/no), use of oral contraceptives for more than 2 months (yes/no), family history of breast cancer in a first degree relative (yes/no), history of benign breast disease (yes/no), obesity on the day of the interview (BMI ≤ 29 kg/m 2 vs. BMI>29 kg/m 2 , median value for the study population) and radiation history of the chest (yes/no). Immunohistochemical study For this study, tumour blocks were successfully retrieved in 378 (98.4%) and in 377 (98.17%) of the 384 interviewed cases for the immunohistochemical detection of HER-2/ neu protein and hormone-receptor expression, respectively. Immunohistochemical detection and scoring of the HER-2/neu protein expression Immunohistochemistry with the monoclonal antibody CB11 (NCLCB11, Novocastra Laboratories, UK 12 8EW), at a dilution of 1/50 with incubation period of 60 min, was performed using the OPTIMAX automated system with the Super Sensitive Link-Label Detection System RTU Multilink AP/Fast Red, QA200OXE (purchased from Biogenex Laboratories, San Ramon CA 94583 USA), following antigen retrieval by microwave pre-treatment at 500 Watts for 3 × 5 min in citrate buffer (0.01 M, pH 6). Sections from breast cancer of known positivity were used as positive controls. Negative controls were processed by omitting the primary antibody and substituting non-immune serum. Scoring was based on the criteria recommended by DAKO A/S for the HercepTest [ 14 ]. Only membrane staining pattern and intensity were scored using the 0–3+ scale: scores of 0–1+ were considered negative, score 2+ was considered weak positive-need for FISH, and score 3+ was considered (strongly) positive. Immunohistochemical detection and scoring of estrogen and progesterone receptors Three (3) μm-thick sections taken on negatively-charged (SuperFrostPlus) slides were dewaxed in xylene, and rehydrated through graded alcohols. Following antigen retrieval by microwave pre-treatment at 500 W for 3 × 5 min in citrate buffer (0.01 M, pH 6), estrogen receptor (ERa) and progesterone receptor (PR) expression was detected by immunohistochemistry using the same automated system and detection kit as above, and primary monoclonal antibodies to ERa (DAKO M7047) and PR (Biogenex code # MU 328-UC) at dilutions of 1/50 and 1/20, respectively, with incubation time 60', at room temperature. Positive and negative controls were processed as above. Positive nuclei counting was performed at a final magnification of 400× (Teaching double-headed NICON, ECLIPSE E400 microscope, equipped with CFI 10X/22 oculars). After scanning at a final magnification of 100× for locating the areas with highest density of ER+ or PR+ carcinoma cell nuclei (hot spots), a 40X/¥/0.17 WD 0.65 objective lens was used for cell counting. All carcinoma cells in three hot spots per immunostained slide were evaluated by two pathologists working simultaneously, though independently, and the mean of the two independent counts was considered the final counting value for each field and hot spot. The ratio of the ER+ or PR+ carcinoma cell nuclei was recorded separately for each of the hot spots. The final immunoreactivity index (score) was calculated as the mean percentage of ER+ or PR+ carcinoma cell nuclei in the three hot spots. Specimens were interpreted as positive for ER or PR if at least 10% of the cells demonstrated nuclear staining of any intensity of reactivity, from 1+ to 3+. Staining intensity was graded as negative (0), weak (1+), intermediate (2+) or strong (3+), and reported separately. A mean value of intensity was assigned for specimens in which the staining intensity varied from field to field, and/or from hot spot to hot spot. Statistical analysis Odds ratios (OR) and 95% CI (confidence intervals), as approximators of relative risk, were calculated to measure the association of the groups of breast cancer and the risk factors, using the chi-square (χ 2 ) test. A p value <0.05 was defined as significant. The potential association between breast cancers stratified by HER-2/ neu status and well known predisposing factors was further investigated by using a stepwise logistic regression analysis (backward LR), testing the independent effect of breast cancer risk factors (independent variables) on breast cancer (dependent variable) for all women and also separately for premenopausal and postmenopausal females. In addition, we undertook further stratification with estrogen receptor status by using the same multivariate logistic regression model. These patient-controls odds ratios helped us to detect the pattern of heterogeneity and to explore plausible aetiological correlations between patient subgroups. Additionally, case-case odds ratios were calculated in order to measure the risk heterogeneity between HER-2/ neu + and HER-2/ neu -tumours. It seems that the departure of the OR from unity can reveal the degree of heterogeneity between these subgroups [ 15 ]. Results Risk factor distributions in breast cancer patients and matched controls are presented in table 1 . The mean age at interview was 56.30 years. When all patients were compared with matched controls, and after adjustment for confounding factors, an increased breast cancer risk was associated with only four of the factors examined: use of oral contraceptives (OR = 4.40, 95%C.I: 1.46–13.28), use of HRT (OR = 7.34, 95%C.I: 2.03–26.53), an age at first full pregnancy more than 23 years (OR = 1.91, 95%C.I: 1.29–2.83) and body mass index more than 29 kg/m 2 (OR = 3.13, 95%C.I: 2.02–4.84). Additionally, a history of abortion or miscarriage (OR = 0.56, 95%C.I: 0.38–0.82) was correlated with a decreased risk of breast cancer. However, the number of oral contraceptive and HRT users was too small for reliable estimates of risk. Table 1 Characteristics of the participants Factors Cases N = 384 n(%) Controls N = 566 n(%) OR 1 (95% CI) OR 2 (95% CI) Age at interview ≤ 50 years 138(36) 178(31) 1.00 >50 years 246(64) 388(69) 0.82(0.62–1.08) NS Area of residence rural 189(49) 292(52) 1.00 urban 195(51) 274(48) 0.91(0.70–1.18) NS Menopausal status Pre/perimenopausal 134(35) 170(30) 1.00 Postmenopausal 250(65) 396(70) 0.80(0.60–1.06) NS Age at menopause 3 ≤ 50 years 144(59) 252(64) 1.00 >50 years 102(41) 140(36) 1.28(0.92–1.77) NS Age of menarche ≤ 12 years 155(40) 150(27) 1.86(1.41–2.45) >12 years 229(60) 412(73) 1.00 NS Use of oral contraceptives no 341(89) 548(97) 1.00 1.00 yes 43(11) 18(3) 3.84(2.18–6.77) 4.40(1.46–13.28) Use of HRT 3 no 231(92) 393(99) 1.00 1.00 yes 19(8) 3(1) 10.78(3.15–36.81) 7.34(2.03–26.53) First degree family history no 341(89) 522(92) 1.00 yes 43(11) 44(8) 1.5(0.96–2.33) NS Age at first full pregnancy <23 years 106(35) 239(50) 1.00 1.00 ≥ 23 years 197(65) 242(50) 1.84(1.37–2.47) 1.91(1.29–2.83) Parity nulliparous 79(20) 78(14) 1.00 1–2 175(46) 255(45) 0.68(0.47–0.98) NS 3 plus 130(34) 233(41) 0.55(0.38–0.80) NS Abortion or miscarriage no 183(57) 247(50) 1.00 1.00 yes 138(43) 248(50) 0.75(0.57–0.99) 0.56(0.38–0.82) Lactation no 67(22) 84(17) 1.00 yes 238(78) 404(82) 0.74(0.52–1.06) NS Medication to suppress lactation no 273(90) 438(90) 1.00 yes 31(10) 50(10) 0.99(0.62–1.60) NS Radiation to the chest no 372(97) 550(97) 1.00 yes 12(3) 16(3) 1.11(0.52–2.37) NS Body mass index ≤ 29 kg/m 2 282(74) 498(88) 1.00 1.00 >29 kg/m 2 97(26) 68(12) 2.52(1.79–3.55) 3.13 (2.03–4.84) Benign breast disease no 315(82) 472(83) 1.00 yes 69(18) 94(17) 1.10(0.78–1.55) NS 1 Adjusted for age. - 2 Adjusted for age, residence, menopausal status, menopausal age, menarche age, use of OC, use of HRT, first degree family history, age at first full pregnancy, parity, abortion, lactation, medication to suppress lactation, radiation to the chest, body mass index and benign breast disease. - 3 Postmenopausal women only. NS : non significant multivariate OR. Bold types: statistically significant values. Tumour characteristics of breast cancer patients are shown in table 2 . Thirty eight percent (145/378) of the tumours showed HER-2/ neu protein overexpression. HER-2/ neu positive tumours were not related with menopausal state, age at interview, tumour size, grade and stage, nodal and estrogen receptor status, but there was a modest positive association between HER-2/ neu and progesterone negative tumours. Table 2 Characteristics of the tumours of breast cancer patients 1 . Tumour characteristics HER-2/ neu + n = 145 (%) HER-2/ neu - n = 233 (%) p value Age at interview 0.533 ≤ 50 years 55 (40) 81 (60) >50 years 90 (37) 152 (63) Staging 0.106 I 30 (45) 36 (55) II 76 (35) 143 (65) III 20 (34) 39 (66) IV 2 (50) 2 (50) Unknown 17 (57) 13 (43) Tumour size 0.161 T1 55 (44) 71 (56) T2 65 (35) 121 (65) >T3 10 (29) 25 (71) Unknown 15 (48) 16 (52) Menopausal status 0.762 Pre/perimenopausal 52 (40) 80 (60) Postmenopausal 93 (38) 153 (62) Grading 0.577 I 7 (33) 14 (67) II 64 (36) 113 (64) III 60 (43) 80 (57) Unknown 14 (35) 26 (65) Node Status 0.119 Negative 55 (33) 112 (67) Positive 89 (43) 118 (57) Unknown 1 (25) 3 (75) Estrogen receptor status 0.108 Er+ 60 (33) 120 (67) Er - 85 (43) 112 (57) Unknown 1 Progesterone receptor status 0.038 Pr+ 49 (49) 52 (51) Pr- 96 (35) 180 (65) Unknown 1 1 Data for HER-2/neu status were missing for 6 of the 384 interviewed cases . Menopausal status and estrogen receptor stratification Subgroups of women stratified by menopausal status were further analysed by a multivariate stepwise logistic regression model adjusted for the remaining variables (table 3 ). In the premenopausal group of women, an increased risk for HER-2/ neu -tumours was observed for those women who reported an age at first full pregnancy ≥ 23 years ( OR = 3.56, 95%C.I: 1.70–7.46 ), a BMI>29 kg/m 2 ( OR = 6.89, 95%C.I: 2.23–21.25 ), first degree family history ( OR = 3.30, 95%C.I:1.10–9.96 ) or use of oral contraceptives ( OR = 11.19, 95%C.I 3.70–33.84 ), while an age at menarche less than 12 years was the only factor which slightly increased the risk in premenopausal HER-2/neu+ patients ( OR = 2.09, 95%C.I 0.99–4.42 ). Abortion played a less protective role ( p = 0.068 ) for HER-2/ neu -breast cancer in premenopausal than in postmenopausal women ( p = 0.038 ). However, the intercase comparison in the premenopausal subgroup showed an evidence of heterogeneity only for the HER-2/ neu + women who had ever had an abortion (ratio of the OR's = 3.12, 95%C.I:1.18–8.24 ), while use of oral contraceptives ( OR = 0.16, 95%C.I: 0.04–0.60, p = 0.007) and a positive first degree family history ( OR = 0.09, 95%C.I: 0.01–0.85, p = 0.035) showed a stronger association for HER-2/ neu negative tumours. Table 3 Multivariate analysis of risk factors and HER-2/neu overexpression according to menopausal status Risk Factors HER-2/neu+ Cases/controls OR (95% CI) HER-2/neu-Cases/controls OR (95% CI) Ratio of the OR's Cases+/cases- OR (95% CI) PREMENOPAUSAL Age at first full pregnancy(≥ 23 years) NS 3.56(1.70–7.46) NS Body mass index(>29 kg/m 2 ) NS 6.89(2.23–21.25) NS Abortion or miscarriage(ever) NS 0.49(0.23–1.05) 3.12 (1.18–8.24) First degree family history(positive) NS 3.30(1.10–9.96) 0.09 (0.01–0.85) Use of oral contraceptives (ever) NS 11.19(3.7–33.84) 0.16 (0.04–0.60) Age of menarche (≤ 12 years) 2.09 (0.99–4.42) NS NS POSTMENOPAUSAL Age at first full pregnancy(≥ 23 years) 2.19(1.23–3.91) 1.66(1.03–2.66) NS Body mass index(>29 kg/m 2 ) 4.83(2.75–8.49) 2.67(1.56–4.55) 2.23 (1.20–4.15) Abortion or miscarriage(ever) 0.50(0.28–0.88) 0.62(0.39–0.97) NS First degree family history(positive) NS 2.23(1.08–4.63) NS Use of estrogens (ever) NS 10.70(2.71–42.31) 0.21 (0.04–1.08) Use of oral contraceptives (ever) NS 6.47(1.89–22.16) NS Age of menarche (≤ 12 years) NS 1.72(1.07–2.75) 0.54 (0.28–1.04) ALL WOMEN Age at first full pregnancy(≥ 23 years) 2.19(1.23–3.91) 1.66 (1.03–2.66) NS Body mass index(>29 kg/m 2 ) 4.83(2.75–8.49) 2.67(1.56–4.55) 2.23(1.20–4.15) Abortion or miscarriage(ever) 0.50(0.28–0.88) 0.62(0.39–0.97) NS First degree family history(positive) NS 2.23(1.08–4.63) NS Use of estrogens (ever) NS 10.70(2.71–42.31) 0.21 (0.04–1.08) Use of oral contraceptives (ever) NS 6.47(1.88–22.16) NS Age of menarche (≤ 12 years) NS 1.72(1.07–2.75) 0.54 (0.28–1.04) Adjusted for age, residence, menopausal status, menopausal age, menarche age, use of OC, use of HRT, first degree family history, age at first full pregnancy, parity, abortion, lactation, medication to suppress lactation, radiation to the chest, body mass index and benign breast disease. NS: non significant. The results of logistic regression were identical for all women and the postmenopausal groups of patients due to the large sample size. Patients with an age of menarche ≤ 12 years ( OR = 1.72, 95%C.I: 1.07–2.75 ), first degree family history ( OR = 2.23, 95%C.I:1.08–4.63 ), use of HRT ( OR = 10.70, 95%C.I: 2.71–42.31 ) or OC ( OR = 6.47, 95%C.I:1.89–22.16 ) were at increased risk of developing HER-2/ neu -breast cancer only, although the significance of the latter two factors was of little value due to the limited sample size. On the other hand, an age at first full pregnancy ≥ 23 years and a BMI greater than 29 kg/m 2 increased breast cancer risk independently of HER-2/ neu status, while a history of abortion decreased risk in the same way. In the case to case comparison only BMI >29 kg/m 2 revealed a relative stronger connection with positive than with negative HER-2/ neu tumours (ratio of OR's = 2.23, 95%C.I: 1.20–4.15, p = 0.011) and this may indicate an evidence of heterogeneity of a rather significant degree for this factor. The stronger association between an age at menarche ≤ 12 years, use of HRT and negative as opposed to positive HER-2/ neu status did not reach statistical significance ( p = 0.067 and p = 0.062, respectively). A different stratification pattern of our study's population is presented in table 4 . This multivariate model, further stratified by estrogen receptor status, confirmed the observed tight connections between HER-2/ neu positivity and obesity already shown in the analysis so far. In more detail, although BMI >29 kg/m 2 elevated risk for both ER negative and positive tumours independently of HER-2/ neu status , the association was significantly stronger for positive HER-2/ neu tumours in the ER negative group ( ratio of OR's = 2.46, 95%CI: 0.97–6.21 ). Additionally, in the same group an age at first full pregnancy >23 years revealed an increase of risk in both HER-2/ neu groups, while first degree family history ( OR = 2.72, 95%C.I: 1.05–7.07, p = 0.040), age at menopause >50 years ( OR = 2.05, 95%C.I: 1.10–3.79, p = 0.023) and birth of 1–2 children ( OR = 2.38, 95%C.I: 1.21–4.67, p = 0.012) elevated risk for HER-2/ neu negative tumours only. In the ER+ group of women the direct comparison between cases revealed no associations with any factor at all, while abortion showed a protective pattern against breast cancer which expressed estrogen receptors independently of HER-2/ neu status. Table 4 Multivariate analysis of risk factors and HER-2/neu overexpression according to ER 1 status. Risk Factors HER-2/neu+ Cases/controls OR 2 (95% CI 3 ) HER-2/neu-Cases/controls OR (95% CI) Ratio of the OR's Cases+/cases- OR (95% CI) ER + cases Body mass index(>29 kg/m 2 ) 5.59 (2.58–12.13) 2.84 (1.52–5.32) NS Age at 1st pregnancy (≥ 23 years) 2.09 (0.97–4.53) NS NS First degree family history (positive) NS 2.18 (0.92–5.14) NS Abortion or miscarriage (ever) 0.44 (0.20–0.95) 0.56 (0.32–0.99) NS ER – cases Body mass index(>29 kg/m 2 ) 5.33 (2.59–10.94) 2.41 (1.15–5.04) 2.46 (0.97–6.21) Age at 1st pregnancy (≥ 23 years) 2.37 (1.08–5.18) 1.78 (0.93–3.42) NS First degree family history (positive) NS 2.72 (1.05–7.07) NS Age of menopause (>50 years) NS 2.05 (1.10–3.79) NS Parity (1–2 children) NS 2.38 (1.21–4.67) NS Adjusted for age, residence, menopausal status, menopausal age, menarche age, use of OC, use of HRT, first degree family history, age at first full pregnancy, parity, abortion, lactation, medication to suppress lactation, radiation to the chest, body mass index and benign breast disease. 1 ER: estrogen receptor. NS: non significant. Discussion This epidemiological study, conducted in a low incidence Mediterranean population, [ 16 ] found that obesity was related with postmenopausal breast tumours that overexpress HER-2/ neu oncoprotein. In fact, increased BMI elevated risk in both groups, but the comparison between HER-2/ neu + and HER-2/ neu- tumours revealed a much stronger association with HER-2/ neu + breast cancers. Very few studies have examined the possibility whether HER-2/ neu status can help discriminate aetiologically distinct subgroups of breast cancer cases, and none of them has identified the effect of increased BMI with HER-2/ neu positive tumours [ 3 , 5 , 7 - 9 ]. More specifically, in contrast with other investigators who have shown an elevated risk for HER-2/ neu + tumours with an early age at first full pregnancy, we found a strong elevated risk with a late age regardless of HER-2/ neu protein expression [ 3 ]. Previous findings suggested an inverse relationship between abortion and HER-2/ neu + breast cancers, while we also found this inverse association but independently of HER-2/ neu status [ 7 ]. Interestingly enough, abortion increased risk for HER-2/ neu + tumours only in the premenopausal group of women. Early contraceptive use has been positively associated with HER-2/ neu + breast cancer in two studies [ 7 , 8 ], but our findings were different, revealing a positive association with HER-2/ neu negative tumours. However, because the number of oral contraceptive (and HRT) users in this study was small, this subgroup analysis was hindered by decreased power to detect associations of any magnitude. The slightly protective effect of parity found in the age-adjusted analysis was diminished after logistic regression and did not reveal any association with HER-2/ neu status, in contrast with previous findings [ 7 ]. Breastfeeding was associated with increased risks for breast cancer in women with HER-2/ neu positive tumours in one study while other investigators reported opposite results [ 3 , 9 ]. Although our study population showed a remarkable lactation incidence (almost 80% of the participants) we found no associations at all. Our findings have similarities and differences with respect to previous reports that examined the associations of breast risk factors with HER-2/ neu status. This inconsistency may reflect differences in study design, populations, and laboratory methodology. In this study we used immunochemistry (CB11 monoclonal antibody) to assess the HER-2/ neu protein overexpression, which is highly correlated with gene amplification according to previous reports [ 2 , 4 ]. Also, the percentage of women with breast cancer and HER-2/ neu protein overexpression found here was within the limits reported elsewhere [ 3 , 5 - 9 ]. This lack of relationship between HER-2/ neu protein overexpression and most of the hormone-related breast cancer risk factors does not completely agree with several hypotheses which have maintained that combined estrogen and HER-2/ neu activation is closely involved in the same pathway in breast cancer carcinogenesis [ 17 , 18 ]. The only hormone-related factor that was found to be related with HER-2/ neu positive tumours in our study was high body mass index, which is an established risk factor that has an estrogen-mediated oncogenic effect on the mammary gland. More specifically, obesity is associated with higher breast cancer risk among postmenopausal women through greater lifetime exposure to higher levels of estrogens produced in adipose tissue and lower SHBG production [ 19 , 20 ]. Higher levels of circulating estrogens enhance the rate of cell division, and this hormone-induced cellular proliferation can result in somatic mutations and finally lead to a malignant change. These alterations involve many genes, including those concerned with hormone metabolism and transport, DNA repair, as well as tumour suppressor genes and oncogenes such as the HER-2/ neu gene [ 18 , 21 ]. According to some investigators, circulating estrogens can stimulate breast cancer cell proliferation, not only through hormone receptors, but also through the HER-2/ neu receptor, and so promote carcinogenesis through common means [ 4 , 17 , 18 ]. Numerous epidemiological and experimental studies have shown the strong relationship between HER-2/ neu -positivity and lack of hormone receptor expression in breast tumours [ 2 , 10 , 18 , 22 ]. In our study, HER-2/ neu positive tumours were weakly related with the absence of estrogen receptors, although this was not statistically significant (see table 2 ). Because different ER status can result in different correlations between risk factors and HER-2/ neu+ breast cancer, it is always important to examine these interactions under ER stratification [ 8 ]. Since antiestrogens can lower HER-2/ neu levels in ER negative tumours, it is possible that an excess of estrogens can stimulate HER-2/ neu in these tumours [ 8 , 18 ]. This mechanism could explain the stronger association between obesity (a situation with an overload of estrogens as mentioned above) and HER-2/ neu -positivity among ER negative patients that was found in the present study (see table 5). The interview was conducted during the subjects' first visit to the unit and before clinical examination or any other intervention took place. This constitutes an advantage, because there was no chance that the subjects (both cases and controls) would be influenced by the diagnosis and might therefore falsely inflate the relative risk. Thus, the likelihood of recall bias is not high, improving the comparability of several covariates in both groups, and the selection bias is lessened since all subjects had taken the same route through the Breast Unit's standard routine procedures. Since each case group was compared with the same control group, any selection bias would be expected to have a similar effect on the estimates in the tumour subgroups. Thus, it is extremely unlikely that recall bias issues would apply only to those cases in a specific HER-2/ neu status subgroup. Some caution regarding our findings is related to the size of the study group. In the analyses stratified by HER-2/ neu and menopausal or ER status numbers are quite small and for some risk estimates the confidence intervals are wide and the estimates of risk unstable. Conclusions In conclusion, our study did not confirm that the established or putative hormonal breast cancer risk factors differ regarding their relations with HER-2/ neu + versus HER-2/ neu -breast tumours, with the exception of increased BMI. Further innovative studies with larger sample sizes are needed to examine how the status of these potentially modifiable breast cancer risk factors interacts with biological markers such as HER-2/ neu oncoprotein. Their findings will provide us with greater insight into breast cancer aetiology and will help us identify any association that would help discriminate subgroups of women at higher risk. Abbreviations EGFR: epidermal growth factor receptor, HRT: hormone replacement therapy, BMI: body mass index, SHBG: sex hormone-binding protein, ER: estrogen receptor, PR: progesterone receptor, OC: oral contraceptives. Competing interests The author(s) declare that they have no competing interests. Authors' contributions NT conceived the study, participated in its design and drafted the manuscript. ES participated in the design of the study, assisted in writing and reviewed the final article. EfS and KM scheduled and performed the laboratory analysis. NA performed the statistical analyses. VG and DDT participated in the design and coordination of the study and reviewed the final article. All authors have read, discussed and approved the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549187.xml

549193

The progressive nature of Wallerian degeneration in wild-type and slow Wallerian degeneration (WldS) nerves

Background The progressive nature of Wallerian degeneration has long been controversial. Conflicting reports that distal stumps of injured axons degenerate anterogradely, retrogradely, or simultaneously are based on statistical observations at discontinuous locations within the nerve, without observing any single axon at two distant points. As axon degeneration is asynchronous, there are clear advantages to longitudinal studies of individual degenerating axons. We recently validated the study of Wallerian degeneration using yellow fluorescent protein (YFP) in a small, representative population of axons, which greatly improves longitudinal imaging. Here, we apply this method to study the progressive nature of Wallerian degeneration in both wild-type and slow Wallerian degeneration (Wld S ) mutant mice. Results In wild-type nerves, we directly observed partially fragmented axons (average 5.3%) among a majority of fully intact or degenerated axons 37–42 h after transection and 40–44 h after crush injury. Axons exist in this state only transiently, probably for less than one hour. Surprisingly, axons degenerated anterogradely after transection but retrogradely after a crush, but in both cases a sharp boundary separated intact and fragmented regions of individual axons, indicating that Wallerian degeneration progresses as a wave sequentially affecting adjacent regions of the axon. In contrast, most or all Wld S axons were partially fragmented 15–25 days after nerve lesion, Wld S axons degenerated anterogradely independent of lesion type, and signs of degeneration increased gradually along the nerve instead of abruptly. Furthermore, the first signs of degeneration were short constrictions, not complete breaks. Conclusions We conclude that Wallerian degeneration progresses rapidly along individual wild-type axons after a heterogeneous latent phase. The speed of progression and its ability to travel in either direction challenges earlier models in which clearance of trophic or regulatory factors by axonal transport triggers degeneration. Wld S axons, once they finally degenerate, do so by a fundamentally different mechanism, indicated by differences in the rate, direction and abruptness of progression, and by different early morphological signs of degeneration. These observations suggest that Wld S axons undergo a slow anterograde decay as axonal components are gradually depleted, and do not simply follow the degeneration pathway of wild-type axons at a slower rate.

Background Wallerian degeneration, the characteristic degeneration sequence of nerve fibres separated from their cell bodies, was described by Waller in 1850 [ 1 , 2 ]. Following various forms of axon injury this rapid degeneration process begins with degradation of axoplasm and axolemma accompanied by development of axonal and myelin debris that is subsequently removed by Schwann cells and invading macrophages. In recent years it became apparent that Wallerian degeneration is initiated by an active process intrinsic to the axon that shares some principles with apoptosis [ 3 - 7 ]. These discoveries were firmly established by studies on the slow Wallerian degeneration (Wld S ) mutant mouse, in which this active process seems to be turned off. Accordingly, this mutant shows a tenfold delay in Wallerian degeneration and synapse breakdown after experimental nerve injury [ 8 , 9 ]. The delay of Wallerian degeneration is an intrinsic property of the axon suggesting that glial cell and macrophage changes are secondary events [ 3 ]. The underlying trait is carried by the autosomal dominant mutation Wld S that arose by spontaneous mutation [ 6 , 10 ]. Genetic analysis has shown that the Wld S mutation on mouse chromosome 4 comprises a stable 85-kb tandem triplication [ 11 , 12 ] encoding the N-terminal 70 amino acids of the multiubiquitination factor Ube4b fused in frame to the nuclear NAD producing enzyme nicotinamide mononucleotide adenylyltransferase 1 (Nmnat 1). Correspondingly, Wld S mice express a novel chimeric protein (Wld S protein) in neuronal nuclei that has full Nmnat 1 activity but seemingly no Ube4b function since the expressed N-terminus lacks ability for multi-ubiquitination within the ubiquitin-proteasome system (UPS) [ 13 , 14 ], the molecular machinery responsible for a major pathway of cellular protein catabolism. Either only one or both parts of the nuclear Wld S protein could be responsible for the phenotype through a nuclear process that has an indirect effect on the axon although recent results foster that the Wld S mechanism is likely to involve a gain of function of NAD synthesis [ 15 ]. However, inhibition of a specific step of the ubiquitin proteasome system or another modifying role of the N-terminal domain of the Wld S protein remains a possibility [ 16 , 17 ]. From a clinical point of view not only traumatic disorders such as nerve, spinal cord or head injury result in Wallerian degeneration [ 18 ] but it is now broadly accepted that Wallerian degeneration is mechanistically related to axon loss in many neurodegenerative disorders such as amyotrophic lateral sclerosis, Charcot-Marie-Tooth disease, toxic neuropathy, multiple sclerosis, and possibly Alzheimer's Disease and Parkinson's Disease [ 7 , 14 , 19 - 23 ]. Protection from neurodegenerative disorders by Wld S is currently under intense investigation. The neuroprotective mutation alleviates diverse PNS axon disorders, including dysmyelination and dying back neuropathy in P0 -/- mutants [ 24 ], motor neuropathy in pmn mutants [ 25 ] and axon degeneration in Vincristine and Taxol toxicity [ 26 - 28 ]. More recently Wld S was reported to attenuate pathology in acute CNS lesions caused by stroke [ 29 ], Parkinson's disease [ 30 ], and in gracile axonal dystrophy ( gad ) mice, a CNS axonal spheroid pathology [ 31 ]. A better understanding of the biological mechanism of delayed axon degeneration in neurological diseases would help to develop therapeutic methods to target axon degeneration. Despite research extending over more than 150 years and its frequent use as tool to detect interneuronal connections in the CNS since the time of Cajal, fundamental issues of Wallerian degeneration remain unresolved and controversial even on a purely morphological level. Among these is the spatiotemporal pattern of Wallerian degeneration along the separated nerve stump. Understanding the exact pattern of spread should provide additional insights into the mechanisms of axon death and may indicate strategies to alter Wallerian degeneration in neurological disease. Shortly after the pioneering investigations of Waller, and in the following decades, there has been much debate as to whether degeneration occurs in an anterograde direction, a retrograde direction or simultaneously along the separated nerve stump axons (reviewed historically in [ 32 - 34 ]). The controversy over the directionality of Wallerian degeneration has arisen partly because appropriate methods to follow axons over considerable distances did not exist until recently but also because the course of Wallerian degeneration varies with many experimental factors. Thus, spatiotemporal evolution of Wallerian degeneration depends on the laboratory animals used [ 32 , 34 - 36 ], on the age of the animals [ 32 , 37 ], on the neuroanatomical locus of study (CNS vs. PNS) [ 35 , 38 , 39 ], on the type of fibre analyzed (e.g. myelinated vs. unmyelinated, thick vs. thin axons) [ 34 , 39 , 40 ], on the type of lesion (axotomy, crush, ligature, intoxication etc.) [ 34 , 41 , 42 ], on the length of the remaining distal nerve stump [ 32 , 43 - 45 ], on the criteria used for identification of fibre degeneration (e.g. myelin breakdown, axon disintegration, decay of electrophysiological activity) [ 32 , 33 , 39 ], on environmental factors (e.g. temperature) [ 6 , 32 , 34 , 36 ] and many more. For example, in more modern experimentation from the last decades, Lubinska [ 46 ] demonstrated with the help of the teased fibre technique on myelinated fibres of the rat phrenic nerve that axonal breakdown into myelin ovoids spreads anterogradely along axons separated from their cell bodies at velocities correlated with fibre diameter and internodal length. George and Griffin [ 47 ] also found anterograde spread of axonal disintegration along dorsal columns of the rat following L 4 L 5 L 6 radiculotomy. Contrary to these views, Lunn and colleagues [ 48 ] showed by means of silver-stained wholemount preparations from the peripheral nerve stump that degeneration after crushing proceeds in a retrograde direction. They also proposed a retrograde progression after sectioning, freezing or ligaturing, although this was less clear because degeneration was more complete at the time-point sampled. Electrophysiological approaches suggested that the spread of failure of conduction in degenerating mammalian nerves runs from proximal to distal after nerve transection [ 49 ]. In cell culture studies using dorsal root ganglions (DRG) explants membrane beading, blebbing, fragmentation and Annexin V staining progressed along interrupted neurites in an anterograde direction with a rate comparable to that of slow axonal transport [ 50 ]. Taking secondary changes after axon disintegration into consideration, Bendszus and colleagues [ 51 ] tracked an anterograde spatiotemporal course of macrophage infiltration after acute peripheral nerve injury in rats. While most of the above mentioned investigators concluded a progressive nature of Wallerian degeneration from the appearance of degeneration gradients along injured nerves other authors did not observe any evidence for an anterograde or retrograde pattern of axonal degeneration [ 52 , 53 ]. In view of the contradictions and anomalies in the previous literature, we have reassessed the directionality of Wallerian degeneration using a recently introduced technique to visualize individual fluorescent axons over cm-long distances during degeneration [ 54 ]. This was made possible by using nerves from transgenic mice expressing Yellow Fluorescent Protein (YFP) in representative subsets of axons, which presents a simplified image of peripheral nerve [ 55 ]. No method existed until recently to follow up individual axons undergoing Wallerian degeneration over a considerable length. Here we compared the progression of Wallerian degeneration along single axons traced over lengths of approximately 2.5 cm. Specifically, we have tested a key prediction of all progressive models: that it should be possible to image axons degenerated at one end but not at the other. We detected such axons and showed that Wallerian degeneration in wild-type peripheral nerve is a rapid, asynchronous, progressive and wave-like process that can change its orientation depending on the lesion type. To our knowledge there have been no reports about the spatiotemporal pattern of the much-delayed axon degeneration in peripheral Wld S nerves that could yield important clues for understanding classic Wallerian degeneration. Therefore, we also report a detailed characterisation of injury-induced axon degeneration in slow Wallerian degeneration mutant mice in order to determine whether axons degenerate with a similar spatial evolution to that in wild-type mice, but in "slow motion", or whether the process is fundamentally distinct. We report a series of differences between axon degeneration in wild-type and Wld S mice, suggesting that irreversible injury in axons where Wallerian degeneration is blocked eventually leads to a different pathway of degeneration. Results YFP labelled wild-type axons fragment abruptly and asynchronously after a latent phase of approximately 36–44 h In preliminary experiments we used conventional light and electron microscopy to investigate whether Wallerian degeneration is progressive in wild-type mouse peripheral nerves. We were never able to find any significant gradients of degeneration along injured nerves that were processed with these traditional methods (data not shown). We then looked for signs of progression in localised observations of degenerating YFP-H nerves because fragmentation of YFP-labelled axons from these mice corresponds to granular disintegration of axoplasm as well as myelin ovoid formation and YFP positive axons represent the whole myelinated axon population [ 54 ]. Axonal fragmentation was first detected at both the proximal and distal ends of the distal nerve stump 37 h after transection (Fig. 1A, B ) and 40 h after crush injury (Fig. 1C, D ). 42 h after transection (Fig. 1E, F ) and 44 h following crush lesion (Fig. 1G, H ) the majority of axons in both locations were fragmented by assessment with conventional fluorescence microscopy. By direct comparison of the separate images of proximal and distal sites in the distal nerve stumps excised at all further time points no apparent difference was visible in the proportion of fragmented axons (data not shown). Figure 1 After a latency period Wallerian degeneration following cut and crush injury starts abruptly in single axons and involves total fragmentation of axons within few hours A-D : Conventional fluorescence micrographs of a ~2.5 cm long peripheral nerve stump (sciatic-tibial nerve segment) wholemount preparation at the proximal (A) and distal site (B) 37 h after cut injury with few individual fluorescent axons broken into fragments. A small number of axons fragmented at the proximal (C) and distal site (D) of a peripheral nerve stump wholemount preparation could also be detected 40 h following crush injury. E-H : Conventional fluorescence micrographs of a ~2.5 cm long peripheral nerve stump (sciatic-tibial nerve segment) wholemount preparation at the proximal (E) and distal site (F) 42 h after cut injury with most YFP labelled axons fragmented. A similar picture with a majority of axons degenerated is evident at the proximal (G) and distal end (H) of a peripheral nerve stump wholemount preparation 44 h after crush injury. YFP fluorescence has been pseudo-coloured green with the applied imaging software (MetaVue, Universal Imaging Corporation). Magnification: 100 × Wallerian degeneration in wild-type nerves progresses anterogradely after nerve transection and retrogradely after nerve crush The failure to observe any gradient of degeneration in the above experiment does not mean that Wallerian degeneration is not progressive. It could propagate so rapidly that it was not detectable by this method, or a gradient might not be detectable because of the considerable statistical noise of the highly asynchronous process. In order to investigate these possibilities we turned to confocal tracing of individual axons in long wholemount YFP-H nerve segments. At 37 h after transection we found 2.0 % of fluorescent axons with extensive proximal fragmentation and intact distal regions indicating an anterograde gradient of Wallerian degeneration in these axons (Fig. 2A ). Around 40 h the proportion of distal axon stumps degenerated at their proximal but not distal ends peaked at 9.3%. An example is presented in Fig. 3 . Of the remaining axons, 77.5 % were intact and 13.2 % were entirely fragmented without an apparent gradient. 42 h after cut injury the proportion of partially fragmented axons decreased to 4.4 % and after 48 h we only found axons that were fragmented over the whole length. In summary, at all investigated time points partially degenerated axons (mean: 5.3 %) always exhibited an anterograde spread of Wallerian degeneration after nerve transection. Figure 2 Quantification of fluorescent axons in wholemount YFP-H peripheral nerve stumps after cut and crush injury at different time points. Depending on the extent of fragmentation, YFP positive axons from peripheral nerve stumps were assigned to the group "intact", "entirely fragmented", "fragmented with anterograde gradient" or "fragmented with retrograde gradient". The chart presents means and standard deviations. A : All partially fragmented axons that could be identified at the time points between 37 h and 42 h after cut injury were fragmented at the proximal end of the distal axon stump but not further distal, indicating an anterograde gradient of Wallerian degeneration ("fragmented with anterograde gradient"). A maximum of 9.3 % YFP positive axons with anterograde fragmentation appeared 40 h after cut injury. B : All partially fragmented axons that could be identified at the time points between 40 h and 44 h after crush injury were fragmented at their distal ends but not further proximal indicating a retrograde gradient of Wallerian degeneration ("fragmented with retrograde gradient"). A maximum of 7.2 % of YFP positive axons with retrograde fragmentation appeared 44 h after crush injury. Figure 3 Wallerian degeneration proceeds in anterograde direction along individual axons after cut injury. Confocal composite picture showing seven consecutive lengths (from top to bottom in overview) of the proximodistal course of an individual YFP labelled axon within a distal nerve stump 40 h after transection demonstrating an anterograde progression of axon fragmentation. Note that this axon has fragmented in its proximal end (upper inset) but not in its distal end (lower inset). Axonal fragments are clearly demarcated by fluorescence interruptions (arrows in upper inset). YFP fluorescence has been pseudo-coloured yellow with the applied confocal imaging software (Biorad LaserSharp 2000). Scale bar: 500 μm In contrast, all partially fragmented axons after crush injury at all investigated timepoints were fragmented in distal tibial nerve but not at the proximal end of the distal stump (Fig. 2B ). Once again, only a small minority (mean: 5.0%) could be detected in this state at any one time. Partially fragmented axons first appeared at 40 h (1.6 % of axons) and the proportion peaked at 44 h (7.2 %). A representative example is shown in Fig. 4 . The remaining 92.8 % of YFP labelled axons at 44 h was entirely fragmented without an apparent gradient. At all earlier investigated time points axons with a retrograde gradient of fragmentation were also observed but in lower proportions and after 48 h we only found axons that were fragmented over the whole length. Figure 4 Wallerian degeneration proceeds in retrograde direction along individual axons after crush injury. Confocal composite picture showing seven consecutive lengths (from top to bottom in overview) of the proximodistal course of an individual YFP labelled axon within a peripheral nerve stump 44 h after crush injury displaying a retrograde progression of axon fragmentation. Note that this axon has fragmented in its distal end (lower inset) but not in its proximal end (upper inset). Axonal fragments are clearly demarcated by fluorescence interruptions (arrows in lower inset). YFP fluorescence has been pseudo-coloured yellow with the applied confocal imaging software (Biorad LaserSharp 2000). Scale bar: 500 μm Summarising all these quantification results of cut and crush lesions at time points where partially fragmented fibres were observed, on average 94.8 % of all axons were either completely intact or fragmented and 5.2 % showed a gradient of Wallerian degeneration, whose orientation depended on the lesion type. A wave of axonal fragmentation propagates rapidly along individual wild-type axons and the axon population degenerates asynchronously To study further the gradients of axonal fragmentation both after cut and crush injury (Fig. 3 , 4 ) we quantified the number of axonal breaks along partially and totally fragmented axons. Firstly, this was a way to distinguish between a locally restricted wave of fragmentation such that entirely intact lengths of axon suddenly change into entirely fragmented lengths, and a more gradual fragmentation process that would result in a few interruptions that become more frequent further along the axon. Secondly, by this approach we tried to get insight on the question of whether axons assigned into the group "entirely fragmented" continue to break into smaller fragments leading to a gradient of fragment size along the nerve. Concerning the first question we found that Wallerian degeneration progresses as a wave, with the wave front defining the point to which fragmentation had spread along the axon. In partially fragmented axons separated from the cell body by transection or proximally compressed by crush lesion, axon regions with no features of degeneration abruptly change into segment lengths with marked breakdown within a transition zone of less than one millimeter (Fig. 5 , 6 ). A short region of intact axon immediately ahead of the wavefront becomes increasingly vacuolated as the wave front approaches, and a newly formed break appears as though a vacuole has filled the entire axon diameter, completely interrupting it (Fig. 5 ). The degeneration wave sequentially affects adjacent regions of the fibre and different lesions cause this wave to progress in different directions. Figure 5 Wave front of Wallerian degeneration in a YFP labelled wild-type axon after crush lesion A : The partially degenerated axon that is bracketed was identified in a 44 h crushed wild-type nerve. All more distal regions of this axon are fragmented and all more proximal regions are intact (data not shown). B-D : higher magnification of this axon from (A) around the transition point between intact and fragmented regions. (D) shows the most proximal breakpoint in this nerve and the inferred retrograde direction of propagation of Wallerian degeneration. Immediately proximal to the breakpoint severe vacuolation occupies almost the entire axon thickness. Slightly further proximal in (C), there are also severe YFP negative vacuoles and fragmentation appears imminent at two points (asterisks). Further proximal still in (B), the degree of vacuolation decreases. YFP fluorescence has been pseudo-coloured green with the applied confocal imaging software (Zeiss LSM Software Release 3.2). Scale bars: 50 μm (A) and 10 μm (B, C, D) Figure 6 Axonal fragmentation progresses asynchronously as a localised wave along individual axons in a anterograde or retrograde direction A-D : Graphs showing the number of axonal breaks along individual YFP labelled axons with anterograde gradient of fragmentation in relation to the distance in mm from the transection point 37 h (A), 40 h (B), 41 h (C) and 42 h (D) after cut lesion. Note that with increasing distance from the transection, axon lengths with marked fragmentation abruptly change into lengths with no or just a few axonal breaks, indicating that Wallerian degeneration progresses with a localised fragmentation wave front. Additionally note the variable localisation of the fragmentation wave front along different axons at one timepoint representing the asynchronity of Wallerian degeneration among the axon population. E-H : Graphs showing the number of axonal breaks along individual YFP labelled axons with retrograde gradient of fragmentation in relation to the distance in mm from the crush point 40 h (E), 42 h (F), 43 h (G) and 44 h (H) after crush lesion. Note that with increasing distance from the crush point axon lengths without any features of fragmentation abruptly change into lengths containing axonal breaks. Asynchronity of progression of Wallerian degeneration along individual axons is also apparent after crush lesion. In order to determine whether the anterograde and retrograde fragmentation wave runs at the same velocity along the axon we next estimated the rate of progression. As the average axon length measured was 24 mm and the majority of axons must have entirely fragmented between 41 h and 42 h after transection (significant difference between percentage of entirely degenerated axons at 42 h and entirely plus partially degenerated axons at 41 h in Student t-test) (Fig. 2A ), the minimal velocity for the degeneration wave is 24 mm/h. Analogously, after crush lesion the calculated velocities of the retrograde degeneration wave is also at least 24 mm/h (Fig. 2B ), as the majority of axons fragmented between 43 h and 44 h after the lesion (significant difference between entirely degenerated axons at 44 h and entirely plus partially degenerated axons at 43 h in Student t-test). Thus, the rates of Wallerian degeneration progression are similar or possibly even equal in these opposite directions, but with a faster initiation of the fragmentation wave after transection. We then tested whether our crush lesions interrupted axon continuity as in nerve transection, because a failure to do so could underlie the different direction of propagation in crushed nerves (see Discussion). In fluorescent wholemount preparations of crushed nerve segments we found continuous longitudinal YFP signals across the crush site, and in teased fibre bundles of osmificated nerve segments after crush injury, a majority of fibres remained continuous across the crush site (see additional data file Add Fig 1.pdf ). These data are consistent with the axonal membrane remaining intact after 30 sec nerve crush, unlike that of a transected nerve. The observation that at early time points after cut and crush injury some axons had already fragmented or started to fragment while the majority is still intact (Fig. 1 , 2 ) together with the variable localisation of the fragmentation wave front along different axons at one time point (Fig. 6 ) indicates that Wallerian degeneration is asynchronous among the population of axons in a peripheral nerve. This probably reflects both differences in the timing of onset of degeneration and varying velocities of propagation in axons of different thickness that cannot be distinguished by our imaging approaches. Furthermore, the observation that the transition between intact and degenerated regions can be 19–21 mm distal to the crush within 44 h rules out regeneration as a possible source of error. Quantification of axonal breaks along entirely fragmented axons (Fig. 7 ) revealed that fragmentation is homogenously dispersed through the whole fibre distance and no gradient is detectable. Thus, once fragmentation begins it is rapidly completed. All these findings obtained in YFP-H mice are summarized schematically in Fig. 13 . Figure 7 The Wallerian degeneration wave runs through individual axons and leaves uniformly degenerated fibres without gradients of fragmentation A : Confocal composite picture showing six consecutive lengths (from top to bottom in overview) of the proximodistal course of an individual completely fragmented YFP labelled axon within a peripheral nerve stump 42 h after transection injury without any features of a degeneration gradient. Note that this axon has fragmented in its proximal (upper inset) and distal (lower inset) site equally. Axonal fragments are clearly demarcated by fluorescence interruptions (arrows in insets). YFP fluorescence has been pseudo-coloured yellow with the applied confocal imaging software (Biorad LaserSharp 2000). Scale bar: 500 μm B, C : Graphs showing the number of axonal breaks along 10 YFP labelled axons without apparent gradient of fragmentation in relation to the distance in mm from the cut point 37 h to 42 h after cut lesion. Means and standard deviations are presented in (B). Note that axonal breaks and therefore fragmentation is homogenously dispersed through the axon lengths. D, E : Graphs showing the number of axonal breaks along 10 YFP labelled axons without apparent gradient of fragmentation in relation to the distance in mm from the crush point 40 h to 44 h after crush lesion. Means and standard deviations are presented in (E). Note that axonal breaks and therefore fragmentation is homogenously dispersed through the axon lengths. Figure 13 Schematic illustration depicting the spatiotemporal pattern of axon degeneration after cut and crush injury of a wild-type and a Wld S peripheral nerve. Each yellow line represents an individual YFP positive axon in wild-type (A, B) and Wld S (C, D) peripheral nerves. Accounting for wild-type peripheral nerves, firstly, both after transection (A) and crush injury (B) axonal fragmentation progresses as a localised wave quickly within a matter of few hours over the individual axon. Thereby, the abrupt shift between preserved and fragmented axon distances along partially fragmented axons represents the wave front. The processes differ only in direction with an anterograde course after cut and a retrograde course after crush lesion. Secondly, axonal fragmentation in the YFP positive axon population is asynchronous with some intact and others entirely or partially fragmented in one nerve at one time point. Thirdly, axonal breaks are dispersed homogenously along totally fragmented fibres. In contrast, in Wld S peripheral nerves, firstly, both after transection (C) and crush (D) injury axonal degeneration progresses in anterograde direction with a velocity similar to that of slow axonal transport. Secondly, the gradients of axon degeneration are uniform with gradual decrease of degenerative changes along the axon from proximal to distal. Thirdly, degeneration happens broadly synchronously among the population of Wld S axons. Fourthly, formation of end bulbs with subsequent swellings at the proximal ends of Wld S axons can be observed especially after crush lesion but also occasionally after transection lesion. In contrast to wild-type nerves injured Wld S sciatic and tibial nerves degenerate anterogradely independent of the lesion type We then extended these studies to Wld S axons, already known to survive 14 days after transection lesion [ 13 ], using light and electron microscopy after prolonged lesion times of 15, 20, 25 and 30 days. In contrast to the analogous experiment in wild-type mice, a significant difference in axon preservation rate was immediately apparent between the proximal sciatic nerve and distal tibial nerve. 20 days after high sciatic nerve transection, 28.1 % of myelinated axons were structurally preserved a few millimetres distal to the lesion site in light and electron microscopy (Fig. 8A, B, C ) but the most distal part of the tibial nerve showed ultrastructural preservation in 85.0 % of axons (Fig. 8A, D, E ). Likewise, at all further time points beside 20 days (15, 25 and 30 days) after transection lesion we found more intact axons in distal tibial nerve than in proximal sciatic nerve close to the point of injury (Fig. 8A ). Overall, these results clearly indicate anterograde progression of axon degeneration along transected Wld S peripheral nerves. Figure 8 Light and electron microscopy revealed an exclusively anterograde gradient of axon degeneration in transected and crushed Wld S sciatic/tibial nerves after prolonged lesion times A, F : Quantification of axon preservation at proximal and distal ends of the peripheral nerve stump after transection (A) and crush (F) injury exposed exclusively anterograde gradients of axon degeneration after 15 to 30 days following injury (15 d lesion time-point only after transection injury). Differences in the number of protected axons between the proximal and distal end of the stump were maximum after 20 days and more moderate prior or later to that, correspondingly. Remarkably, after 30 days following crush lesion considerable numbers of totally intact axons could be counted (63.5 % in distal tibial nerve) pointing to a weaker effect of compression over transection and generally to the longevity of distal Wld S axons. B-E : Light microscopic images (B, D) and corresponding electron micrographs (C, E) taken from the proximal (B, C) and distal (D, E) end of the peripheral nerve stump after 20 days following transection lesion. At the proximal end (sciatic nerve) 28.1 % myelinated axons were structurally preserved while at the distal end (tibial nerve) we could observe 85.0 % preserved axons pointing to an anterograde gradient of axon degeneration. G-J : Light microscopic images (G, I) and corresponding electron micrographs (H, J) taken from the proximal (G, H) and distal (I, J) end of the peripheral nerve stump after 20 days following compression lesion. Similar to the transection lesion also here we identified a clear anterograde degeneration gradient with 70.0 % intact axons at the proximal end and 94.8 % preserved axons at the distal end of the nerve stump. Magnification of light microscopy is 630 × and electron microscopy is 3400 × Remarkably, in view of data reported in wild-type mice, a crush injury of the proximal sciatic nerve also resulted in anterograde progression in Wld S sciatic/tibial nerve that was evident with a more moderate gradient after 20–30 days. Twenty days after sciatic nerve crush 70.0 % of all myelinated axons were preserved a few millimetres distal to the lesion (Fig. 8F, G, H ) while 94.8 % of these fibres were preserved at the distal tibial nerve more than 20 millimetres away from the point of lesion (Fig. 8F, I, J ). Similarly, 25 and 30 days after crush lesion we observed an increase in preserved axon numbers from proximal to distal along the sciatic-tibial nerve distance (Fig. 8F ). Even 30 days after transection or crush lesion there were many protected distal axons in tibial nerve (Fig. 8A, F ). By contrast, the degeneration of all distal fibres in wild-type mice was complete within two days (see above). Thus, while axon degeneration in more proximal regions is delayed approximately tenfold by Wld S after a lesion, the delay in more distal regions is at least twenty fold. In summary, quantification of axon preservation assessed by light and electron microscopy is sufficient to indicate a marked anterograde direction of axon degeneration both after transection and crush injury of peripheral Wld S nerves. Anterograde degeneration of individual YFP labelled Wld S axons is slowly progressive In order to exclude nerve branching as an explanation for the observations above, we carried out detailed longitudinal analysis of individual degenerating Wld S axons labelled with the YFP-H transgene as described previously [ 54 ]. Following transection or crush lesions to the sciatic nerve, long-range confocal YFP axon tracing was performed in 2–3 cm wholemount nerve segments. 15 days after transection almost all (96.4 ± 3.4 %) YFP labelled Wld S axons showed a homogeneous anterograde gradient of degenerative changes along the sciatic and tibial nerve (Fig. 9 ). An example is presented in Fig. 10 . In contrast to partially degenerated wild-type axons after transection lesion, where there were clear interruptions between markedly demarcated YFP-positive fragments, axon fragmentation in proximal Wld S nerves after 15 days was mostly incomplete. Instead of interruptions, there were many constrictions of short regions of the axon or thin axoplasmatic bridges between thicker regions (inset 1 + 2 in Fig. 10 ). Occasionally we observed small swellings (bulbs) at the proximal ends (data not shown). Distal areas of the same Wld S axon lacked these degenerative changes (inset 3 + 4 in Fig. 10 ). The remaining 3.6 % of Wld S axons appeared to be completely intact without any discernible signs of degeneration (Fig. 9 ). 20 days after transection, proximal regions of individual Wld S distal axon stumps were more completely fragmented (inset 1 in Fig. 11 ). However, some millimetres distal the fragments became gradually less frequent and again were often joined by YFP positive material indicating incomplete fragmentation (inset 2 in Fig. 11 ). At this time point such incomplete fragmentation with axonal narrowing occasionally continued up to distal regions of the fluorescent Wld S axons (inset 3 in Fig. 11 ). Altogether at 20 days post operation all axons showed anterograde gradients of complete or incomplete fragmentation (Fig. 9 ). Additionally, we analyzed individual axons that were separated for 12.5, 17.5 and 22.5 days from their parent cell body and found that with increasing lesion time proportionally more axons displayed anterograde degeneration gradients. The gradients became structurally clearer through more marked demarcation of axonal fragments (data not shown). Thus, the initial morphological events in the degeneration of Wld S axons are constrictions or atrophy, followed only considerably later by complete interruptions of the axon. Figure 9 Quantification of fluorescent Wld S axons in whole-mounted peripheral nerve stumps from triple heterozygote mice after transection and crush injury at different time points. Partially degenerated YFP positive Wld S axons that could be identified 15 and 20 days either after transection or crush injury showed axonal constrictions or interruptions in their proximal site but not further distal indicating an anterograde gradient of degeneration. They were assigned to the group "fragmented with anterograde gradient". Few entirely preserved fluorescent Wld S axons could be only seen 15 days after transection and crush injury. They were assigned to the group "intact". The chart presents means and standard deviations. Figure 10 Anterograde degeneration of transected Wld S axons initially involves proximal axonal atrophy with occasional interruptions. Confocal composite picture showing eight consecutive lengths (from top to bottom in overview) of the proximo-distal course of an individual YFP labelled Wld S axon within a peripheral triple heterozygote nerve stump 15 days after transection injury displaying an anterograde progression of axon degeneration. Note that this axon shows predominantly narrowings (red asterisks) and occasionally interruptions (white arrows) in its most proximal end (inset 1) with a gradual decrease of this degeneration signs over a few millimetres more distal (inset 2) while at its distal parts almost no degeneration can be identified (inset 3 and 4). YFP fluorescence has been pseudo-coloured yellow with the applied confocal imaging software (Biorad LaserSharp 2000). Scale bar: 500 μm Figure 11 Anterograde degeneration of transected Wld S axons eventually continues with complete proximal fragmentation. Confocal composite picture showing six consecutive lengths (from top to bottom in overview) of the proximo-distal course of an individual YFP labelled Wld S axon within a peripheral triple heterozygote nerve stump 20 days after transection injury demonstrating a clearer anterograde progression of axon degeneration than in Fig. 10. Note that this isolated axon shows complete break-up (white arrows) with clearly demarcated fragments in its most proximal part among a minority of axonal narrowings (red asterisks) (inset 1). Moving further distal fragmentation accompanied by axonal constrictions becomes gradually weaker (inset 2, 3) while at its most distal end almost no degeneration can be identified (inset 4). YFP fluorescence has been pseudo-coloured yellow with the applied confocal imaging software (Biorad LaserSharp 2000). Scale bar: 500 μm We then followed up the earlier EM experiments using crushed nerves from Wld S /YFP-H double mutant mice to study the directionality of axon degeneration in Wld S nerves using this very different method. Once again, the spatio-temporal pattern of axon degeneration after crush injury in Wld S mice was very similar to that after transection injury, contrasting with wild-type mice where directionality depends on lesion type. Correspondingly, 15 days after crush lesion we counted 84.60 % of all YFP labelled Wld S axons with an anterograde gradient of degenerative changes (Fig. 9 ). An example of such an individual axon is shown in the additional data file Add Fig 2.pdf . However, crushed Wld S axons more frequently showed end bulbs at the proximal end of the distal stump, which were often very large, (red arrow in overview of Add Fig 2.pdf ) and subsequent multiple axonal swellings (inset 1 in Add Fig 2.pdf ). This feature was far more prominent in crushed Wld S axons than in the transection experiment where we observed end bulbs just occasionally. Further distally these swellings disappear with remaining axonal constrictions and breaks (inset 2 in Add Fig 2.pdf ) representing incomplete fragmentation. As in transected nerves, distal parts of the crushed Wld S axon were free of degeneration signs (inset 3 and 4 in Add Fig 2.pdf ). Compared to transection lesion, more axons at 15 days remained entirely intact (15.4 %) (Fig. 9 ). By 20 days after nerve crush of Wld S axons, proximal fragmentation became more prominent with fully separated fragments (inset 1 in additional data file Add Fig 3.pdf ) while more distal regions of the same axons were again incompletely fragmented (inset 2 in Add Fig 3.pdf ) and further distal still lacked any degeneration signs (insets 3 + 4 in Add Fig 3.pdf ). The proximal end bulbs and localised swellings were larger than at 15 days (red arrow in overview of Add Fig 3.pdf ), possibly due to continued accumulation of retrogradely transported material. In summary, quantification 20 days after crush lesion revealed that all axons showed anterograde gradients of complete or incomplete fragmentation (Fig. 9 ). Thus, the YFP-H studies confirmed our light and electron microscopy observations that delayed degeneration in individual Wld S axons is directional with an exclusively anterograde pattern both after transection and crush injury. This pattern of degeneration is qualitatively different from that in wild-type mice, which shows asynchronous, bidirectional fragmentation and degeneration. Wld S axons show a continuous gradient of axon degeneration that moves with a velocity similar to that of slow axonal transport In the above experiments we noted the gradual change from degenerated regions to intact regions in Wld S axons. In contrast to our observations in wild-type mice, there was no clearly delineated boundary or wave front separating degenerated and fully intact regions. In order to quantify this, we counted the number of axonal constrictions and breaks along the length of YFP positive Wld S axons following injury. At all post-lesion time-points randomly chosen axons showed gradually decreasing signs of degeneration (constrictions or interruptions) along their length (Fig. 12 ). This markedly contrasts with the wave-like degeneration observed in wild-type mice where a sharp boundary divided preserved regions of the axon from completely fragmented regions. The closely superimposed curves shown in Fig. 12A, C, E, G also indicate that anterograde axon degeneration is more synchronous among the axon population in Wld S nerves at both 15 and 20 days following transection or crush injury. Only rarely could we observe morphologically normal Wld S axons adjacent to axons with anterograde gradients of degeneration (Fig. 9 ). This quantification allowed us to estimate the rate at which anterograde degeneration progresses along single Wld S axons. Between 15 and 20 days the equivalent stage of degeneration has advanced up to 11 mm further along the nerve (sometimes less), giving a maximum velocity of Wld S degeneration progression of 11 mm / 5 days = 2.2 mm/day. This is similar to the reported velocity of slow axonal transport (0.1–3.0 mm/day) [ 56 - 59 ]. All findings concerning topology of axonal degeneration in Wld S peripheral nerves are summarized schematically in Fig. 13 . Figure 12 Progression of axon degeneration in shape of a continuous degeneration gradient appears roughly synchronous along individual Wld S axons A-H : Graphs showing the number of axonal constrictions and breaks along individual YFP labelled Wld S axons with an anterograde gradient of degeneration in relation to the distance in mm from the transection point 15 days (A, B) and 20 days (C, D) after transection lesion or from the crush point 15 days (E, F) and 20 days (G, H) after crush lesion. Means and standard deviations are presented in B, D, F, H. Note that with increasing distance from the transection and crush point degeneration signs decrease uniformly characterized by the steady decline of the curves. Moreover, degeneration in different Wld S fibres is broadly synchronous as shown by the good superimposition of individual curves in A, C, E, G. Discussion We have shown that the fragmentation of axons undergoing Wallerian degeneration in a mixed wild-type peripheral nerve is a rapid, asynchronous and progressive process. By using a recently developed method to visualise individual axons over cm-long distances, and by targeting a short critical period during which nearly all axons degenerate, we have made the first observations in vivo of partially fragmented individual axons and thus determined the directionality and the wave-like nature of Wallerian degeneration, as well as estimating its velocity. Furthermore, we have shown that nerves of Wld S mutant mice undergo a fundamentally distinct process rather than simply following the same pathway in slow motion. Lesioned wild-type axons remain morphologically normal for a latency period of ca. 36–44 hours, which depends on lesion type and individual axonal properties [ 38 , 39 , 44 , 46 , 60 ]. Each axon then undergoes a catastrophic process in which at least 24 mm of the distal stump fragments entirely within an hour. The propagation rate of at least 24 mm/h is considerably faster than reported in rat dorsal column (3 mm/h) [ 47 ], rat phrenic nerve (up to 10.4 mm/h) [ 46 ] and in primary culture (ca. 0.4 mm/h) [ 50 ], probably reflecting differences in neuronal subtype and context. For example, slower propagation of Wallerian degeneration in the CNS is suggested by the observation of an anterograde spread in the gracile tract following a dorsal root lesion, whereas the spread of degeneration within the root itself was too short-lived to be resolved by the methods used [ 47 ]. Wallerian degeneration may also propagate more slowly in longer axons, which could account for differences between mice and rats [ 44 , 45 ], and there may be many reasons why the propagation rate in vitro could differ from that in vivo. Nevertheless, while the propagation rate may differ, the anterograde degeneration after axon transection is a consistent feature of each of these studies. We have shown that Wallerian degeneration in wild-type nerves is a wave-like process that can travel in either direction along the axon, depending on lesion type. 29 partially fragmented axons were observed, and all showed a sharp boundary between fragmented and non-fragmented zones, such that all axon regions up to the wavefront were degenerated and all regions beyond it remained intact. Fragmentation had reached different points along the nerve in different individual axons (Fig. 6 ), reflecting asynchronicity of the onset and rate of degeneration. The wave-like propagation of Wallerian degeneration has been proposed before [ 44 , 46 , 47 , 49 , 61 ], and especially Lubinska [ 44 , 46 ] has shown that Wallerian degeneration of the distal stump progresses centrifugally by jumping from one internode to another, but this is the first time the wavefront has actually been observed. Of the 29 partially fragmented axons, 17 transected axons were fragmented only at their proximal ends and 12 crushed axons were fragmented only at their distal ends (Fig. 2 , 3 , 4 ). Differences between cut and crushed nerves have been suggested before [ 48 , 53 ] and axons cut at both ends also exhibit a retrograde degeneration component [ 44 ], but this is the first demonstration that two different lesions at the same site in the same nerve cause different directions of degeneration. The mechanistic basis of this surprising observation remains unknown, but some models are outlined below. We discuss here two models to explain the wave-like propagation of Wallerian degeneration in wild-type nerves: one based on fast axonal transport and the other based on calcium influx (see also Fig. 14 ). Numerous reports have proposed that the clearance of a supportive or trophic factor by fast axonal transport processes underlies the anterograde direction of Wallerian degeneration after transection [ 33 , 44 , 46 , 47 , 62 ] based on the observation that anterograde fast axonal transport of proteins continues after axotomy in the peripheral nerve stump in a wavelike manner. Such a factor could be an inhibitor of an axonal destruction programme, likely to be stabilised or upregulated by a downstream effector of the Wld S protein. The fastest reported components of axonal transport move at around 14–25 mm/h, but it is reasonable to expect that some minor, thus far undetected, components may move faster [ 63 , 64 ]. This is just compatible with the spread of the fragmentation wave at a minimum of 24 mm/h that we observe. However, Wallerian degeneration could progress even faster, too fast to be accounted for by fast axonal transport, and anterograde clearance of a factor inhibiting Wallerian degeneration could not explain the retrograde degeneration that we found in crushed nerves. Figure 14 Two models to account for the progressive nature of Wallerian degeneration after transection lesions in wild-type axons. (A) A putative inhibitor of intrinsic self-destruction machinery is constantly delivered from the cell body to the unlesioned wild-type axon (top). After axon transection the inhibitor is no longer supplied and is cleared first from proximal regions of the distal stump by fast axonal transport. This leads to a wave of fragmentation moving proximal to distal along the isolated axon stump. (B) In an alternative model, the wave of fragmentation is propagated not by directional removal of a putative inhibitor but by rapid localised influx of calcium ions beginning at the most vulnerable part of the axon. Once inside, calcium ions not only activate calpains to degrade the local axoplasm, but also diffuse and exceed the threshold of calpain activation in the immediately adjacent region. This leads to further axoplasmic and membrane breakdown and further calcium influx. The pattern is repeated to generate a wave of fragmentation moving along the axon. Model (A) has the attraction that the putative inhibitor would be a good candidate for mediating of the Wld S phenotype (e.g., it could be overexpressed in Wld S ), while model (B) more easily explains why the directionality is reversed in a crush lesion. The calcium influx and diffusion wave could spread also retrogradely (not shown) if the distal end were the first to disintegrate. In model (A), however, it is hard to see how retrograde axonal transport could explain the depletion of an inhibitor that ultimately has to come from the cell body (see text for more details). Thus, we consider also an alternative model in which short regions of the axon membrane become permeabilised to calcium ions and this feature moves rapidly as a wave along the axon. Fibre degeneration requires accumulation of axoplasmic calcium [ 65 , 66 ], which probably activates the cystein protease calpain [ 67 - 70 ]. Once inside the axon, calcium ions could diffuse to the immediately neighbouring axoplasm and activate calpain, leading to degradation of axoplasmic and membrane proteins, and thus permeabilisation of the next segment. Such a membrane associated Ca 2+ influx hypothesis was proposed by Schlaepfer for the first time [ 71 , 72 ] and developed further involving calcium channels in more modern studies [ 73 , 74 ]. LoPachin and Lehning [ 75 ] reported calcium entry linked to membrane depolarisation through reverse Na 2+ -Ca 2+ -exchange, leading to a steady rise in intra-axonal calcium and calcium accumulation has been observed beneath Schmidt Lantermann clefts at distal sites 4 h after injury [ 76 ]. Once the threshold for calpain activation is exceeded, a wave of degeneration could be initiated and then propagate rapidly in either direction as outlined above. There are several possible reasons why degeneration may begin proximally in a transected nerve but distally in a crushed nerve. The proximal end of a transected distal stump is especially vulnerable because of the exposure of the axoplasm to the external ionic environment, and because all extra-axonal structures that normally support the axon have been totally disrupted, e.g., blood vessels, Schwann cells, extracellular matrix, perineurium. One of these factors may cause a calcium entry wave to begin at this point. In contrast, intact endoneural blood vessels can be found close to a nerve crush [ 77 , 78 ], the epi- and perineurium tubes are maintained at the site of crush [ 78 - 81 ], and some nerve crush protocols do not break axon continuity [ 82 - 85 ], so that Wallerian or Wallerian-like degeneration occurred only many days after compression or not at all. More specifically, even application of longer high pressure injuries with a minimum of shear forces may squeeze out axoplasm into adjacent parts of the axon rather than interrupting the axolemma preserving nerve conduction monitored electrophysiologically [ 82 ]. In our nerves we observed continuous longitudinal YFP signals across crush sites immediately after lesioning, indicating that at least some axons were not transected by the direct effect of crushing. We also observed many preserved fibres crossing the crush site when we fixed and then partially teased crushed nerves to generate small bundles where individual fibres were easily identifiable (see additional data file Add Fig 1. pdf ). Thus, the proximal end of the distal stump may be less vulnerable than after transection, and fragmentation may begin instead at the distal end because this is the hardest part to supply with everything the axon needs to survive. We observed a series of differences in the pattern of Wallerian degeneration in Wld S nerves that are incompatible with delayed axon degeneration following a similar mechanism to Wallerian degeneration in wild-type nerves, only slower. The spread of degeneration along Wld S nerves is around 100-fold slower, axon degeneration is more synchronous, at least relative to how long it takes to occur, it progresses in a proximal to distal direction in crushed nerves as well as transected nerves, there is a continuous gradient of degeneration along the length of the axon rather than an abrupt change at a boundary, and the first sign of axon degeneration is a constriction rather than a complete interruption. We therefore propose that the ultimate degeneration of axons in Wld S mice be termed "slow anterograde axon decay" rather than Wallerian degeneration as summarised schematically in Fig. 13 . Based on these differences, we propose that injured Wld S axons eventually undergo a passive process of atrophy, rather than an active process of self-destruction similar to apoptosis that appears to take place in wild-type axons [ 5 , 7 , 15 , 16 , 86 ]. It is likely that preserved axonal proteins will eventually be degraded by catabolic processes and may not be replaced by significant new synthesis, even if Wallerian degeneration is completely prevented. A direct indication of this is our observation in primary neuronal cultures of significant atrophy of distal neurites when their degeneration is delayed by Wld S (data not shown). Even the fact that some proteins are synthesised locally in axons [ 57 , 87 - 89 ] may not be sufficient to prevent the eventual depletion of protein in severed Wld S axons, as it remains unclear which proteins are made there and in what quantities. We discuss two models to explain the slow anterograde progression of degeneration along Wld S axons: one based on slow axonal transport and the other on a temperature gradient along the limb. The gradual nature of axonal atrophy in Wld S makes it difficult to be precise about the rate at which it progresses along the axon, but it is certainly not incompatible with the velocity of slow axonal transport of 0.1–3.0 mm/day [ 56 - 58 , 90 ]. Clearance of structural proteins by slow anterograde transport, added to their gradual depletion by protein turnover, could cause the protein content at the proximal end of the distal stump to drop below the threshold level needed to maintain axon integrity. Bidirectional transport of cytoskeletal components continues in transected Wld S nerves, leading to localised neurofilament-depleted constrictions and terminal and intermediate swellings containing disorganised neurofilaments [ 91 , 92 ]. We have made similar observations in YFP labelled Wld S axons, and additionally report a gradient of such features along the nerve. A net anterograde movement of cytoskeletal proteins could therefore underlie the anterograde gradient of axonal atrophy in Wld S axons injured for many days. Alternatively, a proximal-distal decreasing gradient of temperature along the limb could underlie the observed difference in degeneration rate at different points in the Wld S nerve. A decrease in temperature has been shown to delay degeneration both in wild-type and Wld S axons after injury [ 52 , 74 , 93 - 96 ]. In wild-type nerves, a temperature gradient explains neither the different directions of propagation after transection and crush injury, nor the sharp boundary between intact and degenerated regions. However, in Wld S axons there is a proximal to distal gradient of degeneration regardless of lesion type and the change from intact to degenerated is a gradual one. Thus, a temperature gradient could play a more important role here. The extremely long survival of distal tibial nerve following injury in Wld S is in marked contrast to the presynaptic nerve terminal at the neuromuscular junction, which is the first structure to degenerate in both wild-type and Wld S nerves [ 2 , 8 , 46 ]. Intramuscular nerve also degenerates early, at least in Wld S heterozygotes (L. Fan and R.R. Ribchester, unpublished). This supports the hypothesis of compartmentalised degeneration mechanisms of axons and synapses [ 97 ] and suggests that a clear boundary exists between the two domains. The location and nature of this boundary could hold important clues for determining the mechanism of both Wallerian degeneration and synapse degeneration. Finally, the methods we report here could now be applied to study spontaneous nerve degeneration in 'dying-back' disease. The 'dying-back' model also predicts the transitory existence of partially degenerated axons, but as in Wallerian degeneration such axons have never been directly observed and there is no indication of the speed of 'dying-back' of individual axons [ 98 , 99 ]. There are interesting parallels between axon degeneration after nerve crush and 'dying-back', as both can be delayed by the Wld S gene and the direction of degeneration is also shared [ 24 ]. Thus, it is an intriguing possibility that the speed of propagation is equally rapid and asynchronous. If it is a similarly catastrophic event, what can stop it progressing back to the cell body leading to neuron death? In some cases neuronal death does appear to be the outcome [ 25 , 100 ], whereas in others, proximal axons and their cell bodies somehow survive [ 101 ] and it is important to find out why. Future prospects include direct observation of the progression of a degeneration boundary along YFP labelled axons in vivo after nerve lesion or in disease, once methods for in vivo imaging of single axons become more refined and more readily available, e.g., Cell ViZio, Mauna Kea Technologies [ 102 ]. Conclusions In summary, we report the first direct observation of partially degenerated single axons in lesioned nerves undergoing Wallerian degeneration, indicating that Wallerian degeneration propagates in wild-type nerves as a wave whose speed is at least as fast as the highest reported rate of fast axonal transport. It could be faster still in mouse sciatic and tibial nerve. The direction of degeneration is proximal to distal after a cut, but the reverse after a crush. For now the mechanism remains unknown, but these observations will ultimately need to be explained in any comprehensive model of the Wallerian degeneration mechanism. Injury-induced axon degeneration in Wld S nerves is also progressive, but differences in the topographic pattern and morphology of degeneration indicate a fundamentally different process from that in wild-type nerves. We propose that Wld S axons ultimately undergo atrophy, in a passive process similar to that which Wallerian degeneration was once thought to be. Methods Crossbreeding and genotyping of transgenic mice Crossbreeding and genotyping of the YFP-H line [ 55 ] and triple heterozygote mice carrying the original Wld S mutation, the transgenic Wld S mutation (tg- Wld S ) [ 13 ] and the YFP-H gene was performed as previously described [ 54 ]. Triple heterozygote mice rather than Wld S /YFP-H mice were used purely for reasons of convenience in their breeding. They express similar levels of Wld S protein as homozygous natural mutant Wld S mice and display a similar retarded time-course of axon degeneration. They additionally express YFP in approximately 3% of myelinated motor and sensory fibres in the PNS, equal to the original YFP-H line obtained from the Jackson Laboratories. Sciatic nerve lesions Six- to 10-week-old mice from the YFP-H line and triple heterozygote mice for the second part of the study were anaesthetised by intraperitoneal injection of Ketanest (5 mg/kg; Parke Davis) and Rompun (100 mg/kg; Bayer), sciatic nerves were transected or crushed firmly close to the Foramen intrapiriforme and the wound was closed with a single suture. Complete nerve transection was performed with conventional surgical scissors and crush lesion was achieved with fine watchmaker's forceps (model: micrscopic forceps bent No. 7, Aesculap BD 333R, Germany) for 30 seconds. The continuity of the sciatic nerve was always preserved after crush lesion as checked in situ. For light and electron microscopy we removed distal ~2.5 centimeter long nerve stumps after 15 (transection only), 20, 25 and 30 days in triple heterozygote mice following intracardial perfusion. The first 2 millimeters of the distal nerve stump were discarded to reduce artifacts, and the next 2 mm from the most proximal and distal end of the peripheral nerve stump was prepared for Durcupan embedding and electron microscopy. For conventional fluorescence microscopy and confocal tracing of individual YFP labelled wild-type axons after 34 h, 37 h, 40 h, 41 h, 42 h and 48 h following cut lesion and 37 h, 40 h, 42 h, 43 h, 44 h and 48 h following crush lesion the operated YFP-H mice were sacrificed by cervical dislocation and nerve segments prepared as follows. For confocal tracing of individual YFP labelled Wld S axons from triple heterozygote mice we dissected sciatic-tibial nerve segments after 12.5 d (transection only), 15d, 17.5 d (transection only), 20 d and 22.5 d (transection only). For each investigated time-point 2 – 3 YFP-H or triple heterozygote mice were operated. Assessment of axonal continuity in crushed YFP-H nerve segments Sciatic nerves of mice from the YFP-H line were crushed as described in the paragraph above and the short nerve segment containing the crush site immediately removed after lesion, freed from surrounding connective tissue and subsequently either prepared for wholemount fluorescence embedding as described previously [ 54 ] or for tissue osmification. For the latter the crushed segment was immersion-fixed for four hours in 10% PFA in 0.1 M PBS, washed three times in 0.1 M PBS for 10 minutes and osmificated in aqueous osmium-tetroxide (1 %) solution for 90 minutes. After rinsing in fresh 0.1 M PBS individual axon bundles were teased from the crushed segment using fine syringe needles (Neoject 26 G × 1/2, Dispomed WITT, Germany) and mounted on conventional glass slides. Fluorescence imaging of crushed YFP-H nerve segments and light microscopy of osmificated fibre bundles was carried out using an Olympus IX 81 inverted microscope coupled to a Olympus U-TV0.5XC digital camera system. Intracardial perfusion for light and electron microscopy of semithin and ultrathin preparations After sternotomy under deep anaesthesia mice were killed by cardiac puncture and instantly intracardially perfused first with a solution containing 10 000 i.e./l heparin (Liquemin N 25 000, Hoffmann-La Roche) and 1 % procainhydrochloride in 0.1 M PBS for 30 s and then with fresh half-strength Karnovsky's fixative (4 % paraformaldehyde, 2 % glutardialdehyde in 0.1 M sodium cacodylate, pH 7.3). Light and electron microscopy Nerve samples for light and electron microscopy were embedded in Durcupan and further processed for examination with a Zeiss Axiophot light microscope and Zeiss EM 902 electron microscope as described previously [ 54 ]. Morphological quantification of axon preservation in light and electron micrographs The percentage of preserved myelinated axons in proximal and distal ends of peripheral nerve stumps from triple heterozygote mice after transection and crush injury was determined as described previously [ 13 , 54 ]. Conventional fluorescence microscopy and confocal tracing of individual YFP labelled axons after various lesion times In both YFP-H and triple heterozygote mice after cut or crush lesion the entire nerve distances from the proximal sciatic to the distal tibial nerve were carefully excised, the perineurium and the branch of the commune fibular nerve removed and the remaining ~2.5 centimeter long stumps treated for wholemount fluorescence preparation using Vectashield Mounting Medium (Vector Laboratories) as described previously [ 54 ]. For rough orientation the wholemount preparations were photomicrographed at the proximal and distal end of the excised nerve segment using a Zeiss Axiophot microscope connected to a digital camera system (Universal Imaging Corporation). High resolution confocal composite presentation of the entire intra-nerve course of individual degenerating YFP-labelled wild-type or Wld S axons running through the excised stumps over their whole length was achieved as described previously [ 54 ]. Single images were obtained with LaserSharp 2000 software connected to a Biorad Radiance 2000 laser scanning system (Hemel Hempsted, UK) and composite pictures on a black background were created using Adobe Photoshop. Imaging of restricted transition zones dividing intact and fragmented axon regions in YFP-H nerves was performed under highest possible resolution with a Zeiss LSM 510 META confocal system (LSM Software Release 3.2) coupled to a Zeiss Axiovert 200 microscope. Quantification of intact and degenerated YFP labelled wild-type and Wld S axons in peripheral nerve stumps after cut and crush injury For the first part of the study focusing on sciatic/tibial nerves from the YFP-H line in each wholemount preparation (2 – 3 nerve preparations per time point) YFP labelled wild-type axons running continuously through the excised peripheral nerve stump after cut and crush injury were traced individually with the laser scanning confocal microscope under high resolution and divided into four groups: axons without any features of axonal disintegration were assigned to the group "intact". Axons that showed fragmentation of the longitudinal YFP signal over the whole length of the fibre were assigned to the group "entirely fragmented". Axons without any sign of fragmentation at one end but clear axon breakdown at the other end were assigned either into the group "fragmented with anterograde gradient" or "fragmented with retrograde gradient" depending whether the fragmentation appeared close to the lesion point or at the most distal point of the tibial nerve segment. For the second part of the study dealing with triple heterozygote mice YFP labelled Wld S axons in wholemount peripheral nerve stumps (2–3 nerve preparations for each timepoint) after transection and crush injury were equally traced individually with the BioRad Radiance 2000 laser scanning confocal microscope under high resolution, and divided into two groups: axons without any features of axonal disintegration were counted and assigned to the group "intact". Fibres that showed degeneration signs like constrictions or interruptions of the longitudinal YFP signal at the proximal site were associated into the group "fragmented with anterograde gradient". Means and standard deviations for all experiments were calculated using Microsoft Excel. Quantification of axonal degeneration signs along YFP positive axons The length of degenerated YFP labelled wild-type and Wld S axons was measured during the tracing process with the confocal laser scanning microscope and the distribution of axonal interruptions (both in wild-type and Wld S axons) and/or constrictions (only in Wld S axons) was graphed against the distance in mm from the cut or crush point. Means and standard deviations were calculated using Microsoft Excel. Animal experiments All animal experiments were carried out under appropriate German licences: Tierschutzgenehmigung K 13, 11/00 and Anzeige K30/99. Authors' contributions MPC (corresponding author) and BB jointly conceived of the design for this study, interpreted the experimental results and wrote the manuscript; BB carried out preliminary light and electron microscopy experiments, confocal imaging of YFP labelled axons, axon quantification and statistical analysis; RA: collaborated on the set up of the experiments, improvement of histological techniques and interpretation of results, performed teased fibre experiments, assisted in writing the manuscript; DW and DSG: provided excellent technical assistance in many parts of the experiments, performed sciatic nerve cut lesions and carried out routine work; KA: contributed to the discussion of the experimental results, preliminary light and electron microscopy work was done in his lab; RRR: contributed to the interpretation of the experiments and writing of the manuscript; confocal microscopy of YFP axons was mostly done in his lab; MPC: performed sciatic nerve crush lesions and partially confocal imaging, supervised all aspects of the study and main work was done in his lab. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Many nerve fibres are unbroken directly after a 30 second crush as shown in YFP-H wholemount nerves and osmificated teased fibre preparations. A-C : Sciatic nerves crushed for 30 seconds at maximum pressure and then immediately fixed and imaged by osmium staining (A, B) or YFP fluorescence (C). D: YFP-H nerve crushed for 5 seconds at maximum pressure. Nerves in (A) and (B) were partially teased apart after staining, leading to accidental breakage of a few fibres. However, the majority of fibres clearly cross the crush site unbroken. YFP signal (identifying the axoplasm) at the crush site in (C-D) is weak, probably due to squeezing of the axoplasm longitudinally out of the crush site, or quenching of fluorescence by the crushed tissue, or both. To compensate, the photographs are deliberately overexposed. There is no sign of YFP positive axoplasm escaping laterally into the extracellular space, as would be expected if the axolemma were broken. Instead, some YFP positive axons clearly cross the crush site unbroken. In (C) the overexposure would prevent any broken axons from being identified, but in (D) the signal from most or all axons fades gradually as the axon enters the crush site, rather than stopping abruptly as one would expect if the axon were broken. Scale bars: 20 μm (A, B) and 50 μm (C, D). Click here for file Additional File 2 Degeneration of individual crushed Wld S axons proceeds anterogradely beginning with the formation of end bulbs and axonal swellings at the most proximal end and accompanying proximal axonal atrophy. Confocal composite picture showing seven consecutive lengths (from top to bottom in overview) of the proximo-distal course of an individual YFP labelled Wld S axon within a peripheral triple heterozygote nerve stump 15 days after crush injury displaying an anterograde gradient of axon degeneration. This axon exhibits an end bulb at the most proximal end (red arrow in overview) and subsequent multiple axonal swellings delimited by constrictions (red asterisks) (inset 1). Further distally these swellings disappear with remaining constrictions (red asterisks) and sporadic breaks (white arrow) (inset 2). Distal parts of the crushed Wld S axon are free of degeneration signs (inset 3 and 4). Scale bar: YFP fluorescence has been pseudo-coloured yellow with the applied confocal imaging software (Biorad LaserSharp 2000). 500 μm Click here for file Additional File 3 Anterograde degeneration with formation of massive end bulbs of compressed Wld S axons finally includes complete proximal fragmentation. Confocal composite picture showing six consecutive lengths (from top to bottom in overview) of the proximo-distal course of an individual YFP labelled Wld S axon within a peripheral triple heterozygote nerve stump 20 days after crush injury demonstrating a clearer anterograde progression of axon degeneration than in Add Fig 2.pdf . Likewise after 15 days following crush injury this picture shows a massive end bulb at the most proximal end of the distal stump (red arrow in overview) and subsequent multiple axonal swellings and constrictions. Complete axonal fragmentation is evident in inset 1 (white arrows mark axonal breaks) and gradually gives way to incomplete breakup with axonal constrictions (red asterisks) (inset 2). Distal parts of the crushed Wld S axon lack degeneration signs (inset 3 and 4). YFP fluorescence has been pseudo-coloured yellow with the applied confocal imaging software (Biorad LaserSharp 2000). Scale bar: 500 μm Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549193.xml

545797

The ASRG database: identification and survey of Arabidopsis thaliana genes involved in pre-mRNA splicing

The database of Arabidopsis splicing related genes includes classification of genes encoding snRNAs and other splicing related proteins, together with information on gene structure, alternative splicing, gene duplications and phylogenetic relationships.

Rationale Most eukaryotic genes contain introns that are spliced from the precursor mRNA (pre-mRNA). The correct interpretation of splicing signals is essential to generate authentic mature mRNAs that yield correct translation products. As an important post-transcriptional mechanism, gene function can be controlled at the level of splicing through the production of different mRNAs from a single pre-mRNA (reviewed in [ 1 ]). The general mechanism of splicing has been well studied in human and yeast systems and is largely conserved between these organisms. Plant RNA splicing mechanisms remain comparatively poorly understood, due in part to the lack of an in vitro plant splicing system. Although the splicing mechanisms in plants and animals appear to be similar overall, incorrect splicing of plant pre-mRNAs in mammalian systems (and vice versa) suggests that there are plant-specific characteristics, resulting from coevolution of splicing factors with the signals they recognize or from the requirement for additional splicing factors (reviewed in [ 2 , 3 ]). Genome projects are accelerating research on splicing. For example, with the majority of splicing-related genes already known in human and budding yeast, these gene sequences were used to query the Drosophila and fission yeast genomes in an effort to identify potential homologs [ 4 , 5 ]. Most of the known genes were found to have homologs in both Drosophila and fission yeast. The availability of the near-complete genome of Arabidopsis thaliana [ 6 ] provides the foundation for the simultaneous study of all the genes involved in particular plant structures or physiological processes. For example, Barakat et al. [ 7 ] identified and mapped 249 genes encoding ribosomal proteins and analyzed gene number, chromosomal location, evolutionary history (including large-scale chromosomal duplications) and expression of those genes. Beisson et al. [ 8 ] catalogued all genes involved in acyl lipid metabolism. Wang et al. [ 9 ] surveyed more than 1,000 Arabidopsis protein kinases and computationally compared derived protein clusters with established gene families in budding yeast. Previous surveys of Arabidopsis gene families that contain some splicing-related genes include the DEAD box RNA helicase family [ 10 ] and RNA-recognition motif (RRM)-containing proteins [ 11 ]. At present, the Arabidopsis Information Resource (TAIR) links to more than 850 such expert-maintained collections of gene families [ 12 ]. Here we present the results of computational identification of potentially all or nearly all Arabidopsis genes involved in pre-mRNA splicing. Recent mass spectrometry analyses revealed more than 200 proteins associated with human spliceosomes ([ 13 - 17 ], reviewed in [ 18 ]). By extensive sequence comparisons using known plant and animal splicing-related proteins as queries, we have identified 74 small nuclear RNA (snRNA) genes and 395 protein-coding genes in the Arabidopsis genome that are likely to be homologs of animal splicing-related genes. About half of the genes occur in multiple copies in the genome and appear to have been derived both from chromosomal duplication events and from duplication of individual genes. All genes were classified into gene families, named and annotated with respect to their inferred gene structure, predicted protein domain structure and presumed function. The classification and analysis results are available as an integrated web resource, the database of Arabidopsis Splicing Related Genes (ASRG), which should facilitate genome-wide studies of pre-mRNA splicing in plants. ASRG: a database of Arabidopsis splicing-related genes Our up-to-date web-accessible database comprising the Arabidopsis splicing-related genes and associated information is available at [ 19 ]. The web pages display gene structure, alternative splicing patterns, protein domain structure and potential gene duplication origins in tabular format. Chromosomal locations and spliced alignment of cognate cDNAs and expressed sequence tags (ESTs) are viewable via links to the Arabidopsis genome database AtGDB [ 20 ], which also provides other associated information for these genes and links to other databases. Text-search functions are accessible from all the web pages. Sequence-analysis tools including BLAST [ 21 ] and CLUSTAL W [ 22 ] are integrated and facilitate comparison of splicing-related genes and proteins across various species. Arabidopsis snRNA genes A total of 15 major snRNA and two minor snRNA genes were previously identified experimentally in Arabidopsis [ 23 - 28 ]. These genes were used as queries to search the Arabidopsis genome for other snRNA genes. A total of 70 major snRNAs and three minor snRNAs were identified by this method. In addition, a single U4atac snRNA gene was identified by sequence motif search. We assigned tentative gene names and gene models as shown in Table 1 , together with chromosome locations and similarity scores relative to a representative query sequence. The original names for known snRNAs were preserved, following the convention atUx.y, where x indicates the U snRNA type and y the gene number. Computationally identified snRNAs were named similarly, but with a hyphen instead of a period separating type from gene number (atUx-y). Putative pseudogenes were indicated with a 'p' following the gene name. Pseudogene status was assigned to gene models for which sequence similarity to known genes was low, otherwise conserved transcription signals are missing and the gene cannot fold into typical secondary structure. A recent experimental study of small non-messenger RNAs identified 14 tentative snRNAs in Arabidopsis by cDNA cloning ([ 29 ], GenBank accessions 22293580 to 22293592 and 22293600, Table 1 ). All these newly identified snRNAs were found in the set of our computationally predicted genes. Conservation of major snRNA genes As shown in Table 1 , each of five major snRNA genes (U1, U2, U4, U5 and U6) exists in more than 10 copies in the Arabidopsis genome. U2 snRNA has the largest copy number, with a total of 18 putative homologs identified. Both U1 and U5 snRNAs have 14 copies, U6 snRNA has 13 copies, and U4 snRNA has only 11 copies. Sequence comparisons within Arabidopsis snRNA gene families showed that the U6 snRNA genes are the most similar, and the U1 snRNA genes are the most divergent. Eight active U6 snRNA copies are more than 93% identical to each other in the genic region, whereas active U1 snRNAs are on average only 87% identical. The U2 and U4 snRNAs are also highly conserved within each type, with more than 92% identity among the active genes. Details about the individual snRNAs and the respective sequence alignments are displayed at [ 30 ]. Previous studies identified two conserved transcription signals in most major snRNA gene promoters: USE (upstream sequence element, RTCCACATCG (where R is either A or G) and TATA box [ 24 - 27 ]. All 14 U5 snRNAs have the USE and TATA box. Furthermore, their predicted secondary structures are similar to the known structure of their counterparts in human, indicating that all these genes are active and functional (structure data not shown; for a review of the structures of human snRNAs, see [ 31 ]). Similarly, we identified 17 U2, 10 U1, nine U4, and nine U6 snRNA genes as likely active genes, with a few additional genes more likely to be pseudogenes because of various deletions. U4-10 and U6-7 do not have the conserved USE in the promoter region, but their U4-U6 interaction regions (stem I and stem II) are fairly well conserved. U2-16 is also missing the USE but has a secondary structure similar to other U2 snRNAs. These genes may be active, but differences in promoter motifs suggest that their expression may be under different control compared with other snRNAs homologs. The U2-17 snRNA has all conserved transcription signals, but 20 nucleotides are missing from its 3' end. The predicted secondary structure of U2-17 is similar to that of other U2 snRNAs, with a significantly shorter stem-loop in the 3' end as a result of the deletion. We are not sure if the U2-17 snRNA is functional, but the conserved transcription signals imply that it may be active. Other conserved transcription signals were also identified in most active snRNAs, including the sequence element CAANTC (where N is either A, C, G or T) in U2 snRNAs (located at -6 to -1) [ 23 ], and the termination signal CAN 3-10 AGTNNAA in U snRNAs (U1, U2, U4 and U5) transcribed by RNA polymerase II (Pol II) [ 23 , 24 , 32 ]. The previously identified monocot-specific promoter element (MSP, RGCCCR, located upstream of USE) in U6.1 and U6.26 [ 33 ] is also found in five other U6 snRNA genes (U6.29, U6-2, U6-3, U6-4, U6-5). In all seven U6 snRNAs the consensus MSP sequence extends by two thymine nucleotides to RGCCCRTT. Although the MSP does not contribute significantly to U6 snRNA transcription initiation in Nicotiana plumbaginifolia protoplasts [ 33 ], the extended consensus may imply a role in gene expression regulation in Arabidopsis . Low copy number of minor snRNA genes The minor snRNAs are functional in the splicing of U12-type (AT-AC) introns. Four types of minor snRNAs, which correspond to four types of major snRNAs, exist in mammals. U11 is the analog of U1, U12 is the analog of U2, U4atac is the analog of U4, and U6atac is the analog of U6. The U5 snRNA seems to function in both the major and minor spliceosome [ 34 ]. Two minor snRNAs (atU12 and atU6atac) were experimentally identified in Arabidopsis [ 28 ]. Both have the conserved USE and TATA box in the promoter region. We identified another U6atac gene ( atU6atac-2 ) by sequence mapping. This gene has a USE and a TATA box in the promoter region. The atU6atac-2 gene is more than 90% similar to atU6atac in both its 5' and 3' ends, with a 10-nucletotide deletion in the central region. The putative U4atac-U6atac interaction region in atU6atac-2 is 100% conserved with the interaction region previously identified in atU6atac [ 28 , 35 ]. U11 and U4atac have not been experimentally identified in Arabidopsis . BLAST searches using the human U11 and U4atac homologs as queries against the Arabidopsis genome failed to find any significant hits, indicating divergence of the minor snRNAs in plants and mammals. Using the strategy described below, we successfully identified a putative Arabidopsis U4atac gene. It is a single-copy gene containing all conserved functional domains. We also found a single candidate U11 snRNA gene (chromosome 5, from 17,492,101 to 17,492,600) that has the USE and TATA box in the promoter region. This gene also contains a putative binding site fr Sm protein and a region that could pair with the 5' splice site of the U12-type intron. Identification of an Arabidopsis U4atac snRNA gene Like U4 snRNA and U6 snRNA, human U4atac and U6atac snRNAs interact with each other through base pairing [ 36 ]. The same interaction is expected to exist between the Arabidopsis homologs. Therefore, we deduced the tentative AtU4atac stem II sequence (CCCGTCTCTGTCAGAGGAG) from AtU6atac snRNA and searched for matching sequences in the Arabidopsis genome. Hit regions together with flanking regions 500 base-pairs (bp) upstream and 500 bp downstream were retrieved and screened for transcription signals (USE and TATA box). One sequence was identified that contains both the USE and TATA box in appropriate positions, as shown in Figure 1 . The tentative U4atac snRNA gene contains not only the stem II sequence, but also the stem I sequence that presumably base-pairs with U6atac snRNA stem I. Furthermore, a highly conserved Sm-protein-binding region exists at the 3' end. The predicted secondary structure is nearly identical to hsU4atac, with a relative longer single-stranded region (data not shown). With the highly conserved transcriptional signals, functional domains and secondary structure, this candidate gene is likely to be a real U4atac snRNA homolog. We named it AtU4atac and assigned At4g16065 as its tentative gene model because it is located between gene models At4g16060 and At4g16070 on chromosome 4. Tandem arrays of snRNAs genes Some snRNAs genes exist as small groups on the Arabidopsis chromosomes [ 6 ]. We identified 10 snRNA gene clusters: seven U1-U4 snRNA clusters, one U2-U5 snRNA cluster, and a tandem duplication for both U2 snRNA (U2-10) and U5 snRNA (U5.1) (Figure 2 ). All seven Arabidopsis U1-U4 clusters have the U1 snRNA gene located upstream of the U4 snRNA gene, with a 180-300-nucleotide intergenic region. Five of the U1-U4 arrays are located on chromosome 5 (U1a/U4.1, U1-4/U4-5, U1-8/U4-7, U1-9/U4-8, and U1-13p/U4.3p), and the remaining two on chromosome 1 (U1-10/U4-6 and U1-14p/U4-10). The U2-17 and U5-10 occur in tandem array on chromosome 5, separated by fewer than 200 nucleotides. Arabidopsis splicing-related protein-coding genes Most of the proteins involved in splicing in mammals and Drosophila are known [ 4 , 37 , 38 ]. In addition, recent proteomics studies revealed many novel proteins associated with human spliceosomes (reviewed in [ 18 ]). Using all these animal proteins as query sequences, we identified a total of 395 tentative homologs in Arabidopsis . Sequence-similarity scores and comparison of gene structure and protein domain structure were used to assign the genes to families. Each gene was assigned a tentative name based on the name of its respective animal homolog. Different homologs within a gene family were labeled by adding an Arabic number (1, 2, and so on) to the name. Close family members with similar gene structure were indicated by adding -a, -b, and -c to the name. The 395 genes were classified into five different categories according to the presumed function of their products. Ninety-one encode small nuclear ribonucleoprotein particle (snRNP) proteins, 109 encode splicing factors, and 60 encode potential splicing regulators. Details of EST evidence, alternative splicing patterns, duplication sources and domain structure of these genes are listed in Table 2 . We also identified 84 Arabidopsis proteins corresponding to 54 human spliceosome-associated proteins. The remaining 51 genes encode proteins with domains or sequences similar to known splicing factors, but without enough similarity to allow unambiguous classification. These two categories are not discussed in detail here, but information about these genes is available at our ASRG site [ 39 ]. The majority of snRNP proteins are conserved in Arabidopsis There are five snRNPs (U1, U2, U4, U5 and U6) involved in the formation of the major spliceosome, corresponding to five snRNAs. Five snRNPs (U1 snRNP, U2 snRNP, U5 snRNP, U4/U6 snRNP and U4.U6/U5 tri-snRNP) have been isolated experimentally in yeast or human [ 40 - 45 ]. Each snRNP contains the snRNA, a group of core proteins, and some snRNP-specific proteins. Most of these proteins are conserved in Arabidopsis . All U snRNPs except U6 snRNP contain seven common core proteins bound to snRNAs. These core proteins all have an Sm domain and have been called Sm proteins. The U6 snRNP contains seven LSM proteins ('like Sm' proteins). Another LSM protein (LSM1) is not involved in binding snRNA (reviewed in [ 46 ]). As shown in Table 2 , all Sm and LSM proteins have homologs in Arabidopsis , and eight of them are duplicated. It is likely that these genes existed as single copies in the ancestor of animals and plants, but duplicated within the plant lineage. Only one of the 24 genes ( LSM5 , At5g48870) has been characterized experimentally in Arabidopsis . The LSM5 gene was cloned from a mutant supersensitive to ABA (abscisic acid) and drought ( SAD1 [ 47 ]). LSM5 is expressed at low levels in all tissues and its transcription is not altered by drought stress [ 47 ]. cDNA and EST evidence exist for all other core protein genes, indicating that all 24 genes are active. There are 63 Arabidopsis proteins corresponding to the 35 snRNP-specific proteins used as queries in our genome mapping. Very few of them have been characterized experimentally, including U1-70K, U1A and a tandem duplication pair of SAP130 [ 48 - 50 ]. U1-70K was reported as a single-copy essential gene. Expression of U1-70K antisense transcript under the APETALA3 promoter suppressed the development of sepals and petals [ 51 ]. We identified an additional homolog of U1-70K (At2g43370) and named it U1-70K2 . The U1-70K2 proteins showed 48% similarity to the U1-70K protein according to Blast2 results. Both genes retain the sixth intron in some transcripts, a situation which would produce truncated proteins [ 48 ]. Interestingly, we found that five of the 10 Arabidopsis U1 snRNP proteins, including the U1-70K-coding genes, may undergo alternative splicing. Several genes in U2, U5, U4/U6 and U4.U6/U5 snRNPs, but none in U1 snRNP, occur in more than three copies in the Arabidopsis genome. The atSAP114 family has five members, including two that occur in tandem ( atSAP114-1a and atSAP114-1b ). Three members have EST/cDNA evidence (Table 2 ). Interestingly, the predicted atSAP114p (At4g15580) protein contains a RNase H domain at the amino-terminal end, and thus atSAP114p shares similarity to At5g06805, a gene annotated as encoding a non-LTR retroelement reverse transcriptase-like protein. It is likely that the atSAP114p gene is a pseudogene that originated by retroelement insertion. There are three copies of the gene for the tri-snRNP 65 kilodalton (kDa) subunit, which are clustered on chromosome 4. Both the U4/U6 90 kDa protein and the U4/U6 15.5 kDa protein also have three gene copies, and the 116 kDa and 200 kDa subunits in U5 snRNP have four copies apiece. The yeast U1 snRNP contains several specific proteins that are not present in mammalian U1 snRNPs [ 52 ]. As in mammals, Arabidopsis also lacks homologs of Prp42, a component of U1 snRNP in yeast [ 53 ]. However, Arabidopsis has two copies of the gene for Prp39, which are similar to Prp42. Furthermore, atPrp39a can produce a shorter protein isoform with a novel amino-terminal sequence by exon skipping. It is possible that the duplicates and alternative isoforms of plant U1 snRNP proteins are functional homologs of the yeast-specific proteins. Several proteins specific to the minor spliceosome are also conserved in Arabidopsis . The human 18S U11/U12 snRNP contains several proteins found in U2 snRNP as well as seven novel proteins [ 14 ]. Four of the seven U11/U12-specific proteins (U11/U12-35K, 25K, 65K and 31K) are conserved in Arabidopsis , while the remaining three (59K, 48K and 20K) have no clear homologs. Interestingly, all four Arabidopsis genes are single copy in the genome, and three of them are apparently alternatively spliced (Table 2 ). Splicing factors are slightly different in Arabidopsis than in other organisms We divided the splicing factors into eight subgroups according to recent human spliceosome studies [ 13 , 14 , 16 , 18 ]: splice-site selection proteins; SR proteins; 17S U2 associated proteins; 35S U5 associated proteins; proteins specific to the BΔU1 complex; exon junction complex (EJC) proteins; second-step splicing factors and other known splicing factors. We focused our analysis on the first two subgroups because their functions in splicing are well established. A total of 109 proteins in Arabidopsis were identified, corresponding to 67 human queries from all eight subgroups. Most of the proteins are conserved among eukaryotes, but some human proteins have no obvious homologs in the Arabidopsis genome, and some novel splicing factors appear to exist in Arabidopsis . About 43% of genes encoding splicing factors are duplicated in the genome, whereas some proteins, such as SF1/BBP (branchpoint-binding protein, which facilitates U2 snRNP binding in fission yeast [ 54 ]) and cap-binding proteins (CBP20 and CBP80, possibly involved in cap proximal intron splicing [ 55 ]), derive from single-copy genes [ 56 ]. These single-copy gene products may work with all pre-mRNAs, including the ones with U12-type introns. Surprisingly, mutation of CBP80 ( ABH1 ) is not lethal and is non-pleiotropic. The abh1 plants show ABA-hypersensitive closure of stomata and reduced wilting during drought [ 57 ]. Many splicing factors have been identified previously in Arabidopsis , including two U2AF65, two U2AF35, and 18 SR proteins [ 58 - 67 ]. The U2AF35-related protein atUrp, which could interact with U2AF65 and position RS-domain-containing splicing factors [ 68 ], is also present in the Arabidopsis genome. Although the Urp gene is expressed ubiquitously in human tissues, no ESTs from this gene were found in Arabidopsis . Three copies of PTB/hnRNP-I genes were identified in Arabidopsis . The PTB protein competes for the poly-pyrimidine tract with the U2AF large subunit, thus negatively regulating splicing [ 69 ]. We also identified a homolog related to atU2AF 65 (At2g33440) and an additional SR protein (At2g46610). The U2AF 65 -related protein (atULrp, At2g33440) has 247 amino acids and shares over 40% similarity with the carboxy terminal region of the two atU2AF 65 homologs. Only one RRM can be identified in atULrp, in contrast to three RRMs and one amino-terminal RS domain in atU2AF 65 proteins, and there is no apparent RS domain in atULrp. No animal homolog of atULrp could be identified. The function of this one-RRM U2AF 65 -related protein is not clear. As it lacks other functional motifs, it might act as a competitor of U2AF65. A two-RRM U2AF 65 protein can be produced through alternative splicing. The 11th intron of atU2AF 65 a can be retained (see RAFL full-length cDNA, gi:19310596) to produce a truncated protein with only the first two RRMs. Interestingly, the last RRM in atU2AF 65 a contains several amino-acid variations from the consensus pattern such that it could not be detected by InterPro and NCBI-CDD searches using default values, also suggesting that perhaps only the first two RRMs are essential. The additional SR protein belongs to the atRSp31 family and was named atRSp32 (At2g46610). It shares 70% identity and 78% similarity with atRSp31. The protein is 250 amino acids in length and contains two RRMs and some RS dipeptides in the carboxy-terminal region. The gene structure of atRSp32 is similar to that of atRSp31 . Two other genes ( atRSp40 and atRSp41 ) are in the same family and also have similar exon and intron sizes (see gene structure information at [ 70 ]). Similarly to the previous classification of 18 SR proteins [ 61 ], the 19 SR proteins (including SR45) can be grouped into four large families of four to five members according to sequence similarity, gene structure and protein domain structure. The atRSp31 family (atRSp31, atRSp32, atRSp40 and atRSp41) belongs to a novel plant SR family and has no clear animal ortholog. Other families include the SC35 (or SRrp/TASR2) family, SF2/ASF family, and the 9G8 family. Arabidopsis has a single copy of the SC35 ortholog and four SC35-like proteins (atSR33, atSCL30a, atSCL30 and atSCL28), which appear to have diverged significantly from SC35. It seems that this divergence predates the split of plants and animals because a similar SC35-like gene family exists in the human genome (SRrp35 and SRrp40). The SRrp35 and SRrp40 were found to antagonize other SR proteins in vitro and function in 5' splice-site selection [ 71 ]. SF2/ASF has four copies (atSR1/SRp34, atSRp30, atSRp34a and atSRp34b) with similar gene structures and domains. Human 9G8 protein has five homologs in Arabidopsis , with three (atRSZp21, atRSZp22 and atRSZp22a) containing one CCHC-type zinc finger and two (atRSZ33, atRSZ32) containing two CCHC-type zinc fingers in addition to an RRM and an RS domain. Interestingly, several SR proteins (atRSZp21, atRSZp22, SR45 and SCL33) were found to interact with atU1-70K, and some SR proteins can interact with each other, thus forming a complicated interaction network to facilitate splice-site selection and spliceosome assembly [ 3 , 61 - 63 ]. atSR45 was initially regarded as a novel plant SR protein [ 63 ], but by virtue of sequence-similarity scores it actually may be the ortholog of the human RNPS1 gene, which encodes an EJC protein. Other human SR proteins (SRp20, SRp30c, SRp40, SRp54, SRp55 and SRp75) lack clear orthologs in Arabidopsis . We conclude that SR protein families evolved differently in animals and plants from three to four common ancestors, including SC35, SF2/ASF and 9G8/RSZ. The SRrp (SC35-like in plants) family may not be classical SR proteins but they play important roles in splice-site selection. Proteins in other subgroups, such as 17S U2 snRNP-associated proteins, 35S U5 snRNP-associated proteins, and protein specific to the BΔU1 complex, are also conserved in Arabidopsis . The BΔU1 complex is the spliceosome complex captured immediately before catalytic activation. Most proteins in the 35S U5 snRNP are absent in the BΔU1 complex but present in the active B complex, indicating the important roles of 35S U5 snRNP-associated proteins in spliceosome activation [ 13 ]. Conservation of these proteins in Arabidopsis revealed the same pathway of spliceosome activation in plants. A subcomplex named Prp19 complex in 35S U5 snRNP has a critical role in spliceosome activation [ 13 , 72 ]. All proteins in the human Prp19 complex have homologs in Arabidopsis , including a chromosomal duplication pair of Prp19 genes and a single copy of the CDC5 gene. For the BΔU1 complex, six human genes have homologs, and five of them are single copy in Arabidopsis . Two genes ( NPW38BP/SNP70 and p220 ( NPAT )) in the human BΔU1 complex have no apparent Arabidopsis homologs. Arabidopsis also lacks an SMN protein complex. In human, the SMN protein (survival of motor neurons) can interact with a series of proteins including Gemin2, Gemin3 (a helicase), Gemin4, Gemin5 and Gemin6 to form an SMN complex, which has important roles in the biogenesis of snRNPs and the assembly of the spliceosome through direct interactions with Sm proteins and snRNA [ 73 ]. Although the SMN protein exists in the fission yeast genome (GenBank accession CAA91173), no SMN complex members can be identified in the Arabidopsis genome. Splicing regulators are expanded in Arabidopsis Splicing regulators are proteins that can either modify splicing factors or compete with splicing factors for their binding site. Important splicing regulators are hnRNP proteins and SR protein kinases. The exact role of phosphorylation of SR proteins in splicing is not yet clear, but SR protein kinases are well conserved and exist as multiple copies in Arabidopsis . A total of eight SR protein kinases were identified in Arabidopsis , including three Lammer/CLK kinases (AFC1, AFC2 and AFC3), two SRPK1 homologs, and three SPRK2 homologs. The three Lammer/CLK kinases were identified previously, and AFC2 was shown to phosphorylate SR protein in vitro [ 63 , 74 ]. Overexpression of tobacco AFC2 homolog PK12 in Arabidopsis changed the alternative splice patterns of several genes, including atSRp30 , atSR1 / atSRp34 and U1-70K [ 75 ], indicating that these SR proteins may function to modulate splicing in plants. The heterogeneous nuclear ribonucleoproteins (hnRNPs) bind to splice sites and to binding sites for splicing factors on nascent pre-mRNAs, thus competing with splicing factors to negatively control splicing (reviewed in [ 76 ]). Humans have about 20 hnRNP proteins, many of which function in splicing. A total of 35 potential hnRNP proteins possibly related to splicing was found in Arabidopsis by sequence-similarity searches, including a superfamily of glycine-rich RNA-binding proteins. This family contains 21 members similar to human hnRNP A1 and hnRNP A2/B1. It can be further divided into two subfamilies. One includes eight proteins containing one RRM, and another has 13 members with two RRMs. 12 of these proteins were identified previously, including AtGRP7, AtGRP8, UBA2a, UBA2b, UBA2c and AtRNPA/B1-6 [ 11 , 77 , 78 ]. AtGRP7 was found to be able to influence alternative splicing of its own transcripts as well as AtGRP8 transcripts [ 79 ]. UBA2 proteins can interact with UBP1 and UBA1 proteins, which have three RRMs and one RRM respectively, to recognize U-rich sequences in the 3' untranslated region (UTR) and stabilize mRNA [ 78 ]. Although the overexpression of UBA2 did not stimulate splicing of a reporter gene in tobacco protoplasts [ 78 ], we cannot rule out the possibility that it could be involved in splicing of other genes. Other human hnRNPs related to splicing also have homologs in Arabidopsis . BLAST searches of the human (CUG)n triplet repeat RNA-binding protein (CUG-BP) against all Arabidopsis proteins revealed three putative homologs, including atFCA. atFCA and CUG-BP share similarity within the RRMs and a region approximately 40 amino acids in length. An additional protein (At2g47310) related to FCA was identified and named FCA2, as it shares about 50% similarity with FCA. The FCA proteins have two RRMs and a WW domain, which interact with the FY protein, a homolog of yeast polyadenylation factor Psf2p [ 80 , 81 ]. The FCA-FY complex negatively regulates the FCA protein by favoring a polyadenylation site from the third intron of FCA pre-mRNA [ 80 , 82 ]. FCA may be a multifunctional protein involved in mRNA processing, as human CUG-BP can function in both alternative splicing and deadenylation [ 83 ]. We also list 15 previously identified hnRNP-like proteins and two additional homologs as possible splicing regulators. The UBP1 proteins can strongly enhance splicing of some introns in protoplasts [ 84 ], whereas UBA1, RBP45 and RBP47 proteins have no similar function [ 78 , 85 ]. Unclassified splicing protein candidates In addition to the 260 proteins in the above three categories, there are also 84 Arabidopsis proteins corresponding to human spliceosome-associated proteins identified in recent proteomic studies [ 15 - 18 ]. Some of these proteins function in other processes, such as transcription, polyadenylation and even translation. Their association with spliceosomes provides evidence for the coupling of splicing and other processes. Other proteins have no known functions. Only 35.8% of the proteins in this category are duplicated in Arabidopsis . We also identified a total of 51 Arabidopsis protein-coding genes similar to known splicing proteins. They have conserved domains and some level of sequence similarity to known splicing factors. We did not include these two categories in Table 2 , but detailed information about them is available at ASRG [ 39 ]. Distribution and duplication of Arabidopsis splicing-related genes The distribution of Arabidopsis snRNA and splicing-related proteins across the genome is shown in Figure 2 and at the ASRG website. Overall, the genes appear evenly distributed on the chromosomes, with several small gene clusters. Only four snRNA genes are located on chromosome 2, three of which are U2 snRNA genes. No U4 snRNA gene is located on chromosome 4. For the protein-coding genes, most functional categories have members located on each chromosome. The only exception is the SR protein kinase family, which has no member on chromosome 1. Interestingly, chromosome 1 contains the most snRNP proteins and splicing factors, but has the fewest splicing regulators. Several gene clusters encoding splicing-related proteins were also identified. Some clusters, such as tandemly duplicated gene pairs, include genes from the same category. One cluster located on chromosome 4 includes four genes encoding tri-snRNP proteins (atTri65a, atTri65b, atTri65c and atTri15.5c, homologs of tri-snRNP 65-KD protein and 15.5 KD protein). Two other clusters, atU2A-atCdc5 and atCUG-BP1-atU1C , include genes from different functional categories. No clear clusters of genes for snRNA-splicing-related proteins were identified. Although about one third of snRNA genes are located near other protein-coding genes, none of their neighboring genes is related to splicing. As a caveat, we should point out that our snRNA gene determination strongly suggests annotation errors in overlapping protein-coding gene models. Thus, atU2-1, atU2.3, atU4.2, atU4-11p, atU5-13 and atU6.26 overlap gene models At1g16820, At3g57770, At3g06895, At1g68390, At5g53740 and At3g13857, respectively, but none of these models is well supported by cDNA or EST evidence (see displays linked at ASRG [ 30 ]). The 260 proteins in the first three categories could be grouped into 130 families, 66 of which consist of multiple members. The duplication rate is over 50%, which is higher than the 44% duplication rate of Arabidopsis transcription factors [ 86 ]. As shown in Table 3 , about 50% of genes encoding snRNP proteins, 43% of splicing factors, and 78% of splicing regulators have duplications. The much higher duplication rate of splicing regulators may reflect diversification in splicing control. At least 130 duplication events are required to yield the 260 proteins from 130 families given one single-copy ancestor per family. Thirty-three duplication events (about a quarter of the total) are likely to be the result of chromosome duplications. The chromosomal duplication ratio is 18.9-27.5% among the three groups (see Table 3 ). Some snRNA genes pairs, such as U2-14 / U2-10 , U5-3 / U5-5 and ( U6.1 U6.26 )/( U6-8p U6-9p ), may also have been produced by chromosome duplication. The C.D.2-3 region (chromosome duplication region between chromosomes 2 and 3, see [ 87 ]) has the most splicing-related gene pairs. Six genes in this region on chromosome 2 were duplicated in the same order on chromosome 3. EST evidence shows that all these genes are expressed. Three U5 snRNA genes ( U5.1 , U5.1b and U5-4 ) and four U2 snRNA genes ( U2.2 , U2.3 , U2.4 and U2.6 ) also are located in the same region on chromosome 3. No U5 and U2 homologs exist in the corresponding region on chromosome 2, suggesting that the snRNA duplication events in that region may have happened after the chromosome duplication event, or that the snRNA duplicates were lost subsequent to chromosome duplication. Chromosomal duplication rather than individual gene duplication appears to be the predominant mode of amplification for some types of genes. As shown in Table 2 , the 24 genes encoding core proteins have nine duplication pairs, five of which can be attributed to chromosomal duplications. The 19 SR protein genes include eight duplication pairs, six of which are probably the results of chromosomal duplications. At least five chromosomal duplication events contributed to the superfamily of 21 hnRNP glycine-rich RBD and A/B genes. It is not clear why these functional categories have high chromosomal duplication ratios. It is possible that chromosomal duplication could create positive selection to maintain similar copy numbers of other genes encoding proteins that interact with the products of already duplicated genes. Alternative splicing of Arabidopsis splicing-related genes According to EST/cDNA alignments, 80 of the 260 protein coding genes show 66 alternative splicing events. This rate (30.8%) is much higher than the overall frequency of alternative splicing in Arabidopsis , which is about 13% using the same criteria (2,747 genes out of 20,446 genes with EST/cDNA evidence; B.-B.W. and V.B., unpublished work). As shown in Table 4 , the snRNP protein-coding genes have the lowest alternative splicing ratio (24.2%), whereas the ratios for splicing factor and splicing regulator genes are both over 33%. More than half of the genes encoding EJC proteins, proteins specific for the BΔU1 complex, SR proteins, U11/U12 snRNP-specific proteins and U1 snRNP proteins undergo alternative splicing. Among different types of alternative splicing, intron retention is the most abundant of the alternative transcripts identified for the 260 classified splicing-related genes. As shown in Table 4 , 44 of the total 80 alternative splicing genes (about 55%) involve intron retention, 28 (35%) involve alternative acceptor-site selection and 15 (18.7%) are due to exon skipping. Compared with the corresponding ratio in all Arabidopsis alternative splicing events (55.3% intron retention, 23.4% alternative acceptor-site selection and 6.3% exon skipping; B.-B.W. and V.B., unpublished work), the ratio of intron retention in splicing-related genes is similar and the ratio of exon skipping is higher. Interestingly, only one of the 20 splicing regulator genes processed by alternative splicing (about 5%) shows exon skipping, indicating that exon skipping is an important post-transcriptional method for controlling the expression of splicing factor coding genes but not the splicing regulator genes. Discussion Previous studies had determined 30 snRNA genes and 46 protein-coding genes related to splicing in Arabidopsis (see Tables 1 and 2 ). In this study, we have computationally identified an additional 44 snRNA genes (Table 1 ) and 349 protein-coding genes (Table 2 ) that also may be involved in splicing. Among the five types of U snRNAs, U6 is the most conserved and U1 is the least conserved. We identified seven U1-U4 snRNA gene clusters. We were surprised to see so many U1-U4 clusters in Arabidopsis . In Drosophila , four snRNA clusters were reported [ 4 ], but none of them includes U1-U4 gene pairs. It is likely that a U1-U4 snRNA cluster existed in a progenitor of the current Arabidopsis genome, which was duplicated several times to form the extant seven clusters. The non-clustered U1 and U4 snRNA genes may have arisen by individual gene duplication or gene loss in duplicated clusters. Among the proteins involved in splicing, most animal homologs are conserved in plants, indicating an ancient, monophylytic origin for the splicing mechanism. A striking feature of plant splicing-related genes is their duplication ratio. Fifty percent of the splicing genes are duplicated in Arabidopsis . The duplication ratio of the splicing-related genes increases from genes encoding snRNP proteins to genes encoding splicing regulators. These data strongly suggest that the general splicing mechanism is conserved, but that the control of splicing may be more diverse in plants. The high duplication ratio of Arabidopsis splicing-related genes could be the result of evolutionary selection. Unlike animals, which can move around to maintain more homogeneous physiological conditions, plants are exposed to a larger range of stress conditions such as heat and cold. The duplicates will more probably be maintained in the genome as their functions become diversified, and potentially plant-specific, to ensure the fidelity of splicing under such varied conditions. Chromosome duplication has produced several Sm proteins, SR proteins and hnRNP proteins in Arabidopsis , which in turn could create positive selective pressures influencing the rate of duplication for functionally related genes. Because chromosome duplication occurred differentially within each plant lineage, we would expect different duplication patterns of these genes in, for example, monocots and dicots. To confirm the above hypothesis, we searched the recently sequenced rice genome using the five Arabidopsis SC35 and SC35-like proteins as probes. Eight distinct genome loci were found to encode SC35 and SC35-like proteins, including three homologs of atSC35, two homologs of atSR33/SCL33 and atSCL30a, two homologs of atSCL30, and one homolog of atSCL28. Five of the eight rice genes are currently annotated in GenBank with accession numbers BAC79909 (osSC35a), BAD09319 (osSC35b), AAP46199 (osSR33-1), BAC799901 (osSCL30a/osSR33-2), and BAD19168 (osSCL30-1). As shown in the phylogenetic tree displayed in Figure 3 , the two rice SC35 genes and atSC35 are likely to be orthologs of the animal SC35 gene. The other sequences cluster in SC35-like (SRrp/TASR) clades, indicating that the SC35 and SRrp/TASR genes diverged before the divergence of monocot and dicot plants (the divergence presumably happened even before the divergence of animals and plants, as described earlier). In addition, there are species-specific duplications. Thus, the Arabidopsis chromosomal duplication pair atSR33 - atSCL30a forms a clade, while their rice copies (osSR33-1 and osSCL30a) form another clade. Also there are additional duplications for the rice SC35 and SCL30 genes. We are currently working to identify all rice splicing related genes. The complete sets of these genes in two plant species should provide a good foundation for assessing similarities and differences in splicing mechanisms used by monocot and dicot plants. As introns evolve rapidly, the mechanism to recognize and splice them should either evolve correspondingly or be flexible enough to accommodate the changes. It seems that plants deploy the most economic and practical way by keeping a largely conserved splicing mechanism and a very flexible recognition and control mechanism. Direct evidence comes from the presence of plant-specific splicing proteins, such as the novel SR protein family and the superfamily of hnRNP A/B. The absence of SMN complex and some yeast U1 snRNP proteins in Arabidopsis indicates that other organisms also have integrated new proteins or pathways into the splicing mechanism over the course of evolution relative to other eukaryotes. Other evidence supporting the conserved splicing but flexible regulating mechanism include differential conservation among U snRNAs (U1 snRNAs are less conserved than U6 snRNAs) and high alternative splicing frequency in U1 snRNP proteins, SR proteins and hnRNP proteins. The SR proteins and U1 snRNP proteins are involved in early steps of splicing and 5' and 3' splice-site selection; multiple isoforms of these proteins may be functionally significant in the control of splicing. It is interesting to note that the overall alternative splicing frequency in splicing related genes is much higher than the frequency averaged over all Arabidopsis genes. More than half of SR proteins and U1 snRNP proteins show alternative splicing. Alternative splicing might increase protein diversity derived from splicing-related genes, which would further add flexibility to the splicing mechanism. The high frequency of alternative transcripts from splicing related genes raises another interesting question - how is splicing regulated in these splicing-related genes? One possible answer is that some splicing-related genes may be autoregulated. Accumulation of one transcript would feed back to inhibit/promote other isoforms. Several splicing-related genes have been reported to be regulated in this way. For example, AtGRP7 (hnRNP A/B superfamily) is a circadian clock-regulated protein which negatively autoregulates its expression [ 79 ]. When the AtGRP7 protein accumulates over the circadian cycle, it promotes production of alternative transcripts which use a cryptic 5' splice site. As a result of message instability, the alternative transcripts contain pre-mature stop codons and do not accumulate to high levels, thus decreasing the level of AtGRP7 protein [ 79 ]. atSRp30 has similar effects on its own transcripts [ 65 ]. Another possible answer is that some splicing-related genes might regulate the splicing of other splicing-related genes. For example, overexpression of AtGRP7 and atSRp30 is known to affect the splicing of AtGRP8 and atSR1 , respectively [ 65 , 79 ]. A third possibility is that the environment could affect the alternative splicing pattern. A good example is the SR1 gene. The ratio of two transcripts from the SR1 gene (SR1B/SR1) increases in a temperature-dependent manner [ 67 ]. Generally, heat or cold stress could cause intron retention in some splicing regulators, which could further alter the splicing pattern of other genes. The fourth possible regulators are intronless genes. Combining all these possibilities, a pathway to regulate splicing could be inferred as follows: environmental changes → splicing pattern changes in some specific splicing-related genes and/or intronless genes → expression pattern changes (including splicing pattern changes) in general splicing related genes → changes in splicing patterns for specific genes. Conclusions A large number of Arabidopsis splicing-related genes were computationally identified in this study by means of sequence comparisons and motif searches, including a tentative U4atac snRNA gene containing all conserved motifs, a new SR protein-coding gene ( atRSp32 ) belonging to the atRSp31 family, and several genes related to genes encoding known splicing-related proteins (atULrp and atFCA2). A web-accessible database containing all the Arabidopsis splicing related genes has been constructed and will be expanded to other organisms in the near future. This compilation should provide a good foundation to study the splicing process in more detail and to determine to what extent these genes are conserved across the entire plant kingdom. Our data show that about 50% of the splicing-related genes are duplicated in Arabidopsis . The duplication ratios for splicing regulators are even higher, indicating that the splicing mechanism is generally conserved among plants, but that the regulation of splicing may be more variable and flexible, thus enabling plants to respond to their specific environments. Materials and methods Search for Arabidopsis snRNAs Sequences of the 15 experimentally identified major snRNAs were downloaded from GenBank. The two minor snRNAs sequences were compiled from the literature [ 28 ]. These genes were used to search against the Arabidopsis genome at the AtGDB BLAST server [ 88 ] and at the SALK T-DNA Express web server [ 89 ]. Our initial analysis was based on Release 3.0 of the Arabidopsis genome (GenBank accession numbers NC_003070.4, NC_003071.3, NC_003074.4, NC_003075.3, and NC_003076.4). Local BLAST [ 21 ] was used to derive the locations of the snRNA homologs from more recently sequenced regions of the genome. Criteria used for local BLAST were 'e 1 -F F -W 7' (cutoff eval is 1, dust filter on, with a minimum word size of 7). Human and maize snRNAs were also included as query sequences, and all hits with e-values less than 10 -5 were regarded as possible homologs. A total of 70 major snRNAs and three minor snRNAs were identified by this method. Each major snRNA type has 10-18 copies in the genome. A tentative gene name and gene model were assigned to each snRNA gene after comparison with the snRNAs identified in MATDB [ 90 ]. Sequence-similarity values were based on BLAST alignments. Search for Arabidopsis splicing-related proteins A three-round BLAST search strategy was used to identify Arabidopsis splicing related protein-coding genes. First, sequences of splicing-related proteins from human and Drosophila were downloaded from GenBank according to several recent proteomic studies [ 15 - 18 ] and the website compilation of Stephen Mount's group available at [ 91 ]. Human hnRNP proteins identified in a recent review [ 76 ] were downloaded from GenBank. All these sequences were used as queries in a local BLAST search against Arabidopsis annotated proteins (obtained from TIGR at [ 92 ]). All hits with an e-value less than 10 -10 were collected as candidates. Many of these candidates had highly significant e-values (usually 10 -30 or below and much lower than other hits). These candidates were regarded as true homologs. In the second step, all identified true homologs were used to query the Arabidopsis protein set again. An e-value of 10 -20 was used as a cutoff value to find possible paralogs of the true homologs. Sequences identified in both rounds of BLAST hits were regarded as main candidates for splicing related proteins. Finally, the main candidates were queried against GenPept and all annotated human proteins (obtained from Ensembl [ 93 ]). All candidates with significant similarity to proteins unrelated to splicing were removed from the main candidate list, and all candidates with significant similarity to proteins related to splicing were regarded as true splicing-related genes and were promoted to the status of true homologs. The remaining candidates were regarded as unclassified splicing-related proteins. BLAST results were initially analyzed by MuSeqBox [ 94 ]). Two custom scripts were written to read MuSeqBox output files, largely automating the search procedure. Gene structure and chromosomal locations The gene structure and chromosomal locations for the genes encoding splicing-related proteins were retrieved from AtGDB [ 95 ]. The chromosomal locations of the snRNA genes were inferred from the BLAST results. The location maps (Figure 1 ) were generated using the AtGDB advanced search function [ 96 ]. Spliced alignments of ESTs and cDNAs generated by GeneSeqer [ 97 ] were used to verify gene models. Gene structure information was used as an important criteria to group homologs into gene families. Protein domains InterProScan 3.3 was downloaded from [ 98 ] and was subsequently used to search protein domain databases using default parameters [ 99 ]. A Perl script was written to process the text results from InterProScan. Protein domain information was used in comparisons of homologs from different species. The search of the National Center for Biotechnology Information Conserved Domain Database (NCBI-CDD) [ 100 ] was conducted manually for certain genes to confirm the InterPro results. Duplication source The gene families with multiple copies were inspected to determine whether they were likely to have derived from chromosome-duplication events. Gene models of the duplicated gene were searched against the gene list of each chromosome redundancy region at MATDB [ 101 ]. If the gene and its duplicate were both in the list, they were regarded as a chromosome duplication pair. Otherwise, they were assumed to be produced by random gene duplication. Identification of alternative splicing All Arabidopsis ESTs and cDNAs were aligned against the genome using the spliced alignment program GeneSeqer as made available through AtGDB [ 102 ]. We retrieved the intron and exon coordinates of the reliable cognate alignments from the database. Scripts were written to identify introns that overlap with other introns or exons. We defined the alternative splicing cases as follows: alternative donor (AltD): an intron has the same 3'-end coordinate but different 5'-end coordinate as another overlapping intron; alternative acceptor (AltA): an intron has the same 5'-end coordinate but different 3'-end coordinate as another intron; alternative position (AltP): an intron has different 5'-end and 3'-end coordinates as another overlapping intron; exon skipping (ExonS): an annotated intron completely contains an alternatively identified exon in the same transcription direction; intron retention (IntronR): an annotated intron is completely contained by an alternatively identified exon. Database and interface construction Details about each splicing-related gene were saved in a MySQL database. PHP scripts were written to interact with the database and generate the interface web pages. Text and BLAST searches were implemented by Perl-cgi scripts.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545797.xml

545781

Reconstruction of regulatory and metabolic pathways in metal-reducing δ-proteobacteria

A study of the genetic and regulatory factors in several biosynthesis, metal ion homeostasis, stress response, and energy metabolism pathways suggests that phylogenetically diverse δ-proteobacteria have homologous regulatory components.

Background The delta subdivision of proteobacteria is a very diverse group of Gram-negative microorganisms that include aerobic genera Myxococcus with complex developmental lifestyles and Bdellovibrio , which prey on other bacteria [ 1 ]. In this study, we focus on anaerobic metal-reducing δ-proteobacteria, seven representatives of which have been sequenced recently, providing an opportunity for comparative genomic analysis. Within this group, sulfate-reducing bacteria, including Desulfovibrio and Desulfotalea species, are metabolically and ecologically versatile prokaryotes often characterized by their ability to reduce sulfate to sulfide [ 2 ]. They can be found in aquatic habitats or waterlogged soils containing abundant organic material and sufficient levels of sulfate, and play a key role in the global sulfur and carbon cycles [ 1 ]. Industrial interest in sulfate reducers has focused on their role in corrosion of metal equipment and the souring of petroleum reservoirs, while their ability to reduce toxic heavy metals has drawn attention from researchers interested in exploiting this ability for bioremediation. Psychrophilic sulfate-reducing Desulfotalea psychrophila has been isolated from permanently cold arctic marine sediments [ 3 ]. In contrast to sulfate-reducing bacteria, the genera Geobacter and Desulfuromonas comprise dissimilative metal-reducing bacteria, which cannot reduce sulfate, but include representatives that require sulfur as a respiratory electron acceptor for oxidation of acetate to carbon dioxide [ 4 ]. These bacteria are an important component of the subsurface biota that oxidizes organic compounds, hydrogen or sulfur with the reduction of insoluble Fe(III) oxides [ 5 ], and have also been implicated in corrosion and toxic metal reduction. Knowledge of transcriptional regulatory networks is essential for understanding cellular processes in bacteria. However, experimental data about regulation of gene expression in δ-proteobacteria are very limited. Different approaches could be used for identification of co-regulated genes (regulons). Transcriptional profiling using DNA microarrays allows one to compare the expression levels of thousands of genes in different experimental conditions, and is a valuable tool for dissecting bacterial adaptation to various environments. Computational approaches, on the other hand, provide an opportunity to describe regulons in poorly characterized genomes. Comparison of upstream sequences of genes can, in principle, identify co-regulated genes. From large-scale studies [ 6 - 9 ] and analyses of individual regulatory systems [ 10 - 14 ] it is clear that the comparative analysis of binding sites for transcriptional regulators is a powerful approach to the functional annotation of bacterial genomes. Additional techniques used in genome context analysis, such as chromosomal gene clustering, protein fusions and co-occurrence profiles, in combination with metabolic reconstruction, allow the inference of functional coupling between genes and the prediction of gene function [ 15 ]. Recent completion of finished and draft quality genome sequences for δ-proteobacteria provides an opportunity for comparative analysis of transcriptional regulation and metabolic pathways in these bacteria. The finished genomes include sulfate-reducing Desulfovibrio vulgaris [ 16 ], D. desulfuricans G20 , and Desulfotalea psychrophila , as well as the sulfur-reducing G. sulfurreducens [ 17 ], while the G. metallireducens genome has been completed to draft quality. A mixture of Desulfuromonas acetoxidans and Desulfuromonas palmitatis has been sequenced, resulting in a large number of small scaffolds, the identity of which ( acetoxidans or palmitatis ) has not been determined, and we refer to this sequence set simply as Desulfuromonas . Though draft-quality sequence can make it difficult to assert with confidence the absence of any particular gene, we have included these genomes in our study because they do provide insight as to the presence or absence of entire pathways, they can be compared to the related finished genome of G. sulfurreducens , and because complete genome sequence is not necessary for the methodology we use to detect regulatory sequences. In this comprehensive study, we identify a large number of regulatory elements in these δ-proteobacteria. Some of the corresponding regulons are highly conserved among various bacteria (for example, riboswitches, BirA, CIRCE), whereas others are specific only for δ-proteobacteria. We also present the reconstruction of a number of biosynthetic pathways and systems for metal-ion homeostasis and stress response in these bacteria. The most important result of this study is identification of a novel regulon involved in sulfate reduction and energy metabolism in sulfate-reducing bacteria, which is most probably controlled by a regulator from the CRP/FNR family. Results The results are organized under four main headings for convenience. In the first, we analyze a number of specific regulons for biosynthesis of various amino acids and cofactors in δ-proteobacteria. Most of them are controlled by RNA regulatory elements, or riboswitches, that are highly conserved across bacteria [ 18 ]. In the next section we describe several regulons for the uptake and homeostasis of transition metal ions that are necessary for growth. These regulons operate by transcription factors that are homologous to factors in Escherichia coli , but are predicted to recognize entirely different DNA signals. We then describe two stress-response regulons: heat-shock regulons (σ 32 and HrcA/CIRCE), which operate by regulatory elements conserved in diverse bacteria, and newly identified peroxide stress response regulons that are quite diverse and conserved only in closely related species. Finally, we present a completely new global regulon in metal-reducing δ-proteobacteria, which includes various genes involved in energy metabolism and sulfate reduction. Biosynthesis and transport of vitamins and amino acids Biotin Biotin (vitamin H) is an essential cofactor for numerous biotin-dependent carboxylases in a variety of microorganisms [ 19 ]. The strict control of biotin biosynthesis is mediated by the bifunctional BirA protein, which acts both as a biotin-protein ligase and a transcriptional repressor of the biotin operon. The consensus binding signal of BirA is a palindromic sequence TTGTAAACC-[N 14/15 ]-GGTTTACAA [ 20 ]. Consistent with the presence of the biotin repressor BirA, all bacteria in this study have one or two candidate BirA-binding sites per genome, depending on the operon organization of the biotin genes (Table 1 ). In the Desulfovibrio species, the predicted BirA site is located between the divergently transcribed biotin operon and the birA gene. In other genomes, candidate binding sites for BirA precede one or two separate biotin biosynthetic loci, whereas the birA gene stands apart and is not regulated. All δ-proteobacteria studied possess genes for de novo biotin synthesis from pimeloyl-CoA precursor ( bioF , bioA , bioD , bioB ) and the bifunctional gene birA , but the initial steps of the biotin pathway are variable in these species (Figure 1 ). The Geobacter species have the bioC-bioH gene pair, which is required for the synthesis of pimeloyl-CoA in Escherichia coli . The Desulfuromonas species contain both bioC-bioH and bioW genes, representing two different pathways of pimeloyl-CoA synthesis. In contrast, D. psychrophila is predicted to synthesize a biotin precursor using the bioC-bioG gene pair, where the latter gene was only recently predicted to belong to the biotin pathway [ 20 ]. Both Desulfovibrio species have an extended biotin operon with five new genes related to the fatty-acid biosynthetic pathway. Among these new biotin-regulated genes not present in other δ-proteobacteria studied, there are homologs of acyl carrier protein (ACP), 3-oxoacyl-(ACP) synthase, 3-oxoacyl-(ACP) reductase and hydroxymyristol-(ACP) dehydratase. From positional and regulatory characteristics we conclude that these genes are functionally related to the biotin pathway. The most plausible hypothesis is that they encode a novel pathway for pimeloyl-CoA synthesis, as the known genes for this pathway, bioC , bioH , bioG and bioW , are missing in the Desulfovibrio species. Riboflavin Riboflavin (vitamin B 2 ) is an essential component of basic metabolism, being a precursor to the coenzymes flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN). The only known mechanism of regulation of riboflavin biosynthesis is mediated by a conserved RNA structure, the RFN -element, which is widely distributed in diverse bacterial species [ 21 ]. The δ-proteobacteria in this study possess a conserved gene cluster containing all genes required for the de novo synthesis of riboflavin ( ribD-ribE-ribBA-ribH ), but lack this regulatory element. The only exception is D. psychrophila , which has an additional gene for 3,4-dihydroxy-2-butanone-4-phosphate synthase ( ribB2 ) with an upstream regulatory RFN element. Thiamine Vitamin B 1 in its active form, thiamine pyrophosphate, is an essential coenzyme synthesized by the coupling of pyrimidine (HMP) and thiazole (HET) moieties in bacteria. The only known mechanism of regulation of thiamine biosynthesis in bacteria is mediated by a conserved RNA structure, the THI -element [ 22 ]. Search for thiamine-specific regulatory elements in the genomes of δ-proteobacteria identified one or two THI -elements per genome that are located upstream of thiamine biosynthetic operons (Figure 1 in Additional data file 1). The δ-proteobacteria possess all the genes required for the de novo synthesis of thiamine (Figure 2 ) with the exception of Geobacter species, which lack some genes for the synthesis and salvage of the HET moiety ( thiF , thiH and thiM ), and D. psychrophila , which has no thiF . In most δ-proteobacteria there are two paralogs of the thiamine phosphate synthase thiE , and Geobacter and Desulfuromonas species have fused genes thiED . In D. psychrophila , the only THI -regulated operon includes HET kinase thiM and previously predicted HMP transporter thiXYZ [ 22 ], whereas other thiamine biosynthetic genes are not regulated by the THI -element (Figure 2 ). In most cases, downstream of a THI -element there is a candidate terminator hairpin, yielding regulation by the transcription termination/antitermination mechanism. The two exceptions predicted to be involved in translational attenuation are THI -elements upstream of genes thiED in Desulfuromonas and thiM in D. psychrophila . In the Desulfovibrio species, the thiSGHFE operon is preceded by two tandem THI -elements, each followed by a transcriptional terminator. This is the first example of possible gene regulation by tandem riboswitches. Cobalamin Adenosylcobalamin (Ado-CBL), a derivative of vitamin B 12 , is an essential cofactor for several important enzymes. The studied genomes of δ-proteobacteria possess nearly complete sets of genes required for the de novo synthesis of Ado-CBL (Figure 3 ). The only exception is the precorrin-6x reductase, cbiJ , which was found only in Desulfuromonas but not in other species. The occurrence of CbiD/CbiG enzymes instead of the oxygen-dependent CobG/CobF ones suggests that these bacteria, consistent with their anaerobic lifestyle, use the anaerobic pathway for B 12 synthesis similar to that used by Salmonella typhimurium [ 23 ]. Ado-CBL is known to repress expression of genes for vitamin B 12 biosynthesis and transport via a co- or post-transcriptional regulatory mechanism, which involves direct binding of Ado-CBL to the riboswitch called the B12 -element [ 24 , 25 ]. A search for B12 -elements in the genomes of δ-proteobacteria produced one B12 -element in D. desulfuricans , D. psychrophila and G. metallireducens , two in D. vulgaris and G. sulfurreducens , and four in Desulfuromonas (Figure 2 in Additional data file 1). In Geobacter species these riboswitches regulate a large locus containing almost all the genes for the synthesis of Ado-CBL (Figure 3 ). One B12 -element in the Desulfovibrio species regulates both the cobalamin-synthesis genes cbiK-cbiL and the vitamin B 12 transport system btuCDF , whereas three such regulatory elements in Desulfuromonas precede different vitamin B 12 transport loci. In D. psychrophila , a B12 -element occurs within a large B 12 synthesis gene cluster and precedes the cbiK-cbiL genes. The most interesting observation is that genes encoding the B 12 -independent ribonucleotide reductase NrdDG are preceded by B12 -elements in D. vulgaris and Desulfuromonas . Notably, all δ-proteobacteria have another type of ribonucleotide reductase, NrdJ, which is a vitamin B 12 -dependent enzyme. We propose that when vitamin B 12 is present in the cell, expression of the B 12 -independent isozyme is inhibited, and a relatively more efficient B 12 -dependent isozyme is used. This phenomenon has been previously observed in other bacterial genomes [ 26 ]. Methionine The sulfur-containing amino acid methionine and its derivative S -adenosylmethionine (SAM) are important in protein synthesis and cellular metabolism. There are two alternative pathways for methionine synthesis in microorganisms, which differ in the source of sulfur. The trans -sulfuration pathway ( metI-metC ) utilizes cysteine, whereas the direct sulfhydrylation pathway ( metY ) uses inorganic sulfur instead. All δ-proteobacteria in this study except the Desulfovibrio species possess a complete set of genes required for the de novo synthesis of methionine (Figure 4 ). The Geobacter species and possibly Desulfuromonas have some redundancy in the pathway. First, these genomes contain the genes for both alternative pathways of the methionine synthesis. Second, they possess two different SAM synthase isozymes, classical bacterial-type MetK and an additional archaeal-type enzyme [ 27 ]. Moreover, it should be noted that the B 12 -dependent methionine synthase MetH in these bacteria lacks the carboxy-terminal domain, which is involved in reactivation of spontaneously oxidized coenzyme B 12 . In Gram-positive bacteria, SAM is known to repress expression of genes for methionine biosynthesis and transport via direct binding to the S-box riboswitch [ 28 ]. In contrast, Gram-negative enterobacteria control methionine metabolism using the SAM-responsive transcriptional repressor MetJ. The δ-proteobacteria in this study have no orthologs of MetJ, but instead, we identified S-box regulatory elements upstream of the metIC and metX genes in the genomes of the Geobacter species and Desulfuromonas (see Figure 3 in Additional data file 1). A strong hairpin with a poly(T) region follows all these S-boxes, implying involvement of these S-boxes in a transcriptional termination/antitermination mechanism. Both Desulfovibrio species have genes involved in the conversion of homocysteine into methionine ( metE , metH and metF ), which could be involved in the SAM recycling pathway, but not those genes required for de novo methionine biosynthesis. The ABC-type methionine transport system ( metNIQ ), which is widely distributed among bacteria, was also not found in these δ-proteobacteria. The Desulfovibrio species appear to have the single-component methionine transporter metT [ 28 ]. Lysine The amino acid lysine is produced from aspartate through the diaminopimelate (DAP) pathway in most bacteria. The first two stages of the DAP pathway, catalyzed by aspartokinase and aspartate semialdehyde dehydrogenase, are common for the biosynthesis of lysine, threonine, and methionine. The corresponding genes were found in δ-proteobacteria where they form parts of different metabolic operons. Four genes for the conserved stages of the lysine synthesis pathway ( dapA , dapB , dapF and lysA ) were further identified in δ-proteobacteria, whereas we did not find orthologs for three other genes ( dapC , dapE and dapD ), which vary in bacteria using different meso-DAP synthesis pathways. The lysine synthesis genes are mostly scattered along the chromosome, and in only some cases are dapA and either dapB , dapF or lysA clustered. All δ-proteobacteria studied lack the previously known lysine transporter LysP. However, in D. desulfuricans and D. psychrophila we found a gene for another candidate lysine transporter, named lysW , which was predicted in our previous genomic survey [ 29 ]. In various bacterial species, lysine is known to repress expression of genes for lysine biosynthesis and transport via the L-box riboswitch [ 30 ]. In addition, Gram-negative enterobacteria use the lysine-responsive transcriptional factor LysR for control of the lysA gene. Among the δ-proteobacteria studied, we found neither orthologs of LysR, nor representatives of the L-box RNA regulatory element. In an attempt to analyze potential lysine regulons in this phylogenetic group, we collected upstream regions of all lysine biosythesis genes and applied SignalX as a signal detection procedure [ 31 ]. The strongest signal, a 20-bp palindrome with consensus GTGGTACTNNNNAGTACCAC, was observed upstream of the lysX-lysA operons in both Desulfovibrio genomes and the candidate lysine transporter gene lysW in D. desulfuricans (Table 2 ). The first gene in this operon, named lysX , encodes a hypothetical transcriptional regulator with a helix-turn-helix motif (COG1378) and is the most likely candidate for the lysine-specific regulator role in Desulfovibrio . To find new members of the regulon, the derived profile (named LYS-box) was used to scan the Desulfovibrio genomes. The lysine regulon in these genomes appears to include an additional gene (206613 in D. vulgaris , and 394397 in D. desulfuricans ), which encodes an uncharacterized membrane protein with 14 predicted transmembrane segments. We predict that this new member of the lysine regulon might be involved in the uptake of lysine or some lysine precursor. Metal ion homeostasis Iron Iron is necessary for the growth of most bacteria as it participates in many major biological processes [ 32 ]. In aerobic environments, iron is mainly insoluble, and microorganisms acquire it by secretion and active transport of high-affinity Fe(III) chelators. Under anaerobic conditions, Fe(II) predominates over ferric iron, and can be transported by the ATP-dependent ferrous iron transport system FeoAB. Genomes of anaerobic δ-proteobacteria contain multiple copies of the feoAB genes, and lack ABC transporters for siderophores. Regulation of iron metabolism in bacteria is mediated by the ferric-uptake regulator protein (FUR), which represses transcription upon interaction with ferrous ions. FUR can be divided into two domains, an amino-terminal DNA-binding domain and a carboxy-terminal Fe(II)-binding domain. The consensus binding site of E. coli FUR is a palindromic sequence GATAATGATNATCATTATC [ 33 ]. In all δ-proteobacteria studied except D. psychrophila , we identified one to three FUR orthologs that form a distinct branch (FUR_Delta) in the phylogenetic tree of the FUR/ZUR/PerR protein family (see below). One protein, FUR2 in D. desulfuricans , lacks an amino-terminal DNA-binding domain and is either non-functional or is involved in indirect regulation by forming inactive heterodimers with two other FUR proteins. Scanning the genomes with the FUR-box profile of E. coli did not result in identification of candidate FUR-boxes in δ-proteobacteria. In an attempt to analyze potential iron regulons in this phylogenetic group, we collected upstream regions of the iron-transporter genes feoAB and applied SignalX to detect regulatory signals. The strongest signal, a 17-bp palindrome with consensus WTGAAAATNATTTTCAW (where W indicates A or T), was observed upstream of the multiple feoAB operons and fur genes in all δ-proteobacteria except D. psychrophila (Table 3 ). The constructed search profile (dFUR-box) was applied to detect new candidate FUR-binding sites in these five genomes (Figure 5 and Table 3 ). The smallest FUR regulons were observed in the Geobacter and Desulfuromonas species, where they include the ferrous iron transporters feoAB (one to four copies per genome), the fur genes themselves (one copy in the Geobacter species and two copies in Desulfuromonas ), and two hypothetical porins. The first one, named psp , was found only in G. metallireducens and Desulfuromonas genomes, where it is preceded by two tandem FUR-boxes. The psp gene has homologs only in Aquifex aeolicus and in various uncultured bacteria, and in one of them (a β-proteobacterium) it is also preceded by two FUR-boxes (GenBank entry AAR38161.1). This gene is weakly similar to the family of phosphate-selective porins (PFAM: PF07396) from various Gram-negative bacteria. The second hypothetical porin was found only in G. sulfurreducens (383590), where it is preceded by a FUR-box and followed by feoAB transporter. This gene, absent in other δ-proteobacteria, has only weak homologs in some Gram-negative bacteria and belongs to the carbohydrate-selective porin OprB family (PFAM: PF04966). Thus, two novel genes predicted to fall under FUR control encode hypothetical porins that could be involved in ferrous iron transport. Another strong FUR-box in the G. sulfurreducens genome precedes a cluster of two hypothetical genes located immediately upstream of the feoAB -containing operon. The first gene in this operon, named genX (383594), has no orthologs in other bacteria and the encoded protein has a heme-binding site signature of the cytochrome c family (PFAM: PF00034). The second gene, named genY (383592), encodes a two-domain protein that is not similar to any known protein. In Desulfuromonas , an ortholog of the genY amino-terminal domain (391875) is divergently transcribed from a predicted ferric reductase (391874), and their common upstream region contains a strong FUR-box. Moreover, orthologs of the genY C-terminal domain were identified in Desulfovibrio species, where they are again preceded by two tandem FUR-boxes and form a cluster with the hypothetical gene, genZ , encoding a protein of 100 amino acids with two tetratricopeptide repeat domains that are usually involved in protein-protein interactions (PFAM: PF00515). From genomic analysis alone it is difficult to predict possible functions of these new members of the FUR regulon in δ-proteobacteria. Two Desulfovibrio species have significantly extended FUR regulons that are largely conserved in these genomes and include ferrous iron transporter genes feoAB and many hypothetical genes. Another distinctive feature of the FUR regulon in Desulfovibrio species is a structure of two partially overlapping FUR-boxes shifted by 6 bp. Interestingly, the flavodoxin gene, fld , is predicted to be regulated by FUR in both Desulfovibrio species. In addition to this iron-repressed flavodoxin (a flavin-containing electron carrier), the Desulfovibrio species have numerous ferredoxins (an iron-sulfur-containing electron carrier). One possible explanation is that in iron-restricted conditions these microorganisms can replace ferredoxins with less-efficient, but iron-independent alternatives. A similar regulatory strategy has been previously described for superoxide dismutases in E. coli , Bordetella pertusis and Pseudomonas aeruginosa [ 34 - 36 ] and predicted, in a different metabolic context, for B 12 -dependent and B 12 -independent enzymes [ 26 ]; see the discussion above. Other predicted regulon members with conserved FUR-boxes in both Desulfovibrio species are the hypothetical genes pep (Zn-dependent peptidase), gdp (GGDEF domain protein, PF00990), hdd (metal dependent HD-domain protein, PF01966), and a hypothetical P-type ATPase (392971) that could be involved in cation transport, and a long gene cluster starting from the pqqL gene (Zn-dependent peptidase). The latter cluster contains at least 10 hypothetical genes encoding components of ABC transporters and biopolymer transport proteins ( exbB , exbD and tonB ). In D. vulgaris , the first gene in this FUR-regulated cluster is an AraC-type regulator named foxR , since it is homologous to numerous FUR-controlled regulators from other genomes ( foxR from Salmonella typhi , alcR from Bordetella pertussis , ybtA from Yersinia species, pchR from Pseudomonas aeruginosa ), which usually regulate iron-siderophore biosynthesis/transport operons [ 33 ]. An ortholog of foxR , a single FUR-regulated gene, was identified in D. desulfuricans located about 30 kb away from the FUR-regulated pqqL gene cluster. Given these observations, we propose that this gene cluster is involved in siderophore transport and is regulated by FoxR. A hypothetical gene in D. vulgaris ( 209207 ) has the strongest FUR-box in this genome; however, its orthologs in D. desulfuricans are not predicted to belong to the FUR regulon. Another operon in D. desulfuricans (392971-392970-392969), encoding three hypothetical proteins, is preceded by two candidate FUR-boxes, but these genes have no orthologs in other δ-proteobacteria. Thus, FUR-dependent regulation of these hypothetical genes is not confirmed in other species, and their possible role in the iron homeostasis is not clear. Nickel The transition metal nickel (Ni) is an essential cofactor for a number of prokaryotic enzymes, such as [NiFe]-hydrogenase, urease, and carbon monoxide dehydrogenase (CODH). Two major types of nickel-specific bacterial transporters are represented by the NikABCD system of E. coli (the nickel/peptide ABC transporter family) and the HoxN of Ralstonia eutropha (the NiCoT family of nickel/cobalt permeases). Nickel uptake must be tightly regulated because excessive nickel is toxic. In E. coli and some other proteobacteria, nickel concentrations are controlled by transcriptional repression of the nikABCD operon by the Ni-dependent regulator NikR [ 37 ]. The genomes of δ-proteobacteria studied so far contain multiple operons encoding [NiFe] and [Fe] hydrogenases and Ni-dependent CODH, but lack urease genes. Both known types of nickel-specific transporters are absent in δ-proteobacteria, but these genomes contain orthologs of the nickel repressor nikR . In an attempt to identify potential nickel transporters in this taxonomic group, we analyzed the genome context of the nikR genes. The nikR gene in Desulfuromonas is co-localized with a hypothetical ABC transport system, which is weakly homologous to the cobalt ABC-transporter cbiMNQO from various bacteria. Orthologs of this system, named here nikMNQO , are often localized in proximity to Ni-dependent hydrogenase or urease gene clusters in various proteobacteria (data not shown). Among δ-proteobacteria, the Geobacter species have a complete nikMNQO operon, whereas operons in D. desulfuricans and D. psychrophila lack the nikN component but include two additional genes, named nikK and nikL , which both encode hypothetical proteins with amino-terminal transmembrane segments (Figure 6 ). Desulfovibrio vulgaris has a nikMQO cluster and separately located nikK and nikL genes. Since various other proteobacteria also have the same clusters including nikK and nikL , but not nikN (data not shown), we propose that these two genes encode additional periplasmic components of the NikMQO ABC transporter, possibly involved in the nickel binding. By applying SignalX to a set of upstream regions of the nikMQO operons, we identified de novo the NikR binding signal in all δ-proteobacteria except D. psychrophila (Table 4 ). This signal has the same structure as in enterobacteria (an inverted repeat of 27-28 bp), but its consensus (GTGTTAC-[N 13/14 ]-GTAACAC) differs significantly from the consensus of NikR binding signal of enterobacteria (GTATGAT-[N 13/14 ]-ATCATAC) [ 37 ]. Using the derived profile to scan the genomes of δ-proteobacteria we identified one more candidate NikR-binding site in D. desulfuricans . Thus the nickel regulon in this bacterium includes the hydAB2 operon, encoding periplasmic iron-only hydrogenase. Altogether, D. desulfuricas has three paralogs of [NiFe] hydrogenase and two paralogs of [Fe] hydrogenase. We predict that an excess of nickel represses a nickel-independent hydrogenase isozyme using the Ni-responsive repressor NikR. Regulation of hydrogenase enzymes by NikR has not been described previously. A closer look at the upstream region of the putative nickel transport operon in D. psychrophila revealed similar NikR consensus half-sites but in the opposite orientation to each other (GTAACAC-[N 13/14 ]-GTGTTAC). Searching the genomes with this reversed NikR signal, we observed one more hypothetical gene cluster in D. psychrophila which has two high-scoring NikR-sites in the upstream region, and a NikR-site upstream of the single nikK gene in D. vulgaris (Figure 6 ). Zinc Zinc is an important component of many proteins, but in large concentrations it is toxic to the cell. Thus zinc repressors ZUR regulate high-affinity zinc transporters znuABC in various bacteria [ 38 ]. An orthologous zinc transporter was found in δ-proteobacteria (Figure 7 ). In G. sulfurreducens and the Desulfovibrio species, this cluster also includes a hypothetical regulatory gene from the FUR/ZUR/PerR family, named zur_Gs and zur_D , respectively. Phylogenetic analysis of this protein family demonstrated that ZUR_Gs and ZUR_D are not close relatives and are only weakly similar to known FUR, ZUR, and PerR regulators from other bacteria (see below). The predicted ZUR-binding site located just upstream of the zur-znuABC operon in G. sulfurreducens is highly similar to the ZUR consensus of Gram-positive bacteria (TAAATCGTAATNATTACGATTTA). Another strong signal, a 17-bp palindrome with consensus ATGCAACNNNGTTGCAT, was identified upstream of the znuABC-zur operons in two Desulfovibrio genomes (Table 5 ). Although znuABC genes are present in all δ-proteobacteria, we observed neither candidate ZUR regulators, nor ZUR-binding sites in G. metallireducens , Desulfuromonas and D. psychrophila , suggesting either the absence of zinc-specific regulation or presence of another regulatory mechanism for these genes. Cobalt The previously described cobalt transport system CbiMNQO was found only in the Geobacter species, where it is located within the B 12 -regulated cbi gene cluster close to the cobaltochelatase gene cbiX , responsible for incorporation of cobalt ions into the corrin ring (see the 'Cobalamin' section above). In contrast, other δ-proteobacteria, possessing a different cobaltochelatase ( cbiK ), lack homologs of any known cobalt transporter. It was previously suggested by global analysis of the B 12 metabolism that different types of cobalt transporters are interchangeable in various bacterial species [ 26 ]. From genome context analysis and positional clustering with the cbiK gene, we predicted a novel candidate cobalt transporter in δ-proteobacteria, named cbtX (Figure 3 ), which was previously annotated as a hypothetical transmembrane protein conserved only in some species of archaea (COG3366). Molybdenum Molybdenum (Mo) is another transition metal essential for bacterial metabolism. Bacteria take up molybdate ions via a specific ABC transport system encoded by modABC genes. Mo homeostasis is regulated by the molybdate-responsive transcription factor ModE, containing an amino-terminal DNA-binding domain and two tandem molybdate-binding domains. Orthologs of ModE are widespread among prokaryotes, but not ubiquitous [ 39 ]. All δ-proteobacteria have one or more homologs of the modABC transporter (Figure 8 ). However, full-length modE genes containing both DNA- and molybdate-binding domains were observed only in G. sulfurreducens and Desulfuromonas . In G. sulfurreducens , the molybdate transport operon is co-localized with modE and is preceded by a putative ModE-binding site (Table 6 ), which is similar to the E. coli consensus of ModE (ATCGNTATATA-[N 6 ]-TATATANCGAT). In contrast, we could not identify E. coli -type ModE-binding sites upstream of the mod operons in Desulfuromonas , indicating that these operons may be regulated by a different, unidentified signal. Three other δ-proteobacteria (two Desulfovibrio species and D. psychrophila ) have genes encoding a single DNA-binding domain of ModE (Figure 8 ). Searching with the E. coli -type profile did not reveal candidate binding sites of ModE in these species. To predict potential ModE sites de novo , we collected upstream regions of all molybdate transport operons and applied SignalX. In both Desulfovibrio genomes, we identified a common inverted repeat with consensus CGGTCACG-[N 14 ]-CGTGACCG, which is considerably different from the E. coli consensus of ModE (Table 6 and Figure 8 ). The modABC gene cluster in these species includes an additional chimeric gene encoding a fusion of phage integrase family domain (PF00589) and one or two molybdate-binding domains (MOP). The functions of these chimeric molybdate-binding proteins, and the mechanism of Mo-sensing by DNA-binding ModE domains in the Desulfovibrio species, are not clear. Stress response regulons Oxidative stress Under aerobic conditions, generation of highly toxic and reactive oxygen species such as superoxide anion, hydrogen peroxide and the hydroxyl radical leads to oxidative stress with deleterious effects [ 40 ]. Strictly anaerobic sulfate-reducing bacteria are adapted to survive in transient oxygen-containing environments by intracellular reduction of oxygen to water using rubredoxin:oxygen oxidoreductase (Roo) as the terminal oxidase [ 41 ]. The main detoxification system for reactive oxygen species in aerobic and anaerobic bacteria involves superoxide dismutase (Sod), catalase (KatA, KatG) and nonspecific peroxidases (for example, AhpC). In addition to these enzymes, Desulfovibrio species have an alternative mechanism for protecting against oxidative stress, which includes rubredoxin oxidoreductase (Rbo), which has superoxide reductase activity, rubrerythrin (Rbr) with NADH peroxidase activity, and rubredoxin-like proteins (Rub, Rdl), which are used as common intermediary electron donors [ 42 ]. Searching for orthologs of the oxidative stress-related genes in the genomes in this study revealed great variability in content and genomic organization (Figure 9 ). We also searched for homologs of transcription factors known to be involved in regulation of the peroxide and superoxide stress responses. Lacking orthologs of the E. coli OxyR and SoxR/SoxS regulators, the δ-proteobacteria studied have instead multiple homologs of the peroxide-sensing regulator PerR of B. subtilis [ 43 ]. The PerR-specific branch on the phylogenetic tree of the FUR/ZUR/PerR family contains at least three distinct sub-branches with representatives from δ-proteobacteria (Figure 10 ). In all cases except D. psychrophila , the perR genes are co-localized on the chromosome with various peroxide stress-responsive genes (Figure 9 ). However, the upstream regions of these genes contain no candidate PerR-binding sites conforming to the B. subtilis PerR consensus TTATAATNATTATAA. Applying the SignalX program to various subsets of upstream regions of peroxide stress-responsive genes resulted in identification of candidate PerR operators in δ-proteobacteria (Table 7 ). In the Desulfovibrio species, a common palindromic signal was found upstream of the perR and rbr2 genes. In D. vulgaris , perR forms an operon with rbr and rdl genes [ 42 ]. Searching for genes with the derived profile identified additional candidate members of the PerR regulon, alkyl hydroperoxide reductase ahpC in D. vulgaris ( D. desulfuricans has no ortholog of ahpC ), and a hypothetical gene of unknown function in both Desulfovibrio species (206199 in D. vulgaris and 395549 in D. desulfuricans ). The perR-rbr-roo operon in both Geobacter species is preceded by a conserved palindromic region (Table 7 ) which overlaps a candidate -10 promoter element (Figure 11 ). The second perR paralog in G. sulfurreducens (named perR2 ), which is followed by a gene cluster containing two cytochrome peroxidase homologs ( hsc and ccpA ), glutaredoxin ( grx ) and rubrerythrin ( rbr ), has a close ortholog in the Desulfuromonas species, where it precedes the rbr gene (Figures 9 , 10 ). For these gene clusters we found a common palindromic signal, which is not similar to other predicted PerR signals in δ-proteobacteria (Table 7 ). Two other perR paralogs in Desulfuromonas ( perR2 and perR3 ) probably result from a recent gene duplication (Figure 10 ), and both are co-localized on the chromosome with the peroxide stress-responsive genes katG and rbr2 , respectively (Figure 9 ). A common new signal identified upstream of the katG and perR3 genes is probably recognized by both PerR2 and PerR3 regulators in this organism (Table 7 ). The PerR regulons in δ-proteobacteria are predicted to include only a small subset of all peroxide stress-related genes identified in these genomes. In addition to the mainly local character of the predicted regulation, these regulons seem to be highly variable between different species, both in their content and DNA signals. Heat shock In bacteria, two major mechanisms regulating expression of heat-shock proteins are positive control by alternative sigma factor σ 32 , encoded by the rpoH gene, and negative control by binding of the repressor protein HrcA to palindromic operators with a consensus TTAGCACTC-[N 9 ]-GAGTGCTAA called CIRCE [ 44 ]. The rpoH gene was identified in the genomes of all δ-proteobacteria studied. Though the HrcA/CIRCE system is conserved in very diverse taxonomic groups of bacteria, it is not universal, as some γ-proteobacteria lack it [ 45 ]. We detected the hrcA genes and CIRCE sites in all genomes studied except D. psychrophila (Table 8 ). We then searched the genomes of δ-proteobacteria with previously constructed profiles for σ 32 promoters and CIRCE [ 45 ]. As was observed previously for other bacteria, the only constant member of the HrcA regulon in δ-proteobacteria is the groESL operon. In addition, CIRCE sites are present upstream of the hrcA-grpE-dnaKJ operons in the Geobacter and Desulfuromonas species and upstream of the rpoH gene in G. sulfurreducens . In contrast to the highly conserved CIRCE signal, the σ 32 promoters identified in multiple copies in various proteobacteria are less conserved [ 45 , 46 ]. Among δ-proteobacteria, we identified σ 32 -like promoters upstream of some heat-shock-related genes encoding chaperons (GroE, DnaJ, DnaK, GrpE) and proteases (ClpA, ClpP, ClpX, Lon) (Table 9 ). Thus, in δ-proteobacteria, as in most proteobacteria, σ 32 plays a central part in the regulation of the heat-shock response, although detailed regulatory strategies seem to vary in different species. The alternative HrcA/CIRCE system controls expression of groE and other major chaperons. Central energy metabolism The CooA regulon for carbon monoxide utilization in Desulfovibrio species Growth using carbon monoxide (CO) as the sole energy source involves two key enzymes in the γ-proteobacterium Rhodospirillum rubrum - CO dehydrogenase (CODH) and an associated hydrogenase - which are encoded in the coo operons and induced by the CO-sensing transcriptional activator CooA [ 47 ]. Among the sequenced δ-proteobacteria, only Desulfovibrio species have coo operons and the CooA regulator. D. vulgaris has two separate operons encoding CODH and the associated hydrogenase, whereas D. desulfuricans has only one operon encoding CODH (Figure 12 ). The strongest identified signal, a 16-bp palindrome with consensus TGTCGGCNNGCCGACA, was identified upstream of the coo operons from both Desulfovibrio species and R. rubrum (Table 10a ). This consensus conforms to the experimentally known CooA-binding region at the R. rubrum cooFSCTJ operon [ 48 ]. New CRP/FNR-like regulon for sulfate reduction and prismane genes Sulfate-reducing bacteria are characterized by their ability to utilize sulfate as a terminal electron acceptor. To try to identify the regulatory signals responsible for this metabolism, we applied the signal detection procedure SignalX to a set of upstream regions of genes involved in the sulfate-reduction pathway in Desulfovibrio species. A conserved palindromic signal with consensus sequence TTGTGANNNNNNTCACAA was detected upstream of the sat and apsAB operons, which encode ATP sulfurylase and APS reductase, respectively. This novel signal is identical to the E. coli CRP consensus, and we hypothesized that a CRP-like regulator might control the sulfate-reduction regulon in Desulfovibrio . Scanning the Desulfovibrio genomes resulted in identification of similar sites upstream of many hypothetical genes encoding various enzymes and regulatory systems (Table 10b and Figure 12 ). One of them, the hcp gene in D. vulgaris , encodes a hybrid-cluster protein (previously called the prismane-containing protein) of unknown function [ 49 ], which is coexpressed with a hypothetical ferredoxin gene, named frdX* : new gene names introduced in this study are marked by asterisk. In both Desulfovibrio species, the hcp-frdX* genes are co-localized with a hypothetical regulatory gene from the CRP/FNR family of transcriptional regulators, named HcpR* for the Hcp regulator (Figure 12 ). Close HcpR* orthologs were detected in two other δ-proteobacteria, D. psychrophila and Desulfuromonas ; however, the same CRP-like signals were not present in their genomes. Examination of a multiple alignment of the CRP/FNR-like proteins revealed one specific amino acid (Arg 180) in the helix-turn-helix motif involved in DNA recognition, which is changed from arginine (for example, in E. coli CRP and Desulfovibrio HcpR*) to serine and proline in these two δ-proteobacteria (data not shown). As both these species have multiple hcp and frdX paralogs, we applied SignalX to a set of corresponding upstream regions and obtained another FNR-like palindromic signal with consensus at ATTTGACCNNGGTCAAAT, which is notably distinct from the CRP-like signal in the third position (which has T instead of G). Such candidate sites were observed upstream of all hcp and frdX paralogs identified in D. psychrophila and Desulfuromonas , as well as upstream of some additional genes in Desulfuromonas , for example those encoding polyferredoxin and cytochrome c heme-binding protein (Table 10 and Figure 12 ). The HcpR regulon was also identified in other taxonomic groups, including Clostridium , Thermotoga , Bacteroides , Treponema and Acidothiobacillus species, and in all cases candidate HcpR sites precede hcp orthologs (data not shown). Moreover, the hcpR gene is often co-localized with hcp on the chromosome. In clostridia, frdX orthologs are also preceded by candidate HcpR sites. These data indicate that the main role of HcpR is control of expression of two hypothetical proteins - hybrid-cluster protein and ferredoxin - which are most probably involved in electron transport. However, the HcpR regulon is significantly extended in some organisms. Additional members of this regulon that are conserved between the two Desulfovibrio species include two operons involved in sulfate reduction ( apsAB and sat ), a hypothetical cluster of genes (206515-206516) with similarity to dissimilative sulfite and nitrite reductases, polyferredoxin, a hypothetical gene conserved in Archaea (209119), and the putative thiosulfate reductase operon phcAB (209106-209105). Notably, both CooA and HcpR candidate sites precede the cooMKLXUHF operon for CODH-associated hydrogenase, which is present only in D. vulgaris . Because regulators from the CRP/FNR family are able to both repress and activate gene expression, it was interesting to predict the mode of regulation of the HcpR regulon members. To this end, we investigated the positions of candidate HcpR sites in pairwise alignments of orthologous regulatory regions from the two Desulfovibrio species. These two closely related genomes are diverse enough to identify regulatory elements as conserved islands in alignments of intergenic regions. For the sat and apsAB operons, the HcpR sites were found within highly conserved parts of alignments and in both cases the site overlaps the -10 box of a site strongly resembling a promoter (Figure 13a,b ), suggesting repression of the genes by HcpR. In contrast, positive regulation by HcpR could be proposed for the hcp-frdX , 206515-206516 and 209119 operons, which have HcpR sites upstream or slightly overlapping the -35 box of predicted promoters (Figure 13c ). In the case of the cooMKLXUHF operon in D. vulgaris , the HcpR site is located upstream of the candidate site of the known positive regulator CooA; thus it is also predicted to be an activator site. By analysis of the functions of genes co-regulated by HcpR, it is difficult to predict the effector for this novel regulon. The physiological role of the hybrid iron-sulfur cluster protein Hcp, the most conserved member of the HcpR regulon, is not yet characterized despite its known three-dimensional structure and expression profiling in various organisms. In two facultative anaerobic bacteria, E. coli and Shewanella oneidensis , the hcp gene is expressed only under anaerobic conditions in the presence of either nitrate or nitrite as terminal electron acceptors [ 50 , 51 ]. More recent expression data obtained for anaerobic D. vulgaris have showed strong upregulation of the hcp-frdX* and 206515-206516 operons by nitrite stress (J. Zhou, personal communication). While HcpR is predicted to activate these two hypothetical operons, as well as the CODH-associated hydrogenase operon, it most probably represses two enzymes from the sulfate reduction pathway, APS reductase and ATP sulfurylase. We hypothesize that HcpR is a key regulator of the energy metabolism in anaerobic bacteria, possibly controlling the transition between utilization of alternative electron acceptors, such as sulfate and nitrate. The absence of the dissimilatory sulfite reductase DsrAB in the predicted HcpR regulon of Desulfovibrio could be explained by its experimentally defined ability to reduce both sulfite and nitrite [ 52 ]. Discussion Regulation of biosynthesis pathways Because the organisms considered in this study are commonly identified on the basis of their catabolic capabilities, comparatively little is known about the regulation of their biosynthetic pathways. In this study, we identified a number of previously characterized regulatory mechanisms (involved in biotin, thiamine, cobalamin and methionine synthesis), all of which, excluding the biotin regulon, are mediated by direct interaction of a metabolic product with a riboswitch control element (summarized in Table 11 ). Of particular interest in this set was observation of a dual tandem THI -element riboswitch in Desulfovibrio species. Multiple protein-binding sites are a common regulatory feature and often imply cooperative binding of multiple protein factors. Although true riboswitch units do not interact with trans -acting factors, it is theoretically possible for independently acting sites to yield a cooperative effect when ligand binding derepresses transcription. For switches that are repressed by ligand binding, however, tandem sites would simply lower the concentration threshold at which a response is seen, but not affect cooperativity unless some more complicated interaction of the sites were allowed. On the one hand, independently acting sites is a simpler mechanism to explain, while on the other hand, it seems unusual that duplicate sites would have evolved to adjust the concentration response instead of simply changing the binding affinity for the ligand at the sequence level. Moreover, it seems unlikely that a tandem switch would be preserved across a large evolutionary distance without offering some other advantage such as cooperativity. It would be interesting to investigate the biochemical behavior of these tandem THI -elements in the laboratory to resolve whether their genomic organization reflects a more sophisticated mode of regulation, or is simply an evolutionarily convenient way to adjust the concentration response, or is perhaps just a recombination remnant that has persisted in these genomes by chance. Another interesting finding was the absence of complete machinery for the de novo synthesis of methionine in the Desulfovibrio species. These organisms have the necessary genes to form methionine from homocysteine, but no apparent process by which to produce homocysteine. Although the enzymatic pathway of cysteine synthesis has been studied in Desulfovibrio vulgaris [ 53 ], its ability to synthesize methionine has not been characterized. Growth in minimal medium using sulfate as the only source of sulfur is routine, however, and suggests that these bacteria use a previously uncharacterized mechanism for assimilation of sulfur into methionine. On the basis of genomic context analysis we also predicted that the Desulfovibrio species contain a novel set of genes involved in biotin synthesis. Regulation of metal-ion homeostasis A number of regulators believed to be involved in metal-ion homeostasis were identified on the basis of orthology with known factors from E. coli or B. subtilis . However, in almost all cases, with the possible exception of ZUR and ModE in G. sulfurreducens , which appear to have signals similar to the B. subtilis and E. coli consensus respectively, similarity to known binding signals was not observed (Table 11 ). The presence of similar sets of target genes in the δ-proteobacteria studied allowed us to apply the signal detection procedure to elucidate novel regulatory signals, to expand core regulons, and to observe species-specific differences in regulation. Interestingly, the FUR/ZUR/PerR family of transcriptional regulators was found to be ubiquitous in these bacteria and responsible for a broad range of functions including iron and zinc homeostasis as well as oxidative stress response. In some cases, multiple paralogous factors were found, perhaps indicating previously uncharacterized functions for this versatile gene family. The large number of iron-containing proteins predicted from the genome sequence of these organisms, and their ability to use ferric iron anaerobically as a terminal electron acceptor, makes iron homeostasis a key target for analysis. A number of new genes were identified that may belong to the FUR regulon of these organisms. First, uncharacterized porins with upstream FUR boxes were identified in the Geobacter and Desulfuromonas genomes, which we speculate might be involved in iron transport. Additionally, a two-domain protein with no homologs of known function was identified in all species except D. psychrophila . In G. sulfurreducens , this gene occurred downstream of another gene with a cytochrome-type heme-binding motif, while in Desulfuromonas it was divergently transcribed with a ferric reductase, and was associated with a tetratricopeptide repeat protein in the Desulfovibrio genomes. In both Desulfovibrio species, we identified an additional regulon, possibly under FoxR control, which might be involved in siderophore transport. This finding was particularly surprising because we did not identify any known siderophore biosynthetic pathway. A possible explanation is that these bacteria use a novel siderophore biosynthesis pathway, or alternatively, take up siderophores released by other bacteria in the environment. Stress response Oxidative stress is one of the most common environmental stressors for these organisms, especially in the metal-contaminated sites of interest for bioremediation. The bacteria in this study are unusual in that they contain both the aerobic superoxide dismutase (Sod)/catalase-type oxidative response as well as the anaerobic Sor/rubrerythrin-type response as previously noted for D. vulgaris [ 54 ]. Analysis of the signal peptides in these proteins indicates that the Sod/catalase system acts periplasmically, whereas the Sor/rubrerythrin system acts cytoplasmically [ 54 ]. While these organisms have no homologs of the OxyR or SoxRS regulators known to respond to changes in oxygen levels in E. coli , they do contain homologs of the PerR regulator of B. subtilis , known for its involvement in peroxide stress (Table 11 ). Clustering of PerR homologs with oxidative stress genes, as well as their grouping with known Bacillus PerR genes in a phylogenetic analysis of the FUR/ZUR/PerR family of transcription factors, allowed the inference that they may, in part, be responsible for the control of the oxidative stress response of these organisms. Although we did not identify conserved regulatory elements for some known oxidative stress genes such as the Rbo/Rub/Roo operon in Desulfovibrio species, it has been observed that the Rub/Roo operon of Desulfovibrio gigas shows strong constituitive expression from a previously identified σ 70 promoter, indicating that additional factors may not be involved [ 55 ]. The heat-shock response of these bacteria was found to be mediated by two regulons previously described in other species (Table 11 ). First, the σ 32 regulon was identified, with a consensus signal similar to that characterized for E. coli . The second observed regulon was the HrcA/CIRCE regulon known in B. subtilis and other bacteria, but not present in E. coli . These two regulons include a partially overlapping set of genes. Notably, CIRCE elements were identified in all of the genomes used in this study with the exception of D. psychrophila . It is tempting to speculate that the constant and cold temperatures encountered by this species in its environmental niche have removed the need for this particular heat-shock response. Similarity of regulatory signals with those in other bacteria Comparison with well studied bacterial model organisms has shown that δ-proteobacteria share regulatory components with both Gram-positive and Gram-negative microorganisms (Table 11 ). For example, the use of NikR and ModE for the regulation of, respectively, nickel and molybdenum uptake and utilization is consistent with E. coli -like regulation. However, the presence of PerR, CIRCE elements and S-box motifs is reminiscent of B. subtilis -like regulation. Moreover, in the case of FUR, although the regulon structure showed overlap with known downstream targets in model organisms, the sequence of the FUR box, which is conserved in both E. coli and B. subtilis , was observed to be different in the metal-reducing δ-proteobacteria. We recognize that this is one of the first direct studies comparing entire regulons in δ-proteobacteria. Two recent computational works, considering either a single D. vulgaris or two Geobacter species, used the AlignACE signal detection program, which is based on a Gibbs-sampling algorithm, to derive large sets of conserved DNA motifs without linking them to specific regulatory systems [ 56 , 57 ]. Unfortunately, the predicted regulatory signals based on single genomes turned out not to be conserved across genomes, and could not be used for functional gene annotation. In this comparative work, we tried to extensively describe a set of biologically reasonable regulons in δ-proteobacteria. The regulatory sites predicted here were not detected in the other two computational studies by Hemme and Wall and by Yan et al . [ 56 , 57 ]. Previously published experimental studies of sulfate-reducing δ-proteobacteria have focused mostly on the biochemistry unique to these organisms, and little is known about the regulation of gene expression. In part, this has been due to difficulties in genetically manipulating these strictly anaerobic bacteria. Recent advances in microarray technologies provide genome-scale expression data for D. vulgaris under various conditions. In support of our findings, all operons predicted to be co-regulated by the peroxide-responsive regulator PerR in D. vulgaris are significantly downregulated by oxygen stress (J. Zhou, personal communication). Furthermore, recent microarray data obtained for G. sulfurreducens in iron-limiting conditions confirm our prediction of the FUR regulon in this genome (R. O'Neil, personal communication). It is interesting to observe the extent to which regulatory motifs are conserved between δ-proteobacteria. Although riboswitches and some DNA signals (that is, CIRCE, σ 32 and BirA) seem to be conserved across vast spans of evolutionary time, in many cases we observe divergence in binding signals even when the core components of a regulon are conserved (NikR, FUR, PerR, ModE). These findings raise, but do not answer, questions such as what circumstances cause transcription factor binding specificities to change or remain conserved, and whether those changes reflect genetic drift, or active selection to alter the regulatory action of the factor. Energy metabolism We identified two regulons involved in the control of energy metabolism (Table 11 ). The first, controlled by the CooA protein, was present only in the Desulfovibrio genomes. It is orthologous to a known regulon in R. rubrum , and regulates genes involved in the oxidation of CO. The second regulon is novel and distributed widely among anaerobic and facultatively anaerobic bacteria. The primary downstream target of this newly identified regulator, which we called HcpR*, is the hybrid-cluster protein Hcp. Upregulation of the hcp gene in response to growth on nitrate or nitrite in Shewanella oneidensis , E. coli and D. vulgaris indicates that Hcp is likely to be involved in the utilization of alternative electron acceptors. Consistent with this hypothesis, we predicted positive regulation of Hcp and the associated ferredoxin FrdX by HcpR, and negative regulation of the sulfate-reduction genes by HcpR in the Desulfovibrio genomes, based on the position of the candidate HcpR-binding sites relative to the predicted promoters. Thus, HcpR is predicted to be responsible for switching between alternative electron acceptors during anaerobic respiration in these species. Interestingly, we found an HcpR site upstream of the CO-dependent hydrogenase that was also predicted to be under the control of CooA. This hydrogenase was recently proposed to play a key role in sulfate reduction [ 16 ], and it is tempting to speculate that its inclusion in a common regulon with known sulfate-reduction genes supports this hypothesis. The position of the binding site, however, suggests that it activates rather than represses transcription, contrary to predictions for other known sulfate-reduction genes, so its regulation is likely to be complex, and further experiments will be needed to determine whether it plays the role of the cytoplasmic hydrogenase necessary for the proposed 'hydrogen cycling' of sulfate reduction [ 58 ]. The ubiquitous phylogenetic distribution of the HcpR regulon indicates that it has a central role in facilitating an anaerobic life style, yet very little is known about its specific function. We hope our elucidation of the core components and regulator of this important regulon will inspire future experimental studies to determine its cellular role. Regulatory motifs for alternative cofactor adaptation In the course of this study we identified several cases in which different variants of genes were predicted to be regulated according to the availability of required cofactors or nutrients. Three examples were observed in which an alternative enzyme, not requiring a given cofactor, was repressed by the availability of that cofactor: B 12 -independent ribonucleotide reductase was repressed by the availability of B 12 ; [Fe] hydrogenase was repressed by the availability of nickel (and presumably replaced by [NiFe] hydrogenase); and Fe(II) was predicted to repress a flavodoxin gene which we suspect may be used as an alternative to ferredoxins present in the genome. This mode of regulation for B 12 -independent isozymes of ribonucleotide reductase and methionine synthetase has been previously described [ 26 ]. Moreover, a similar regulatory strategy has been reported for one of the alternative superoxide dismutases and for paralogs of ribosomal proteins [ 34 - 36 , 38 , 59 ]. Taken together, these data suggest that this flexible strategy may represent a common theme in the adaptation of bacteria to their environment. Indeed, similar mechanisms may, in part, explain some of the apparent genetic redundancy in many genomes. Materials and methods The genomes of δ-proteobacteria that were analyzed in this study are Desulfovibrio vulgaris Hildenborough (DV); Desulfovibrio desulfuricans G20 (DD); Geobacter metallireducens (GM); Geobacter sulfurreducens PCA (GS); Desulfuromonas species (DA); and Desulfotalea psychrophila (DP). Complete genomic sequences of DV and GS were downloaded from GenBank [ 60 ]. Draft sequences of DD, GM and DA genomes were produced by the US Department of Energy Joint Genome Institute and obtained from [ 61 ]. Draft sequence of the DP genome was provided by the Max Planck Institute for Marine Microbiology in Bremen, Germany [ 62 ]. Numerical gene identifiers from the Virtual Institute for Microbial Stress and Survival (VIMSS) Comparative Genomics database [ 63 ] are used for hypothetical genes without common names. New gene names introduced in this study are marked by an asterisk. For de novo definition of a common transcription factor-binding signal in a set of upstream gene fragments, a simple iterative procedure implemented in the program SignalX was used [ 31 ]. Weak palindromes were selected in each region, and each palindrome was compared to all others. The palindromes most similar to the initial one were used to make a profile. The positional nucleotide weights in this profile were defined as W ( b,k ) = log[ N ( b , k ) + 0.5] - 0.25Σ i = A,C,G,T log[ N ( i , k ) + 0.5], where N ( b , k ) is the count of nucleotide b in position k [ 10 ]. The candidate site score Z is defined as the sum of the respective positional nucleotide weights Z ( b 1 ... b L ) = Σ k = 1... L W ( b k , k ), where k is the length of the site. These profiles were used to scan the set of palindromes again, and the procedure was iterated until convergence. Thus a set of profiles was constructed. The profile with the greatest information content [ 64 ] was selected as the recognition rule. Each genome was scanned with the profile using the GenomeExplorer software [ 65 ], and genes with candidate regulatory sites in the 300-bp upstream regions were selected. The upstream regions of genes that are orthologous to genes containing regulatory sites were examined for candidate sites even if these were not detected automatically. The threshold for the site search was defined as the lowest score observed in the training set. Sets of potentially co-regulated genes contained genes that had candidate regulatory sites in their upstream regions and genes that could form operons with such genes (that is, located downstream on the same strand with intergenic distances of less than about 100 bp). A complete description of the GenomeExplorer software, including the SignalX program, is given at [ 65 ]. The RNApattern program [ 66 ] was used to search for conserved RNA regulatory elements (riboswitches) in bacterial genomes. The input RNA pattern for this program describes an RNA secondary structure and sequence consensus motifs as a set of the following parameters: the number of helices, the length of each helix, the loop lengths, and a description of the topology of helix pairs. The latter is defined by the coordinates of helices. For instance, two helices may be either independent or embedded helices, or they could form a pseudoknot structure. This definition is similar to the approach implemented in the Palingol algorithm [ 67 ]. Orthologous proteins were identified as bidirectional best hits [ 68 ] by comparing the complete sets of protein sequences from the two species using the Smith-Waterman algorithm implemented in the GenomeExplorerprogram [ 65 ]. When necessary, orthologs were confirmed by construction of phylogenetic trees for the corresponding protein families. Phylogenetic analysis was carried out using the maximum likelihood method implemented in PHYLIP [ 69 ]. Large-scale gene cluster comparisons were carried out using the VIMSS Comparative Genomics database [ 63 ]. Multiple sequence alignments were done using CLUSTALX [ 70 ]. The COG [ 68 ], InterPro [ 71 ], and PFAM [ 72 ] databases were used to verify the protein functional and structural annotation. Note added in proof Recently it has been demonstrated by in vitro experiment that the glycine-specific riboswitch consists of two tandem aptamer sequences that appear to bind target molecules cooperatively [ 73 ]. This indirectly confirms our hypothesis of a cooperative effect of ligand binding to tandem THI -elements in Desulfovibrio spp. Also we have recently shown that Geobacter spp. have a modified HcpR regulon, which uses a signal similar to that found in DA and DP, but contains multiple nitrate/nitrite reductase genes. Additional data files An additional data file (Additional data 1 ) containing three figures with detailed description of DNA- and RNA-type regulatory sites is available with the online version of this paper and on our website [ 74 ]. Supplementary Material Additional data file 1 Three figures with detailed description of DNA- and RNA-type regulatory sites Click here for additional data file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545781.xml

350672

Small RNA Pathways in Plants

xx

Since small RNA molecules were discovered just over ten years ago, it's become clear that these once overlooked bits of genetic material play a decidedly large role in controlling gene expression. Though typically just 21 to 24 nucleotides long, small RNAs regulate a diverse array of cellular processes, from developmental patterning and genome rearrangement to antiviral defense. They typically accomplish these tasks by targeting specific nucleotide sequences to shut down gene expression. A. thaliana at the rosette stage Found in both plants and animals, small RNAs come mainly in two classes—microRNA (miRNA) and short interfering RNA (siRNA). miRNAs arise from nonprotein-coding transcripts that adopt extended “fold-back” structures, which are then cleaved by enzymes called Dicer or Dicer-like (DCL). siRNAs, on the other hand, arise from perfectly base-paired double-stranded RNA, which are also cleaved by Dicer. Some siRNAs require enzymes called RNA-dependent RNA polymerases (RdRp). miRNAs and many types of siRNAs function post-transcriptionally—that is, they affect genes that have been expressed, or transcribed, into RNA—to guide cleavage or prevent translation into protein. In plants and some animals, this post-transcriptional RNA interference (RNAi) acts as an adaptive antiviral response, among other things. siRNAs can also “silence” gene expression by altering chromatin—the DNA-protein complex into which chromosomes assemble—and preventing transcription. It is thought that chromatin silencing acts as a genome defense mechanism, guarding against potential damage from mobile genetic elements or invasive DNA (say, from a virus) by keeping genes in the tightly coiled, and thus inaccessible, “heterochromatic” state. While much remains to be learned about the mechanisms and pathways that govern small RNAs, it's becoming clear that they add an important layer of regulation and flexibility to gene expression. Now a team led by James Carrington at Oregon State University and Steve Jacobsen at the University of California at Los Angeles demonstrates that plants have evolved multiple systems to produce distinct classes of small RNAs with specialized regulatory and defensive functions. The first generates miRNAs; the second produces siRNAs that regulate chromatin structure; and the third generates siRNAs in response to viral infections. Each system requires a unique spectrum of functions of three different DCL proteins; the siRNA systems each function in coordination with one of several RdRp proteins. The researchers propose that the expansion and subsequent diversification of these proteins, which occurred in plants but not in many animals, has contributed to the diversification of specialized small RNA-directed pathways. Working in Arabidopsis thaliana , a favorite model organism for plant biologists, Zhixin Xie et al. analyzed a series of mutants with nonfunctional dcl and rdr genes, as well as a few other mutants of interest, to determine how the small RNAs responded to loss of these proteins. Two mutations (one in a dcl gene and one in another gene) affected the miRNAs, either impairing their function or reducing their populations. None of the RdRp proteins had any effect on miRNAs. The researchers performed the same type of genetic analyses on siRNAs and found that a different DCL mutant caused a reduction in one class of siRNAs and that an RdRp mutant nearly eliminated these populations of siRNAs. The diversity of siRNAs produced by the Arabidopsis genome reveals an important role in genome maintenance, expression, and defense, the authors conclude. Given that large numbers of siRNAs arise from highly repeated sequences—such as those introduced by viruses or mobile genetic elements—it may be that the cell senses such “invasive” sequence duplication events and enlists siRNAs to run interference by silencing these potentially damaging sequences. In this way, chromatin-associated siRNAs may offer an additional line of defense against invasive sequences, on top of that offered by post-transcriptional RNAi—a dual adaptive advantage since a fast-spreading virus or over-proliferating transposon (also known as a jumping gene) could wreak havoc on a plant population. Whatever other roles small RNAs may play in genome regulation—they have also been implicated in regulating growth and development—their primary responsibility appears to be blocking gene expression. Whether they accomplish that by controlling chromosome activity to prevent gene transcription or by inhibiting or degrading RNA transcripts to block translation into protein, small RNAs appear to make wide-ranging contributions to the overall gene expression program of the cell.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC350672.xml

534811

Using Biology to Create Complex Patterns

In his seminal exploration of the properties of living organisms, What Is Life? , Erwin Schroedinger concluded that life depends in large part on storing and processing information. For genetic material to carry the diverse instructions required for living processes, he proposed, it must be stored in an aperiodic crystal. Just nine years later, it was clear that DNA is indeed an aperiodic crystal and that genetic information is conveyed through this irregular pattern. Much like computers, biological systems are programmed to follow a precise set of rules, or algorithms, to store information and solve problems. These biological algorithms direct all manner of biochemical processes to create complex patterns and structures by chemically modifying and assembling individual components. Of course, cells use biochemical circuits not electronic circuits. Single tubulin proteins, for example, follow precise rules of chemistry and physics to spontaneously self-assemble, or polymerize, into the microtubules essential for cell transport and motility. The proteins' binding interactions effect rules that specify how the pieces fit together to form the resulting structure. They also specify when and how tubulins assemble from a nucleation complex—a molecular algorithm governing the logic of polymerization. These complex structures self-assemble with remarkably few mistakes. Though considered quite simple, little is understood about the principles that govern programmable structural order underlying this type of spontaneous self-assembly. In crystals, the simplest example of spontaneous self-assembly, subunits of the whole are arranged in a repeating pattern that extends indefinitely in all directions. If you know the position of one unit in the pattern, you can tell the exact position of every other unit. In a new study, Rothemund and colleagues use DNA to show that crystal growth can be programmed to create specific aperiodic patterns. Inspired by a model of crystal growth as a computational process, they have programmed DNA molecules to act as molecular building blocks, arranging themselves according to local rules that in turn create a complex global pattern. The resulting two-dimensional structures, which self-assemble from knotted DNA complexes (called tiles), grow to create a fractal pattern known as a Sierpinski triangle. These DNA structures—neither periodic (as in quartz), nor random (as in glass), nor pseudorandom (as in quasicrystals with “forbidden” five-fold symmetries)—demonstrate a form of self organization in crystalline materials determined by programmable growth rules, and are hence dubbed “algorithmic crystals.” How can such growth algorithms be encoded in biological molecules? The rules of chemical base-pairing follow regular, predictable patterns, allowing the authors to use DNA to determine the tiles' binding interactions. Fractal patterns from DNA Desired binding interactions between tiles were programmed by endowing each tile with single-stranded “sticky ends” whose sequence was complementary to the sticky ends of tiles it should stick to. Each tile was either white (0) or black (1): a black tile can fit at any site where the two neighboring tiles are opposite colors, while a white tile can fit at any site where the two neighboring tiles are the same color. Logically, the new tile's color is the exclusive-or (XOR) of the tiles in the previous layer. That such logical layer-by-layer iteration of XOR computations will produce the Sierpinski triangle is well known. What's remarkable is that DNA molecules can be programmed to grow according to this logic. With this programmable algorithmic crystal, Rothemund and colleagues demonstrate a method for designing DNA molecules capable of implementing any pattern of abstract logical tiles. What's more, the authors argue, any algorithmic crystal growth process can, in principle, be experimentally investigated using DNA self-assembly. So how is algorithmic self-assembly related to biology? Like the algorithmic crystals, many of the self-assembled structures in biology are ordered but aperiodic. The hope is that the theoretical insights of computer science—well-honed for describing, analyzing, and programming computational systems—can direct investigations of biochemical self-assembly and information processing. And with a method for demonstrating how simple chemical and physical elements can create complex organization, Rothemund and colleagues have added a concrete experimental framework to bolster that work. (For more on DNA and complexity, see “The Emergence of Complexity: Lessons from DNA” by Chengde Mao.)

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC534811.xml

548272

What do hospital decision-makers in Ontario, Canada, have to say about the fairness of priority setting in their institutions?

Background Priority setting, also known as rationing or resource allocation, occurs at all levels of every health care system. Daniels and Sabin have proposed a framework for priority setting in health care institutions called 'accountability for reasonableness', which links priority setting to theories of democratic deliberation. Fairness is a key goal of priority setting. According to 'accountability for reasonableness', health care institutions engaged in priority setting have a claim to fairness if they satisfy four conditions of relevance, publicity, appeals/revision, and enforcement. This is the first study which has surveyed the views of hospital decision makers throughout an entire health system about the fairness of priority setting in their institutions. The purpose of this study is to elicit hospital decision-makers' self-report of the fairness of priority setting in their hospitals using an explicit conceptual framework, 'accountability for reasonableness'. Methods 160 Ontario hospital Chief Executive Officers, or their designates, were asked to complete a survey questionnaire concerning priority setting in their publicly funded institutions. Eight-six Ontario hospitals completed this survey, for a response rate of 54%. Six close-ended rating scale questions (e.g. Overall, how fair is priority setting at your hospital?), and 3 open-ended questions (e.g. What do you see as the goal(s) of priority setting in your hospital?) were used. Results Overall, 60.7% of respondents indicated their hospitals' priority setting was fair. With respect to the 'accountability for reasonableness' conditions, respondents indicated their hospitals performed best for the relevance (75.0%) condition, followed by appeals/revision (56.6%), publicity (56.0%), and enforcement (39.5%). Conclusions For the first time hospital Chief Executive Officers within an entire health system were surveyed about the fairness of priority setting practices in their institutions using the conceptual framework 'accountability for reasonableness'. Although many hospital CEOs felt that their priority setting was fair, ample room for improvement was noted, especially for the enforcement condition.

Background Priority setting, also known as rationing or resource allocation, occurs at all levels of every health care system, including governments, funded provincial/territorial agencies, pharmaceutical benefit-management organizations, hospitals and clinical programs [ 1 ]. Countries with very different health care systems and levels of health care are all grappling with the problem of how to reconcile growing demands and constrained resources [ 2 ]. Hospitals, in particular, are struggling to meet growing demands, affordably, without compromising delivery of services [ 3 , 4 ]. Daniels and Sabin have proposed a framework for priority setting in health care institutions called 'accountability for reasonableness' [ 5 - 7 ], which links priority setting to theories of democratic deliberation, operationalizing the ethical concept of fairness. Fairness is a key goal of priority setting. According to 'accountability for reasonableness', health care institutions engaged in priority setting have a claim to fairness if they satisfy four conditions of relevance, publicity, appeals/revision, and enforcement (described in Table 1 ). Table 1 The four conditions of 'accountability for reasonableness' Relevance Rationales rest on evidence, reasons, and principles that fair-minded parties can agree are relevant to deciding how to meet the diverse needs of a covered population under necessary resource constraints. Publicity Limit-setting decisions and their rationales must be publicly accessible. Appeals/Revision There is a mechanism for challenge and dispute resolution regarding limit-setting decisions, including the opportunity for revising decisions in light of further evidence or arguments. Enforcement There is either a voluntary or public regulation of the process to ensure that the first three conditions are met. 'Accountability for reasonableness' may be an effective framework for describing the components of priority setting, evaluating the fairness of priority setting in hospitals and identifying 'good' practices and opportunities for improvement. It can help hospitals improve their priority setting practices and can be an effective driver of health care reforms [ 8 - 10 ]. 'Accountability for reasonableness' has been used to evaluate priority setting in health systems [ 11 ]. However, only a limited number of studies have addressed how priority setting occurs in hospitals [ 12 - 17 ], and no study has attempted to survey the views of hospital decision makers throughout an entire health system about the fairness of priority setting in their institutions. The purpose of this study is to elicit hospital decision-makers' self-report of the fairness of priority setting in their hospitals using an explicit conceptual framework, 'accountability for reasonableness'. Methods Design We conducted a survey of Chief Executive Officers, or their designates, of Ontario hospitals. The survey questionnaire covered 102 items, including hospital profile information (e.g. hospital name, number of beds, operating budgets). In this paper we focus on the results of 9 questions concerning priority setting, fairness, and the four conditions of 'accountability for reasonableness' (refer to Table 2 ). Table 2 Survey Questions – Rating* and Open Scale Overall, how fair is priority setting at your hospital? (Rating Scale). Please explain your response (Open-ended). How well does your hospital meet its priority setting goal(s)? (Rating Scale). Please explain your response (Open-ended). What do you see as the goal(s) of priority setting in your hospital? (Open-ended) How well is the relevance condition met at your hospital? (Rating Scale) How well is the publicity condition met at your hospital? (Rating Scale) How well is the appeals condition met at your hospital? (Rating Scale) How well is the enforcement condition met at your hospital? (Rating Scale) * Ratings were on a five-point scale, from 1 (not well) to 5 (very well) or from 1 (not fair) to 5 (very fair). Setting and sample scope With 12 million people, Ontario is the largest province in Canada. Like the rest of the country, it is a single payor, predominately publicly-funded health care system with some privately funded services and products (e.g. dental services, drugs). There are 160 hospitals in Ontario and all were invited to participate in this study. Participants 160 Ontario hospital Chief Executive Officers, or their designates, were asked to participate. 86 Ontario hospitals completed this survey, for a response rate of 54%. The average bed count of the responding hospitals was 250.4, with a range of 18 to 1265 beds. The average operating budget of the hospital sample was $75.8 million, ranging from $3.3 million to $733 million. The sampled hospitals represented a blend of teaching, small, community and specialized service facilities across the province. Data collection The survey was pre-tested and mailed to Chief Executive Officers of all Ontario hospitals in January 2001. Data from returned surveys were entered into an electronic database for further analysis. Of the 9 questions analyzed in this study, 6 asked for a rating on a 5-point scale ranging from 'Not Well' or '1' to 'Very Well' or '5', and 3 were open-ended. Data analysis Close-ended ratings were analyzed statistically [ 18 ], with ratings equal to and below the mid-point combined to identify proportions of respondents suggesting improvement, using a conservative bivariate cut-off point to discriminate among reported responses. P values for all hypothesis tests were two tailed. Open-ended responses were analyzed using a modified thematic analysis involving open and axial coding techniques [ 19 , 20 ]. In open coding, data was segmented and coded with a label identifying parts of text relating to a concept or idea. For axial coding, concepts were organized into overarching themes, and compared, both within and between questions, in search of patterns in responses and to ensure consistency. Results Overall, 60.7% of respondents indicated their hospitals' priority setting was fair, while 79% stated their hospitals met their priority setting goals. With respect to the 'accountability for reasonableness' conditions, respondents indicated their hospitals performed best for the relevance (75.0%) condition, followed by appeals/revision (56.6%), publicity (56.0%), and enforcement (39.5%). (Refer to Table 3 ). Table 3 Summary of Decision Maker Responses a,b How well does your hospital meet its priority setting goal(s)? Overall, how fair is priority setting at your hospital? How well is the relevance condition met at your hospital? How well is the publicity condition met at your hospital? How well is the appeals condition met at your hospital? How well is the enforcement condition met at your hospital? # Respondents 81 84 84 84 83 81 5 'Very Well' 15 (18.5%) 15 (17.9%) 21 (25.0%) 14 (16.7%) 12 (14.4%) 9 (11.1%) 4 49 (60.5%) 36 (42.8%) 42 (50.0%) 33 (39.3%) 35 (42.2%) 23 (28.4%) 3 12 (14.8%) 28 (33.3%) 17 (20.2%) 26 (30.9%) 23 (27.7%) 34 (41.9%) 2 3 (3.7%) 4 (4.8%) 3 (3.6%) 9 (10.7%) 10 (12.0%) 9 (11.1%) 1 'Not Well' 2 (2.5%) 1 (1.2%) 1 (1.2%) 2 (2.4%) 3 (3.6%) 6 (7.4%) (SD) 3.89 (0.84) 3.71 (0.86) 3.94 (0.84) 3.57 (0.97) 3.52 (1.00) 3.25 (1.04) a Due to rounding, frequencies may not add up to 100% b Based on survey scale from 1 to 5 Respondents rated the relevance ( = 3.94) condition significantly higher than the publicity ( = 3.57), appeals ( = 3.52), and enforcement ( = 3.25) conditions (p < .05) (paired-samples t test). While publicity and appeals conditions were not significantly different, respondents rated these conditions significantly higher than the enforcement condition (p < .05). The distribution of ratings suggested that there was ample opportunity for improvement, with most room for improvement in meeting the enforcement condition. Priority setting and fairness ratings were positively correlated (r = .51, p < .01). Fairness ratings were also positively correlated with each of the 4 'accountability for reasonableness' conditions (p < .01), as were meeting priority setting goals (p < .05). The bivariate correlation between each of the accountability variables of relevance, publicity, appeals/revision and enforcement was strongly positive (p < .01). The Pearson correlation coefficients among the 4 conditions ranged from 0.41 to 0.57. The internal consistency (Cronbach's α) of all of the 'accountability for reasonableness' conditions was very good (α = .78). Bed size (r = -.24, p <.05) and operating budget (r = -.32, p < .01) were negatively correlated with the rating of fairness in priority setting. Analysis of respondent comments In this section we describe participants' responses to the open-ended questions organized according to themes. We have included verbatim responses to help illustrate the themes. What do you see as the goals of priority setting in your hospital? Decision makers emphasized four priority setting goals which were complex, interdependent, and required balancing. Some respondents said that (1) meeting needs existed in relation to delivering (2) quality services, or was constrained by resource availability. Examples of these goals include "quality care within available resources" and "access to high quality service in areas of greatest need". While (3) meeting budgets was identified as a goal by some decision makers, it did not exist in isolation from other goals such as meeting strategic directions or needs. One decision maker said his goal was "ensuring a balanced situation at year end while meeting the direction set out in our strategic plan". Achieving (4) organizational goals was expressed as a limiting, or organizational focusing process, which involved alignment or balancing of multiple goals and values. How well does your hospital meet its priority setting goals? Please explain your response Decision makers described a deliberative process by which their priority setting goals were operationalized, providing examples of good practices and challenges or barriers for achievement. Overall, decision makers appeared confident that their priority setting goals were the correct ones, in no case did they comment that expressed goals were unrealistic, so contributing to lack of achievement. Decision makers pointed to five factors of importance in meeting their priority setting goals. Review processes (1) were described in which different review processes are brought to bear pointing to areas requiring additional resources. For example, a decision maker said "annually we review and ensure that resources are being allocated to priority programs and services". Leadership (2) was implicated as an important factor in meeting priority-setting goals. For example, a decision maker said "We do not set goals that are pie in the sky, they must be achievable". With respect to (3) stakeholder consultation, some respondents pointed to the need to involve the wider community in decision making. Decision-makers felt that improvements in priority setting were contingent upon (4) access to relevant information. Finally, some decision-makers emphasized the importance of (5) decision making tools, or benchmarking, to improve decision making. Overall, how fair is priority setting at your hospital? Please explain your response Respondents gave information to support their self-evaluation, and provided examples to demonstrate the fairness of priority setting processes from their perspective. In evaluating the fairness of priority setting, decision makers emphasized five themes. Stakeholder consultation raises the level of (1) inclusivity, bringing different points of views and interests to bear in solving common problems. For example, a decision maker said "we have a service providers network in our community that has direct input into resource allocation and program planning". Respondents described review processes (2) involving the deliberation of concerned parties, using relevant data, based on need, cost and other values (e.g. evidence). Reporting systems (3) provide an institutional feedback loop, pointing to areas where limits may be fairly set. With respect to (4) revision/appeals, revising arguments or values based on iterative processes involving various stakeholders is a feature of fairness, and helps to improve buy-in, but may increase organizational tensions. Finally, (5) governance may also be involved in developing decision-making models, and in ensuring the conditions of relevancy, appeals/revision, and publicity are met. For example, a decision maker said "The Board and senior management have developed decision-making models for key program/service changes...our appeals process is informal, but open to access including the Board". Discussion To our knowledge, this is the first published survey of hospital decision makers covering an entire health system, using Daniels and Sabin's framework of 'accountability for reasonableness'. It provides data to understand how fair priority setting processes are, what decision makers' priority setting goals are, and how well these goals are met. Several studies, using either survey or qualitative methodologies sometimes in combination with hypothetical "tradeoffs", have examined priority setting from theoretical perspectives aimed at capturing the public or professional's preferences or values in priority setting [ 21 - 24 ]. Relatively few studies have described what is happening in real-world contexts with a view of evaluating and improving priority setting activities [ 25 - 31 ]. These have described actual priority setting in individual hospitals, focusing on strategic planning [ 12 ], surgery [ 13 ], critical care [ 14 ], new technologies [ 15 ], and the rationing of new drugs [ 17 ], with no study reviewing priority setting in hospitals across the health system. This study makes five contributions to knowledge on priority setting. First, there is ample room for improvement in fair practices within Ontario hospitals as described by their Chief Executive Officers. Although many hospital CEOs felt that their priority setting was fair, ample room for improvement was noted, especially for the enforcement condition, followed by publicity, appeals/revision and relevance. While 79% of decision makers felt their hospitals met their priority setting goals, only 60.7% felt their processes were fair. According to decision makers, the perceived 'gap' was greater in meeting fairness than in priority setting goals. Second, it is feasible for decision makers to assess the fairness of priority setting in their institutions on a quantitative basis according to the 'accountability for reasonableness' framework, with greater than half of respondents saying their processes are fair. There is a high degree of association between 'fairness' and the internal components of 'accountability for reasonableness', lending support that fairness can be operationalized through 'accountability for reasonableness'. As well, the internal conditions of 'accountability for reasonableness' are highly related, and positively correlated, suggesting relevance, publicity, appeals/revision, and enforcement are measuring comparable aspects of fairness. Consistent with this, the high Cronbach's α finding is promising for the development of an 'accountability for reasonableness' scale of perceived fairness of priority setting in health care institutions. Measuring decision makers' self-report of their perception in meeting the various conditions of 'accountability for reasonableness', the scale would provide a practical tool for decision makers to assess self-reported views over time, noting progress in meeting conditions, while providing a reference point for needed improvement (i.e. enhance meeting 'publicity' condition). Finally, the finding of a negative relationship between number of hospital beds or budgets and self-report on fairness is consistent with the view that smaller hospitals may be perceived to have fairer processes, with greater involvement of their local communities in decision making and emphasis on transparency and "trust". Third, decision makers' views were surveyed at a health system level to determine what their actual priority setting practices were. Four complex, interrelated priority setting goals were described, suggesting a balance of need, quality, budget and organizational goals in priority determination. In addition, the study points to a close alignment of factors required in the meeting of priority setting goals and in the elements of fairness. In providing decision maker perspectives on characteristics of fairness, this study also adds to previous empirical work suggesting fair priority setting depends on a fair priority setting process [ 31 ]. Additional research is required to understand the capacity of organizations to improve such practices. Four, this study has shown, from a blend of qualitative and quantitative approaches, that it is feasible to operationalize the 'accountability for reasonableness' framework, through the design of survey instruments to facilitate self-evaluation, and identification of good practices. This data can be shared with decisions makers to improve the fairness of their priority setting processes in meeting the four conditions of 'accountability for reasonableness'. Finally, this study expands on the likely relationship between leadership and priority setting in which leadership was found to contribute to perceptions of fairness in two committees engaged in priority setting for new health care technologies. Study results indicate that greatest room for improvement exists in meeting the 'enforcement' condition, with decision-makers describing critical leadership success factors, including: the role of governance in establishing policy and in meeting this condition, the need to set achievable corporate goals, and the significance of a lobbying funding function. Further study is needed to clarify the nature of this leadership contribution. Limitations The study is limited, first, in the response rate from hospitals. Only 54% of Ontario's 160 hospital CEO's responded to the survey. It is possible that hospitals responding to this survey were those doing, or perceived doing, better on self-report than others. However, there is no evidence to suggest that this is the case, and the sample did include small, medium and larger teaching hospitals, as well as specialized facilities to mitigate against representative selection bias. Second, social desirability bias was possible in that the views of decision makers expressed in the survey may not have corresponded to what they actually believed, or did. However, unpublished data [ 32 ] based on in-depth interviews with Ontario hospital decision makers suggest self-reported views in relation to fairness were similar in both survey respondents and non-respondents, and that reasons for non-response were related to convenience, workload and priority, and not social desirability. Third, corroborative evidence of 'fairness', obtained through in-depth interviews with staff or community stakeholders, or through review of other relevant information (e.g. meeting minutes, publication of rationales on web sites) was not done. The study design was a survey, focusing on the self-report of fairness from the perspective of Chief Executive Officers, or delegates. In-depth interviews with Chief Executive Officers, or additional individual hospital case studies would be helpful in understanding actual priority setting practices, building on these survey results. Conclusions In this first survey of Chief Executive Officers within a health region reporting on their assessment of the fairness of priority setting practices in their institutions according to 'accountability for reasonableness', ample room for improvement was observed, especially for the enforcement condition. Additional study is required to understand how application of 'accountability for reasonableness' will vary within and across institutions, and is shaped or influenced by various parameters, including the nature of corporate leadership. These evaluations can help to identify good practice opportunities for improvement that can be shared between local institutions or used within government to drive health care reforms. Competing interests The author(s) declare they have no competing interests. Authors' contributions DR was the primary analyst and principal author of the manuscript. DKM, CK and PAS were involved in study conception, analysis and drafting the manuscript. CK and DKM were involved in data collection. 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/PMC548272.xml

548514

Relevance of tissue Doppler in the quantification of stress echocardiography for the detection of myocardial ischemia in clinical practice

In the present article we review the main published data on the application of Tissue Doppler Imaging (TDI) to stress echocardiography for the detection of myocardial ischemia. TDI has been applied to stress echocardiography in order to overcome the limitations of visual analysis for myocardial ischemia. The introduction of a new technology for clinical routine use should pass through the different phases of scientific assessment from feasibility studies to large multicenter studies, from efficacy to effectiveness studies. Nonetheless the pro-technology bias plays a major role in medicine and expensive and sophisticated techniques are accepted before their real usefulness and incremental value to the available ones is assessed. Apparently, TDI is not exempted by this approach : its applications are not substantiated by strong and sound results. Nonetheless, conventional stress echocardiography for myocardial ischemia detection is heavily criticized on the basis of its subjectivity. Stress echocardiography has a long lasting history and the evidence collected over 20 years positioned it as an established tool for the detection and prognostication of coronary artery disease. The quantitative assessment of myocardial ischemia remains a scientific challenge and a clinical goal but time has not come for these newer ultrasonographic techniques which should be restricted to research laboratories.

Pharmacologic stress echocardiography is an established cost-effective technique for the detection of coronary artery disease [ 1 ]. The widespread use in the clinical practice has become possible only after evidence collected through large scale multicenter studies that demonstrated its feasibility, safety, diagnostic and prognostic accuracy [ 4 - 8 ]. According to the guidelines of ACC/AHA – pharmacological stress echocardiography with either dobutamine or dipyridamole is a class I indication (of documented effectiveness and usefulness) for the diagnosis of coronary artery disease and for the prognostic stratification of patients with known coronary artery disease [ 2 , 3 ]. Its major limitation is related to a high inter-observer variability and to operator-dependent expertise that might be overcome by an appropriate training and the use of strict reading criteria [ 9 - 11 ]. Nonetheless the hunt for an objective, operator-independent technique to be applied to the conventional black and white regional wall motion analysis remains a major goal in stress echocardiography. Tissue Doppler Imaging (TDI) provides a quantitative analysis of regional myocardial function through the analysis of myocardial velocities [ 12 , 13 ]. Since velocity imaging is confounded by influence from velocities in other segments, the TDI – based modalities strain and strain rate imaging have been introduced to measure regional shortening fraction and shortening rate, respectively [ 14 ] Is the application of Tissue Doppler Imaging to stress echocardiography the technique that will solve it all? According to major journals the answer is yes: the diagnostic accuracy of stress echocardiography improves with TDI when analyzed in comparison with visual assessment of wall motion analysis for the detection of inducible ischemia. Inducible ischemia quantified in a number without the approximations of visual assessment. However the enthusiasm showed by some investigators is not substantiated by scientific results. In fact, a careful analysis of the data published so far raises more doubts than certainties. What we talk about when we talk about TDI The TDI modalities include myocardial velocity imaging, displacement imaging, strain rate imaging and strain imaging. TDI measures velocities by the Doppler shift of reflected ultrasound. Velocities are measured in the conventional imaging planes, from apical views as longitudinal velocities and from parasternal views as radial velocities. When we employ TDI, the velocities within a myocardial segment are the net result of motion caused by contractions in that segment, motion due to tethering to other segments, and overall motion of the heart. This tethering effect is the reason why longitudinal velocities increase progressively from the apex toward the base, when measured in an apical window. Therefore ischemia in the apical region reduces myocardial velocities not only in the apex, but also in the nonischemic basal segments [ 15 ]. In practical terms, the reduction of TDI velocities in basal segments is not synonymous of reduction of function in the same segments. The opposite effect, tethering of nonischemic segments might induce increase in velocity of adjacent ischemic segments. These limitations could be overcome by the employment of strain and strain rate. Strain rate reflects how fast regional myocardial shortening or lengthening occurs measured at two locations separated by a distance. This is the reason why some authors define strain rate as a spatial velocity gradient. Strain is calculated as the time integral of strain rate and is a dimensionless quantity. The limitations of TDI have been widely outlined [ 16 - 18 ] and this is beyond the scope of the present review but they may be synthesized into two main problems: 1 – the amplitude of the estimated velocity is dependent on the angle at which the region is imaged; 2 – the overall cardiac motion, rotation, and contraction of adjacent segments will influence regional velocity estimates. Therefore, a more critical approach to this technology would have avoided the inconsistencies of scientific results when it was applied in the clinical arena. TDI and Stress echocardiography for myocardial ischemia detection Feasibility studies have been published demonstrating the applicability of TDI to stress echocardiography [ 19 - 27 ] but only few studies addressed the issue of its diagnostic accuracy in a clinical context [ 28 - 31 ]. Cain et al [ 28 ] applied myocardial Doppler velocity to dobutamine stress echocardiography in order to assess its diagnostic accuracy when compared to conventional visual assessment. They first identified the normality ranges of myocardial velocities in patients with normal coronary arteries or with a very low pretest probability of having coronary artery disease. Then they selected 114 patient with coronary artery disease assessed at coronary angiography and evaluated the diagnostic accuracy: see Table 1 ( Additional file 1 ). Neither overall nor vascular territory accuracy was better for myocardial velocity when compared to visual wall motion scoring. The MYDISE Study [ 29 ] was the first multicenter study on the absolute value of TDI applied to dobutamine stress echocardiography. The study enrolled 289 patients separated in 3 groups: group 1 (n = 92) healthy volunteers or patients with normal coronary arteries, group 2 (n = 48) patients with known coronary artery disease and group 3 (n = 149) consecutive patients with known or suspected coronary artery disease. Exclusion criteria were: atrial fibrillation, previous myocardial infarction (Q waves on the electrocardiogram, or akinetic segments on the resting echocardiographic images), previous revascularization, unstable angina, complete bundle branch block, significant heart valve disease, contraindication to dobutamine or atropine). The diagnostic criteria were developed by comparing 92 normal subjects with 48 patients with coronary artery disease and applied in a prospective series of 149 patients referred to stress echo laboratory for the suspect of coronary artery disease. Velocity cut-off points were tested and discarded since they did nor perform well when compared to logistic regression models, using systolic velocities at peak stress in 7 myocardial segments and after adjusting for heart rate, age and gender [ 29 ]. The main concerns refer to strict stress echocardiographic issues:1.the lack of a comparison between conventional visual assessment of regional wall motion and TDI analysis. In absolute terms the diagnostic accuracy is not striking: see Table 1 ( Additional file 1 ) and Fig 1 . Even if we accept the hypothesis of a non-inferiority analysis of TDI versus dobutamine stress echocardiography we have to take into consideration some major limitations outlined by the authors: the optimal diagnostic accuracy was obtained by using peak systolic velocity after adjusting for maximal heart rate, age and gender: "ignoring these factors reduces both sensitivity and specificity" [ 29 ]. Moreover, authors applied a very complex regression model for diagnostic accuracy assessment. A recent meta-analysis on dobutamine stress echocardiography showed an overall sensitivity of 80% and a specificity of 87% [ 32 ] (see fig 2 ). 2 – The extent and severity of myocardial ischemia as defined by the number of ischemic segments and the pharmacologic load is never provided. The protocol was interrupted only in the presence of secondary criteria, but never for development of myocardial ischemia since the quantitative analysis was performed off-line. It has been demonstrated that diagnostic and prognostic accuracy of stress echocardiography increases when the response is stratified in the time and space domain, i.e. number of ischemic segments, severity of ischemia induced, the time of onset of ischemia and the pharmacologic dose. 3 – the apical segments have been excluded by the analysis since the systolic velocity is not reliable making the analysis possible only in 11 segments. Nonetheless, the apex and the apical segments are very often the site of stress echocardiographic positivity unless very proximal atherosclerotic lesions are present 4 – the time for performing analysis is never reported. We are informed that the comparison between systolic velocities at rest and at peak stress is disregarded since it is time consuming and increases the potential for observer variability without increasing diagnostic accuracy. 5 -, apparently, TDI cannot be applied to patients with wall motion abnormalities at rest. The exclusion of this group of patients makes this quantitative approach quite unfeasible for routine clinical application. Figure 2 Sensitivity and specificity of dobutamine stress echocardiography Voigt et al. [ 30 ] used a more realistic approach to the application of TDI to stress echocardiography. They first demonstrated in 44 patients with known or suspected coronary artery disease that strain rate quantitatively differentiates ischemic and nonischemic regional myocardial response to dobutamine stress echocardiography [ 30 ] and compared it with conventional visual assessment. The ratio of postsystolic shortening to maximal strain was the best quantitative parameter to identify dobutamine stress induced-ischemia. This quantitative analysis improved sensitivity from 81 (visual assessment) to 86% and specificity from 82% to 89%. The statistical significance is not provided in the manuscript. Then, in the same population of 44 [ 31 ], they compared the visual assessment of wall motion abnormalities with different parameters derived from TDI application such as peak systolic velocity, systolic displacement and strain rate imaging. They employed simultaneous perfusion scintigraphy as a gold standard of myocardial ischemia. The stress echocardiographic methodology employed was not a standard one for segmentation of the left ventricle (18 segments instead of 16 or 17), pharmacologic protocol (up to 2 mg atropine instead of 1 mg) and criteria for ischemia (worsening of wall motion only in 1 segment). Also in this case, the overall accuracy is not striking: Table 1 ( Additional file 1 ). On the basis of these results TDI reduces significantly the diagnostic accuracy of dobutamine stress echocardiography whereas strain rate imaging equals the diagnostic accuracy but it does not improve it. Interestingly enough, the sensitivities and specificities of strain rate imaging are slightly different in the two papers even though the analysis was conducted in the very same set of patients. Moreover, since coronary angiography was performed in all patients, the diagnostic accuracy should have been calculated on this real gold standard instead of perfusion scintigraphy. In a recent [ 33 ] article Marwick et al. questioned the hypothesis that false-negative results of dobutamine stress echocardiography reflect the underinterpretation of regional left ventricular function. On the opposite, the quantitative parameters such as strain rate and peak systolic strain rate were no different between false and true negative tests, suggesting that false-negative results are related to lack of ischemia in a functional sense. On the basis of this observation, quantitative markers are unlikely to increase the sensitivity of dobutamine stress echocardiography. Conclusions The quantitative interpretation of stress echocardiography is not superior to expert wall motion assessment. Open issues in the quantitative analysis remain at stake: which technique to be employed among systolic velocities, strain and strain rate, the assessment of normality criteria of myocardial velocities and how to interpret their values, the management of patients with regional and global left ventricular dysfunction, the analysis of the apical segments, the complexity of the analysis in a real clinical environment, the applicability to unselected populations, its unsuitability to exercise, the most widely used stressor in the clinical practice [ 34 ]. What is presented as a breakthrough technology should have already answered to these issues and when exported into the clinical arena should have provided an incremental value to the established and more easily accessible methods. It is at this point that we are lost in clinical translation: authoritative journals provide data that cannot be transferred into the daily life of a busy stress echocardiographic laboratory, although the general message is optimistic and tend to ignore flaws and limitations of the technique [ 35 , 36 ]). The advantage/disadvantage balance of a new technology should clearly be stated. The potential advantages should always outweigh the disadvantages related to the higher costs and higher complexity of analysis. Perhaps, the shape of the quantitative technology to come has not been designed yet [ 37 , 38 ]. TDI is one of the tools in our hands but apparently this is not its time. At least not on the basis of this evidence. Figure 1 In the left panel sensitivities and specificities in the three main vascular districts and overall sensitivities and specificities without correcting results for age, gender and heart rate. In the right panel, the same data corrected for age, gender and heart rate (modified from ref.29). Supplementary Material Additional File 1 Table 1.TDI vs visual assessment of myocardial ischemia during dobutamine stress echocardiography. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548514.xml

539306

GPs' decisions on drug treatment for patients with high cholesterol values: A think-aloud study

Background The purpose was to examine how General Practitioners (GPs) use clinical information and rules from guidelines in their decisions on drug treatment for high cholesterol values. Methods Twenty GPs were presented with six case vignettes and were instructed to think aloud while successively more information about a case was presented, and finally to decide if a drug should be prescribed or not. The statements were coded for the clinical information to which they referred and for favouring or not favouring prescription. Results The evaluation of clinical information was compatible with decision-making as a search for reasons or arguments. Lifestyle-related information like smoking and overweight seemed to be evaluated from different perspectives. A patient's smoking favoured treatment for some GPs and disfavoured treatment for others. Conclusions The method promised to be useful for understanding why doctors differ in their decisions on the same patient descriptions and why rules from the guidelines are not followed strictly.

Background The medical decision examined in our study is whether or not to initiate drug treatment for high cholesterol values. The topic has been the focus of much debate on the grounds that the proportion of individuals with elevated cholesterol values is high in most Western populations, and that the costs for treating all these people with drugs life-long would be enormous, with a marginal benefit in risk reduction for the majority of them [ 1 - 3 ]. The current Swedish guidelines [ 4 ] from 2003 are national applications of the recommendations on coronary prevention of the Third Joint European Task Force [ 5 ]. In sum, the national guidelines define a total cholesterol value above 5 mmol/l and/or an LDL value above 3 mmol/l as hypercholesterolaemia and the same values as the goals for treatment. As a comparison, the American guidelines are more aggressive in terms of treatment goals for patients with established coronary heart disease and they are more focused on the LDL levels [ 6 ]. The Swedish guidelines identify two group of patients that in general should be offered pharmacological treatment after a sufficiently long trial of life style intervention: the individuals with already established cardiovascular disease (so called secondary prevention cases) or diabetes. A third group that in general should receive medication are patients with familial hyperlipidaemia (FH). For the remaining individuals with a total cholesterol above 5 mmol/l and/or LDL above 3 mmol/l (primary prevention), the same guidelines suggest that the decision to recommend a drug or not should be based on an estimate of the combined risk stemming from the individual's different risk factors. More specifically, the recommendation is to make a numerical risk estimate of the risk for coronary heart disease (CHD) within the next ten years (or the risk projected to 60 years) with a cut-off value at 20%. Based on the results from epidemiological studies, algorithms for arriving at such risk estimates (e.g. Anderson et al. [ 7 ]) have been developed and are available, for instance, as charts in pocket format for doctors. Thus, the decision-making task can be carried out as follows. The first step is to decide whether the patient case is an instance of secondary prevention, diabetes or FH, and if not, to estimate the numerical risk for coronary disease within ten years. A risk above 20% would justify pharmacological treatment, given that life style intervention has been tried for a sufficiently long period. In this study we address the question of how General Practitioners (GPs), who manage most of the cholesterol testing and treatment in Sweden, make such decisions when guidelines are not physically available to them. We will try to highlight the decision-making by examining how it is affected by clinical variables describing the patient and by medical knowledge and decision rules on behalf of the doctors. The reason for studying decisions without access to written guidelines is that as experienced GPs (in the case of three of us), we have found that this is how decisions on cholesterol treatment are usually made. Furthermore, in a previous study concerning the ability of GPs to make numerical estimates of future cardiovascular risks, we asked the GPs if they had access to any tool for making numerical risk estimates at their clinic [ 8 ]. Only nine out of 84 respondents said they had such a tool. GPs' judgments regarding cholesterol treatment have been studied previously using Clinical Judgment Analysis, CJA (for a description of this research paradigm, see Cooksey [ 9 ], for an overview of the medical applications, see Wigton [ 10 ]). The variation among doctors with respect to which information about the patient influenced their decisions most (strategies) seemed to be high [ 11 ], and the strategies were not in accordance with the guidelines for a substantial proportion of the doctors [ 12 ]. The number of patient variables (cues) that influenced the decisions was two or three for most doctors [ 11 , 12 ]. About one-fourth of the doctors did not include coronary heart disease in the irjudgments [ 12 ], in spite of the central role of this risk factor according to the guidelines. The statistical modelling with CJA describes individual doctors' judgment strategies added together for a set of cases, but does not give direct information about the kind of rules or medical knowledge that the participants use in their decisions. In the present study we used the think-aloud technique [ 13 ] in order to learn more about the use of medical knowledge and rules. In addition, with this technique we are better able to study the decisions for individual patient cases. In a previous paper, based on the present data [ 14 ], we coded the think-aloud protocols for preferences concerning two decision alternatives, prescribing a cholesterol lowering drug or not doing so. The codings proved to be reliable. They also appeared to be valid in the sense that there was good agreement on how think-aloud data and rating data, both concerning preferences for prescribing and not prescribing, described the decision process over time for different simulated patient cases. In the present study we link such preference data from think-aloud protocols to different kinds of information describing the patient cases. We also investigate how the use of rules and guidelines can be inferred from the think-aloud protocols. Our first set of research questions concerned how different kinds of information about the patient (e.g. age, sex, previous diseases and laboratory tests) relate to the decision to prescribe a drug or not to do so. First, we estimated the importance of the individual information categories by counting the total number of times they had, according to the coding of the verbal protocols, been valued in a positive or negative direction in relation to the decision at hand. Second, in order to get an idea about why different doctors reach different decisions when presented with identical case information, we made the following analyses. For each of the patient cases separately, the subgroup who decided to prescribe was compared with those with an opposite decision regarding how often they valued different information categories. Third, to further understand how the participants differed in their judgments, we examined which kinds of specified information about a patient (e.g. male sex) are the most likely to lead to disagreement, i.e. to be judged in a positive direction by some participants and in a negative direction by others. Disagreement about the evaluation of data on a given variable may result from different cut-off values, e.g. for the cholesterol variable. A certain value can be considered high for one participant and thus speaking for drug treatment. The same cholesterol value may be considered almost normal by another participant and thereby speaking against drug treatment for the same patient. The age variable may also be associated with disagreement due to different cut-off values. A higher age is generally associated with a higher risk, but there is a lack of evidence for the potential benefit of treating the oldest age groups, and this may introduce different cut-off values for different doctors. Disagreement may also be caused by what might be called different perspectives. If we take smoking as an example all doctors should recognize smoking as a factor associated with an increase in future cardiovascular risk, and should accordingly make statements with a positive directionality for drug treatment. On the other hand, some doctors in some situations may regard actions aimed at smoking cessation as more beneficial than cholesterol reduction, which may give smoking a negative directionality in relation to drug treatment. Overweight can be regarded in the same way, i.e. as an indicator of drug treatment or as indicating change of life style as preferable to drug treatment. Thus, there are two alternative treatment philosophies – drug treatment or life style change – which in turn may be associated with opposite evaluations of the same data in relation to drug treatment. To the extent that these philosophies in fact are associated with different evaluations, one may regard them as different perspectives where certain data (e.g. smoking) are seen from different angles, as risk indicators or as entities that could be changed through patient's own efforts (i.e. by changing life style) as a means to treat the his or her health problem. The latter perspective may also be associated with somewhat moralistic evaluations, e.g. that overweight or smoking is the patient's own choice or own "fault", which in turn would decrease the inclination to initiate drug treatment. Some evidence for this conjecture comes from a CJA study by Evans et al [ 11 ]. They interviewed the doctors after the case presentations regarding which factors they thought had influenced their judgment most. The GPs stated that they were generally less likely to treat people who were overweight. Most were also less likely to treat smokers, but some had the opposite policy. Those less likely to treat smokers were also less likely to treat obese patients. The traditional medical risk factors like diabetes and hypertension may also be associated with either the risk increase perspective or an alternative perspective, where other variables than the cholesterol level are in focus for treatment. Such an alternative perspective should be more likely with a poor control of the blood pressure or diabetes parameters. As this was hardly the case with our case vignettes, and as the moralistic perspective is more likely with life style factors, we expected disagreement to be more frequent with life style variables than with traditional medical factors. Our second set of research questions concerned the use of rules, and the concept of risk as shown in the verbal protocols. Six patient cases were chosen that included two high-risk patients (secondary prevention or diabetes) for whom the guidelines can be transferred to a simple decision rule (e.g. "patients with elevated cholesterol values and previous coronary heart disease should have drug treatment recommended"). Our question was how frequently such decision rules were in the verbal protocols and their content in relation to practice guidelines for elevated blood lipids. For the remaining four cases (primary prevention) no such simple guideline-based rule can be applied and instead, a numerical risk calculation is suggested. We examined the extent to which references to risk estimates were made in the think-aloud protocols. For both secondary and primary prevention cases we were interested in determining how the decisions corresponded to what is indicated by guidelines and risk algorithms. In sum, our research questions concerned: Importance of information (which categories of information about the patients seem to be most important for the decisions?). Patterns of importance for "Yes" and "No" decisions (when each case is analyzed separately, can we get an idea of the reasons behind different decisions by comparing the information evaluation for doctors who chose prescription with those with the opposite decision?). Disagreement (which categories of information give rise to disagreement?). Use of rules (their frequency and contents). Risk estimation (for cases that should be decided by use of a numerical risk calculation, according to the guidelines, how frequent is a referral to the concept of risk estimation in the verbal protocols?). Our approach to analyse think-aloud protocols in a medical decision task for the relative importance of different information categories and the amount of disagreement in the evaluation of these categories has not been tried before as far as we know. We believe that the results can be useful for understanding why doctors reach different decisions in response to the same patient cases and why they often do not act in accordance to guidelines. This knowledge should be useful as an aid to design guidelines and teaching. Methods Design The 20 participants received the same six patient cases and the order of the cases was the same for all participants. Cases with "Yes"- and "No" decisions as the recommended treatment according to the guidelines were mixed as evenly as possible. Ten of the participants were randomly assigned to a condition where, in addition to thinking aloud, they also rated their willingness to prescribe a drug at regular intervals during each case. As was described in a previous paper [ 14 ], this group did not differ from the group without the rating task as regards the prescription decision or the directionality in the think-aloud data. In the present paper the think-aloud data from these two groups have been compiled, whereas the analyses of the rating data are confined to this previous paper. Participants Twenty GPs working in the southern Stockholm area participated. There were 10 males and 10 females. Their ages varied between 34 and 60 years (M = 48.3) and they had practiced between one and 22 years (M = 11.4) as specialists in family medicine. A total of 36 doctors were contacted by telephone. They were selected so as to have a relatively even distribution across different districts in the area and according to gender, but the selection was not random. Twenty-four agreed to participate, but before the session four of these later declined to participate. Cases Six clinical cases were selected from an original set of 40 authentic cases with cholesterol values above normal (at least 5.5 mmol/L). The original set was used in a Clinical Judgment Analysis design with a different sample of doctors and is described in Backlund et al [ 12 ]. Two of the cases, GM (with diabetes mellitus) and AR (with angina pectoris), were obvious high-risk patients and it would be reasonable to use the guidelines in a straightforward manner and recommend treatment for these cases (under the assumption that lifestyle modification had already been tried). Case SH was distinguished by the absence of risk factors other than a moderate increase in cholesterol level, and it would therefore be reasonable to refrain from drug treatment. In the remaining three cases (IS, TW and PU) additional risk factors existed such as smoking and hypertension, and the recommended line of management would be to let the decision be guided by a numerical estimate of the future risk for coronary heart disease. The presently available risk-charts that are referred to in the Swedish [ 4 ] and European guidelines [ 5 ] indicate a 10–20% ten-year risk for case IS (treatment not justified), and a 20–40% risk for case TW (treatment justified). For case PU the calculated risk is 10–20%, which would suggest refraining from drug treatment. However, case PU had a strong family history of coronary heart disease, which is an important risk factor although it is not directly included in the chart. A decision to prescribe a drug is probably justified for this case. Case SH had a calculated risk of 5–10%. For all six cases lifestyle intervention as well as advice concerning diet and exercise had been tried for at least six months before the visit in question. The different kinds of clinical information presented on the six successive screens were divided time-wise in the same way for all six cases. The order in which this information was presented was arranged so as to be as realistic as possible in relation to clinical practice (including how patient cases are described in written referrals to other clinics and in clinical conferences and tutoring). Table 1 shows case IS as it was presented to participants in the study. All previously shown information about a case was repeated on the later screens to reduce and control for memory effects. This part of the text was placed at the top of the screen and was a different colour. Procedure The study was conducted at the doctor's office or in a room nearby. All visits and recordings were made by one of the authors (LB). The cases were presented on a computer screen (Software Question Asker™ (QA) [ 15 ]. In the course of six screens, more clinical information was gradually added to the case. The participants were instructed that authentic cases of hypercholesterolaemia would be presented and that their task was to voice aloud all their thoughts about the case, and that each case would end with the question as to whether or not they would prescribe a drug for this patient. They controlled the shift to a new screen by using a mouse click. When the participant had finished a screen, he or she clicked on a "continue" button. If a participant was silent for more than 10–15 sec, he or she was reminded to voice aloud all thoughts about the information presented. The study was approved by the local ethics committee. Response measures and coding of data Decision Each case ended with a screen with the following question: "Would you prescribe a cholesterol-lowering drug for this patient?" The participant responded by clicking on one of two response alternatives, "Yes" or "No". Think-aloud protocols The sessions were tape-recorded. A secretary then transcribed the recorded sessions into a written, word-by-word format. The protocols were segmented into statements. The next step was to categorize the statements into one of ten categories concerning the general characteristic of the statement (cp. Cognition Categories): Attention, Evaluation, Rule, Explanation, Action pharmacological treatment, Action non-pharmacological treatment, Action other, Want of information, Rating (valid for the ten participants with an additional rating task) and Other. The set of categories is described in more detail in Backlund et al [ 14 ]. Each statement was also assigned one of the values +, -, 0 or x, denoting directionality in relation to the decision task (to prescribe or not to prescribe). The majority of statements that could be assigned a positive or negative directionality had been categorized into either "evaluation" or "application of a rule". However, the most frequent outcome of the categorization was "attention" (the participant read the information aloud, or retrieved it from memory, with neutral or no reformulation), and no directionality could be assigned (i.e. coded as "x"). Zero directionality indicated an explicit statement that the information was neutral in relation to the decision. Finally, each statement was coded with respect to the information referred to. A certain information category could be coded more than once for a given doctor and a given case, unless we regarded the statement as a mere repetition (close in time and identically or almost identically phrased as an earlier statement). The original set contained 21 information categories within the areas of background data, medical conditions, previous diseases, lifestyle factors, physical examination and laboratory tests. Results Reliability of the coding system Two of the authors (LB and YS) independently coded the protocols from the first six participants. Reliability was computed separately for directionality (+, -, 0 or x) and for information category (one of 21) as the percentage of statements that were coded into the same directionality/information category. For these first six participants, the inter-judge reliability was 92% for directionality and 94% for information category. The reliability measures were considered to be satisfactory and therefore only one of the authors (LB) performed the remaining coding. Information categories The original set of 21 different information categories was reduced to eleven. When we ranked the information categories with regard to the frequency of positive or negative evaluations there was a great leap between weight (frequency 14 and rank eleven) and triglycerides (frequency four and rank twelve). We therefore excluded triglycerides and information categories with fewer evaluations. Examples of such excluded categories were information about physical examination of the heart and lungs (normal outcome for each of the cases) and test results concerning liver or thyroid function. The remaining eleven information categories were Cholesterol, LDL (low density lipo-protein), HDL (high density lipo-protein), weight, smoking, CHD (coronary heart disease), diabetes, hypertension, heredity, sex and age. Treatment decisions Table 2 summarizes the information for each case as regards these eleven categories and shows the number of doctors who decided to prescribe a drug as well as the recommended decisions according to the Swedish guidelines. The frequency varied from zero (case SH) to 17 (case AR). "Yes"-decisions and "No"-decisions were equal in frequency when summarized over participants and cases. Figure 1 shows that the majority of doctors chose to prescribe for two, three or four of the six cases. No participant decided to prescribe for all six cases and one participant chose not to prescribe for any of the cases. Importance of information Figure 2 shows that information about cholesterol was evaluated most frequently, both in the positive and the negative direction. It can also be seen that positive evaluations were more frequent than negative ones, except for sex, age and weight, which in part can be due how the final question was worded. For each of the 20 participants we calculated the number of evaluative statements (i.e. with a positive or negative directionality) for each of the eleven information categories as an index of the importance of the information category. A 2 (Direction: Positive/Negative) × 11 (Information category 1–11) × 6 (Case 1–6) ANOVA with three within-group variables and the number of evaluated statements as dependent variable was performed. The main effect of Information category was highly significant, F (4.4; 84.1) = 8.80, p < .01, as was Information category × Case interaction, F (9.1; 172.2) = 11.80, p < .01. (In ANOVA with repeated measures in this study, the degrees of freedom were adjusted according to the Greenhouse-Geisser Epsilon). Thus, as could be expected, the different information categories were evaluated unequally often, and the pattern of relative importance differed in the six individual patient cases. All other main effects and interaction effects were also significant with p < .01. Patterns of importance for "Yes" and "No" decisions In the following, the six patient cases will be analysed separately. This may allow us to detect possible differences in the pattern of importance between different information categories for participants who decided to prescribe a drug and those who made the opposite decision. For each case the number of evaluative statements was the dependent variable in a 2 (Decision: Yes/No) × 11 (Information category 1–11) × 2 (Direction: positive/negative) ANOVA, with the first as a between-group variable and the latter two as within-group variables. The statistical effects are summarized in Table 3 . The main effect of Decision was not significant in any of the cases, indicating that there was no evidence of an association between the number of evaluative statements and decision outcome. For four of the cases the effect of Direction was significant, indicating that positive and negative statements were of unequal frequency. Except for case SH (the case for which all 20 participants decided not to prescribe), positive statements were more frequent than negative ones. For all six cases the different information categories were evaluated unequally often (i.e. main effect of Information). A significant Decision × Direction Interaction, indicating that the decision to prescribe or not to prescribe was associated with different distributions between positive and negative statements, was found in only two of the cases. Direction × Information Interaction was significant or nearly significant for all five cases, suggesting that the distribution of positive and negative directionality was unequal across the different information categories. The most interesting part of these analyses, however, is whether different decisions were associated with different evaluative patterns across the information categories. Statistically, this corresponds to two- or three-way interaction effects including Decision and Information. A significant Decision × Information interaction for a case would indicate that participants with a "Yes"-decision had their number of evaluative statements differently distributed across Information categories compared to participants with a "No"-decision, regardless of whether the direction was positive or negative. The three-way interaction includes the directionality of the statements as well. As can be seen from Table 3 , most of these interaction effects were non-significant. However, the number of evaluations per patient case is probably too small to give enough power for such interaction effects. Three of the cases will be selected to illustrate how this approach may provide hypotheses about information strategies. Case IS represents a 67-year-old female with hypertension as a central risk factor in addition to her cholesterol elevation. She also had a modest heredity. As Figure 3 shows, the 12 participants who decided to prescribe had more positive evaluations of the central risk factor hypertension. In addition, the group who decided not to prescribe seemed to make negative evaluations of information about heredity, suggesting that this information may have been an "argument" against pharmacological treatment for some of the participants in the "No"-subgroup. Case TW (Figure 3 ) represents a case with several risk factors in addition to cholesterol elevation (e.g. smoking and hypertension). An important difference between the groups may be that the "No"-group evaluated the patient's relatively low cholesterol level more often and in the negative direction in relation to pharmacological treatment. Case AR (Figure 3 ) represents a case with CHD (in this case angina pectoris). A comparison between the response groups suggests that greater emphasis was put on CHD by the "Yes"-group, and at the same time there was a negative evaluation, as regards pharmacological treatment, of the patient's (over-) weight by some participants in the "No"-group. Thus, an analysis of the response patterns for these three cases suggests that the "Yes" and "No"-groups differed not only in how much they evaluated the central risk factor(s) (in addition to cholesterol elevation) as favouring drug treatment but also in that the "No"-group seemed to have identified at least one information category as evidence against treatment. As regards the remaining cases, the evaluative pattern for case PU (young woman with severe heredity for CHD) could be interpreted in a similar way, with the patient's (young) age as an argument against drug treatment, whereas case GM (diabetic case) did not invite any such interpretation. For Case SH, no comparison between response groups could be made as all participants decided not to prescribe. Disagreement We defined agreement as the degree to which the same information about a patient case was evaluated with the same directionality. We hypothesized that disagreement would be more common for information about lifestyle-related factors like smoking and weight than would be the case for medical conditions like hypertension and diabetes. There was only one case with a clear overweight and one case where the patient smoked, and the number of evaluative statements concerning these two information categories was therefore rather low. Regarding case TW's smoking, there were 21 statements with a positive direction and ten with a negative direction. For case AR's overweight there were three positive and six negative statements. For hypertension, each of the cases had either only positive or only negative directions (or no statements with directionality at all). The same pattern was found for diabetes and CHD, except that for diabetes two statements concerning case GM were negative compared to 24 statements with positive directionality, and for CHD (case AR) one statements was negative whereas 31 were positive. Thus, with minor exceptions the participants agreed on the evaluations of these three information categories. The data were consequently in line with our hypothesis. There were few evaluative statements concerning the sex of the cases. With one single exception all statements were negative and concerned female cases, which is in line with the known lower risk for female patients to suffer from cardiovascular diseases. A corresponding tendency towards positive evaluations of the male cases was not clearly demonstrated in this material, which could suggest a possible bias in how sex is evaluated as a risk factor. As far as the age variable is concerned, there were both positive and negative statements (i. e. disagreement) for four of the six cases, which was in accord with our expectations. Among the information categories concerning laboratory values, cholesterol was evaluated most often by far, with a fairly even distribution of positive (total 46) and negative (total 39) statements. For at least four of the cases, there were approximately the same numbers of positive and negative evaluations of the same cholesterol value. In other words, according to our definition there is evidence of disagreement among participants in the evaluation of the different cholesterol values. At the level of individual participants, there were eleven instances where doctors made both positive and a negative evaluation(s) of one case. Four of these eleven concerned smoking, two cholesterol, two LDL and one each of hypertension, coronary heart disease and diabetes. Use of rules A total of 32 statements (i.e. not more than 1.6 per participant) were coded as rules. According to our judgment, 18 of these 32 statements were derived from or were compatible with the guidelines (including those rules that were not entirely correct in detail, e.g. regarding the cut-off limits for LDL and HDL) and twelve of these 18 were a more or less directly referring to secondary prevention or diabetes (e.g "He has angina pectoris and should be below 5 in cholesterol"). Examples of other contents for the statements coded as rules were age limit for cholesterol treatment, importance of looking for secondary hypercholesterolaemia, the role of LDL/HDL ratio, priority of smoking vs. pharmacological treatment, the desired blood pressure value for diabetics blood pressure and the cut-off value for ten-year risk for primary prevention. For two of the patient cases (case GM with diabetes mellitus, and case AR with a history of angina pectoris), the guidelines allow a simple decision rule to be applied. Of the 32 instances of reference to a rule, 24 were in connection with these two patient cases. Risk estimation For the four primary prevention cases, IS, TW, SH and PU, a number of statements referring to numerical risk estimate (guidelines say 20% within the next ten years) could have been expected. Only two participants referred to numerical risk estimates. Discussion We discuss first how the doctors evaluated the available information in relation to the decision to be made (i.e., in terms of directionality). When each case was analyzed separately, there was some evidence of different patterns of information use shown by prescribers and non-prescribers. The non-prescribers seemed to evaluate central risk factors with a positive directionality less often than prescribers, and they also appeared to identify at least one information category that was given a negative directionality. This is compatible with theories that describe decision-making as search for arguments or reasons for one or the other decision alternative [ 16 , 17 ]. Paradoxically, the information categories that some doctors used as arguments against treatment were used as arguments for treatment by other doctors. We interpret this finding as showing that prescribing and not prescribing doctors evaluate given information from different perspectives, i.e., from different viewing angles that will put different aspects of the given information in the foreground and background, respectively [ 18 ]. If we take smoking or overweight as examples, they could be seen as risk indicators for CHD (which is what is naturally seen from a drug treatment perspective) or as possibilities for life-style change, which in turn will reduce the patient's future coronary risk (which is what is naturally seen from a life-style change perspective). It may be noted that the doctors did in general not consider both ways of evaluating the information to assess their relative weight for and against a decision to prescribe. Instead, only the aspects supporting this decision or an alternative decision were focused, which is in line with the assumption that the doctors viewed the information from a certain perspective that favoured the decision to be made. For example, we can compare the protocol by participant 6 (decision Yes) regarding Case AR: "He has angina and he has overweight so I will treat him" with the protocol from participant 3 (decision No) regarding the same case: "I would like him to reduce his weight first". The use of life style factors as arguments for prescription decisions was further illustrated in a separate analysis based on the same verbal protocols that also include a task where the doctors were asked to describe freely how they usually reason when they meet patients with high cholesterol values [ 19 ]. The protocols were coded for knowledge of guidelines content and for arguments for the decision to prescribe or not. In several instances the doctor seemed to be fully aware of the contents of the guidelines but still decided to refrain from a strict application of it. The arguments for the decisions in these cases often concerned life style factors like smoking or overweight – either as risk increasing factors or as alternative strategies for intervention. Disagreement was also shown for the age variable. Age is generally considered as positively and monotonically related to risk for future cardiovascular events. At the same time, the guidelines make the reservation that the benefit of giving drugs to very old people is unclear. As far as young patients are concerned, the perspective of ten-year risk appears to be too narrow. The recommended procedure is to increase or project the age to 60 years in order to estimate the risk [ 4 , 5 ]. For doctors with limited experience in using the risk charts, this might be confusing. Cholesterol was another variable that was ambiguous which could be explained in part by the selection of patient cases. Four of the six cases had cholesterol values in the range of 5.0–6.5 mmol/L, which is often labeled as a mild elevation. This might have formed the basis for negative evaluations, i.e. when a value was close to normal a decision to refrain from drug prescription could have been favored. A few participants also commented that the cholesterol values were lower than they had expected, or lower than those of their own patients. Ambiguity in the decision situation due to seeing the situation in terms of different treatment perspectives (relevant for life style factors) or different ideas about the optimal cut-off points (relevant for age and laboratory values) could possibly be reduced by clearer guidelines, which in turn accentuate the need for more research on a number of issues. These issues include the role of life-style factors for coronary heart disease, as well as how patients should be motivated to change their life style, and cost-benefit outcomes of using drug treatment of patients in different age groups and with different cholesterol values. We will now consider the second set of research questions addressed in the present study, viz., the extent to which the participants used certain rules as a basis for their decisions. Based on the verbal protocols, the frequency of statements classified as a rule was rather low, on average 1.6 per participant, and most of the rules concerned secondary prevention. Part of the explanation for the low number of rules might be that the participants were uncertain about the contents of the guidelines and were therefore less willing to talk about them. However, from our separate coding for knowledge of guidelines content referred to above [ 19 ], the conclusion was that the doctors were in general well aware of the distinction between primary and secondary prevention. The low frequency of statements containing a reference to the risk concept could be explained in the same way, since the participants had no immediate access to an aid for calculating risk and were possibly unsure about the general content of such an aid (e.g. the numerical value for ten-year risk that would put the patient into a high-risk category and justify drug treatment). Another reason for the low number of rules might be that the instructions did not encourage the participants to explain their decisions, but simply to state aloud their thoughts about the presented information, which is generally considered as the best method to ensure that the verbal protocols reflect the cognitive processes of interest [ 13 ]. A third possible influence on the use of rules may be that cases that could be handled in a straight-forward way by applying rules from the guidelines were at he same time characterized by having cholesterol values that were only marginally increased above normal, which might have introduced conflict in the decision situation. From the view of evidence-based decisions and quality of care, we can say that many of the cases were difficult and that a considerable spread in the decisions was to be expected. In fact, most of the participants found it difficult to decide about several of the cases, which was evident from interviews after the sessions. On the other hand, the only case (SH) with a mild risk (5–10%) was correctly judged by every participant as not being a candidate for drug treatment. Case AR with angina pectoris represents a decision situation where the guidelines could justify pharmacological treatment in a straightforward manner, and 17 out of 20 chose to prescribe. The presence of diabetes in Case GM could similarly justify drug treatment, but this was the choice for only half the participants. The reason could be that the recommendations concerning diabetes as a risk factor in parity with coronary heart disease is rather new. The Swedish guidelines were published in 1999 and the study was conducted in 2000. One limitation of the present analyses is that most of the conclusions are based on pooled data from groups of participants, while the principal interest is in strategies at the individual level. For example, the opposing evaluations of the same patient data could only be demonstrated between doctors due to the low number of patient cases. After completion of the six cases, the participants were asked to relate in their own words how they usually reason regarding pharmacological treatment when confronted with patients with high cholesterol values. In a forthcoming paper these narratives will be analyzed at the individual level as "scripts" for dealing with cholesterol treatment. We will then have a better understanding of how knowledge and guidelines in this area of medicine are represented in memory, and how these cognitive structures are related to actual decisions and to the individual doctor's think-aloud protocols from processing the cases. Conclusions In this study we have used a new method to analyse a medical treatment decision. Verbal protocols were coded with respect to how different patient variables seemed to favor or not to favor the decision to prescribe a drug or not. The method promised to be fruitful for understanding why doctors reach different decisions in response to the same patient descriptions and why guidelines are not followed. Competing interests The author(s) declare that they have no competing interests. Authors' contributions All authors participated in the design of the study. LB carried out the data collection. LB and YS performed the coding of the protocols. LB performed the rest of the data analyses and drafted the manuscript. All authors participated in the discussion of the draft. 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/PMC539306.xml

555551

Increased daily physical activity and fatigue symptoms in chronic fatigue syndrome

Individuals with chronic fatigue syndrome (CFS) have been shown to have reduced activity levels associated with heightened feelings of fatigue. Previous research has demonstrated that exercise training has beneficial effects on fatigue-related symptoms in individuals with CFS. Purpose The aim of this study was to sustain an increase in daily physical activity in CFS patients for 4 weeks and assess the effects on fatigue, muscle pain and overall mood. Methods Six CFS and seven sedentary controls were studied. Daily activity was assessed by a CSA accelerometer. Following a two week baseline period, CFS subjects were asked to increase their daily physical activity by 30% over baseline by walking a prescribed amount each day for a period of four weeks. Fatigue, muscle pain and overall mood were reported daily using a 0 to 100 visual analog scale and weekly using the Profile of Mood States (Bipolar) questionnaire. Results CFS patients had significantly lower daily activity counts than controls (162.5 ± 51.7 × 10 3 counts/day vs. 267.2 ± 79.5 × 10 3 counts/day) during a 2-week baseline period. At baseline, the CFS patients reported significantly (P < 0.01) higher fatigue and muscle pain intensity compared to controls but the groups did not differ in overall mood. CFS subjects increased their daily activity by 28 ± 19.7% over a 4 week period. Overall mood and muscle pain worsened in the CFS patients with increased activity. Conclusion CFS patients were able to increase their daily physical activity for a period of four weeks. In contrast to previous studies fatigue, muscle pain, and overall mood did not improve with increased activity. Increased activity was not presented as a treatment which may account for the differential findings between this and previous studies. The results suggest that a daily "activity limit" may exist in this population. Future studies on the impact of physical activity on the symptoms of CFS patients are needed.

Introduction Chronic fatigue syndrome (CFS) is characterized by persistent debilitating fatigue often accompanied by a complex of other symptoms (e.g., impaired memory, sore throat, post-exertional fatigue, aching and stiffness in muscles) lasting at least six months that is unresolved with rest or medication [ 1 , 2 ]. A primary component of the case definition of CFS is the inability of patients to maintain their own pre-illness level of activity. Using both self-report as well as accelerometers, previous research on activity levels in this population suggests that individuals with CFS have physical activity levels that are 15% to 40% reduced from those of otherwise healthy sedentary individuals [ 3 - 6 ]. These data suggest that most CFS patients are on the lowest end of the activity spectrum. It is widely accepted that a sedentary lifestyle may greatly increase the risk of development of cardiovascular disease and Type II diabetes, as well as contributing to bone loss, and an age-related loss of function in a person's ability to perform daily activities [ 7 - 9 ]. Recent recommendations by the Surgeon General suggest that accumulating thirty minutes of moderate intensity physical activity per day could provide positive health benefits [ 10 ]. Individuals are encouraged to increase their daily physical activity not only by traditional, structured exercise programs, but also by increasing the amount of unstructured physical activity they perform each day (i.e. the amount they walk each day). The CFS population, with its low daily activity, could derive many positive health benefits from increasing its daily activity, and given their extremely sedentary nature the needed exercise stimulus could easily be met by increasing the amount of walking each subject performs daily. Few randomized controlled trials designed to assess the efficacy of exercise training have been conducted in the CFS population. Two studies used graded exercise performed several days-a-week over a period of 12–26 weeks and both reported an improvement in fatigue-related symptoms and aerobic capacity [ 11 , 12 ]. Additionally, a single subject performed graded aerobic exercise as well as strength training and reported moderate improvements in fatigue-related and other CFS symptoms [ 13 ]. Results from these studies suggest that exercise, in addition to providing positive health outcomes, could also provide beneficial effects to CFS symptomology. Although exercise has been shown to be beneficial, it is unknown to what extent a structured, formal exercise program may alter daily activity in CFS patients. It is common for CFS patients to report exacerbation of their symptoms when too much physical activity is undertaken. CFS patients may have to "rest" more often and/or decrease other types of physical activity in order to compensate for their periods of exercise. The purpose of this study was to assess normal daily physical activity levels in CFS patients using accelerometers over a two-week period and compare these values to a group of healthy sedentary individuals. Additionally, the CFS patients were asked to increase their daily physical activity approximately 30% over a four week period by walking a prescribed amount each day while maintaining their non-exercise daily physical activity. Daily ratings of mood, fatigue, and muscle pain intensity were assessed and compared between the groups as well as in response to increased daily activity in the CFS patients. Methods Participants All experimental procedures were approved by the Institutional Review Board at the University of Georgia, and informed consent was obtained from each participant. All participants were recruited from the general community and either responded to a newspaper ad, responded to a flyer placed around campus, or were referred to the study by their physician. Seventeen CFS and twenty-one controls responded and were screened as possible study participants. A physician's diagnosis of CFS was required for inclusion. Additionally, CFS patients were required to confirm a self-report of decreased physical activity compared to pre-CFS levels, and self-reported inability to sustain high levels of physical activity without a subsequent exacerbation of CFS symptoms for study inclusion. CFS participants with a self-reported history of depression or other psychiatric illness were excluded. Control participants were chosen to be similar in age, height, and weight to the CFS patients and were defined as sedentary by self-report of one bout of regular exercise per week or less. The most sedentary participants (those with desk jobs, etc.) were given first priority for study inclusion. Control participants were also apparently healthy and reported no illnesses or disease conditions. Medications were monitored in all participants. The CFS participants were found to be taking many medications, both prescription and over-the-counter. Analgesics such as Vioxx, Celebrex, Advil, and Aleve were common. Study Design Initially, participants received instructions for wearing the activity monitors and for completing a daily activity log. A "pre" score from the 30-item Profile of Mood States short form questionnaire (POMS) was obtained (The Educational and Industrial Testing Service, San Diego, CA). Participants proceeded to wear the monitors for two weeks during which time they were instructed to maintain normal daily activity. After the two weeks, data were collected from the activity monitors, and the monitors were reset. The participants then wore the monitors for an additional four weeks. The CFS patients were asked to increase their daily physical activity (30% above baseline) during this four week period by walking a prescribed amount each day in order to approximate the daily physical activity of a healthy sedentary person. This was based upon averaging the findings of others that suggested CFS patients had activity levels that were 15% to 40% reduced from healthy but sedentary individuals [ 3 , 4 , 6 ]. CFS patients were given neutral instructions as to whether or not increasing their daily physical activity would alter their mood and fatigue symptoms. Control participants maintained their normal activity for a six week period. Daily activity logs were completed each day. Participants recorded all daily activities, time spent in each activity, as well as time periods when the monitor was not worn (e.g., bathing). Participants also completed a series of questions documenting their daily mood, perceptions of fatigue and muscle pain intensity, and the duration of time fatigue and muscle pain were experienced each day. Objective Measurement of Physical Activity Daily physical activity was assessed by a CSA accelerometer (Computer Science Associates Inc., Fort Walton Beach, FL). To ensure accurate measurements, the procedure recommended by CSA was followed – the monitors were positioned over the subject's anterior superior iliac spine with the belt fitting snuggly so as to limit extraneous monitor movement. Participants were asked to wear the monitors at all times of the day, including sleep. Two minute epochs were used. Data were retrieved from the monitors using a specially designed docking module that input data into a computer. Recommended percent increases in daily activity were calculated based on each subject's average daily counts during their two-week baseline activity period. Counts are arbitrary units assigned to movements detected by the accelerometer. Counts are assigned based upon the magnitude of a change in velocity during a given time period. The number of counts needed to raise daily activity approximately 30% was calculated. Participants were then asked to walk on a treadmill at what they considered to be a comfortable walking pace. Walking speed was recorded, and used to calculate the recommended daily walking time. The approximate number of counts per minute for various walking speeds was assessed prior to the onset of the study (unpublished observations). Additionally, a pedometer was also used to aid participants in achieving the desired daily activity increase. Participants were given an approximate number of steps to take each day during their walk based upon their prescribed walking pace and time. Steps per minute for various walking speeds were assessed in a similar manner as counts per minute prior to the study (unpublished observations). Self-report of Daily Activity and Feelings Data concerning daily activity, mood, fatigue, and muscle pain were obtained from each subject via daily self-report. Participants were asked to report all daily activities and time spent engaged in each. Time periods when the activity monitor was not worn were also reported. A series of five questions concerning daily mood, fatigue intensity, and muscle pain intensity were also answered. Participants ranked, using a 10 cm (0 to 100 mm) visual analog scale their general overall daily mood for that day (with 0 being their best possible overall mood and 100 being their worst possible overall mood), the intensity of their fatigue that day (with 0 being no fatigue and 100 being the highest intensity fatigue imaginable), and the intensity of their muscle pain that day (with 0 being no pain and 100 being the worst imaginable pain). A similar visual analog scale has been used previously in CFS patients to rate daily fatigue [ 13 ]. Additionally, participants reported the amount of time each day they experienced fatigue as well as muscle pain. Once each week participants completed a Profile of Mood States (POMS) short form questionnaire consisting of 30 questions (The Educational and Industrial Testing Service, San Diego, CA) in which participants reported how they had been feeling during the prior seven days. These forms were scored for both fatigue and vigor ratings. Statistical analysis Independent samples t-tests were conducted to compare for differences in subject characteristics between the groups. A repeated measures ANOVA was conducted to determine differences between activity level and mood in CFS and control participants. When a significant interaction was observed a one-way repeated measure ANOVA was performed to analyze simple effects with planned comparisons performed to analyze differences in treatment means. All values are reported means ± SD. Analyses were conducted and significance was assumed at an α level of 0.05. Results Participants No participants (CFS or controls) reported adverse events associated with the increased activity program. Data were obtained from six CFS patients as well as seven sedentary control participants. Additionally, two other CFS patients began the testing protocol but were removed from the study at their own request. One was removed on doctor's recommendation due to an injury (unrelated to the study) and the second was removed due to a change in residence. The physical characteristics of all participants are presented in Table 1 . Mean age, height, and weight, were not different between the CFS patients and the sedentary controls. Five of the six CFS patients also had a physician's diagnosis of fibromyalgia. Table 1 Participants characteristics for six CFS and seven sedentary control participants, values are mean ± SD. Age (Years) Height (cm) Weight (kg) CFS 43 ± 4.6 164 ± 7.3 73 ± 21.2 Control 43 ± 6.5 167 ± 7.0 70 ± 16.7 Daily Physical Activity Individual as well as mean group daily activity counts are presented in Table 2 . All 24-hour periods in which the monitor was not worn for at least 23 hours were removed from the analysis. No trends were observed in daily activity counts in either subject group across all monitoring periods. Day-to-day counts were also relatively stable within a given activity period. Based on this, activity levels are presented as average daily counts during a given activity period. During baseline activity, CFS participants demonstrated 39% lower daily activity counts compared to controls ( P = 0.017). All six of the CFS participants were successful in increasing their daily physical activity. Their daily activity counts increased 28%, on average, during the four-week training period ( P < 0.001). However, it should be noted that 4 of the 6 CFS participants did not reach the prescribed 30% increase in daily activity. Interestingly, following their activity increase the CFS participants had activity levels that were still 24% reduced from those of the control group ( P = 0.08). Table 2 Average daily activity counts for CFS and control participants. Data are mean ± SD. Subject Average Daily Activity (Coun'ts × 10 3 ) %Difference CFS Baseline Increase 1 88.3 ± 24.1 143.4 ± 42.5 + 62.3 2 126.7 ± 19.5 179.1 ± 44.9 + 41.4 3 199.8 ± 38.7 226.7 ± 57.5 + 13.5 4 167.4 ± 32.8 197.4 ± 32.5 + 17.9 5 234.2 ± 37.2 263.8 ± 31.2 + 12.6 6 158.3 ± 32.9 193.5 ± 33.0 + 22.2 Mean 162.5 ± 51.7 200.6 ± 41.2 Control Baseline 1 Baseline 2 1 415.1 ± 79.4 360.7 ± 67.3 - 13.1 2 284.1 ± 87.1 289.5 ± 81.3 + 1.9 3 150.6 ± 61.6 146.2 ± 48.7 - 3.0 4 254.9 ± 60.8 281.9 ± 76.5 + 10.6 5 223.7 ± 73.5 190.9 ± 68.8 - 14.7 6 263.7 ± 60.7 233.9 ± 58.1 - 11.3 7 278.1 ± 106.2 274.0 ± 113.9 - 1.5 Mean 267.2 ± 79.5 253.9 ± 70.0 Self-Report Mood/Feeling Ratings Figures 1 , 2 , and 3 contain daily ratings of overall mood, fatigue intensity, and muscle pain intensity averaged over two week periods. Days where missing data were found (i.e. ratings were not completed), were checked against the activity monitor data to ascertain if the participants simply forgot to fill out the form or if some problem was present that could prevent them from filling out the ratings. Out of 692 possible days, missing data were found for only 19 days. Activity monitor data appeared normal on all 19 of these days. This suggests that the missing data were likely the consequence of participants forgetting to fill out the form rather than due to any adverse medical event. Figure 1 Overall mood ratings. Each time point represents an average of the scores from the previous two weeks. "0" represents the best possible mood and "100" represents the worst mood imaginable. For CFS participants, the two week time point is from baseline activity and the four and six week time points are from increased activity. For control participants, all time points are from baseline activity. * Significant group × time interaction ( P = 0.016). Values are mean ± SD. Figure 2 Ratings of fatigue intensity with "0" being a complete lack of fatigue and "100" being the highest intensity fatigue imaginable. Each time point represents an average of the daily scores from the previous two weeks. For CFS participants, the two week time point is from two weeks of baseline activity and the four and six week time points are from increased activity. For the control participants all time points are from baseline activity. CFS participants did not change significantly over time with increased activity. # Significant difference between CFS and control participants ( P < 0.001). Values are mean ± SD. Figure 3 Ratings of muscle pain intensity with "0" being a complete lack of muscle pain and "100" being the worst muscle pain imaginable. Each time point represents an average of the daily scores from the previous two weeks. For CFS participants, the two week time point is from two weeks of baseline activity and the four and six week time points are from increased activity. For the control participants all time points are from baseline activity. * Significant group by time interaction ( P = 0.030). Values are mean ± SD. Figure 1 demonstrates a significant group-by-time interaction between the CFS and control participants ( P = 0.016, Eta 2 = 0.311) in overall mood. CFS participants reported a worsening of overall mood over time compared to controls. Neither an interaction nor a time main effect was observed in ratings of fatigue intensity. A significant group difference was observed between the CFS and control participants (Fig. 2 , P < 0.001, Eta 2 = 0.892). Although not statistically significant, ratings of fatigue intensity in the CFS group did increase from 58.2 ± 8.5 to 67.0 ± 17.5 (indicating a worsening of symptoms). A significant group-by-time interaction ( P = 0.03, Eta 2 = 0.295) was seen between the CFS and control participants in their ratings of muscle pain intensity (Fig. 3 ). As their daily activity was increased, the CFS participants reported higher intensity muscle pain compared to controls. The amount of time spent with fatigue each day as well as the amount of time spent with muscle pain each day was also reported by both groups. During baseline activity, the CFS participants reported experiencing a significantly greater amount of time spent with fatigue per day compared to the control participants (930 ± 397 min/day. vs. 43 ± 73 min/day; P < 0.001). Additionally, the CFS participants also reported experiencing a significantly greater amount of time spent with muscle pain each day (552 ± 505 min./day vs. 9 ± 22 min./day; P = 0.011). A significant time main effect was found for time spent with fatigue each day ( P = 0.047, Eta 2 = 0.243). A significant difference was found between baseline activity, 451 ± 528 min/day, compared to the final two weeks of increased activity 521 ± 566 min/day ( P = 0.048, Eta 2 = 0.287). The CFS participants also demonstrated a non-significant increase in time spent with muscle pain each day during baseline activity, the first two weeks of increased activity and the final two weeks of increased activity (554 ± 507 min/day vs. 642 ± 546 min/day vs. 713 ± 557 min/day). Control participants demonstrated no change over time in time spent each day with fatigue or muscle pain. Figures 4 5 shows the mean weekly scores on the POMS fatigue and vigor scale for the CFS and control participants. A large and significant difference was observed between the CFS participants and the control participants with respect to their fatigue scores ( P < 0.001, Eta 2 = 0.916). No change was observed in the fatigue scores of the CFS participants as their daily physical activity was increased. The control participants also demonstrated no change over time. Similarly, a significant difference was also observed between the CFS and control groups when vigor scores were compared ( P = 0.036, Eta 2 = 0.343). No change was observed in the vigor scores over time or in response to increased activity in either group. Figure 4 Mean POMS short form ratings (taken once a week) scored for fatigue. Higher scores represent greater fatigue. For CFS participants scores were obtained prior to beginning the study (pre), at the end of each week of baseline activity (1–2) and after each of four weeks of increased activity (3–6). For control participants scores were obtained prior to beginning the study and at the end of each of six weeks of baseline activity (1–6). * Significant difference between the CFS and control participants at all time points ( P < 0.001). Scores did not change over time in either group. Values are mean ± SD. Discussion A primary finding of this study was that individuals with CFS can increase their daily physical activity 28% on average over a four week period. Average daily activity counts, measured by accelerometer, increased in all six of the CFS participants who participated in the study. The magnitude of the observed increases in daily activity ranged from approximately 13% to 60%. These results are consistent with previous studies that have shown that CFS patients can sustain training programs lasting 12 to 26 weeks [ 11 , 12 ]. The major difference between this study and previous training studies was that our goal was to increase total daily physical activity opposed to participating in an aerobic training program several days per week. Participation in a traditional program, even if the subject is compliant, does not insure that the subject increased total daily physical activity. With a traditional training program, it is conceivable that the CFS patients could "rest" between training sessions and consequently not experience a net gain in daily activity. This study demonstrates that with encouragement CFS patients can not only exercise daily, but also sustain enough of their non-exercise daily activities to result in sustained increases in daily physical activity over a period of four weeks. Another interesting aspect of this study was the fact that even after increasing their daily activity 28%, our CFS participants were still approximately 25% less active than our sedentary control participants. We attempted a 30% increase in daily physical activity with the hope of bringing the daily activity levels of our CFS participants up to those of sedentary individuals. This was based upon the findings of others that suggested CFS patients had activity levels that were 15% to 40% reduced from healthy but sedentary individuals [ 3 , 4 , 6 ]. Our results suggest that previous estimates of daily activity in this population may have been overestimated. However, given our small sample size it is possible that our CFS participants were more inactive than those examined in previous studies. It is also possible that our sedentary control group may not have been as inactive as those used in other studies, even though great care was taken to insure the sedentary nature of the group. While additional accelerometer data on the daily physical activity of sedentary women may exist, comparison of data between studies may prove difficult. Variables such as epoch period, monitor location, monitor type (brand), and calibration may all contribute differences in observed daily physical activity. For that reason, we feel it best to limit our comparison to only those subjects in our own study. To our knowledge, this study was one of the first to obtain daily ratings of fatigue intensity from a group of CFS participants under normal daily physical activity conditions as well as during periods of increased daily physical activity. Consistent with their diagnosis, CFS patients reported much greater daily ratings of fatigue intensity, time spent each day with fatigue, as well as fatigue recalled during the prior week compared to healthy sedentary participants. This large difference was expected based on the diagnosis and demonstrated the usefulness of ratings of this type to confirm fatigue symptoms. We found that overall mood, muscle pain intensity, and time spent each day with fatigue worsened following increased activity in our CFS participants compared to controls. Additionally, daily ratings of fatigue intensity (VAS) increased moderately over time while weekly fatigue (POMS) remained stable but elevated in the CFS patients. These observations are contrary to previous studies of increased exercise in this population where exercise training has been shown to reduce fatigue-related symptoms [ 11 , 12 ]. These data demonstrate that, at least in our CFS participants, increases in daily physical activity had no beneficial effects on self-rated fatigue over the measured four week time period. It is not clear why our study did not find increased daily physical activity to reduce symptom severity, as reported in previous studies [ 11 - 13 ]. One possible explanation is that our fatigue ratings provided a more accurate and discriminatory measure of our participants' fatigue symptoms. By providing them with a 0 to 100 visual analog scale to rate fatigue as well as asking them to assess their symptoms each day we may have obtained a more thorough and accurate picture of their fatigue symptoms than those obtained in previous studies using different rating scales. This is an important distinction between the present study and prior related studies. In our assessment of fatigue, we included daily fatigue intensity and duration. Previous studies have employed measures of fatigue-related symptoms that incorporated more symptoms than simply fatigue intensity or duration. For example, Wearden et al. [ 12 ] measured fatigue using the Chalder Fatigue Scale which includes items beyond the scope of fatigue intensity or duration such as sleepiness and "slips of the tongue". Additionally, the reliability of the factor structure of the Chalder Sale has been questioned in the CFS population [ 14 ]. Presentation of exercise as a possible treatment for CFS symptoms could have played a role in the improvements in symptom severity observed with exercise training [ 11 - 13 ]. Cognitive behavior therapy has been shown to be beneficial in CFS patients [ 15 ], and if CFS participants were made to believe that exercise would be beneficial to them then the observed improvements in other studies could represent some form of a placebo effect. Exact instructions given in previous studies were not reported. Care was taken in this study to present the increase in daily physical activity as a neutral intervention. A third explanation of our findings is that there may have been something inherent in our exercise protocol that prevented fatigue symptoms from improving. Our method of activity increase, self-paced walking, was no more strenuous than the exercise used in other studies [ 11 - 13 ]. However, the participants were asked to walk every day, as opposed to 2–4 times per week as in other studies, and to attempt to maintain all non-exercise daily activities [ 11 - 13 ]. It is possible the marked increase in activity each day over several weeks had a cumulative effect. By not providing "rest" days, it is possible that the CFS patients were approaching their daily "activity limit." Conversely, while using larger training volumes, previous training studies in this population may not have observed evidence of an "activity limit" due to the fact "rest" days were provided and that 24-hour physical activity levels were not assessed. This hypothesis is based on self-reports that CFS is associated with an inability to sustain normal daily activity levels without a subsequent worsening of symptoms. If the prescribed increase in daily physical activity caused the CFS patients to approach their tolerable activity limit, this could result in a worsening of their fatigue symptoms. Our data suggests that the CFS participants were able to sustain an increase in daily activity over the course of the study. However, it is worth noting that four of our six CFS patients did not reach the goal of a 30% increase in daily activity. Whether this simply represents non compliance is unclear. However, the CFS patients with the lowest baseline daily activity were able to sustain the greatest increase while the patients with the highest baseline activity experienced the smallest increase. This perhaps points toward an activity limit in the CFS patients. An "activity limit' hypothesis might be consistent with the athletic overtraining syndrome in which athletes report heightened feelings of fatigue and loss of energy [ 16 - 19 ]. In addition to fatigue and energy loss, other features of overtraining such as immune dysfunction are similar to those observed in CFS [ 1 , 2 , 20 ]. Our prescribed exercise was less intense and occurred over a shorter time period than that of previous studies [ 11 - 13 ]. It is possible that more intense exercise over a longer time period is needed to see improvements in fatigue symptoms, and that we simply did not exercise our participants enough to see any improvements. This seems unlikely to us. Although not objectively measured, our CFS participants did subjectively report that maintaining the increased daily activity was difficult for them and they expressed doubts about the possibility of a further increase their daily activity levels. Five of our six CFS participants also reported having fibromyalgia. Fibromyalgia is a related syndrome with a primary symptom of muscle pain and tenderness [ 21 ]. Interestingly, our CFS-FM participants reported no reduction in the intensity of their daily muscle pain as their daily physical activity increased. Similar to their ratings of fatigue intensity, muscle pain showed a trend toward worsening as activity was increased. Exercise programs have been shown to have modest benefits with respect to muscle pain in FM patients [ 22 , 23 ], and it is possible that our muscle pain results should be interpreted in a similar manner to our fatigue results. It is important to note that our small sample size is a key limitation in our study. The study is under powered to detect differences between our groups. Most of the measured variables had small to moderate effect sizes (Eta 2 values from 0.101 to 0.377), suggesting that more participants would be needed to detect differences between the subject groups. Consequently, additional studies will be needed to confirm our results. In conclusion, this study found that individuals with CFS were able to increase their daily physical activity by approximately 28% for four weeks without serious health complications. At baseline activity our CFS participants exhibited significantly lower daily activity than sedentary controls. This reduced level of daily activity was larger than previously reported values[ 3 , 4 , 6 ]. Large differences were seen between CFS and controls in several fatigue ratings as well as ratings of muscle pain (consistent with an additional diagnosis of FM). Ratings of overall mood, muscle pain, and time spent each day with fatigue worsened as daily activity increased, and ratings of fatigue intensity did not improve. These findings are in contrast to those of prior exercise studies in this population, which have suggested exercise as a possible clinical treatment for CFS. It is possible that our CFS participants were approaching their daily "activity limit" and this prevented improvements in fatigue symptoms. Future studies are needed to understand the complex interaction between daily physical activity and fatigue symptoms, and to determine if a daily "activity limit" can be quantified in CFS patients. Figure 5 Mean POMS short form ratings (taken once a week) scored for vigor. Higher scores represent greater vigor. For CFS participants scores were obtained prior to beginning the study (pre), at the end of each week of baseline activity (1–2) and after each of four weeks of increased activity (3–6). For control participants scores were obtained prior to beginning the study and at the end of each of six weeks of baseline activity (1–6). * Significant difference between CFS and control participants ( P < 0.036). Values are mean ± SD.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555551.xml

514525

Diagnostic accuracy of ultrasonography compared to unenhanced CT for stone and obstruction in patients with renal failure

Background To determine accuracy of ultrasound (US) kidney, ureter and bladder (KUB) compared to un-enhanced helical CT (UHCT) in patients with renal failure in the diagnosis of stone and obstruction. Methods This is a case controlled study conducted in the period from June 2000 to July 2003 at a university hospital. All patients had both US and UHCT scan. Patients with serum creatinine ≥ 1.8 mg/dl were included in the study. Only direct visualization of stone was considered as confirmatory. In both the studies, UHCT and US, presence of stone and obstruction were noted. The relevant biochemicals, radiological and clinical records of all the patients were analyzed. Data was analyzed using commercially available software. Results During the period of study 864 patients had UHCT for evaluation of the urinary tract in patients presenting with flank pain. Out of these 34 patients had both UHCT and US done within a span of one day and had serum creatinine of ≥1.8 mg/dl. Mean age was 48 ±15.8 years and 59% of patients were males. UHCT identified renal stones in 21 (62%), whereas 17 of these were identified on US, with a sensitivity of 81%. Of the four patients with renal stones missed on US, three were identified on plain x-ray; the mean size of stones missed was 6.3 mm. Of the 22 (65%) patients with ureteric stone on UHCT, US could only identify 10; a further 7 were identified on x-ray KUB, giving a sensitivity of 45% (US alone) and 77% (US with x-ray KUB). Conclusions US is sensitive and specific for renal stones, 81% and 100% and for hydronephrosis, 93% and 100%, respectively. Its sensitivity to pick ureteric stone (46%) and to identify hydroureter (50%) is low. Addition of x-ray KUB abdomen increases the sensitivity for ureteric stones to 77%.

Background Intravenous urography (IVU) has been the gold standard for the radiological survey of intra renal collecting system, ureter and bladder. Choice of imaging for urinary tract in patients with raised serum creatinine is limited to non-contrast enhanced studies. These considerations have led to the use of other modalities like combination of plain abdominal radiography (KUB) and gray scale ultrasound (US) kidney, ureter and bladder. More recently use of non-contrast enhanced CT (UHCT) and magnetic resonance urography (MRU) in the evaluation of flank pain has received increasing attention [ 1 , 2 ]. Work in the past decade has shown UHCT to be highly sensitive and specific [ 1 , 3 , 4 ]. It is highly sensitive for both renal and ureteric stone [ 3 ]. The probability of misdiagnosis in distal ureter with multiple phleboliths is still a significant problem. Presences of tissue rim [ 3 , 5 ] and comet tail [ 5 ] signs along with secondary signs of obstruction are helpful in these situations. Ultrasound has many inherent advantages, which includes lack of radiation, universal availability, in expensive and non-invasive. It is useful in the diagnosis of renal and ureteric calculi. Stones on US are characteristically demonstrated as highly echogenic foci with distinct acoustic shadowing. The greatest challenge with regard to US is the identification of ureteral calculi, particularly in it's abdominal, and upper pelvic course. This limitation of US is due to its inability to scan retroperitoneum due to overlying bowel loop, and bony structures [ 4 , 6 ] Plain abdominal radiograph also lacks specificity, as phleboliths are not readily differentiated from ureteric calculi. Plain radiographs are also not sensitive to radiolucent calculi and non-calculus obstruction. In the present study we have compared the diagnostic accuracy of UHCT with US with x-ray KUB for the diagnosis of renal and ureteric stones in patients with raised serum creatinine precluding the use of contrast enhanced study. Methods This is a case controlled study conducted in the period from June 2000 to July 2003 at a university hospital. All patients who had both US and un-enhanced helical CT (UHCT) scans performed within a span of 24 hours and a serum creatinine ≥ 1.8 mg/dl were included in the study. Serum creatinine of 1.8 mg/dl is considered as a cut off for use of intravenous contrast by our radiology department. The radiologist's reports on CT, CT films and medical records of patients for suspected renal/ureteral colic were reviewed. The UHCT were obtained on a Cti/pro single slice helical CT scanner (General Electrical medical systems, Milwaukee, WI). The exposure factors setting were KVp 130 and mAS 200–250. All scans were obtained from the upper border of T12 vertebral body to the lower border of symphysis pubis using 5–7 mm collimation, without the use of oral or intravenous contrast material. Patients were placed in supine position with full urinary bladder at the time of the UHCT. Additional prone films were taken whenever the radiologist needed a better description of suspected distal ureteric calculi. Ultrasound KUB was done using 3.75 MHz surface probe. All ultrasounds were seen and reported after being reviewed by a senior radiologist. Secondary signs of obstruction, like hydronephrosis, hydroureter, nephromegaly, perinephric and periureteric stranding were also noted but only direct visualization of stone was considered confirmatory. The relevant biochemicals, radiological and clinical records of all the patients were analyzed. In the studies, UHCT, and US presence of stone and obstruction were noted. Data was analyzed using commercially available software (statistical package for social sciences version 8.0). Results During the 38-month period of study 864 patients had UHCT for evaluation of the urinary tract in patients presenting with flank pain. Out of these 34 patients had both UHCT and US done within a span of one day and had serum creatinine of ≥ 1.8 mg/dl. UHCT was considered as a reference point in the study as all stones identified on the CT were subsequently reconfirmed with interventional treatment or history of spontaneous passage. Mean age was 48 ± 15.8 years (range 20–76 years), 59% of patients were males. UHCT identified renal stones in 21 and ureteric stones in 22 patients. Forty-two (98%) of these stones were confirmed clinically (history of spontaneous passage), or during treatment with ureteroscopy, percutaneous nephrolithotomy and extracorporeal shock wave lithotripsy. Of the 21 renal stones, only 17 were identified on US, with a sensitivity of 81%, specificity and positive predictive value of 100% and negative predictive value of 77%. Of the four patients with renal stones missed on US, three were identified on x-ray KUB; the mean stone size of stones missed on US was 6.3 mm. In all cases US and x-ray KUB were performed prior to the UHCT. Of the 22 patients with ureteric stone, on UHCT, US could only identify 10. Twelve patients with ureteric stones identified on UHCT were missed on US. The mean size of stones missed was 6.1 mm (range 3–15 mm). The sensitivity, specificity, positive and negative predictive values were 46, 100, 100 and 50% respectively. A further 7 patients, missed on US, were identified on x-ray KUB. The overall sensitivity of US and x-ray KUB was 77%. The impact of location of stones missed on US is shown in table 1 and 2 . Table 1 The impact of location on detection of stone and hydronephrosis by UHCT and US. Upper ureter Middle ureter Distal ureter n 6 14 2 Identified on CT 6 14 2 Identified on US 4 5 1 Hydrouretero-nephrosis on CT 6 14 2 Hydrouretero-nephrosis on US 4 8 1 CT un enhanced helical CT US gray scale ultrasound Table 2 Site and size of stones missed on US, the mean size of stones missed was 6.1 mm. Stone identified/total Mean size Upper ureter 2/6 (33%) 6 mm Middle ureter 9/14 (64%) 5 mm Distal ureter 1/2 (50%) 7 mm Discussion IVU has been the traditional imaging modality of choice for evaluation of patients suspected of having urolithiasis and obstruction. Choice of imaging for urinary tract in patients with renal insufficiency and renal failure is limited to non-contrast enhanced studies. Gray scale ultra sonography is the most effective way to exclude sub acute or chronic obstruction. However, regular gray scale US is not accurate in minimally dilated obstruction, such as with partially obstructing ureteric stone; in one series, 4–5% of patients with obstruction showed minimal or no upper tract dilatation [ 7 ]. Duplex Doppler is less effective in acute and incomplete obstruction since obstruction for longer than six hours is necessary to show a consistently elevated resistive index (RI) [ 8 ]. Therefore, we did not evaluate RI values or ureteric jets in our study. Others have also recently examined the role of RI with disappointing results. Cronan showed that the addition of RI did not improve the 77% sensitivity of gray scale US in that series [ 9 ]. US has high sensitivity for renal stones and presence of hydronephrosis. But its sensitivity for ureteral calculi is low. In one study, where IVU was compared with US, the sensitivity of US for ureteral calculi was only 37% (direct visualization) and when hydronephrosis was included as positive sign for ureteral calculi the sensitivity increased to 74% [ 10 ]. Recent studies have demonstrated that UHCT is an excellent method for demonstrating urolithiasis and obstruction in patients presenting with flank pain [ 1 - 3 ]. Smith et al [ 3 ] showed UHCT to be more effective than IVU in identifying ureteral stones. In another comparative study, Sommer et al [ 4 ] noted that reformatted (see Figure 2 ), UHCT images are superior to US and plain radiographs. Data from our institution showed that UHCT has a sensitivity of 99% and specificity of 98% in the diagnosis of ureteric calculi [ 1 ](US and plain radiograph Figure 1 and Figure 2 ). Additionally UHCT could also suggest additional, non-urinary tract abnormalities as cause of flank pain in 12% of patients [ 11 ]. Sensitivity of US is reported to be 96 % for renal stones and is 100% sensitive for stones larger than 5 mm in reported literature [ 1 , 12 ]. In our study US had sensitivity and negative predictive value of 81 and 77%. If x-ray KUB is added the sensitivity increased to 95%. The four patients with renal stones missed on US had a mean stone size of 6.3 mm. Lower sensitivity in our work could be due to small sample size. In the present study US alone had a sensitivity of only 46% for direct visualization of ureteric stones, in combination with x-ray KUB it increased to 77%. The 12 stones missed on US had a mean size of 6.1 mm (range 3–15 mm). Majority of stones missed on US were in the middle ureter (n = 9), 2 were in the proximal ureter and one in the distal ureter. X-ray KUB identified 7 of the 12 stones missed on US. Of these 12 patients with ureteric stones missed only 2 had hydroureter. Presence of hydroureter in patients with ureterolithiasis is valuable as it allows the ureter to be traced to the level of obstruction. Majority (9 out of 12) of stones missed on US were in the middle ureter, an area often obscured by bowel gas. Conclusions In summary, US is the first imaging study for evaluating the patients with previously undiagnosed renal failure. It helps the clinician to separate end stage renal disease from potentially reversible obstructive uropathy secondary to urolithiasis. US is highly sensitive and specific for renal stones in patients with renal failure, it lacks sensitivity for ureteric calculi particularly when they are in the middle ureter. Even addition of x-ray KUB to US misses about a quarter of ureteric stones; we therefore recommend using UHCT if ureterolithiasis is clinically suspected or US and x-ray KUB examinations are equivocal. Due to small sample size, findings of this study should be validated by other studies on a larger cohort of patients. Competing interest None declared. Authors' contribution MHA conceived the idea, analyzed data and drafted the manuscript. AHJ collected the data and analyzed the results. MNS analyzed results and drafted the manuscript. Figure 1 US of 65 years old male presented to emergency room with bilateral flank pain, nausea and vomiting for the past 1 week. He had an ultrasound in a peripheral hospital, which identified hydronephrosis on the right side, and percutaneous nephrostomy tube was placed. His left kidney showed hydronephrosis with renal stone (upper picture). This scan shows small-scarred right kidney (middle picture), pigtail catheter could be identified (arrow) and a proximal ureteric stone could also be seen (lower picture). Figure 2 Reformatted unenhanced helical CT image of the same patient showing small-scared right kidney with proximal ureteric calculus and hydroureter. Left kidney shows hydronephrosis and renal calculus. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514525.xml

534805

High sensitivity 1H-NMR spectroscopy of homeopathic remedies made in water

Background The efficacy of homeopathy is controversial. Homeopathic remedies are made via iterated shaking and dilution, in ethanol or in water, from a starting substance. Remedies of potency 12 C or higher are ultra-dilute (UD), i.e. contain zero molecules of the starting material. Various hypotheses have been advanced to explain how a UD remedy might be different from unprepared solvent. One such hypothesis posits that a remedy contains stable clusters, i.e. localized regions where one or more hydrogen bonds remain fixed on a long time scale. High sensitivity proton nuclear magnetic resonance spectroscopy has not previously been used to look for evidence of differences between UD remedies and controls. Methods Homeopathic remedies made in water were studied via high sensitivity proton nuclear magnetic resonance spectroscopy. A total of 57 remedy samples representing six starting materials and spanning a variety of potencies from 6 C to 10 M were tested along with 46 controls. Results By presaturating on the water peak, signals could be reliably detected that represented H-containing species at concentrations as low as 5 μM. There were 35 positions where a discrete signal was seen in one or more of the 103 spectra, which should theoretically have been absent from the spectrum of pure water. Of these 35, fifteen were identified as machine-generated artifacts, eight were identified as trace levels of organic contaminants, and twelve were unexplained. Of the unexplained signals, six were seen in just one spectrum each. None of the artifacts or unexplained signals occurred more frequently in remedies than in controls, using a p < .05 cutoff. Some commercially prepared samples were found to contain traces of one or more of these small organic molecules: ethanol, acetate, formate, methanol, and acetone. Conclusion No discrete signals suggesting a difference between remedies and controls were seen, via high sensitivity 1 H-NMR spectroscopy. The results failed to support a hypothesis that remedies made in water contain long-lived non-dynamic alterations of the H-bonding pattern of the solvent.

Background The mechanism of action of homeopathic remedies has baffled practitioners and scientists for two centuries. A widely accepted premise of those doing research in this field is that if remedies are more than placebos, then the process of making remedies by alternating dilution and succussion must alter the solvent, encoding in it a "memory" or "information" that biological systems can detect. Experiments attempting to measure or document solvent alteration through direct study of the physical and chemical properties of remedies have so far failed to yield any independently replicated positive effects. Proton nuclear magnetic resonance ( 1 H-NMR, or simply NMR) is among the techniques that have been used to look for differences between remedies and control samples. The term "NMR" encompasses both solvent mobility studies (results are given as a pair of relaxation times denoted T 1 and T 2 ) and analytical studies. In analytical studies, also called spectroscopy, the results are displayed as a graph or spectrum plotting concentration against a variable called chemical shift. There are also more complex applications of NMR such as imaging and two-dimensional NMR that are not relevant to the study of discrete remedy samples. The chemical shift of a proton in a molecule in a sample reflects the (time-averaged) amount of magnetic shielding provided by the electrons making up the covalent or hydrogen bond(s) in which the proton participates, with greater electron density generally correlating with lower chemical shift numbers. Chemical shifts are measured in units of parts per million (ppm) deviation from a reference shift. Recent literature reviews by Baumgärtner [ 1 ] and by Becker-Witt et al .[ 2 ] identified 18 published articles on the use of NMR to study remedies, of which 9 projects used NMR spectroscopy [ 3 - 11 ]. Of these nine, eight [ 3 - 10 ] reported finding differences between remedies and controls when focusing on the relative height, chemical shift, or width of one or more of the peaks due to H's in the solvent. However Aabel and coworkers' recent study [ 11 ] found no differences between remedies and controls. It should be noted that all studies had methodological weaknesses according to the criteria developed by Becker-Witt et al .[ 2 ] One hypothesis to explain how UD remedies might be different from controls states that remedies contain long-lived stable clusters of solvent molecules that are not present in controls [ 12 - 14 ]. If this hypothesis is correct, then the H's making up H-bonds in the stable clusters would experience a bonding environment different from that of the ambient solvent, and they should generate a separate signal on the NMR spectrum. There would be one discrete signal or peak for each symmetry-distinct H in the cluster, along with the expected large peak at 4.8 ppm for the ambient water. "Unexplained" discrete peaks have never been reported in NMR spectroscopy studies of remedies, but we wondered if this might be because cluster concentrations are very small and the methodology has been insufficiently sensitive to detect them. A technical but relevant point must be introduced here. One version of the "cluster theory" posits that once a solvent molecule becomes part of a stable cluster it will stay in that cluster indefinitely. We call such a cluster "non-dynamic". A trickier but more believable hypothesis is that solvent molecules cycle back and forth between the ambient solvent and the clusters. If only one molecule at a time leaves a cluster and it is replaced quickly by another solvent molecule, then the pattern or geometry of the cluster could be sustained without indefinitely tying up any individual solvent molecules. Exchanges of surface hydrogens could also occur rapidly without losing the cluster geometry. A cluster whose components exchange with the ambient solvent would be called "dynamic". The key concept is a parameter called "dwell time", which is the average time a molecule spends in a cluster. If the dwell time is shorter than 10 -3 sec or so, NMR will not be able to "see" it because the chemical shift will be the average over some milliseconds, and the discrete signals will blend back in with the ambient water signal. Dwell times are considerably shorter than 10 -3 sec for a variety of processes, such as ion solvation [ 15 - 17 ] and protein association [ 18 - 22 ]. Therefore NMR spectroscopy is a good method for testing the cluster hypothesis only if, as part of the hypothesis, we postulate that clusters are non-dynamic or that dwell times are on the order of several milliseconds or longer. The project described in this article used a high-sensitivity NMR method to test this possibility, i.e., do remedies made in water contain low concentrations of long-lived non-dynamic regions of structured H-bonding? A good starting point was to quantify the detection limits of previous studies and of the present study. Without a detection limit a negative finding is hard to interpret. The methodology of Aabel et al .[ 11 ], which did not include individual shimming of tubes and collected the standard 16 scans per tube, might be expected to have a detection cutoff around 1 mM on a 500 MHz magnet. That is, peaks representing concentrations of H smaller than 0.001 mol/L would be lost in the noise. This project undertook to improve the detection limit by conducting individual shimming, increasing the number of scans collected, and most importantly, utilizing a high sensitivity method called presaturation. With presaturation, the H's contributing to the 4.8 ppm peak are excited in advance in a manner which causes their contributions to nearly cancel each other out rather than combine into a huge peak. As a result the receiver gain can be increased and signals that would otherwise be overwhelmed become detectable. By combining these techniques, peaks belonging to H-containing compounds at concentrations as low as 5 μM were consistently detected. Assuming unsplit peaks, this means that if a remedy contained a population of a particular stable cluster species at a concentration of 5 μM, the method would be able to detect it. Assuming peaks are split as doublets, the cutoff rises to 10 μM. Although the H's on an H 2 O molecule are normally NMR-equivalent, they need not be equivalent if the H 2 O is embedded in a fixed stable cluster. Then each H of an H 2 O could split the other's peak, yielding doublets. Protons on distinct H 2 O units are far enough apart that their coupling constants can be expected to be too small to generate further peak splitting. Based on this reasoning we assume doublets as the norm for a hypothetical stable cluster, and we take 10 μM as the detection cutoff. For perspective, at 10 μM, we could detect structuring if just 1 in every 10 7 protons were involved in a fixed H-bond. Methods Chemicals Chemicals including 100.0% D 2 O, 100.0%-d DMSO-d6, and 98% 1,4-cyclohexanedione (C 6 H 8 O 2 ) were obtained from Sigma-Aldrich [ 23 ]. Distilled deionized water was either obtained from Sigma-Aldrich or prepared on site via a purifier capable of producing 18 MΩ·cm water. Ultra-pure water was kept in tightly closed polypropylene bottles. Measurements at the point of use found that the water had conductivity no greater than 2 μS (or 0.5 MΩ·cm). The rise in conductivity occurs within minutes as ultra-pure water picks up CO 2 from exposure to air and ions from exposure to glass, and we considered it to be unavoidable. Remedies Two animal ( Sepia and Lachesis ), two plant ( Ignatia and Lycopodium ) and two mineral remedies ( Natrum Muriaticum and Argentum Nitricum ), all of which occur commonly in clinical homeopathy, were used for the project. Commercial remedies were obtained from Helios Pharmacy [ 24 ] and from Washington Homeopathic [ 25 ]. Both pharmacies are widely considered to make quality products that give good clinical results. Helios remedies are regarded by many to be among the best homeopathic remedies in the world. Each pharmacy provided all six remedies at the 12 C, 30 C, 200 C, 1 M, and 10 M potencies. Staff at both pharmacies made the 12 C and 30 C remedies via the Hahnemann process in washed and rinsed vials using distilled deionized water at each dilution, starting from a standard mother tincture (MT). For the 200 C potencies, a 195 C potency in ethanol-water "off the shelf" was used as the starting point for five serial dilutions and succussions in deionized distilled water. Likewise, 1 M and 10 M potencies were derived from "library" 995 C and 9995 C potencies made in ethanol-water. It was a consensus among homeopathic practitioners that despite the switch from ethanol-water to water, this was a valid way of generating 200 C and higher potency remedies. Upon receipt, commercial remedies were stored in the bottles in which they were sent, at room temperature, in a box in a dark cupboard. Each pharmacy also provided an unsuccussed water control. We also made our own mineral remedies up to the 12 C potency starting from MT's consisting of a hand-made 1 M NaCl solution and a stock bottle of 0.1 N AgNO 3 . To make a remedy series, twelve 12 mL capped borosilicate glass vials were labeled 1 C through 12 C, and to each was added 4 mL water. Two drops of MT were added to the first and it was succussed, then 2 drops of the 1 C were added to the next vial and it was succussed, and so on. Transfers were via sterilized Pasteur pipettes. Vials and pipettes were not re-used. Succussion consisted of 120 strokes of forcefully pounding the closed vials held in the fist against a rubber mouse pad on a counter top. Succussed water to be used as a control was made the same way. Over the course of the project, seven series of Nat Mur potencies and six of Arg Nitr potencies were made. Remedies were made less than 24 hours in advance of their scheduled testing times. They were placed into NMR tubes and readied for analysis within an hour of being produced. The tubes generally waited overnight in a light-resistant foil-wrapped container at room temperature before undergoing analysis. NMR Methodology To prepare a sample for analysis, a borosilicate glass NMR tube rated for 500 MHz (Wilmad Lab Glass [ 26 ]) was primed with 50 to 90 μL of locking agent (either D 2 O or DMSO-d6) and 20 μL of a dilute water solution containing a known concentration of a marker molecule (markers tried were acetone (CH 3 COCH 3 ) and 1,4-cyclohexanedione (C 6 H 8 O 2 )). The remedy or control sample was then added to fill the tube to the 700 μL mark, followed by gentle tilting and turning to mix. The marker served several purposes: it provided a reference line for zeroing the chemical shift scale, it provided a reference peak for comparing concentrations, and its sharpness and shape gave feedback about the accuracy of the shimming process. By carefully varying the marker concentration we also determined the method's limits of detection. [C 6 H 8 O 2 was considered an "ideal" marker in that it met all of these criteria: it is available cheaply at high purity; it dissolves in water without altering pH and does not evaporate over time; its 1 H-NMR spectrum has a single unsplit line; its unique peak occurs at a distinctive location far from the water peak and is not easily confused with other common peaks (2.77 ppm); and it will not normally occur as a contaminant or from other sources, so one can be sure its concentration is exactly what one intends.] Proton NMR spectra were obtained at the Department of Chemistry Instrumentation Facility at M.I.T. ("DCIF"). All spectra reported here were obtained on a Varian INOVA 500 tuned to 499.759 MHz and equipped with an inverse probe. Tubes were run with the temperature clamped at either 20°C or 21°C and were not spun. Shimming was done manually for Z1 through Z7 and for first and second order XY magnets, using the lock signal. Presaturation used the standard presat pulse sequence (satdly = 1.5 sec), with the optimal presat frequency being determined to within 0.1 Hz via an array method. Locking, shimming and presat frequency optimization were repeated each time a tube was inserted. Between 128 and 200 transients were collected for each sample. Start-to-finish time for each tube was typically around 35 minutes. Spectrum analyses were done via the standard Varian programs running on an SGI workstation provided by DCIF. Randomization Randomization and blinding are among the recommendations in Ref. [ 2 ] for how to conduct high quality studies on homeopathy. We did not conduct strict randomization, but we did intentionally "mix up" the samples in some respects. We labeled all NMR tubes at the outset and we deliberately changed around which tubes were used for controls and for remedies, in case tube-specific effects were to occur. On most days when we were analyzing remedies we included at least one control, and we always ran the control neither first nor last, in case a time-related trend or drift were to influence the results. Our plan was this: if any signals appeared which seemed to be occurring significantly more often in remedies than in controls, the studies would be followed up with strictly randomized, blinded trials to verify those particular signals. If no significant differences were found, then strict randomization could not alter the outcome, and the follow-up step would be unnecessary. Results We analyzed a total of 57 remedy samples, 5 succussed water controls, and 41 unaltered water controls. Of this total, 28 samples were non-commercial (i.e. made on site) Nat Mur and Arg Nitr remedies of potency 6 C to 12 C. Because the details of the protocol evolved over the course of the project, it could be argued that this total represents the combination of many different experiments. Realizing this, we made a point toward the end of the project, of applying what we considered to be our best methodology to screen a set of what were arguably the best remedies. Specifically, we screened 18 Helios remedies, namely the 12 C, 30 C and 10 M potencies of the six MT's. The screening protocol consistently used 70 μL of D 2 O for locking and concentrations of 12.5 μM or less of C 6 H 8 O 2 as marker. These 18 trials are included in the 57 for the purposes of discussion. Spectra can be conveniently split into those obtained during 2002 (listed by sample and date in Table 1, Additional File 1 ) and the Helios screening run (listed in Table 2, Additional File 1 ). Sensitivity As indicated above, when marker concentration was 5 μM of H or greater (e.g. 0.62 μM of C 6 H 8 O 2 ), a peak was always seen above noise, using the 3σ criterion [ 27 ]. This included four samples where the concentration was just 5 μM. A peak was seen in two of three samples where the concentration was 4 μM and in neither of two samples where it was 3 μM. Thus 5 μM was taken as the detection cutoff for this methodology. Expected and unexpected signals Expected signals that were seen in all spectra included the large water signal peaking at 4.81 ppm, which dominated the spectrum despite presaturation. The water signal is so dominating that small peaks between approximately 4.4 and 5.2 ppm could be "lost" in it, and this interval must be viewed as an inaccessible region of the spectrum for our methodology. Marker peaks for CH 3 COCH 3 and C 6 H 8 O 2 were observed respectively at 2.22 [ 28 ] and 2.77 ppm. 13 C satellites were seen as expected when marker concentrations were high enough and are not listed separately in Table 3, Additional File 1 . When DMSO-d6 was the locking agent, a residual DMSO-d5 quintuplet centered at 2.68 ppm was always present. The focus of this project was to look for discrete peaks other than these expected peaks. Combining the 103 spectra, there were 35 positions where a discrete signal occurred that was not among these expected signals. These positions are listed in Table 3, Additional File 1 . Signals that had the same structure (i.e. singlet, doublet, etc.) and occurred at the same position (within ± 0.01 ppm) in multiple spectra were assumed to have the same genesis. (This assumption could be challenged, but it would not affect our overall conclusions.) We therefore examined each of the 35 "unexpected" positions to see if we could offer an explanation for why signals occurred there. Artifacts Fifteen of the 35 signals were classified as artifacts, i.e. machine-generated spectrum "glitches" that were unrelated to the sample. A signal was dismissed as artifact if it could not be phased consistent with the marker peak. Artifacts were further classified as either "consistent" or "intermittent". The consistent artifacts occurred in more than half of all samples and throughout the eighteen months when our data was collected. Consistent artifacts were seen at 3.60, 8.17, and 11.53 ppm. Signals at positions 8.17 and 11.53 were explained as mid-spectrum artifact and as a "reflection" of the water signal respectively (Δ = 8.17 - 4.81 = 3.36, reflection across midpoint is at 8.17 + Δ = 11.53), but no rationale for the 3.60-ppm artifact was identified. We labeled the consistent artifacts as C1, C2, and C3. Intermittent artifacts were labeled R1 through R12 and are listed in Table 3, Additional File 1 . These artifacts tended to be small, often little more than bumps or wiggles in the baseline. With the marker peak phased to go up (i.e. wholly above the baseline), artifacts might be entirely below the baseline (downgoing), entirely above (upgoing), or both (biphasic). An artifact that occurred repeatedly could have different behaviors in different spectra. Table 3, Additional File 1 indicates whether each artifact was downgoing (d), biphasic (b), or upgoing (u), or whether more than one behavior was seen. The criterion for classification as artifact was that a signal was downgoing or biphasic in at least some of the spectra where it was seen. Half of the intermittent artifacts were seen in only one spectrum each. Those that occurred repeatedly demonstrated a "burst" pattern in the sense that they were present for spectra collected on a single day or during an interval of weeks or months, but were not seen at other times. For instance, the artifact R5 occurred only on 4-Jun-03 and was seen in every spectrum obtained that day (the control as well as the remedies: Table 2, Additional File 1 ). Contaminants Signals that were consistently upgoing could be artifacts or they could be measuring something actually present in the sample. Among such signals, some were evidence of contamination by small organic molecules. Seven signals were positively identified, and an eighth was given a probable assignment. Zacharias [ 29 , 30 ] raised the issue that some amount of contamination was probably unavoidable in remedies, and that it could affect the outcome of studies of remedies. In our spectra signals representing contamination by small organic molecules were frequently seen. Non-commercial remedies and controls often contained acetate (CH 3 COO - , 1.90 ppm [ 28 , 31 ]) or formate (HCOO - , 8.44 ppm [ 28 ]) ions at concentrations ranging from barely detectable (1.7 μM for acetate, 5 μM for formate) to 30 μM. Long soaking of the NMR tubes in water between uses and exercising extra care to avoid hand contact with the remedies decreased but did not eliminate the occurrence of these two contaminants. Lactate (CH 3 CHOHCOO - , 1.32 ppm [ 31 ]) occurred at barely detectable levels in 3 non-commercial samples. The fact that its signal is a doublet with the characteristic 5 Hz coupling constant assists with the identification, and improves one's confidence that one is seeing a real signal and not just noise that happens to occur at the chemical shift of lactate. The α hydrogen signal of lactate, expected around 4.2 ppm, is a quartet and would be predicted to have 1/4 the height of the methyl peak. The quartet would have been a nice confirmation of the lactate identification, but it could not be seen above noise. One signal, an upgoing singlet at 1.28 ppm, occurred in all but one spectrum run using DMSO-d6 between April and November 2002, and did not occur in spectra run using D 2 O or outside this time interval. We pegged it as coming from a contaminant in the DMSO-d6, which was not present in different bottles of DMSO-d6 that were used before April and after November. Its base was broader by a factor of four than the other singlets seen, and despite its trace size we could often make out that it had a symmetric stepped shape like a ziggurat. A good guess, which fits all of these facts, is that it comes from ethylmethylsulfoxide-d7 (CD 3 SOCD 2 CD 2 H), which could plausibly be introduced during the manufacture of DMSO-d6 (CD 3 SOCD 3 ). We did not find anywhere listed the chemical shift of ethylmethylsulfoxide, but the ethyl's methyl group of the very similar molecule ethylmethylketone resonates at 1.26 ppm [ 28 ] in D 2 O. We gave it the identifier 'X' in Table 3. A DMSO-d6 control run on 1-Aug-02 yielded two peaks in the vicinity of 1.3 ppm, one of which may have been 'X'. For the 18 Helios remedies screened with the C 6 H 8 O 2 protocol, results are given in Table 2, Additional File 1 . All but one had measurable quantities of ethanol, with concentrations typically around 300 μM but in one case as high as 3.6 mM (range 116 – 3632 μM, median 310 μM). The ethanol signal, a triplet at 1.17 ppm paired with a quartet at 3.65 ppm, was unmistakable. The ratio between the peak areas of ethanol's methyl triplet at 1.17 ppm and its methylene quartet at 3.65 ppm is theoretically 3:2. In Table 2, Additional File 1 the range of ratios is 1.37 to 1.66. The proximity of the variably-shaped artifact near 3.60 ppm sometimes interfered with accurate determination of the area of the methylene signal, and ethanol concentrations were taken to be 1/3 of the 1.17-ppm peak area. Repeat measurements of a single sample showed that the concentration figures obtained this way have an experimental error of 5 – 10 %. The 18 Helios remedies of Table 2, Additional File 1 all contained CH 3 COO - (range 22 – 214 μM, median 55 μM) and HCOO - (range 8 – 75 μM, median 44 μM). None contained detectable lactate, but we sometimes saw acetone or methanol (CH 3 OH, 3.34 ppm [ 28 ]). Six of the 18 contained detectable acetone (range 3 – 21 μM, median 5 μM) and a different but overlapping set of six held detectable methanol (range 2 – 10 μM, median 4 μM). Because the samples prepared on site never contained detectable ethanol, methanol, or acetone (excepting the added acetone markers), and all samples were analyzed by the same procedure, it is safe to conclude that the Helios samples came with these contaminants. While we admit that our procedures and lab technique were apparently introducing some extraneous acetate and formate, the Helios remedies' levels of these ions typically ran higher than the levels seen in remedies prepared on site (medians of 55 and 44 versus a maximum of about 30 for on site remedies). We deduce that the source of these contaminants in Helios remedies was at least partially from the remedies themselves, i.e. most or all of the Helios remedies came with some acetate and formate in them. The three Helios remedies that we examined using DMSO-d6 contained ethanol, acetate, and formate as well, and one had methanol, but for those runs we did not add a measured marker and precise concentrations were not determined (Table 1, Additional File 1 ). Of the eight Washington Homeopathic remedies that we analyzed and the Washington control, all had ethanol and acetate, all but the control had formate, seven had acetone, two had methanol, and six had lactate. The high rate of lactate occurrence in Washington's samples (6 of 9 vs 3 of 94 for non-Washington) was not due to chance (p < 10 -6 ), but the fact that the six runs exhibiting lactate were all done on one day with no non-Washington control means that we cannot rule out an extraneous source for the lactate. Unexplained signals Twelve positions for upgoing singlets could neither be ruled out as artifacts nor assigned definitely to any known small organic contaminant. These are the "unexplained" peaks, labeled U1 through U12 in Table 3, Additional File 1 . Six of the unexplained group occurred in just one spectrum each. Those that occurred repeatedly exhibited the same kind of temporal burst distribution that was seen for the intermittent artifacts. (The only exception was U1, which was seen on 2-Jan-02 and again on 20-May-02; this may be an example where two different signals coincidentally had similar shapes and chemical shifts and were lumped together.) The ppm range was from 0.47 to 10.60; among those that occurred more than once the range was 0.47 to 6.80. The three upfield-most (i.e. lowest ppm) signals were "broad", i.e. width at half height was between 0.02 and 0.1 ppm, in contrast to the typical "sharp" singlet whose width at half height was 0.002 to 0.004 ppm. For each of the unexplained signals and artifacts we made a 3 × 2 matrix of their occurrence vs non-occurrence in remedies, succussed controls, and unsuccussed controls. None of the matrices had a p-value below .05. The occurrence counts and p-values are listed in Table 3, Additional File 1 . Spectrum example Figure 1 shows a portion of the spectrum of Helios' Ignatia -30 C sample (Line 27 of Table 2). Figure 2 shows a magnification of the same spectrum between 1.8 and 4.0 ppm. Numbers below the x-axis represent integrated peak areas, relative to C 6 H 8 O 2 peak at 2.77 ppm being set to its known value of 100 μM of H. The complex signal between approximately 4.3 and 5.3 ppm is the presaturated water signal. This spectrum contains five contaminants, the three consistent artifacts (artifact at 11.53 not shown), and two intermittent artifacts. Note that the artifacts are biphasic, i.e. have a component below as well as above the baseline, whereas all of the signals due to actual molecules have signals that stay above the baseline. Figure 1 1 H-NMR Spectrum of Ignatia -30 C (Helios): expansion of 1 – 9 ppm region. Key to peaks [position – interpretation]: 1.17 – ethanol, methyl triplet; 1.47 – artifact (R4); 1.90 – acetate; 2.22 – acetone; 2.77 – C 6 H 8 O 2 (1,4-cyclohexanedione marker); 3.35 – methanol; 3.60 – artifact (C1); 3.65 – ethanol, methylene quartet; 3.88 – artifact (R5); 4.3 to 5.3 – water (presaturated); 8.17 – artifact (C2); 8.44 – formate; 11.53 (not shown) – artifact (C3). Figure 2 1 H-NMR Spectrum of Ignatia -30 C (Helios): expansion of 1.8 – 4.0 ppm region. Key to peaks [position – interpretation]: 1.90 – acetate; 2.22 – acetone; 2.77 – C 6 H 8 O 2 (1,4-cyclohexanedione marker); 3.35 – methanol; 3.60 – artifact (C1); 3.65 – ethanol, methylene quartet; 3.88 – artifact (R5). Discussion Contaminants Concerning the organic contaminants, acetate, formate and lactate are present on human skin and can be introduced at the trace levels seen here through ordinary handling. Acetate and formate derive respectively from the β- and α-oxidation of long-chain fatty acids by fibroblasts [ 32 ], and lactic acid is the principal organic component of eccrine sweat (and is found at higher concentrations on the palms [ 33 ]). They are frequently encountered in high sensitivity work. Methanol and acetone are commonly used for rinsing glassware, and a reasonable but unprovable guess is that they represent a residuum from a vial rinse. A plausible but again unprovable guess is that the source of the ethanol was incomplete flushing after prior use of the vials to contain ethanol-based remedies, or perhaps ethanol was also used in the rinsing process. The source of the ethanol for the 200 C and higher potencies is not the 195 C or other library potency: if there were no other source of ethanol, five 1:99 dilutions in water would reduce the level to at most 17 × 10 -10 M. Despite the implication that commercial remedies may be "impure", we consider the levels of detected impurities to be extremely low. Large doses of methanol are hepatotoxic, but the trace amounts here represent no threat to health. Our findings certainly support the widely accepted view that homeopathic remedies are safe. We also believe that the contaminants found should be unlikely to interfere with remedies' clinical effectiveness. The median value of 310 μM corresponds to less than one part ethanol in 50,000 parts water. Working in the early 19 th century, Hahnemann would have used locally obtained well water or spring water to make remedies, and his water would have been far less pure than our most contaminated sample. (Hahnemann also used wine and brandy to make his remedies – hardly the precision solvents of the modern lab.) While high purity solvents are necessary for scientific validity and reproducibility, there is no a priori reason to think they matter for remedies' effectiveness. Indeed, one of the hypotheses on the mechanism of homeopathy is that the "information" is in the geometry of the solvation shells of the low levels of impurities that are present in the solvent, or in some structuring of how solvent impurities group together or interact with each other. Unexplained peaks Six of the unexplained signals occurred in just one spectrum each and could easily be artifacts that happened to be in phase with the rest of the spectrum. There is little to be gained by trying to explain these one-time events, five of which were from unsuccussed controls. The signals U1 to U3, being broad and upfield, could be consistent with complex aliphatic mixtures such as finger grease on the outside of the NMR tubes [ 28 ]. We have no conjecture as to the identity of U5 (2.80 ppm, 3 occurrences) or U6 (4.12 ppm, 2 occurrences). Signal U8 at 6.80 ppm occurred in 7 of 8 D 2 O runs starting in November 2002, which was when a fresh bottle of D 2 O was brought into use, and it was not seen in the DMSO-d6 runs from the same period. This pattern suggests it was a contaminant in the D 2 O, even though it disappeared when the same D 2 O bottle was used in April 2003. There was also one sporadic occurrence in April 2002 (a DMSO-d6 spectrum). A likely assignment is quinone (p-benzoquinone, C 6 H 4 O 2 ), whose spectrum consists of one singlet at 6.80 ppm [ 34 ]. Quinone sublimes at room temperature, and the approximately one micromole of quinone in the 20-mL D 2 O bottle may have simply evaporated during four months. In Table 3, Additional File 1 we also list the occurrence counts for "any unexplained signal" in remedies and in controls, and its p-value of 0.08. Lest this appear as a "trend", we point out that the reason for this relatively low p-value is that the "trend" was for controls to show more unexplained signals than remedies. We do not believe the difference is meaningful. Other aspects of 1 H-NMR spectra We found no discrete signals that occurred significantly more frequently in remedies than in controls. Does this mean that NMR spectroscopy cannot show differences between remedies made in water and controls? Earlier NMR studies focused not on additional signals but on peak positions and shape. For samples in water there is only one peak, the water peak at 4.8 ppm, along with the expected peaks for marker and/or DMSO-d5. We did not examine the shape of the water signal. Presaturation completely destroys any information that may have been present in the water signal shape, and the shape of the water signal after presaturation is sensitive to so many variables that it would be hard to track down the effect, if any, of solvent alteration. Still, this is an avenue that could conceivably be pursued. Similarly, we indicated that any information in the range of 4.4 to 5.2 ppm was essentially lost. It might be possible to examine this range closely, perhaps by subtracting off a smoothed spectrum, to look for signals between 4.4 and 5.2. We have not attempted this. We were primarily looking for separate peaks that could be from long-lived stable clusters, and did not find any. For this reason we couch our results carefully by saying merely that "no discrete signals suggesting a difference were seen." Implications for homeopathic mechanism The absence of any evidence of stable H-bonds above the 5 μM or 10 μM detection limit, in 57 remedy samples, certainly casts doubt on the hypothesis that remedies in water might contain regions of long-lived non-dynamic H-bond structuring. Is this hypothesis salvageable? We do not believe so. We made a point of screening low, medium and high potency remedies, as well as remedies from six common MT's. While it could be argued that even 10% DMSO could disrupt H-bonding and thereby denature (i.e. destroy) H-bond structure [ 35 ], this cannot be said about D 2 O. Nor do we believe that the addition of the marker could have interfered with the detection of structuring. Consider that Helios's remedies, reputedly among the best anywhere, contained comparable or higher concentrations of organic contaminants to start with, than we added. If one wants to maintain the position that homeopathy works (i.e. clinical effects are attributable to it), and that Helios's and Washington's remedies work, then the fact that their remedies come with traces of small organic molecules tells us that such traces cannot be something that interferes with homeopathy's mechanism. Thus it should not "hurt" a remedy to add a tiny bit of organic marker. Could there be signals from fixed H-bonds but we didn't see them because they fell in the inaccessible 4.4- to 5.2-ppm region? The chemical shift of an H in a fixed O-H - - O setup depends on the O - - O separation and the details of the geometry, but studies of ice [ 36 ], organic H-bonds [ 37 , 38 ] and ab initio simulations [ 39 , 40 ] all show that chemical shifts in the range 7 – 16 ppm can be anticipated. For this reason we do not believe that obliteration by the water signal of the 4.4- to 5.2-ppm interval represents important data loss. Our 103 spectra were particularly sparse in signals in the downfield (i.e. > 7 ppm) region. Besides the two "explained" artifacts C2 and C3 and the identified contaminant formate, they contained just 3 single-occurrence downfield unexplained signals and just 3 occurrences of downfield intermittent artifacts. Thus our results were very different from the constellation of downfield signals predicted by the non-dynamic stable cluster hypothesis. We have ruled out, or more precisely we have rendered highly improbable, only this one hypothesis on the nature of the "active ingredient" of homeopathy. Although a positive finding would certainly have been interesting, our negative findings should not be taken as evidence against clinical homeopathy. In particular, the possibility of dynamic alterations of solvent structuring remains open, but this hypothesis will need to be studied by methods that take a much faster "snapshot" of what is going on in samples. Finally, there are also several non-cluster-based hypotheses that have been proposed to explain homeopathy. These include isotopic patterning, coherence, and chaos-based explanations [ 41 ]. These explanations do not require any long-lived H-bonds and do not predict that "unexpected" discrete peaks would be seen in the NMR spectra of remedies. Conclusion We used a high sensitivity 1 H-NMR spectroscopy method to look for discrete signals that could provide evidence of pockets of fixed H-bonding in water-based homeopathic remedies. No such evidence was found. The method did reveal the presence of some small common organic molecules, at levels deemed too low to be problematic. We hope that we have made a contribution to homeopathic research, both by answering a particular question, and by setting a standard for quality hypothesis-driven research that others will be inspired to follow. Competing interests The author(s) declare that they have no competing interests. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional file 1 Table 1: Results of runs during 2002, Table 2: Results of Helios remedies screening, Table 3: List of peak positions Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC534805.xml

555579

Witch-hunt

Beginning two years ago, the US Dept of Health and Human Services began "special reviews" of all current research grants that involved harm reduction, sex and drugs, and continues its ban on funding of needle exchange. With Bush's second term, the campaign was extended to all US funded international programs that dealt with these issues and populations. And, most recently, the US has again undertaken to dominate the discourse within international organizations charged with drug control and AIDS policies – especially those of the UN. But the international harm reduction and human rights community is fighting back in several important ways, including "An Open Letter to the delegates of the Forty-eighth session of the Commission on Narcotic Drugs (CND) of the UN" prepared by a group of 334 well respected public health experts and human rights advocates, protesting U.S. pressure on the U.N. to withdraw its support from harm reduction. This editorial includes the letter and signatures as well as French, Spanish, and Russian versions of the letter as additional files.

"This is a sharp time, now, a precise time - we live no longer in the dusky afternoon when evil mixed itself with good and befuddled the world. Now, by God's grace, the shining sun is up, and them that fear not light will surely praise it." Arthur Miller, The Crucible, Act III "I do not believe that the meaning of our Eighth Amendment, any more than the meaning of other provisions of our Constitution, should be determined by the subjective views of five members of this court and like-minded foreigners" Justice Antonin Scalia, in his dissent from the US Supreme Court majority decision barring capital punishment for crimes committed by minors . It is indeed "a sharp time" in the US for those of us who agree with "like-minded foreigners". This is especially so regarding matters of drug and AIDS policies based on harm reduction (HR) and public health – decriminalization of drug users, the need for safer injections, low thresholds for access to care, sex education and social supports that work to reduce risk. Today more people are imprisoned for drug use in the US than are incarcerated in the European Union for all crimes. US conservatives have, in the same lethal moralistic tradition of our Salem witch-hunts of the 1600's and the McCarthy era, effectively obstructed and undermined our HR efforts at home for two decades. But now the Bush administration, emboldened by its re-election and in full warrior mode, has undertaken a newly invigorated global jihad against harm reduction. Americans who support HR are now to be made to feel "foreign", their moral compass, patriotism, and loyalty to "American core – values" placed in doubt. Parallels to the death penalty are not casual: the failed punitive drug policies of the war on drugs are also official "death sentences", with many more lives lost to them each year than to all the judicially sanctioned executions of all the countries on earth. And, because preventable death is more then an analogy when it comes to public health, we can learn a lot from this Supreme Court case. Justice Anthony M. Kennedy, writing for the courts majority, recognized the "evolving standards of decency" that should shape our judgment of what constitutes a violation of our Constitution's Eighth Amendment and its prohibition against cruel and unusual punishments: "it is fair to say that the United States now stands alone in a world that has turned its face against the juvenile death penalty." But Justice Scalia, significant as the most likely nominee for Chief Justice with the ailing incumbents imminent departure, saw it differently, the NY Times reporting that he reserved "his strongest dissent for (the majority's) reference to international developments that have left the United States alone in supporting juvenile executions". For while the majority opinion said the court was not bound by foreign developments, "it is proper that we acknowledge the overwhelming weight of international opinion" for its "respected and significant confirmation for our own conclusions", Scalia objected that this position implied that "the views of our own citizens are essentially irrelevant," and had wrongly given "center stage" to the "so-called international community." Assumedly this would be that same " international community" that our Declaration of Independence refers to in its famous opening paragraph where, attempting to justify our nations throwing off British rule, we are told that " a decent respect to the opinions of mankind requires that " we explain our course of action. But as the worlds "sole super- power" (tell China that) it appears that we no longer have to show a decent respect for any other nations opinions about human rights, nor it would seem for the relentlessly insistent biology of HIV. The US harm maximization drug policies, which violate both human rights and the realities of infectious diseases, are immoral and dangerously misguided – sustained by demagogic politicians and mad moralists now in near absolute power in our country. As surely as capital punishment, these policies mete out death sentences on a massive scale to our most vulnerable citizens. These failed policies account for the continued annual incidence of 40,000 new HIV infections in the US. Beginning two years ago, the US Dept of Health and Human Services (the parent agency of our National Institutes of Health, which funds most AIDS and drug research in the world) began "special reviews" of all current research grants that involved sex and drugs. Washington re-asserted the drive for mandated "abstinence only" and "faith based" programs, and continues the Federal ban on funding of needle exchange. Don't even bother applying for work on gay sex. With Bush's second term, the campaign was extended to all US funded international programs that dealt with these issues and populations. And, most recently, the US has again undertaken to dominate the discourse within international organizations charged with drug control and AIDS policies – especially those of the UN. Now the US seeks to impose them as extra judicial capital punishment on the rest of the world. But the international harm reduction and human rights community is not going quietly – it is fighting back in several important ways and we can cite many successes that have already saved hundreds of thousands of lives. Henceforth we will chronicle this struggle in this journal. As a start we are publishing An Open Letter to the delegates of the Forty-eighth session of the Commission on Narcotic Drugs (CND) of the UN prepared by a group of well respected public health experts and human rights advocates, protesting U.S. pressure on the U.N. to withdraw its support from harm reduction (see Additional file 1 ). The letter garnered 334 individual and organizational endorsements from fifty-six countries. The organizers of the letter are in the process of sending it to all country missions in Vienna as well as to UNODC Executive Director Antonio Maria Costa and representatives from UNICEF, WHO, UNAIDS, and the UN Office of the High Commissioner for Human Rights. For more information about this letter contact Jonathan Cohen at Human Rights Watch – cohenj@hrw.org We at HRJ welcome your views, which can be submitted online (as Comments) at harmreductionjournal.com • E Drucker PhD is Director, Division of Public Health and Policy Research, Professor of Epidemiology and Social Medicine and Professor of Psychiatry, and Family Medicine at Montefiore Medical Center/Albert Einstein College of Medicine in New York City. Dr. Drucker was a founder of the International Harm Reduction Association and Founding Chairman of Doctors of the World / USA. He is a currently a senior Soros Justice Fellow. Appendix An Open Letter to the delegates of the Forty-eighth session of the Commission on Narcotic Drugs (CND). In a year when the United Nations Office on Drugs and Crime (UNODC) is chair of the governing body of the UN's Joint Programme on HIV/AIDS (UNAIDS), we write to express concern about U.S. efforts to force a UNODC retreat from support of syringe exchange and other measures proven to contain the spread of HIV among drug users. Injection drug use accounts for the majority of HIV infections in dozens of countries in Asia and the former Soviet Union, including Russia, China, all of Central Asia, and much of Southeast Asia. In most countries outside Africa, the largest number of new infections now occurs among injection drug users. As UNODC director Antonio Maria Costa noted at the July 2004 International AIDS Conference, effective responses to injection driven AIDS epidemics require expanded HIV prevention, including syringe exchange, rather than policies that accelerate HIV infections through widespread and indiscriminate imprisonment. Unfortunately, recent events suggest that UNODC – under pressure from the United States – is being asked to withdraw support from proven HIV prevention strategies at precisely the moment when increased commitment to measures such as syringe exchange and opiate substitution treatment is needed. It is particularly alarming that the silencing of UNODC is occurring in a year when the agency is chair of UNAIDS' Committee of Co-sponsoring Organizations and in a year when HIV prevention is a focus of thematic debate at the 48th meeting of the CND. Among the events that have particularly heightened our concern are: * Mr. Costa, who last year expressed support for positive changes in the Russian criminal code, expansion of syringe exchange in countries facing injection driven epidemics and other measures to reduce drug-related harm, has apparently been rebuked by the U.S. State Department. Following a meeting with Robert Charles, U.S. Assistant Secretary for International Narcotics and Law Enforcement Affairs, Mr. Costa pledged to review all UNODC electronic and printed documents for references to "harm reduction" and to be "even more vigilant in the future." * In Southeast Asia, UNODC has suspended a program that sought reduce drug users' vulnerability to HIV prevention through approaches that emphasized public health and drug users' human rights, rather than punishment. * Even syringe exchange, affirmed as an effective and essential part of HIV prevention by UNAIDS, WHO, and UN member nations, has become politically unpalatable. A November e-mail from a senior UNODC staff member asked junior staff to "to ensure that references to harm reduction and needle/syringe exchange are avoided in UNODC documents, publications and statements." We recognize that UNODC is dependent on contributions from donor nations, and that the U.S. is the single largest donor to UN drug control. At the same time, the lives of hundreds of thousands depend on sound, scientific approaches to HIV prevention. Numerous studies, including U.S. government studies, have found that strategies such as syringe exchange and methadone maintenance demonstrably diminish HIV transmission and other health risks. The fact that U.S. delegates declare the evidence in support of syringe exchange "unconvincing," as they did in last year's CND session, should not be allowed to determine the course of the UN drug control and HIV prevention efforts, which are inextricably and essentially linked. Nor should UNODC – a co-sponsor of UNAIDS, and an agency with an essential role to play in the course of the HIV epidemic – be asked to refrain from public statements about needle exchange simply because they do not fall within the realm of what the U.S. deems acceptable. Strategies that attempt solely to achieve abstinence from drug use do not constitute an acceptable alternative to programs, such as syringe exchange, that help active drug users protect themselves from HIV/AIDS. Experience has shown that "zero tolerance" drug control efforts can have the effect of driving injection drug users underground and away from drug treatment and other health services. This is particularly true where, as in many countries, counter-narcotics efforts lead to false arrest, beatings and extortion by police, prolonged detention without trial, forced drug treatment, disproportionate incarceration in cruel conditions and, in some cases, extrajudicial execution. Programs such as syringe exchange and opiate substitution, by contrast, both prevent HIV infection and can provide a bridge to other health services. Restricting these programs is a blatant infringement of drug users' human right to health. As you gather this year to debate HIV/AIDS prevention and drug abuse, we respectfully urge you to support syringe exchange, opiate substitution treatment and other harm reduction approaches demonstrated to reduce HIV risk; to affirm the human rights of drug users to health and health services; and to reject efforts to overrule science and tie the hands of those working on the front lines. No less than the future of the HIV epidemic is at stake. cc: Joint United Nations Programme on HIV/AIDS World Health Organization Office of the High Commissioner for Human Rights International Narcotics Control Board Organizations and individuals who have signed this letter as of March 1, 2005 are listed in Additional file 1 . For the French version of this Open Letter please see Additional file 2 . For the Spanish version of this Open Letter please see Additional file 3 . For the Russian version of this Open Letter please see Additional file 4 . Supplementary Material Additional File 1 Open Letter to the delegates of the Forty-eighth session of the Commission and other additional information. Click here for file Additional File 2 French version of the Open Letter Click here for file Additional File 3 Spanish version of the Open Letter Click here for file Additional File 4 Russian version of the Open Letter Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555579.xml

555545

Antibiotic activity of telithromycin and comparators against bacterial pathogens isolated from 3,043 patients with acute exacerbation of chronic bronchitis

Background Antimicrobial therapy is considered an important component in the medical management of most patients with acute exacerbation of chronic bronchitis (AECB). The three predominant bacterial species isolated are nontypeable Haemophilus influenzae , Moraxella catarrhalis , and Streptococcus pneumoniae . Staphylococcus aureus is also frequently isolated while atypical bacteria are thought to cause up to 10% of exacerbations. Antibacterial resistance is increasing worldwide and little surveillance data exist concerning pathogens isolated from patients with AECB. Methods This study examines the prevalence of antibacterial resistance in isolates obtained from patients with clinically diagnosed AECB. A total of 3043 isolates were obtained from 85 centres in 29 countries, between 1999–2003, and were tested against the new ketolide telithromycin and a panel of commonly used antibiotics. Results and Discussion Of the S. pneumoniae isolates, 99.9% were susceptible to telithromycin, but only 71% were susceptible to erythromycin and 75.3% to penicillin. Of the H. influenzae isolates, 99.6% were susceptible to telithromycin. 11.7% of these isolates produced β-lactamase. Almost 10% of S. pneumoniae were multidrug-resistant; 99.0% of these isolates were susceptible to telithromycin. Telithromycin also demonstrated good in vitro activity against M. catarrhalis (MIC 90 = 0.12 mg/L) and was the most active compound against methicillin-susceptible S. aureus (98.9% susceptible). Conclusion Telithromycin demonstrated similar or better activity against the bacterial species investigated than the other agents, with the most complete coverage overall. These species are the predominant causative bacterial pathogens in AECB and thus the spectrum of activity of telithromycin makes it a potential alternative for the empirical treatment of AECB.

Introduction The World Health Organization (WHO) estimates that chronic obstructive pulmonary disease (COPD) is the fourth leading cause of death worldwide. In the year 2000, it was estimated that 2.74 million people died from COPD worldwide [ 1 ]. COPD is defined by the presence of irreversible or partially irreversible airway obstruction in patients with chronic bronchitis or emphysema [ 2 , 3 ]. The disease is characterized by recurrent (1–4 per year) acute exacerbations of chronic bronchitis (AECB), defined by a subjective increase from baseline of one or more symptoms including shortness of breath, cough, sputum production, and sputum purulence [ 4 ]. The precipitating factors for AECB have been extensively researched and determined to be heterogeneous with complex aetiology [ 5 - 10 ]. Results from a number of placebo-controlled clinical investigations have demonstrated that antibacterial agents are of significant clinical benefit in the treatment of AECB, particularly for those patients with at least two of the three cardinal symptoms of AECB (worsening dyspnoea, increased sputum volume, and increased sputum purulence) and/or severe airway obstruction [ 11 - 13 ]. Other clinical trials measuring non-traditional endpoints have shown that antibiotic therapy reduces the time to symptom resolution and has long-term benefits including greater intervals between episodes of exacerbation [ 14 , 15 ]. Consequently, antibiotic therapy is considered an important component in the medical management of patients with AECB. Bacteria can be isolated from 40–60% of sputum samples of patients experiencing AECB [ 16 ]. The three predominant bacterial species isolated are non-typeable Haemophilus influenzae , Moraxella catarrhalis , and Streptococcus pneumoniae . Other less frequently isolated potential pathogens are Gram-negative enterobacteria, Haemophilus parainfluenzae , Staphylococcus aureus , and Pseudomonas aeruginosa . Gram negative enterobacteria and Pseudomonas aeruginosa are more frequently isolated in patients with severe underlying disease [ 10 ]. Viral infections are present in approximately 30% of exacerbations, Mycoplasma pneumoniae in 1–10%, and Chlamydophila pneumoniae in 4–5% (serologically identified) [ 6 - 10 ]. Amoxycillin, ampicillin, sulfamethoxazole-trimethoprim (trimethoprim-sulphamethoxazole), tetracyclines, and erythromycin are considered first-line antimicrobial therapy for AECB [ 17 ]. The clinical utility of these agents is, however, being hampered by the increasing global spread of pathogens with resistance to one or more of these agents. Up to 40% of H. influenzae isolates and more than 90% of M. catarrhalis isolates produce β-lactamase and this limits the value of penicillins and some other β-lactams [ 18 ]. Furthermore, resistance to penicillin and macrolides has spread rapidly among isolates of S. pneumoniae [ 19 ]. Other agents used include extended spectrum cephalosporins, amoxycillin/clavulanate, azithromycin, clarithromycin, and levofloxacin. Telithromycin is the first ketolide available for clinical use. Derivatives of erythromycin-A, the ketolides, like the macrolides, exert their antimicrobial action by binding to the bacterial ribosome. Although both macrolides and ketolides bind strongly to a region of domain V in the 23S rRNA of the ribosome, telithromycin has additional strong binding to a region in domain II to which the macrolides bind weakly [ 20 ]. Ketolides are also poor substrates for the efflux pump (mefA) responsible for macrolide resistance in S. pneumoniae [ 21 ]. Consequently, telithromycin has been found to have potent activity against macrolide resistant S. pneumoniae with methylase, efflux or ribosomal mutations as the mechanisms of resistance [ 22 , 23 ]. There is a need for alterative therapeutic options for the treatment of AECB and surveillance data are needed to help determine the suitability of new agents. The PROTEKT (Prospective Resistant Organism Tracking and Epidemiology for the Ketolide Telithromycin) study is an international, longitudinal, antibacterial resistance surveillance study, which was initiated in 1999 to monitor the spread of resistance among respiratory tract pathogens worldwide. Here we analyze the in vitro antimicrobial activity of bacterial isolates obtained from patients clinically diagnosed with AECB in 3 consecutive years of the PROTEKT study. Using these data, and previously published clinical data, the potential role of telithromycin in the treatment of AECB will be discussed. Materials and Methods Patients and bacterial isolates Details of the study design, including the selection of patients and the methodology for the identification of isolates and their storage in the PROTEKT study has been described previously [ 24 ]. Isolates in this study were obtained from patients diagnosed with AECB from in 85 centres in 29 countries (Table 1 ). To be included in this analysis, an isolate was deemed pathogenic in AECB by clinical and laboratory findings. Isolates were only acceptable if the patient was ≥ 30 years old and the specimen was obtained from blood, bronchoalveolar lavage (BAL), or sputum. Isolates from patients diagnosed with AECB obtained from other sites (e.g., ear, throat, nasopharynx) and isolates obtained from patients <30 years of age were excluded from this analysis because AECB is more likely to be present in patients ≥ 30 years of age and the responsible bacterial pathogen is more likely to be correctly isolated from the blood, BAL, or sputum. Table 1 Geographical distribution of isolates from AECB patients used in this study Area Countries Centres Isolates North America 2 4 319 South America 6 14 427 Europe 13 41 1847 Australasia 6 19 437 South Africa 1 6 13 Totals 29 85 3043 In Year 1 (1999–2000), each centre had a quota of 60 isolates of S. pneumoniae , 40 H. influenzae , 15 H. parainfluenzae , 20 M. catarrhalis , 25 Streptococcus pyogenes and 20 S. aureus to collect. In years 2 (2000–2001) and 3 (2001–2002), H. parainfluenzae were not collected and 15 extra isolates of S. pneumoniae were collected instead. Antimicrobial testing The comparator agents used were four β-lactams; penicillin (for S. pneumoniae and S. aureus ), ampicillin (for H. influenzae , H. parainfluenzae and M. catarrhalis ), amoxycillin/clavulanate, and cefuroxime, three macrolides/azalides; erythromycin, clarithromycin, and azithromycin, the folate synthesis inhibitor; trimethoprim-sulphamethoxazole, the tetracycline; tetracycline and a fluoroquinolone, levofloxacin. Minimum inhibitory concentrations (MIC) of each antibacterial were determined using the National Committee for Clinical and Laboratory Standards (NCCLS) broth microdilution methodology and lyophilised microtitre plates (Sensititre, Trek Diagnostics) at a central laboratory (GR Micro Ltd., London, UK) [ 26 ]. NCCLS breakpoints [ 25 , 26 ] were used to interpret the MIC data and to determine susceptibility status. The NCCLS breakpoints for telithromycin for S. pneumoniae and for S. aureus are ≤ 1 mg/l is susceptible, 2 mg/l is intermediate, and ≥ 4 mg/l is resistant, and for H. influenzae ≤ 4 mg/l is susceptible, 8 mg/l is intermediate, and ≥ 16 mg/l is resistant [ 27 ]. Results A total of 3043 bacterial pathogens were isolated from patients in 29 countries around the world, with by far the largest number of specimens (1841, 60.5%) coming from Europe (Table 1 ). Percentage of isolates by country were as follows: Argentina 8.0%, Australia 1.1%, Austria 0.6%, Brazil 4.0%, Canada 9.1%, China 1.7%, Colombia 0.1%, Ecuador 0.6%, Eire 0.03%, France 3.4%, Germany 14.3%, Hungary 1.2%, Indonesia 0.03%, Italy 18.9%, Japan 10.0%, Mexico 1.3%, Poland 10.2%, Portugal 2.8%, Russia 0.2%, South Africa 0.4%, South Korea 0.9%, Spain 5.3%, Sweden 0.4%, Switzerland 0.6%, Taiwan 0.7%, Turkey 0.3%, United Kingdom 2.4%, United States 1.4%, Venezuela 0.1%. Of these isolates identified as causative pathogens for bacterial AECB, S. pneumoniae and H. influenzae formed the majority (1075 and 1037 respectively), followed by M. catarrhalis (536) (Table 2 ). Patients were predominantly male (63.5%), with 47.5% of patients belonging to the (30–64) year age group and 52.5% in the >64 year old age group. No difference in the distribution of pathogens by age group was observed (data not shown). Table 2 Distribution of specimen types by species for the 3043 bacterial pathogens described in this study Specimen S. pneumoniae H. influenzae M. catarrhalis S. aureus H. parainfluenzae 1 Total [n (%)] Sputum 832 895 492 219 43 2481 (81.5) BAL 2 144 135 44 66 17 406 (13.4) Blood 99 7 0 50 0 156 (5.1) Total [n (%)] 1075 (35.3) 1037 (34.1) 536 (17.6) 335 (11.0) 60 (2.0) 3043 (100) 1 Only isolated in the first year of the study 2 Bronchoalveolar lavage Table 3 shows the range of MIC values, the MIC 50 and MIC 90 of the various agents against the five species. Where breakpoints were available the percentage of isolates to the various agents is also included. Telithromycin had similar or better in vitro susceptibility than the comparator agents against all of these species. Activity against S. pneumoniae was particularly good, with telithromycin being the most active agent; 99.9% of isolates were classified as susceptible and the MIC 90 (0.12 mg/L) was substantially lower than all other compounds tested. Table 3 In vitro activity of antibacterial agents against 3043 bacterial pathogens isolated from patients with AECB and % susceptibilities to antibacterial agents. Organism Antibacterial MIC mg/l % susceptible range 50 90 total MSSA MRSA S. pneumoniae N = 1075 Telithromycin 0.004–2 0.015 0.12 99.9 Azithromycin 0.03->64 0.12 >64 71.2 Clarithromycin 0.015->32 0.03 >32 71 Erythromycin 0.03->64 0.06 >64 71 Penicillin 0.008–8 0.03 2 75.3 Amox/clavulanate 1 0.015–8 0.03 2 96.1 Cefuroxime 0.015–16 0.03 2 82.2 Trimethoprim-sulphamethoxazole 0.12–32 0.25 8 62 Tetracycline 0.12–32 0.25 32 69.6 Levofloxacin 0.5->32 1 1 98.9 H. influenzae N = 1037 Telithromycin 0.002–16 1 2 99.6 Azithromycin 0.06–32 1 2 99.7 Clarithromycin 0.25->64 8 16 82.4 Erythromycin 0.25->64 4 8 - 2 Ampicillin 0.12–32 0.25 16 87.3 Amox/clavulanate 1 0.12–4 0.5 1 100 Cefuroxime 0.12–16 1 2 99.5 Trimethoprim-sulphamethoxazole 0.03–32 0.06 4 80.7 Tetracycline 0.12–32 0.5 1 97.4 Levofloxacin 0.008–8 0.015 0.015 99.8 M. catarrhalis N = 536 Telithromycin 0.004–0.5 0.06 0.12 Azithromycin 0.06–0.25 0.06 0.06 Clarithromycin 0.25–0.5 0.25 0.25 Erythromycin 0.25–1 0.25 0.25 Ampicillin 0.12–32 8 16 Amox/clavulanate 1 0.12–0.5 0.12 0.25 Cefuroxime 0.12–16 1 2 Trimethoprim-sulphamethoxazole 0.06–4 0.25 0.5 Tetracycline 0.12–32 0.25 0.5 Levofloxacin 0.008–0.06 0.03 0.03 S. aureus N = 335 Telithromycin 0.03->32 0.06 >32 85.1 98.9 24.2 Azithromycin 0.12->64 1 >64 70.4 84.2 9.7 Clarithromycin 0.03->32 0.25 >32 70.4 84.2 9.7 Erythromycin 0.12->64 0.25 >64 70.4 84.6 9.7 Penicillin 0.008–8 4 8 23.6 28.9 0 Amox/clavulanate 1 0.06–8 0.5 8 83.1 100 0 Cefuroxime 0.12–16 1 16 81.2 100 0 Trimethoprim-sulphamethoxazole 0.12–32 0.12 0.25 94.9 97.4 83.9 Tetracycline 0.12–32 0.5 32 84.8 92.7 50 Levofloxacin 0.5–64 0.5 8 76.7 81.7 6.5 H. parainfluenzae N = 60 Telithromycin 0.06–4 1 2 100 Azithromycin 0.06–2 0.5 1 100 Clarithromycin 0.25–16 4 8 93.3 Erythromycin 0.5–8 2 4 - 2 Ampicillin 0.12–32 0.25 1 90 Amox/clavulanate 1 0.12–2 0.5 1 100 Cefuroxime 0.12–4 0.25 0.5 100 Trimethoprim-sulphamethoxazole 0.03–32 0.03 1 88.3 Tetracycline 0.12–16 0.5 4 88.3 Levofloxacin 0.008–8 0.015 0.06 98.4 1 Amox/clavulanate = Amoxycillin/clavulanate 2 No CLSI interpretive criteria for erythromycin and Haemophilus spp. One hundred and three (9.6%) S. pneumoniae isolates (from 51 and 53 patients in the 30–64 and >64 year old age groups respectively)) were resistant to both penicillin (MIC ≥ 2 mg/L) and erythromycin (MIC ≥ 1 mg/L) and this was reflected in resistance to amoxycillin, cefuroxime, clarithromycin and azithromycin also (Table 4 ). These isolates were found in 35 centres in 16 countries. Sixty of these resistant isolates were also resistant to both trimethoprim-sulphamethoxazole and tetracycline. Both telithromycin and levofloxacin had good activity against these isolates, 99% susceptibility to telithromycin and 98.1% to levofloxacin. The MIC 50 and MIC 90 values for telithromycin in this population were 0.06 mg/L and 0.5 mg/L, respectively. Table 4 Antibacterial activity against 103 Streptococcus pneumoniae isolates with combined macrolide and penicillin resistance Antibacterial % susceptible % intermediate % resistant Telithromycin 99.0 1.0 0.0 Azithromycin 0 0 100 Clarithromycin 0 0 100 Erythromycin 0 0 100 Penicillin 0 0 100 Amoxycillin 0 0 100 Amoxycillin-clavulanate 72.8 13.6 13.6 Cefuroxime 1.0 1.0 98.0 Trimethoprim-sulphamethoxazole 19.4 16.5 64.1 Tetracycline 11.7 0.0 88.3 Levofloxacin 98.1 0.0 1.9 Over 99% of H. influenzae isolates were susceptible to amoxycillin-clavulanate, cefuroxime, telithromycin, azithromycin, and levofloxacin. Tetracycline also had good activity with 97.4% of isolates susceptible. Only 11.7% of H. influenzae isolates produced β-lactamase. There were only 60 isolates of H. parainfluenzae and 100% of these were susceptible to four of the eight compounds tested, telithromycin, amoxycillin/clavulanate, cefuroxime and azithromycin. Trimethoprim-sulphamethoxazole and tetracycline were the least active compounds. In terms of MICs, levofloxacin, azithromycin and telithromycin were the most potent compounds against M. catarrhalis with MIC 90 values of 0.03 mg/l, 0.06 mg/l and 0.12 mg/l respectively. There are currently no interpretative NCCLS guidelines available for M. catarrhalis to allow classification into susceptible or resistant categories. The total number of isolates of S. aureu s was 335 and of these only 62 were resistant to methicillin (MRSA). Trimethoprim-sulphamethoxazole was the most active compound overall, with 94.9% of all isolates being susceptible. Telithromycin and tetracycline were the next most active with 85.1% and 84.8% of all isolates susceptible. Telithromycin was the most active compound against the MSSA isolates, with 98.9% being susceptible. The susceptibility of MSSA to tetracycline and trimethoprim-sulphamethoxazole was 92.7% and 97.4% susceptible respectively. These three compounds were the only ones to have activity against the MRSA isolates (trimethoprim-sulphamethoxazole 83.9%, tetracycline 50% and telithromycin 24.2%). Less than 10% of the MRSA isolates were susceptible to the remainder of the compounds. Discussion The primary cause of COPD is exposure to tobacco smoke, the major risk factor being cigarette smoking. The demography of the disease in this study and others reflects this, as the majority of patients in this analysis were male and half were elderly (>64 yrs of age) (2). S. pneumoniae is most frequently isolated in the least severe cases of AECB, whereas H. influenzae is more commonly isolated from moderate to severe cases, with P. aeruginosa occurring in severe hospitalised cases [ 28 ]. Telithromycin does not have good activity against Pseudomonas spp. (GR Micro Limited, data on file, internal report number 141-02-99) and hence may not be an appropriate empirical therapeutic option for AECB patients with severe underlying disease who are hospitalized for an acute exacerbation. Whether the isolation of a pathogen during AECB represents an infection responsible for the exacerbation has been debated for many years [ 29 - 31 ]. Bacteria have been isolated almost as frequently from patients with stable COPD as those with an AECB, and clinical trials of antibiotic therapy in AECB show contradictory and sometimes unconvincing results [ 30 ]. The presence of bacteria in the lower airways is, however, regarded as abnormal since these airways are sterile in healthy adults, and it has been hypothesized that the presence of bacteria in stable COPD represents a low-grade smouldering infection. In addition, a recent study has shown that infection with different strains of pathogens that are new to the patient is associated with development of exacerbation [ 32 , 33 ]. Amoxycillin-clavulanate, azithromycin, and levofloxacin have been shown to be effective in the treatment of AECB, however, there is concern regarding their long-term usefulness, because of the development of resistance to these agents among the causative pathogens [ 34 , 35 ]. Telithromycin has a more focused spectrum of activity than the β-lactams and the fluoroquinolones; it is specifically targeted against pathogens causing community-acquired respiratory disease, including those most commonly associated with AECB. In addition, it is active against penicillin- and macrolide-resistant strains of S. pneumoniae and hence offers a viable potential option for the empiric treatment of AECB in non-hospitalised patients [ 36 ]. The data in this study demonstrate that telithromycin has high in vitro activity against the commonest bacterial pathogens causing AECB. These data also show that telithromycin has the highest overall activity against bacterial isolates from patients with AECB, regardless of species. Almost 10% of S. pneumoniae isolated were resistant to penicillin, macrolides, and at least one of the other antibiotics tested, with only telithromycin and levofloxacin retaining high activity against these isolates (99.0% and 98.1%, respectively). The validity of this finding is strengthened as the isolates were obtained from a large number of patients over a wide geographical distribution. Although atypical pathogens were not examined in the PROTEKT study, telithromycin has been shown to have superior activity in vitro against Chlamydophila pneumoniae to the other macrolides with the exception of clarithromycin and has similar activity to the fluoroquinolones [ 37 ]. In guinea pig models, telithromycin had better activity than erythromycin against Legionella pneumophila infections [ 38 ]. In vitro , the activity of telithromycin against L. pneumophila was similar to levofloxacin but better than erythromycin [ 38 ]. β-lactams and cephalosporins have no activity against Mycoplasma pneumoniae as this species lacks a typical bacterial cell wall, the site of activity for these drugs. Telithromycin has been found to have higher activity than doxycycline and levofloxacin against M. pneumoniae [ 39 ]. As the atypical pathogens can represent up to 10% of infections associated with AECB, the efficacy of telithromycin against these pathogens could be a consideration in the selection of empiric therapy for AECB. Telithromycin has been shown to penetrate into respiratory tissues well [ 40 ]. The concentration of telithromycin in alveolar macrophages and epithelial lining fluid exceeds that of plasma markedly and remains at therapeutic levels for 24 hours after dosing. Bactericidal levels are also maintained in plasma. A good post-antibiotic effect has also been observed [ 41 ]. Telithromycin causes only moderate ecological disturbance to oral and intestinal flora comparable to that associated with clarithromycin and it does not significantly increase the development of resistance in the normal flora, although the MIC of oral streptococci can be slightly raised [ 42 ]. Telithromycin can be administered once a day for AECB. Clinical studies have demonstrated that 800 mg administered once daily for 5 days was as effective and well tolerated as a 10-day course of amoxycillin/clavulanate (500/125 mg 3 times daily for 10 days), cefuroxime axetil (500 mg twice daily for 10 days) or clarithromycin (500 mg twice daily for 10 days) [ 43 ]. Other clinical studies have also confirmed the safety and tolerability of telithromycin 800 mg administered for 5 – 10 days [ 44 ]. Once a day dosing schedules and shorter courses may promote patient adherence to therapy, and this in turn could delay the development of resistance. Although this study provides valuable information on the overall antimicrobial profile of bacteria causing AECB, care should be taken when interpreting data related to specific demographics. The prevalence of species could not be calculated in this study as a major limitation, inherent to most surveillance studies, is the requirement for collecting centres to fulfil a specified quota of isolates over a defined time period (1 year). If, for instance, a centre managed to fulfil the quota for S. pneumoniae isolates from patients with community-acquired pneumonia, it could then only send H. influenzae from patients with AECB to fulfil the quota for this organism. In addition, atypical pathogens were not sampled and they can represent up to 10% of the causative pathogens [ 28 ]. In summary, the data presented here demonstrate that telithromycin has good in vitro activity against H. influenzae , S. pneumoniae , and M. catarrhalis , respiratory pathogens commonly isolated in AECB. It is as active as or more active than agents that are currently used in this clinical setting. Additionally, although not shown here, telithromycin has better in vitro activity against atypical pathogens than other agents; an important advantage in this clinical setting as these pathogens may represent 10% of AECB associated infections. The development of resistance will always be a threat to the usefulness of antibacterial compounds, however surveillance studies such as PROTEKT allow the rapid detection and characterization of resistance mechanisms and highlight the need for and examine the in vitro efficacy of newer antibacterial agents. Providing careful surveillance for the development of resistance is maintained, telithromycin currently offers a useful agent in the treatment of AECB.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555545.xml

548266

Risk adjustment methods for Home Care Quality Indicators (HCQIs) based on the minimum data set for home care

Background There has been increasing interest in enhancing accountability in health care. As such, several methods have been developed to compare the quality of home care services. These comparisons can be problematic if client populations vary across providers and no adjustment is made to account for these differences. The current paper explores the effects of risk adjustment for a set of home care quality indicators (HCQIs) based on the Minimum Data Set for Home Care (MDS-HC). Methods A total of 22 home care providers in Ontario and the Winnipeg Regional Health Authority (WRHA) in Manitoba, Canada, gathered data on their clients using the MDS-HC. These assessment data were used to generate HCQIs for each agency and for the two regions. Three types of risk adjustment methods were contrasted: a) client covariates only; b) client covariates plus an "Agency Intake Profile" (AIP) to adjust for ascertainment and selection bias by the agency; and c) client covariates plus the intake Case Mix Index (CMI). Results The mean age and gender distribution in the two populations was very similar. Across the 19 risk-adjusted HCQIs, Ontario CCACs had a significantly higher AIP adjustment value for eight HCQIs, indicating a greater propensity to trigger on these quality issues on admission. On average, Ontario had unadjusted rates that were 0.3% higher than the WRHA. Following risk adjustment with the AIP covariate, Ontario rates were, on average, 1.5% lower than the WRHA. In the WRHA, individual agencies were likely to experience a decline in their standing, whereby they were more likely to be ranked among the worst performers following risk adjustment. The opposite was true for sites in Ontario. Conclusions Risk adjustment is essential when comparing quality of care across providers when home care agencies provide services to populations with different characteristics. While such adjustment had a relatively small effect for the two regions, it did substantially affect the ranking of many individual home care providers.

Background In both Canada and the United States efforts are underway to develop systems to assess the quality of health care as a first step to improving services. In the US nursing home sector, the implementation of the Minimum Data Set has been linked to improvements in quality of care [ 1 - 4 ]. Recently, the Centers for Medicare and Medicaid Services (CMS) has developed web pages to give consumers, and the public at large, further information about the quality of nursing homes and home care services across the country. The "Nursing Home Compare" and more recent "Home Health Compare" web sites allow individuals to view several quality indicators (QIs) for individual providers[ 5 , 6 ]. When comparing health care providers across a set of QIs there is a concern that they may differentially admit clients with a greater likelihood of triggering on quality issues. Since the indicators are defined as events that are preferable to avoid (e.g., skin ulcers, untreated pain, weight loss), in the absence of risk adjustment, these providers would appear to be delivering poorer quality of care. Risk adjustment attempts to adjust for different populations of clients who may be at greater risk for experiencing quality issues that is a function of their clinical status rather than the quality of care. In the US nursing home sector, the CMS recently funded a large-scale project to review all potential long-term care QIs and the risk adjustment process. The evaluation of risk adjustment included a review of both client-level and agency-level covariates. The research team revised the original set of client covariates to create a set of new models and assessed the effects of using an agency-level covariate, the Facility Admission Profile (FAP). The FAP represents the prevalence of the quality issue for residents newly admitted to the facility and was intended to adjust for potential selection bias (i.e., facilities preferentially admitting clients with a greater likelihood of triggering on the indicator) and ascertainment bias (i.e., ability of providers to detect a quality issue that might increase the QI rate). However, after extensive analyses, the research team did not recommend the use of the FAP, given that the FAP did not appear to be a particularly useful measure of ascertainment bias [ 7 ]. Nevertheless, in October 2002, the CMS included the FAP in three quality measures reported on the Nursing Home Compare web page[ 8 ]. As part of this same initiative, Kidder et al.[ 9 ] examined the Nursing Case Mix Index from the RUG-III grouping system as well as seven of the RUG scales as potential risk adjusters. In their final list of quality measures, 12 indicators for chronic care residents and four indicators for post-acute care residents, included a covariate related to the RUG-III system. Home Care Quality Indicators (HCQIs) The current research was part of a project to develop a set of 22 Home Care Quality Indicators (HCQIs) based on items in the Minimum Data Set for Home Care (MDS-HC). Implementation of the MDS-HC is complete or underway in 15 states and in 7 Canadian provinces and territories, where it is being used mainly as a clinical assessment instrument. In Michigan, the MDS-HC is being used as part of the MI-Choice waiver program to reduce nursing home admissions. In Ontario, Community Care Access Centre (CCAC) case managers use the instrument to determine needs, allocate services and make placement decisions. The HCQI derivation was accomplished using data from Ontario and Michigan since these were the regions with the largest scale implementation in Canada and the US, respectively. In their paper describing the HCQI derivation, Hirdes et al.[ 10 ] recommended the use of client-level risk adjustment for all but four indicators, and suggested the use of the Agency Intake Profile (AIP) as another method for risk adjustment. These various approaches to risk adjustment are the focus of this study. The set of HCQIs used in this study were developed by members of inter RAI , a non-profit multinational organization dedicated to the development and refinement of assessment instruments for older adults and persons with disabilities, and their related applications. The HCQIs represent outcome measures that document an agency's rate for triggering on a quality issue. For example, one indicator measures the prevalence of weight loss among individuals who are not considered to be palliative clients. The HCQIs include a mix of prevalence measures (i.e., measured on a cross-section of clients at one point in time) and incidence measures (i.e., failure to improve or incidence of an event measured across two points in time). All HCQIs are defined as events to be avoided such that a higher rate on the indicator is indicative of poorer performance. This paper uses a dataset from two Canadian provinces to further explore these three types of risk adjustment for the inter RAI HCQIs. It explores the effects of risk adjustment at both the level of the region and at the level of the individual home care provider. Methods Data The RAI Health Informatics Project was a two and a half year research study begun in 1999 in which fourteen of Ontario's 43 CCACs used the MDS-HC as part of usual practice for all adult (18 years and older) home care clients. CCACs provide several different services: information and referral, case management and placement in long-term care facilities. CCACs purchase in-home services from external contracted agencies through a "request for proposal" process. Services provided include in-home physiotherapy, occupational therapy, personal support and homemaking, medical supplies and equipment, and care by other professional such as social workers, dietitians and speech-language pathologists. Case managers oversee the assessment process, make referrals to service providers and then monitor the care provided in the home. Funding for CCACs is based on an annual budget provided by the Ontario Ministry of Health and Long-term Care. CCACs are governed by independent, non-profit boards of directors, accountable to the Ministry of Health and Long-term Care. In Ontario, it is now mandatory for all long stay home care clients to be assessed with the MDS-HC. However, at the time of the RAI Health Informatics project, the MDS-HC was not mandated for use. As such, some clients (e.g., those expected to be on service less than two weeks) did not receive an MDS-HC assessment and were not included in the current project. At this same time, Manitoba conducted a pilot implementation of the MDS-HC in the fourteen offices of the Winnipeg Regional Health Authority (WRHA). The WRHA is one of twelve health authorities in the province, and is responsible for providing health services to the approximately 646,000 residents of Winnipeg and the surrounding suburban area. The WRHA receives annual funding from the government of Manitoba. The home care program of the WRHA was established in 1974 with a mandate to provide effective and responsive health care services in the community to support independent living and facilitate admission into LTC facilities when independent community living is no longer an option. Home care services include personal care, nursing, counselling, occupational therapy and physiotherapy assessment, referral to other agencies and coordination of services. In the WRHA region, an MDS-HC assessment was completed by care coordinators only if the client was anticipated to be on service for at least 90 days. As a result, the study sample was focused on a long stay population of home care clients. All participating sites in the WRHA and in Ontario identified a set of case managers who received a two-day training session led by a member of the research team. They then used the MDS-HC instrument as part of their usual in-home assessment. The data used for this study represent a cross-sectional cohort of home care clients assessed between November, 1999 and December, 2002. Two CCACs in Ontario and four WRHA sites, were removed from the database because they submitted fewer than 20 assessments, resulting in a total of ten WHRA and twelve Ontario sites. Of these, three Ontario CCACs and eight offices of the WRHA also submitted client reassessments at approximately 90 days, which allowed for the calculation of failure to improve/incidence HCQIs. The protocol for data collection was reviewed and received ethics clearance through the Office of Research Ethics at the University of Waterloo, Canada. Risk adjustment Risk adjustment attempts to adjust for differences in client populations that may bias the HCQI rates. Organizations that provide care to more impaired clients will tend to have higher unadjusted rates, regardless of the quality of care they provide. As such, risk adjustment methods are used to maximize the ability to make fair comparisons across providers. With any type of risk adjustment, caution must be exercised to prevent over adjustment (i.e., adjusting away poor practice). The choice of risk adjustment covariates is therefore highly important and should not include variables that would be considered to reflect suboptimal clinical care. There are two generic types of risk adjustment that have been recommended for the HCQIs. The first adjusts for differences in the population at the client-level[ 10 ]. Potential adjusters can include both individual assessment items and summary scales embedded within the MDS-HC. The HCQI developers evaluated a large range of potential covariates, considering their distributional properties, strengths of association with the outcomes of interest, consistency of findings across jurisdictions and potential for clinically inappropriate adjustment (e.g., benzodiazepine use was not considered a reasonable adjuster for falls). As a result, from zero to five risk adjusters were recommended for the 22 HCQIs and these client covariates were used in the current project[ 11 ]. The second type of risk adjustment was intended to control for two types of bias at the agency level: a) the agency's ability to identify differences in clients' clinical characteristics and b) differences in who an agency selects for admission. These risk adjustments are performed at the agency level after the individual-level risk adjustments are applied. The Agency Intake Profile (AIP) was used following the methodology outlined by Morris et al., for the MDS 2.0 quality indicators[ 7 ]. The AIP was calculated for each agency based on clients for whom the MDS-HC assessment was their intake assessment or for clients who had been on service for no more than 30 days. This group of clients was considered to be the intake cohort. More recently, an alternative form of agency-level risk adjustment has been proposed, to control for potential selection and ascertainment bias. This adjustment employs the Case Mix Index (CMI) associated with the RUG-III/HC methodology[ 9 ]. In this instance, the CMI for the intake cohort was calculated, and adjustments to the HCQIs were performed similar to those used for the AIP. The process followed for adjusting HCQI rates was the same as that used by Morris et al. in adjusting the US nursing home QI rates[ 7 ]. Three sources of information are required: the agency-level observed rate, the agency-level expected rate and the grand mean across all agencies. The first data element was simply the raw observed HCQI rate for a given home care agency. The next data element involved the creation of an expected rate for each client within a given agency, based on output from a logistic regression model. In this model, each HCQI acts as a dichotomous dependent variable (i.e., triggering on the HCQI or not) and the client-level and/or agency-level covariates are entered simultaneously as independent variables. The expected value for each client is then pooled to create an expected value for the agency. Three separate risk adjusted HCQI rates are thus computed, based on which method is used for adjustment: the expected rate based on the client covariates only (CC), on client covariates plus AIP (+AIP) and on client covariates plus CMI (+CMI). The final adjusted value can be thought of as an estimate of an agency's HCQI rate if the agency had clients with an average level of risk[ 7 ]. This risk adjustment method is similar to the concept of indirect standardization, in which the ratio of the observed to expected events is calculated then multiplied by the crude rate in the standard population[ 12 ]. In the current project, the standard population used was the combined set of agencies from both the Winnipeg Regional Health Authority (WRHA) and from Ontario. HCQI rates For each participating agency, the unadjusted HCQI rates were calculated for all 22 indicators. These rates represent the average rate across all eligible clients for a given agency. For prevalence indicators, the rates were calculated only for clients who had been on service for at least 30 days to avoid penalizing an organization for quality issues that were newly recognized. Several methods were utilized to assess the impact of risk adjustment, both at the regional and at the agency-level. The unadjusted and three adjusted rates were compared between Ontario and the WRHA, to assess the effects at the regional level. At the level of the individual agency, ranks were calculated for each HCQI and a count was created to examine how often a given agency was among those with the four highest rates. Since higher rates on each HCQI were indicative of higher prevalence or incidence of undesirable outcomes, agencies ranked within the four highest rates were considered to be among the "worst performers." The range between agencies with the highest and fourth highest rates was also examined for both unadjusted and adjusted rates to assess the degree of variation among the group of worst performers. Results The two regions were very similar on average age and sex. The WRHA clients were significantly less likely to have some level of cognitive impairment, as measured by the Cognitive Performance Scale (CPS)[ 13 ], compared to Ontario clients, although the actual difference was only 2.6% (WRHA: 37.0% vs. Ontario: 39.6%; p < 0.0001). They were also significantly less likely than clients in Ontario to require some assistance with ADLs (22.0% vs. 27.1%, respectively; p < 0.0001), as measured by the ADL Self-performance Hierarchy Scale[ 14 ]. Severe daily pain was experienced by 17.2% of the Ontario clients compared to 14.2% of the WRHA clients (p < 0.0001). Although the differences were statistically significant, with Ontario having significantly lower rates of both arthritis and hypertension, the absolute difference between the regions on the most common diagnoses was less than 5% (Table 1 ). Table 1 Characteristics of home care clients in the WRHA and Ontario* WRHA (n = 6704) ON (n = 5063) p value % % Age Mean (95% CI**) 76.0 (74.9, 77.0) 75.6 (75.2, 76.0) 0.49 18–64 13.4 16.5 <0.0001 65–74 16.5 19.2 75–84 39.9 40.7 85 and older 30.0 23.6 Sex Female 69.2 70.5 0.14 Marital status Never married 11.5 8.7 <0.0001 Married/widowed 79.4 82.7 Separated/divorced 8.2 7.9 Other 0.8 0.7 Primary language English 85.6 85.1 0.63 French 3.5 3.8 Other 10.9 11.1 Aboriginal status Origin is Inuit, Metis or North American Indian 3.2 1.6 <0.0001 Cognitive Performance Scale Mean (95% CI) 0.8 (0.8, 0.8) 0.8 (0.8, 0.9) 0.07 0 – intact 63.0 60.4 <0.0001 1 – borderline intact 15.0 18.4 1 – borderline intact 15.0 18.4 2 – mild impairment 8.4 7.3 3 – moderate impairment 10.9 10.1 4 – moderate to severe impairment 0.8 1.0 5 – severe impairment 1.5 2.5 6 – very severe impairment 0.5 0.4 ADL Self-Performance Hierarchy Scale Mean (95% CI) 0.5 (0.5, 0.6) 0.7 (0.6, 0.7) <0.0001 0 – independent 78.0 72.9 <0.0001 1 – supervision required 5.2 7.9 2 – limited impairment 8.9 8.5 3 – extensive assistance required (level I) 4.7 5.4 4 – extensive assistance required (level II) 1.6 2.6 5 – dependent 1.1 2.0 6 – total dependence 0.5 0.7 Pain Scale Mean (95% CI) 1.2 (1.2, 1.2) 1.3 (1.3, 1.4) <0.0001 0 – no pain 39.8 34.9 <0.0001 1 – less than daily pain 14.4 14.1 2 – daily pain but not severe 31.6 33.8 3 – severe daily pain 14.2 17.2 Top 3 medical diagnoses*** Arthritis 49.4 44.8 <0.0001 Hypertension 42.2 37.4 <0.0001 Diabetes 19.1 20.1 0.24 * client characteristics and scale values based on first submitted MDS-HC assessment ** CI = confidence interval *** disease was present and was or was not being treated or monitored by a home care professional Unadjusted rates When comparing the two regions, there were statistically significant differences for five of the 22 unadjusted HCQIs (Table 2 ). In only one of these five cases (ADL rehabilitation potential and no therapies) was the rate in the WRHA significantly higher than the rate in Ontario. However, the actual size of the absolute difference between regions was small (0.3% on average). Among the prevalence HCQIs, the largest unadjusted difference was for disruptive/intense daily pain, which was 9.1% higher in Ontario. There were no statistically significant differences between the regions for any of the incidence HCQIs. Table 2 Unadjusted HCQI rates comparing the WRHA and Ontario* WRHA Offices Ontario CCACs p value** n = 10 n = 12 Prevalence HCQIs Mean (sd) 95% CI Mean (sd) 95% CI Inadequate meals 2.7 (1.2) 1.9, 3.6 4.4 (1.9) 2.1, 7.1 0.02 Weight loss 5.7 (3.4) 3.2, 8.1 7.6 (2.2) 5.1, 10.3 0.13 Dehydration 0.6(0.7) 0.08, 1.1 2.4 (1.6) 1.0, 5.0 0.003 Not receiving a medication review by a physician 3.8 (1.8) 2.5, 5.0 10.3 (4.3) 6.6, 17.1 0.0002 No assistive device among clients with difficulty in locomotion 13.8 (16.4) 2.1, 25.5 10.2 (5.2) 4.2, 15.6 0.52 ADL/rehabilitation potential and no therapies 84.0 (7.4) 78.7, 89.3 73.7 (5.5) 69.6, 78.5 0.001 Falls 20.8 (7.8) 15.2, 26.3 24.4 (4.5) 19.7, 30.3 0.18 Social isolation 21.6 (4.5) 18.4, 24.8 24.3 (5.9) 17.6, 32.4 0.24 Delirium 5.1 (2.3) 3.4, 6.7 5.9 (2.6) 3.4, 9.0 0.45 Negative mood 7.9 (3.2) 5.6, 10.1 11.6 (6.2) 5.2, 18.4 0.10 Disruptive or intense daily pain 30.5 (4.0) 27.6, 33.3 39.6 (6.3) 33.6, 44.3 0.0008 Inadequate pain control among those with pain 26.6 (5.3) 22.8, 30.4 29.7 (7.4) 22.3, 35.8 0.27 Neglect/abuse 3.0 (2.6) 1.2, 4.9 2.5 (2.1) 0.8, 5.2 0.58 Injuries 13.6 (6.4) 9.0, 18.2 11.0 (3.7) 8.6, 14.3 0.26 No influenza vaccine 28.4 (7.1) 23.2, 33.5 30.3 (9.8) 21.1, 44.0 0.61 Hospitalization 32.0 (7.2) 26.8, 37.2 36.4 (4.9) 31.1, 40.4 0.11 Incidence HCQIs n = 8 n = 3 Failure to improve/incidence of bladder incontinence 29.5 (9.7) 21.4, 37.5 23.1 (7.4) 4.8, 41.4 0.33 Failure to improve/incidence of skin ulcers 3.4 (4.54) 0.0, 7.1 3.4 (3.2) -5.0, 11.3 0.96 Failure to improve/incidence of decline on ADL long form 41.5 (12.5) 31.1, 51.9 30.3 (4.4) 19.3, 41.3 0.17 Failure to improve/incidence of impaired locomotion in the home 11.0 (9.0) 3.5, 18.5 16.5 (9.2) -6.5, 38.4 0.44 Failure to improve/incidence of cognitive decline 44.6 (13.7) 33.1, 56.0 36.1 (3.9) 26.3, 45.9 0.33 Failure to improve/incidence of difficulty in communication 13.7 (8.1) 6.9, 20.4 14.4 (3.4) 6.0, 22.9 0.88 * in the event that the lower value of the 95% confidence interval was less than zero, the value was set to zero. ** based on a t-test of independent means. Agency-level covariates The AIP was calculated for each home care agency for each of the 19 HCQIs for which client-level risk adjustment was recommended. The AIP values shown represent the HCQI rate among an admission cohort within each region and were used in the risk adjustment models that included client-level covariates together with the AIP covariate. Ontario CCACs had a significantly higher AIP value for eight indicators, demonstrating that they admit or at least assess individuals as more likely to have these conditions on admission (Table 3 ). Table 3 AIP values for the WRHA and Ontario for each of the 19 risk-adjusted HCQIs* WRHA n = 10 Ontario n = 12 p value Prevalence HCQIs mean (95% CI) Inadequate meals 2.9 (1.7, 4.1) 6.8 (5.3, 8.2) 0.0003 Weight loss 8.2 (4.7, 11.7) 12.8 (10.7, 14.8) 0.02 Dehydration 0.6 (0.2, 1.1) 4.0 (2.5, 5.5) 0.0005 No assistive device among clients with difficulty in locomotion 13.4 (10.8, 16.0) 19.1 (12.7, 25.4) 0.09 Falls 24.4 (20.6, 28.2) 31.4 (27.5, 35.3) 0.01 Social isolation 22.4 (19.4, 25.3) 25.3 (21.2, 29.4) 0.23 Delirium 5.2 (4.0, 6.4) 10.2 (7.8, 12.6) 0.0008 Negative mood 7.7 (5.5, 9.9) 12.8 (8.7, 17.0) 0.03 Disruptive or intense daily pain 31.5 (28.1, 34.9) 42.3 (37.5, 47.0) 0.0008 Inadequate pain control among those with pain 30.6 (26.6, 34.6) 31.4 (26.5, 36.4) 0.79 Neglect/abuse 2.5 (1.5, 3.5) 2.9 (1.3, 4.6) 0.65 Injuries 11.5 (8.5, 14.5) 12.8 (9.1, 16.5) 0.55 Hospitalization 35.3 (31.6, 39.1) 52.0 (47.4, 56.6) <0.0001 Incidence HCQIs n = 8 n = 3 Failure to improve/incidence of bladder incontinence 0.9 (0.8, 1.0) 0.9 (0.5, 1.2) 0.81 Failure to improve/incidence of skin ulcers 6.3 (5.6, 7.0) 7.6 (4.8, 10.3) 0.07 Failure to improve/incidence of decline on ADL long form 1.9 (1.6, 2.3) 2.8 (0.0, 6.9) 0.45 Failure to improve/incidence of impaired locomotion in the home 0.3 (0.2, 0.4) 0.6 (0.0, 1.6) 0.36 Failure to improve/incidence of cognitive decline 0.8 (0.7, 0.9) 1.0 (0.2, 1.7) 0.21 Failure to improve/incidence of difficulty in communication 0.5 (0.4, 0.6) 0.6 (0.0, 1.6) 0.65 * for prevalence HCQIs and failure to improve/incidence of skin ulcers, the AIP value represents the mean rate (expressed as a percent) of the HCQI on an intake cohort of clients and for the remaining incidence HCQIs, the AIP value represents the mean value for the MDS item(s) involved in calculating the HCQI. Among new home care clients in Ontario, 52% of clients triggered on the HCQI for the prevalence of hospitalization versus 35% in the WRHA (difference of 17%). Ontario also had a significantly (p < 0.0001) higher Case Mix Index (CMI) on intake (i.e., the CMI covariate), at 0.84 (95% CI: 0.81, 0.86), compared with the WRHA at 0.70 (CI: 0.68, 0.71). Risk adjustment at the regional level In general, the risk adjustment process minimized the differences between the two regions compared with the unadjusted rates. For example, the unadjusted difference between regions for the prevalence of disruptive/intense daily pain was 9.1%, however, the difference was reduced to 8.0% with the CC adjustment, 5.7% with the +CMI adjustment and 2.8% with the +AIP adjustment (Table 4 ). The +AIP adjusted rates showed the most variability, so that for five HCQIs, the direction of the difference was reversed. For example, the unadjusted rate of falls was 3.6% higher in Ontario than in the WRHA. Following the +AIP adjustment, Ontario had a rate that was 1.4% lower than in the WRHA. Table 4 Difference between Ontario and WRHA HCQI unadjusted and adjusted rates* Unadjusted Client covariates only AIP + client covariates CMI+ client Covariates Difference (ON-WRHA) expressed as a % Prevalence HCQIs Inadequate meals 1.7 1.7 0.2 1.5 Weight loss 1.9 1.9 0.2 1.5 Dehydration 1.8 1.9 1.1 1.7 No assistive device among clients with difficulty in locomotion -3.6 -1.9 -2.8 -3.3 Falls 3.6 1.7 -1.4 0.7 Social isolation 2.7 0.8 1.1 0.6 Delirium 0.8 1.1 -0.4 0.8 Negative mood 3.7 3.1 -1.0 1.8 Disruptive or intense daily pain 9.1 8.0 2.8 5.7 Inadequate pain control among those with pain 3.1 2.8 1.7 3.3 Neglect/abuse -0.5 -0.4 -0.5 -1.0 Injuries -2.6 -3.0 -3.3 -1.8 Hospitalization 4.3 2.8 -1.4 2.1 Incidence HCQIs Failure to improve/incidence of bladder incontinence -6.4 -5.7 -5.6 -2.3 Failure to improve/incidence of skin ulcers 0.0 0.0 -1.2 -0.8 Failure to improve/incidence of decline on ADL long form -11.2 -11.0 -10.7 -6.6 Failure to improve/incidence of impaired locomotion in the home 5.0 5.3 -0.1 4.2 Failure to improve/incidence of cognitive decline -8.5 -6.1 -7.7 -4.1 Failure to improve/incidence of difficulty in communication 0.7 1.3 0.0 1.7 Mean 0.3 0.2 -1.5 0.3 * differences were always calculated as Ontario rate minus WRHA rate Agencies with the highest rates A summary measure was created to count the frequency with which an agency was ranked among the worst performers. Overall, only three of the ten offices within the WRHA did not change their ranking when rates were adjusted (Table 5 ). For example, Office 6 was ranked within the four highest (i.e., worst) agencies for three out of the 19 HCQIs both for unadjusted and for the three adjusted rates. In another six offices, the effect of risk adjustment was a decline in their standing (i.e., poorer quality) so that after at least one type of risk adjustment they were ranked among the worst performers. For example, in Offices 2, 3 and 7, the number of times they were ranked within the four highest rates increased as a result of the +AIP adjustment. The most dramatic effect was evident in Office 7 such that the unadjusted rates ranked this group among the worst performers only once, but the +AIP adjustment increased this frequency to five. There was no instance of an office in the WRHA consistently benefiting from risk adjustment. Table 5 Number of times a given agency was ranked within the four highest rates Number of times agency ranked within the four highest rates Unadjusted Client covariates AIP + client covariates CMI + client covariates WRHA 1 8 9 10 9 2 1 1 4 3 3 2 2 5 3 4* 0 0 0 0 5 3 5 4 5 6 3 3 3 3 7 1 2 5 2 8 0 0 2 0 9 7 6 8 6 10* 4 4 4 4 Ontario 100* 3 3 2 4 200 5 5 4 4 300* 1 1 2 2 400* 5 5 6 4 500* 1 2 0 2 600* 4 3 3 1 700* 1 1 2 1 800* 5 4 2 4 900* 6 6 6 6 1000* 1 1 0 1 1100 10 9 3 8 1200 6 4 1 4 * indicates agencies for which the count was based on 13 HCQIs since incidence HCQIs could not be calculated In Ontario, only CCAC 900 experienced no change in their ranking following risk adjustment (Table 5 ). In another nine cases, at least one type of adjustment resulted in an improved standing, with the agency ranking among the four highest rates less often. For example, CCAC 1100 ranked among the worst performers ten times across the 19 HCQIs prior to risk adjustment, but the +AIP adjustment reduced this value to three. A similar effect was observed for CCAC 1200 which was ranked in the worst four performers six times across the unadjusted rates, but only once following +AIP adjustment. Examining only the ranking of agencies may be misleading if very small changes in the actual rate resulted in an increase or decrease in the rank for a given agency. For example, the change in the rate for a given agency could be very small, say 5%, but could result in the organization moving from the fifth highest rate (i.e., fifth worst performer) to the third highest rate. Examining only the ranking tells one little about the actual magnitude of change among the worst performers. Therefore, it is also important to explore the range in the rates across the four highest agencies. The unadjusted and CC adjusted rates had the largest degree of variation between the highest and fourth highest agency, with a mean of 11% across the 19 HCQIs. The +CMI adjustment had the next largest amount of variation at 10% and the +AIP adjustment at 8% (Table 6 ). These results further reinforce the finding that the +AIP method of adjustment consistently had the largest impact compared with the other types of risk adjustment. Table 6 Range in the HCQI rates comparing agencies with the highest rates Unadjusted Client covariates AIP + client covariates CMI + client covariates Range between 1 st and 4 th highest agency Difference between 1 st and 4 th agency Range between 1 st and 4 th highest agency Difference between 1 st and 4 th agency Range between 1 st and 4 th highest agency Difference between 1 st and 4 th agency Range between 1 st and 4 th highest agency Difference between 1 st and 4 th agency % % % % Prevalence HCQIs Inadequate meals 7.4 – 4.9 2.5 8.3 – 5.4 2.9 6.4 – 4.6 1.8 8.2 – 5.4 2.9 Weight loss 13.3 – 9.3 4.0 13.3 – 9.2 4.1 10.7 – 9.2 1.5 13.6 – 9.8 3.8 Dehydration 6.1 – 2.6 3.5 6.6 – 3.0 3.6 3.8 – 2.3 1.5 6.5 – 2.8 3.7 No assistive device among clients with difficulty in locomotion 60.0 – 15.3 44.7 50.4 – 14.8 35.6 53.8 – 15.6 38.2 52.0 – 14.3 37.7 Falls 38.4 – 28.8 9.6 38.4 – 28.8 9.6 36.5 – 28.4 8.1 39.1 – 29.1 10.0 Social isolation 33.1 – 28.6 4.5 31.8 – 30.2 1.6 31.5 – 30.3 1.2 31.9 – 30.2 1.7 Delirium 11.0 – 8.1 2.9 14.3 – 12.4 1.9 12.8 – 11.9 0.9 14.2 – 12.9 1.3 Negative mood 26.2 – 13.4 12.8 30.4 – 16.1 14.3 18.0 – 15.0 3.0 23.4 – 15.9 7.5 Disruptive or intense daily pain 54.1 – 42.5 11.6 51.6 – 40.8 10.8 43.7 – 39.4 4.3 44.5 – 41.2 3.3 Inadequate pain control among those with pain 46.5 – 32.6 13.9 47.6 – 33.8 13.8 36.6 – 33.0 3.6 47.4 – 34.2 13.2 Neglect/abuse 9.7 – 3.9 5.8 9.5 – 4.0 5.5 6.7 – 3.7 3.0 9.8 – 4.4 5.4 Injuries 22.4 – 18.9 3.5 22.8 – 18.3 4.5 21.9 – 16.1 5.8 22.0 – 19.0 3.0 Hospitalization 46.8 – 40.1 6.7 44.6 – 40.8 3.8 47.7 – 38.4 9.3 46.2 – 41.2 5.0 Incidence HCQIs Failure to improve/incidence of bladder incontinence 48.0 – 29.0 19.0 48.4 – 28.4 20.0 44.5 – 31.9 12.6 47.3 – 27.9 19.4 Failure to improve/incidence of skin ulcers 12.9 – 4.2 8.7 13.2 – 4.3 8.9 11.1 – 5.5 5.6 14.1 – 4.9 9.2 Failure to improve/incidence of decline on ADL long form 56.0 – 42.9 13.1 55.1 – 43.4 11.7 54.9 – 43.3 11.6 53.9 – 41.2 12.7 Failure to improve/incidence of impaired locomotion in the home 29.0 – 14.1 14.9 33.4 – 19.1 14.3 36.6 – 22.2 14.4 34.1 – 18.4 15.7 Failure to improve/incidence of cognitive decline 72.0 – 48.4 23.6 70.1 – 46.7 23.4 69.5 – 46.3 23.2 69.6 – 46.0 23.6 Failure to improve/incidence of difficulty in communication 23.8 – 18.3 5.5 30.7 – 22.3 8.4 29.3 – 22.4 6.9 30.6 – 22.4 8.2 Mean difference 11.1 10.5 8.2 9.9 Discussion Ontario CCACs exhibited several key differences when compared to the WRHA. Ontario sites had significantly higher unadjusted rates across four HCQIs. However, the degree of variation between the sites was small, with a mean difference across indicators of less than 1%. Ontario as a region had significantly higher AIP values for eight HCQIs, indicating a greater propensity in Ontario to admit clients exhibiting quality issues, and they also had a higher mean Case Mix Index. Overall, the change in the actual rates at the regional level was small following risk adjustment. When risk adjustment was applied, for the client covariates alone or the client covariates with the CMI covariate, there was little effect on the mean difference between the two regions. In each case, the mean difference was positive (i.e., Ontario had a higher rate on average) and less than 1%. The +AIP adjustment, however, resulted in a negative mean difference (i.e., WRHA higher than Ontario on average) across the set of risk-adjusted HCQIs. At the agency level, there was a greater influence of risk adjustment, as assessed by agency rankings across the set of indicators. In general, sites in Ontario benefited from risk adjustment and were less likely to be ranked among the worst performers. The opposite was true within the WRHA. Again, the +AIP adjustment had the largest influence when compared to the other two types of risk adjustment. It is possible that the level of variation between agencies on the AIP covariate was greater than the level of variation between clients on the individual-level covariates. Although a detailed analysis is beyond the scope of this paper, use of a single HCQI example may prove beneficial. When examining the prevalence of falls (an HCQI that showed the largest changes in the agency rankings following risk adjustment), the coefficient of variation (CV) for the AIP value was 23.0. The CV is a relative measure of dispersion about the mean and is calculated by dividing the standard deviation by the mean and multiplying by 100. This CV value was much lower than any of the corresponding CV values for the various MDS outcome scales that reflect differences at the level of the individual client. For example, the CV for the Cognitive Performance Scale was 158.5 and for the ADL Self-performance Hierarchy Scale, the CV was 195.6 (data not shown). Clearly, the +AIP adjustment had the effect of minimizing differences between the individual providers and it also resulted in rates that had less variability, as demonstrated by the drop in the CV value when comparing the unadjusted and +AIP adjusted rates for the prevalence of falls indicator (CV of 27.7 and 19.1, respectively; data not shown). However, this effect cannot be explained by differences in the variances of the adjusters since the AIP had a lower level of variance than the individual covariates. It is also possible that the AIP covariate serves as a proxy for regional differences in practice patterns. If the Ontario agencies are grouped into four main geographic regions, the AIP value for the falls HCQI ranged from 24.1% to 34.0%. The degree of variation appears modest, but cannot be discounted as one factor in the ability of the AIP adjustment to minimize differences between providers and between regions. Although there continues to be support for the conceptual notion of risk adjustment, the potential to over adjust remains a concern. In a recent publication, Mor et al[ 15 ] recognize the possibility for over adjustment and conclude that in the absence of a simple solution to this problem, researchers must carefully consider each performance measure on an individual basis in an attempt to minimize this issue. In the current project, the AIP covariate represents a continuous, numeric value that can range from zero to one. Thus, it does not simply serve as an agency identifier, but represents the rate among an intake cohort. Furthermore, in the US, Morris et al. determined that the Facility Admission Profile, the conceptual cousin to the AIP, did not substantially improve the risk adjustment models and they did not recommend its use. This decision was not based on a fear of over adjustment, but rather a concern that the FAP added increased complexity without significantly improving the risk adjustment process[ 7 ]. It is also important that to develop a clearer understanding of what is meant by "over adjustment". For example, it would seem reasonable to differentiate between instances of truly inappropriate risk adjustment where variance is due to poor practices (e.g., adjusted for benzodiazepine use for a falls indicator) and "over adjustment" due to the excessive use of spurious risk adjusters resulting in suppression of variance. Another possible example of over adjustment is the use of individual level covariates that are too closely related to the quality indicator. For example, one might argue that using dressing of the upper body as a risk adjuster for a QI on dressing the lower body is a form of over adjustment because the dependent variable is represented on both sides of the regression equation. This area of research continues to present many challenges. Given the current results and those from the long-term care sector, it appears advantageous to undertake some form of risk adjustment, even though the definitive method may not yet be best understood. Given that these HCQIs are new and so little is known about the risk adjustment process, it seems appropriate that all three type of risk adjustment be considered by researchers and policy makers. The +AIP adjustment represents the most conservative approach and may therefore be most appropriate for public reporting of HCQI results. The assessment of quality of care, and ongoing refinement of risk adjustment methods, is ultimately intended to provide information to different audiences to assist in continuously improving the quality of care. To date, many initiatives in the US have taken place to provide additional information on the quality of LTC and home health services[ 6 ]. Similar efforts have not begun in the Canadian home care sector. Several important issues should be addressed prior to moving towards more public reporting of these types of quality data. For example, there needs to be a discussion of the relative importance of the HCQIs. The current project made no attempt to prioritize the indicators, although clearly some would have higher clinical priority than others. To some degree, individual agencies must determine their own priorities. However, a simple system based on prevalence, severity and modifiability may be useful in determining relative importance. For example, pain is a prevalent condition in this population, can be severely debilitating for clients and is clearly modifiable. One might argue that it is of higher importance to address than declining cognitive performance, which is less prevalent and in many cases not modifiable. On the other hand, cognitive decline can have extremely serious consequences for the individual In addition, a decision is needed regarding the calculation of the HCQIs, in particular, which pairs of assessment are to be used in the calculation of failure to improve/incidence HCQIs and what time period in between assessments is acceptable (e.g., a minimum of 90 or 120 days). There will need to be further analysis to explore the relationship between the length of stay and triggering on the HCQIs. It is anticipated that clients who have been on service longer will show important clinical differences from short-stay clients and would therefore be expected to have differential HCQI rates. Tracking of the HCQI rates over time will be essential to determine the level of stability in the indicators. It would be useful to measure the amount of variation and to assess methods to summarize the rates over time in order to maximize their stability. For example, it may be important to report annual HCQI rates for a given organization or province if in fact the rates are found to be highly variable over shorter periods of observation. A shorter time frame may also lead to unstable rates if the number of observations is small. In this paper, a minimum of 20 observations was required and a decision would also be needed for the minimum sample size for public reporting of HCQIs. Finally, it will be essential to assess possible methods to create summary measures across the set of HCQIs. This is not a simple task and one for which little research has been conducted to date. Previous research has pointed to the lack of correlation between measures of quality, but has also suggested that summary measures would be useful for consumers who would likely prefer less complicated information[ 15 ]. Knowledge about the home care sector in Canada remains rudimentary, and our knowledge and understanding of quality assessment in this sector is still in its infancy. The results from the current project serve to enhance the overall understanding of the issues involved in quality assessment and our ability, through the risk adjustment process, to create fair comparisons across providers. Conclusions Risk adjustment of quality indicators is important to enable fair comparisons across geographic regions or across home care providers. To date, little research has examined the quality of home care services. This project, using a set of HCQIs developed by inter RAI , provides an important first step in assessing quality and the variable effects of different types of risk adjustment. Competing interests The author(s) declare that they have no competing interests. Authors' contributions DD participated in the coordination of the study, assisted in the conceptual design of the study, carried out the data analysis and took the lead in developing the manuscript. JH designed the original study, oversaw the development of the indicators, gave input into the data analysis and development of the draft manuscript. BF was the Co-Principal Investigator in the development of the indicators and provided feedback in terms of data analysis and choice of statistical methods. 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/PMC548266.xml

423160

Combining Measures to Characterize Subcellular Machinery

Understanding how the cell functions—or breaks down—implies an understanding of the assembly lines, transportation systems, and powerhouses that keep it running. Global approaches are needed to identify the numerous proteins essential to each cellular machine. But which techniques are best? Lars Steinmetz and colleagues applied and evaluated a variety of methods to define mitochondrial proteins, and report that sets of complementary approaches are needed to characterize a cellular subsystem. Zooming in on mitochondria (Image by Peter Seibel, design by Shayna Roosevelt) Yeast mitochondria make an ideal subject for study. About two-thirds of the estimated 700 mitochondrial proteins have been identified to date, leaving fertile ground for new finds. The researchers previously compiled a list of 477 proteins with compelling evidence for mitochondrial involvement. This list provides a well-defined reference set against which to test protein-finding methods. Moreover, the well-studied and accessible yeast genome is well-suited for exploration and genetic manipulation. Since mitochondria are very similar among all eurkaryotes (organisms whose cells have nuclei), the results will prove relevant across species. Steinmetz and colleagues triangulated results from multiple techniques to identify new candidate mitochondrial proteins. They compared the reference protein list to their new data from protein, mRNA, and gene knockout studies, and to 19 published datasets from other researchers, to evaluate the success of different techniques at finding known mitochondrial proteins. Then they combined evidence across studies to identify a set of proteins that likely characterizes most of the mitochondrial machinery. The researchers first identified proteins from yeast mitochondria using a technique called liquid chromatography mass spectrometry, which separates the proteins by water insolubility (also called hydrophobicity), then identifies each by the mass and molecular charge of its constituents. By comparing this approach with others, the authors show that this proteomic technique alone is by no means comprehensive, nor error-free. Mass spectrometry is biased toward finding more abundant proteins, and the purified mitochondria can contain contaminants from elsewhere in the cell. To address these issues, the authors compared their protein data with a protein study from another group, the reference protein set, and a recent subcellular localization study. Potential mitochondrial proteins identified by more than one protein approach were more likely to localize to mitochondria in the localization study than were proteins identified by only one approach. This finding suggests that, compared to either method alone, a combination of protein and localization measures can more robustly identify proteins residing in mitochondria. But since mitochondria, as the cell's power plants, are integrated into other cellular machinery, the authors argue, methods targeting proteins that are physically located to mitochondria should be complemented with functional approaches. Proteins with mitochondrial roles, regardless of concentration or location in the cell, are better identified by approaches that associate mRNA expression or gene deletion—which removes proteins or renders them inoperable—with changes in mitochondrial function. By comparing results from multiple methods against the reference protein list, the researchers evaluated the likelihood that each protein was mitochondrial. They compiled a list of 691 top candidates. This multi-technique analysis easily outperformed any single study in terms of its ability to identify proteins in the reference set, and of the proportion of known versus unconfirmed proteins located. As mitochondria are well-conserved across species, the results provide a candidate gene list for finding human counterparts that might be associated with mitochondrial disorders. Future studies can use this analysis to evaluate which research methods are likely to be most informative in other cell systems. This paper demonstrates the power of combining techniques with differing strengths in order to zero in on proteins that might elude any single approach, resulting in a more complete parts list for specific cellular machinery.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC423160.xml

543455

Unpacking analyses relying on area-based data: are the assumptions supportable?

Background In the absence in the major Australian administrative health record collections of a direct measure of the socioeconomic status of the individual about whom the event is recorded, analysis of the association between the health status, use of health services and socioeconomic status of the population relies an area-based measure of socioeconomic status. This paper explores the reliability of the area of address (at the levels typically available in administrative data collections) as a proxy measure for socioeconomic disadvantage. The Western Australian Data Linkage System was used to show the extent to which hospital inpatient separation rates for residents of Perth vary by socioeconomic status of area of residence, when calculated at various levels of aggregation of area, from smallest (Census Collection District) to largest (postcode areas and Statistical Local Areas). Results are also provided of the reliability, over time, of the address as a measure of socioeconomic status. Results There is a strong association between the socioeconomic status of the usual address of hospital inpatients at the smallest level in Perth, and weaker associations when the data are aggregated to larger areas. The analysis also shows that a higher proportion of people from the most disadvantaged areas are admitted to hospital than from the most well-off areas (13% more), and that these areas have more separations overall (47% more), as a result of larger numbers of multiple admissions. Of people admitted to hospital more than once in a five year period, four out of five had not moved address by the time of their second episode. Of those who moved, the most movement was within, or between, areas of similar socioeconomic status, with people from the most well off areas being the least likely to have moved. Conclusion Postcode level and SLA level data provide a reliable, although understated, indication of socioeconomic disadvantage of area. The majority of Perth residents admitted to hospital in Western Australia had the same address when admitted again within five years. Of those who moved address, the majority had moved within, or between, areas of similar socioeconomic status. Access to data about individuals from the Western Australian Data Linkage System shows that more people from disadvantaged areas are admitted to a hospital, and that they have more episodes of hospitalisation. Were data to be available across Australia on a similar basis, it would be possible to undertake research of greater policy-relevance than is currently possible with the existing separations-based national database.

Background The majority of work in Australia describing the association between the health status, use of health services and socioeconomic status of the population uses an area-based measure of socioeconomic status. It is necessary to use a proxy measure (the socioeconomic status of the population in the area) because there is no direct measure in the major administrative health record collections of the socioeconomic status of the individual about whom the event is recorded. However, the application of an area-based measure requires a number of assumptions, including that people who move do so between, or within, geographic areas of similar socioeconomic status; and that the (often large) areas used in these analyses provide a reliable indication of the socioeconomic status and health service utilisation of the individuals in the area about whom the event is recorded. Area level socioeconomic status can also be considered as an independent predictor. For example, an individual with low socioeconomic status in an area of higher socioeconomic status is more likely to have better health outcomes than their counterpart in an area of lower socioeconomic status [ 1 , 2 ]. This aspect is not addressed in this paper. In relation to this latter point, Hyndman et al [ 3 ] found that "Misclassification of individuals to SES groups based on the basis of postcode caused an underestimation of the true relationship between SES and health-related measures. A reduction of this misclassification by using smaller spatial areas, such as CD or census enumeration districts, will provide improved validity in estimating the true relationship." A reduction in strength of correlation with increasing size of area is consistent with the results of this paper. In a study of hospitalisations in Michigan, USA, Hofer et al [ 4 ] found that 'the impact of socioeconomic characteristics on hospitalization rates is consistent when measured by individual or community-level measures'. This is an encouraging finding for those limited to using area-based data. Another limitation of the majority of Australian health-related datasets is that they record events (eg., hospital inpatient separations, services by general medical practitioners), rather than individuals. The analysis in this report uses the Western Australian Data Linkage System to explore the reliability of area data as a proxy for socioeconomic disadvantage when analysed for the relatively large geographic units often used in health-related research: it also addresses the limitations of using data about events, rather than individuals. It does this by examining the extent to which hospital inpatient separation rates vary, both overall and by socioeconomic status of area of residence, when calculated at various levels of aggregation, from Census Collection District (CD) – the smallest area level for which a measure of socioeconomic status is available – to the larger units of postcode and Statistical Local Area (SLA). Methods applied include the calculation of correlation coefficients and examination of hospital separation rates by quintile of socioeconomic disadvantage of area, separately for events and individuals. The report also examines the reliability of the socioeconomic status of the address over time, by examining the extent of change in socioeconomic status of area of residence for individuals with repeat hospital episodes over a five year period. The analysis shows that aggregating data to larger area reduces the gap between the index scores for the most disadvantaged and least disadvantaged areas, with the greatest impact on the scores for the most disadvantaged areas. This results in an understatement of the extent of disadvantage in the most disadvantaged areas, as well as an understatement in inequality between the most well off and the poorest areas. Results Individuals Over the five years from 1994 to 1998, a total of 358 948 residents of Perth were admitted to a hospital in Western Australia on one or more occasion, an average of 71 750 individuals admitted per annum. Just over half (53.6%) the individuals admitted were females; 46.4% were males. The rate of individuals admitted was 16.4% higher for females (247.6 separations per 1000 population) than for males, (212.7 separations per 1000 population) (Table 1 ). As can be seen in Figure 1 , the rates of males and females admitted vary notably by age. For females, the highest rate is in the 30 to 34 year age group (with a further three of the five highest female rates between ages 20 and 39 years), with the second highest rate in the 80 years and over age group. The highest male rate, in the 80 years and over age group, is substantially above the next highest male rates, in the 50 to 69 year age groups. Table 1 Perth residents admitted to hospital, by age and sex, at first admission, 1994–98 Rate per 1000 Age Males Females Persons 0–4 185.3 147.2 166.8 5–9 207.1 168.4 188.3 10–14 163.5 135.5 149.9 15–19 201.0 244.4 222.4 20–24 204.3 288.2 245.6 25–29 207.8 315.7 261.4 30–34 212.9 328.1 270.9 35–39 214.3 282.5 248.8 40–44 213.4 250.4 232.3 45–49 214.3 242.4 228.2 50–54 243.6 270.2 256.5 55–59 242.6 254.7 248.6 60–64 252.7 257.5 255.2 65–69 240.5 241.4 241.0 70–74 232.2 237.9 235.3 75–79 237.5 254.8 247.6 80+ 291.7 283.5 286.2 Total 212.7 247.6 230.3 Figure 1 Perth residents admitted to hospital, by age and sex at first admission, 1994–98. Perth population is at 30 June 1996. Per cent shown is of males and females separately, not for persons. A total of 358 768 Perth residents had one admission to a Western Australian hospital over the five years from 1994 to 1998, with a further 298 805 people admitted on two or more occasions (Table 2 ). The number of people with two or more admissions in any period is higher in the earlier years, as the more time that passes the greater the opportunity for a second admission. That is, those with a first admission in 1994 have had more time to record a second admission than have those with a first admission in 1995: thus the greater number with two or more admissions in 1994. Table 2 Perth residents admitted to hospital, by number of admissions and year of separation, 1994–98 Year Individuals One admission Two or more admissions Total 1994 71 566 118 039 189 605 1995 68 400 75 830 144 230 1996 68 989 52 577 121 566 1997 71 917 34 497 106 414 1998 77 896 17 862 95 758 Total 358 768 298 805 657 573 Just over half (54.6%) those admitted to hospital had one admission over this period, and more than one third (36.0%) had between two and four admissions, together comprising 90.6% of those admitted (Table 3 ). Table 3 Residents of Perth admitted to hospital, 1994–1998, by number of admissions per person Admissions per person Number Per cent 1 358 769 54.6 2–4 236 611 36.0 5–9 46 377 7.1 10+ 15 821 2.4 Total 657 578 100.0 Females accounted for just over half (53.6%) of those admitted once, compared with 59.7% of those admitted more than once. For males, the proportions were 46.4% and 40.3%, respectively. Separations There were 1 665 308 separations of Perth residents from Western Australian hospitals, an average of 2.53 separations per person admitted over the five years from 1994 to 1998. Over half (55.1%) of the separations were of females and 44.9% were of males. Figure 2 shows the profiles of males and females, by age, for both individuals admitted (as in Figure 1 ) and separations. For males, the proportion of individuals admitted is highest at ages 20 to 49 years, dropping away at younger and older ages, with the latter exhibiting a particularly marked drop. Total separations for males are generally highest at older ages (the highest at ages 70 to 74 years), reflecting the higher number of separations per person. The notable exception is the high proportion of separations in the 0 to 4 year age group. The profile of the proportion of females admitted is similar to that for males, although it is somewhat distended at ages 20 to 39 years. The proportion of separations of females at ages 25 to 54 years closely follows that for females (individuals) admitted. Figure 2 Perth residents admitted to hospital and total separations, by age and sex, 1994–98. Perth population is at 30 June 1996. Per cent shown is of males and females separately, not for persons. The main differences in the profiles of male and female separations are evident at the youngest ages (higher proportions of males), from ages 20 to 44 years (higher proportions of females) and from 50 to 79 years (higher proportions of males). The ages at which the highest rates of admissions of individuals and of multiple admissions (the gap between the separations and admitted profiles) occur are clearly visible in the chart. Unlike the rates for individuals admitted (Table 1 , above), the highest rates for separations of both males and females occur in the oldest age groups (Table 4 ). The five highest rates for both males and females are in the age groups 60 to 64 years and over, with male rates higher (and often substantially so) than female rates. Also of note is the high rate of separations for females at ages 30 to 34 years (1,672.2 admissions per 1000 population): this is the sixth highest rate for females, and is more than twice the rate for males at the same age (729.5 separations per 1000 population). Table 4 Separations of Perth residents, by age and sex, 1994–98 Rate per 1000 Age Males Females Persons 0–4 989.4 697.9 847.6 5–9 513.2 379.4 448.0 10–14 375.3 310.5 343.8 15–19 443.7 693.4 567.2 20–24 505.2 1 151.9 823.5 25–29 630.1 1 562.6 1 093.4 30–34 729.5 1 672.2 1 203.9 35–39 747.8 1 411.6 1 083.2 40–44 780.6 1 175.8 982.2 45–49 947.6 1 253.0 1 099.0 50–54 1 289.3 1 528.1 1 405.5 55–59 1 686.8 1 666.8 1 676.9 60–64 2 210.8 1 957.6 2 082.6 65–69 2 859.5 2 356.6 2 598.8 70–74 3 991.2 2 661.3 3 268.6 75–79 4 723.2 2 979.7 3 706.7 80+ 4 823.9 3 086.5 3 667.4 Total 1 099.0 1 345.0 1 223.0 Discussion Effect of aggregation of areas on disadvantage scores As noted, the majority of the analysis by socioeconomic status undertaken in the health sector in Australia is area based, and uses the postcode or SLA as the unit of analysis. This raises the question of the extent to which area based analyses at the postcode or SLA level provide a reliable indication of the socioeconomic status and health service utilisation of the individuals admitted. This report explores the reliability of postcode or SLA level data by examining the extent to which rates of individuals admitted and separations vary when calculated at various levels of aggregation (CD, postcode and SLA). Ideally, the comparison would be between the socioeconomic status of individuals and of areas; however, the smallest area level for which a measure of socioeconomic status is available is the CD. Variation in the minimum and maximum Index of Relative Socio-economic Disadvantage (IRSD) scores when calculated at the CD, postcode and SLA level is striking and clearly shows the value of the smaller unit in area based analyses (Table 5 ). The range at the CD level is from a minimum index score of 532 to a maximum index score of 1221, a differential of 2.3 times. When individuals and separations are analysed by postcode, the range in the IRSD scores is narrower, from 863 to 1168 (a differential of 1.4). At the SLA level it is slightly lower again (a differential of 1.3). The effect of aggregation to the larger areas is most noticeable in the minimum IRSD score, increasing the minimum score by 70.5% from the CD level to the SLA level. At the other end of the scale, the maximum score varies little, dropping by 4.0%. That is, the greatest loss in specificity in the IRSD score is in the most disadvantaged areas. Table 5 Range of IRSD scores for area of address of individuals and separations Variable Median for individuals Minimum Maximum Ratio: Maximum/minimum for separations Collection District (1) 1012 532 1221 2.30 Postcode (2) 1015 863 1168 1.35 Statistical Local Area (3) 1017 907 1174 1.29 Ratio of IRSD scores in area (3) to area (1) 1.00 1.70 0.96 .. Thus, the use of larger area aggregates reduces the gap between the index scores for the most disadvantaged and least disadvantaged areas (thus lessening the extent of inequality between these areas), with the greatest impact on the scores for the most disadvantaged areas (thus understating the extent of inequality in these areas). Notably, the difference between the maximum and minimum scores, and the absolute level of the scores, is much less marked between the postcode and SLA. There was a strong association between the IRSD scores for CDs and those for postcode of usual address at the first admission (a Spearman correlation coefficient of 0.74). The correlations were between CDs grouped to quintiles and postcodes grouped to quintiles, ranked by the IRSD, and not between individual CDs and postcodes. A weaker association was found between the quintiles for CDs and those for SLAs (0.64 for people with one separation and 0.63 for people with more than one separation) (Table 6 ). There was little difference in correlation coefficients for those who had moved address. Similar Spearman correlation coefficients were calculated for raw IRSD scores. Table 6 Spearman correlation coefficients between IRSD of address for individuals (at first discharge) and area level Variable Area level of first discharge CD Postcode SLA Individuals: one separation 1.00 0.74 0.64 more than one separation 1.00 0.74 0.63 more than one separation & moved address 1.00 0.73 0.62 Effect of aggregation of areas on separation rates Data at the CD level for the five years from 1994 to 1998 show a variation in rates of individuals admitted from 51 442 admissions per 100 000 population in the most advantaged areas to 58 130 admissions per 100 000 population in the most disadvantaged areas (Table 7 ). This is a differential of 13%. The differential in separation rates is substantially higher, at 47%, reflecting multiple admissions. Table 7 Residents of Perth admitted to hospital, 1994–1998, by socioeconomic disadvantage of area for selected area levels Quintile Individuals admitted Separations CD Postcode SLA CD Postcode SLA Number Q1: Least disadvantaged 126 615 123 380 138 127 294 130 303 131 340 294 Q2 130 907 123 465 114 244 294 307 326 652 279 537 Q3 133 073 126 770 142 107 316 066 328 999 363 908 Q4 124 279 128 863 123 199 327 228 328 630 313 879 Q5: Most disadvantaged 142 704 155 100 139 901 433 577 377 896 367 690 Total 657 578 657 578 657 578 1 665 308 1 665 308 1 665 308 Rate (per 100 000 population) Q1: Least disadvantaged 51 442 48 247 51 950 119 813 120 567 127 986 Q2 53 343 48 239 52 235 120 582 129 945 127 810 Q3 53 889 48 789 52 656 127 995 126 618 134 841 Q4 50 919 51 263 52 564 133 342 128 400 133 920 Q5: Most disadvantaged 58 130 61 691 58 491 176 157 147 734 153 728 Total 53 547 53 547 53 547 135 607 135 607 135 607 Rate ratio: Ratio of rate in Q5 rate in Q1 1.13 *** 1.28 *** 1.13 *** 1.47 *** 1.23 *** 1.20 *** The extent of any inequality is shown by the rate ratio, which expresses the ratio of the rate in Quintile 5 to the rate in Quintile 1; rate ratios indicating differing significantly from 1.0 are shown with * p < 0.05; ** p < 0.01; *** p < 0.001. When data are aggregated to postcode area or SLA, the differentials in separation rates between Quintile 5 and Quintile 1 areas are smaller (differentials of 1.23 and 1.20, respectively) than at the CD level (a differential of 1.47) (Table 7 ). In the case of postcodes, this is largely because of the lower separation rate in Quintile 5 areas (likely to be a result of the process of aggregating CDs), whereas for SLAs it is a combination of a lower separation rate in Quintile 5 areas and a higher rate in Quintile 1 areas (likely to be a result of the aggregation process, exacerbated by the variable size of SLAs – see section titled 'Methods, Area' under 'Methods.' The differential in rates of individuals admitted is the same for data at the SLA and CD level, but higher for postcode areas. These results again reflect the difficulty inherent in producing groups of approximately equal populations. While just over half (54.6%) those admitted to hospital had one separation over this period, the proportion varied from 56.3% in Quintile 1 to 51.9% in Quintile 5 (Table 8 ). This is as expected, with people from the most disadvantaged areas representing a smaller proportion of those with one separation and a larger proportion with more than one separation. Table 8 Number of separations per individual, by socioeconomic disadvantage of area, Perth residents, 1994–1998 Separations per person Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 Total Number 1 69 485 69 118 69 960 69 709 80 497 358 769 2–4 43 274 43 776 45 566 46 747 57 244 236 607 5–9 7 907 7 902 8 449 9 220 12 899 46 377 10+ 2 714 2 668 2 793 3 187 4 459 15 821 Total 123 380 123 465 126 770 128 863 155 100 657 578 Per cent 1 56.3 56.0 55.2 54.1 51.9 54.6 2–4 35.1 35.5 35.9 36.3 36.9 36.0 5–9 6.4 6.4 6.7 7.2 8.3 7.1 10+ 2.2 2.2 2.2 2.5 2.9 2.4 Total 100.0 100.0 100.0 100.0 100.0 100.0 There is a substantial difference in the proportion of the population in Quintiles 5 and 1 having two or more separations (a difference of 38.8%, from a rate of 30 389 separations per 100 000 persons in Quintile 5 to 21 897 separations per 100 000 persons in Quintile 1): the differential for people having one separation is lower, although still notable at 16.1% (a rate of 32 790 separations per 100 000 persons in Quintile 5 and 28 231 separations per 100 000 persons in Quintile 1) (Table 9 ). Table 9 Separations per individual, by socioeconomic disadvantage of area, Perth residents, 1994–1998 Separations per person Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 Total Ratio of rates in Q5/Q1 Per cent 1 19.4 19.3 19.5 19.4 22.4 100.0 .. 2+ 18.0 18.2 19.0 19.8 25.0 100.0 .. Total 18.8 18.8 19.3 19.6 23.6 100.0 .. Rate per 100 000 population 1 28 231 28 165 28 331 28 561 32 790 29 215 1.16*** 2+ 21 897 22 146 23 006 24 236 30 389 24 332 1.39*** Total 50 128 50 311 51 337 52 797 63 180 53 547 1.26*** Average admissions per person with two or more admissions Number 4.2 4.1 4.3 4.4 4.7 4.4 1.12*** The extent of any inequality is shown by the rate ratio, which expresses the ratio of the rate in Quintile 5 to the rate in Quintile 1; rate ratios differing significantly from 1.0 are shown with * p < 0.05; ** p < 0.01; *** p < 0.001. The average number of admissions per person for people admitted to hospital on more than one occasion over the five years to 1998 was 4.4; this varied from 4.2 separations per person admitted in the least disadvantaged areas to 4.7 in the most disadvantaged areas. Reliability over time of address as a proxy for socioeconomic status Studies using the address of usual residence as a proxy for socioeconomic status require two important assumptions. They are that: • people who move do so within, or between, areas of similar socioeconomic status; and that • the areas used in an area based analysis (which can vary in size and are quite often large) provide a reliable indication as to the socioeconomic status and use of health services of the individuals in the area. Data from the 1996 Census show that 53.5% of Perth's population at the 1996 Census reported that they had a different address to that at the previous Census, five years earlier [ 5 ]. Data were not available to compare the IRSD of the first and last SLA of address of the Perth population who moved. However, almost one quarter (24.0%) of Perth residents who moved between the 1991 and 1996 Censuses moved to an address within the same SLA. That is, some 59.3% of the population were in the same SLA after five years (either moved within the SLA, or did not move). This is an encouraging statistic for area based analyses. Similarly, almost four out of five people admitted to hospital more than once in a five year period had not moved (out of the CD of their address at the first separation) by the time of their second separation. For example, of the 298 809 people admitted to a Perth hospital more than once over the five year period 1994 to 1998, over three quarters (78.6%, 64 075 people) had the same address at the time of the second separation. People were recorded as having 'moved' if the CD of their address changed between the first and last separation over the period from 1994 to 1998. Movement to a different address within a CD was not included. The following table illustrates, for people with multiple admissions, the extent of movement by socioeconomic status. For this part of the analysis, the CD of first and last separation have been allocated to quintiles of socioeconomic disadvantage of area, to provide a comparison of the extent of movement between different levels of socioeconomic status. The construction of the quintiles is described in the section titled 'Methods, Measurement of socioeconomic status' under 'Methods.' Table 10 shows, for people who moved to an address in another CD, that: Table 10 Residents of Perth admitted to hospital more than once, 1994–1998, who changed address, by socioeconomic disadvantage of area CD of first separation CD of last separation (%) Total Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 Total Number Quintile 1 40.2 22.8 16.4 15.9 4.7 100.0 9 537 Quintile 2 21.5 24.4 22.9 23.6 7.5 100.0 10 551 Quintile 3 12.7 20.3 24.1 32.5 10.5 100.0 11 730 Quintile 4 7.8 14.6 22.0 40.3 15.3 100.0 13 298 Quintile 5 4.6 9.2 15.0 40.7 30.5 100.0 18 875 Total 14.8 16.9 19.6 32.6 16.0 100.0 63,991 • people from the most well off areas are less likely to have moved to areas of greatly different socioeconomic status (ie, changed quintiles) than are those from the most disadvantaged areas – 40.2% of people in the most advantaged areas (Quintile 1) remained there, despite moving from the CD of their first separation. The proportion in the most disadvantaged (Quintile 5) areas was a lower 30.5%; • while there is movement right across the socioeconomic profile, most movement is between adjacent quintiles. For example, of the 18 875 people who lived in the most disadvantaged areas at their first separation (and moved before a subsequent admission), 71.2% had moved to a CD in the same or next ranked quintile (Quintiles 5 or 4), with just 4.6% moving to the most advantaged areas. Similarly, of the 9 537 people in the most well off areas at their first separation, 63.0% had moved to a CD in the same or next ranked quintile (Quintiles 1 or 2), with a similarly low proportion (4.7%) moving to the most disadvantaged areas; • the most substantial movement between quintiles was of people moving from an address rated as Quintile 5 to one rated as Quintile 4 (40.7%); this was marginally higher than the proportions moving within Quintiles 4 or 1 (40.3% and 40.2%, respectively). There is a strong association between the quintile of socioeconomic disadvantage of area at the first and the last discharge when analysed by CD (a correlation coefficient of 0.88) or SLA (a correlation coefficient of 0.89) of usual address (Table 11 ). This supports the earlier finding that people admitted to hospital who had moved between episodes, moved to or within areas of similar socioeconomic status. The weaker correlations between CD and SLA (see table) highlight the loss in specificity of the index score when aggregated to the (larger) SLA level. Table 11 Correlation coefficients between quintile of socioeconomic disadvantage of area of address of first and last separation, 1994–98 Area of address CD of SLA of first separation last separation first separation last separation CD of first separation 1.00 0.88 0.66 0.60 CD of last separation 0.88 1.00 0.60 0.65 SLA of first separation 0.66 0.60 1.00 0.89 SLA of last separation 0.60 0.65 0.89 1.00 Table includes people admitted more than once, who had moved from the CD of their address at their first separation. Area of address shown at various levels of aggregation of areas. Conclusions The analysis shows that, for Perth residents admitted to hospital, the use of larger area aggregates reduces the gap between the index scores for the most disadvantaged and least disadvantaged areas, thus understating the extent of inequality between these areas. The greatest impact of aggregation of areas is on the scores for the most disadvantaged areas. This results in an understatement of the extent of disadvantage in the most disadvantaged areas, as well as an understatement in the extent of inequality between the most well off and the poorest areas. Further, the analysis shows that a more people from the most disadvantaged areas are admitted to hospital than from the most well-off areas (13% more), and that these people have more separations overall (47% more), as a result of larger numbers of multiple admissions. As regards the extent of movement, four out of five people admitted to hospital more than once in a five year period had not moved (out of the CD of their address at the first separation) by the time of their second separation. In addition: • people from the most well off areas are less likely to have moved to areas of greatly different socioeconomic status than are those from the most disadvantaged areas; • while there is movement right across the socioeconomic profile, most movement out of a quintile is to areas in adjacent quintiles; and • the most substantial movement between quintiles was of people moving from an address rated as Quintile 5 to an address rated as Quintile 4, although this was only marginally higher than the proportions moving within Quintiles 4 or 1. In summary, postcode level and SLA level data provide a reliable indication of socioeconomic disadvantage of area, when compared with CD-level data. That is, the association between rates of total separations and individuals admitted and socioeconomic disadvantage of area evident at the smallest area level (CD) is also evident in the higher level area aggregates of postcode and SLA. It is reasonable to assume that similar relationships exist in other Australian cities, as well as in other health-related activity (eg. visits to general medical practitioners). Given the widespread use in Australia of area based analyses at the postcode and SLA level, and the limitations of CDs an area level for the analysis of most health datasets, it is important to know that such analyses provide a reliable indication of the direction and underlying strength of the influence of socioeconomic factors in hospital admissions rates. This is not to imply that the postcode or SLA are the ideal areal unit for analysis, nor that data for Collection District would be. The ideal population size for area-level analysis is likely to vary dependent on the number of cases in the dataset under analysis. For datasets with a large number of cases per capita (eg. services by general medical practitioners) the number will be smaller than those with a small number of cases per capita (eg. deaths), even with aggregation of data over a number of years. May SLAs have much larger populations than are necessary to produce reliable results; and the populations of most CDs are too small (see Table 12 ). HealthWIZ [ 6 ], the National Social Health Database, comprising among the most widely available small area datasets in Australia, seeks to provide health service use and health status data for areas with populations of approximately 10 000. This is a useful benchmark. Table 12 Number of areas and average population for CDs, postcodes and SLAs in Perth, 1996 Area Number Population In smallest In largest Average CD 2,297 15 1 861 535 Postal area 105 42 49 551 11 780 SLA 37 876 103 736 33 631 It is also clear that data as to socioeconomic position at the smallest area level possible or, more importantly, of individuals, would also be of value. Were data to be available across Australia on a similar basis to that from the Western Australian Data Linkage System, it would be possible to undertake research of greater policy-relevance than is currently possible with the existing separations-based national database. Such moves are under consideration in several Australian States. Further, linking data (eg, using probabilistic linkage) for individuals in the Western Australian Data Linkage System to the Australian Bureau of Statistics Population Census has the potential to add considerable value to such analyses. For example, it would be possible to examine an individual's characteristics of education, occupation, labour force status, housing tenure etc., and to more directly examine the relationships between the number of individuals admitted and total separations and these important socioeconomic variables. Linkage to death registration data would also be valuable in understanding more about outcomes related to socioeconomic status. This latter example is a possibility under recently announced plans for the ABS to test the linking of 2006 Census of Housing and Population data to other datasets, such as deaths registrations, held under their Act. This is similar to the approach elsewhere, including New Zealand [ 7 ]. It is to be hoped that such arrangements can be put in place in Australia in the near future. Methods Terminology The report addresses differences in the number of individuals admitted and the number of separations they incurred. These are described as 'individuals', or individuals admitted' and separations (the total number of separations, where an individual may have had one or more episodes of hospitalisation over the period of the analysis). 'Separation' is the term describing a completed hospital episode: it is defined in the section titled 'Glossary, Separation' under 'Glossary.' Data sources Details of all separations to public and private hospitals in Western Australia for the five years from 1994 to 1998 were extracted from the Western Australian Hospital Morbidity Database (HMDS). Any separation records thought to belong to the same person had previously been linked together within the Data Linkage System, permitting analyses to be performed for both separations and individual persons. The population used in calculating rates is the 1996 Census population. The analysis has been limited to separations of residents of Perth, but includes separations occurring at any public acute or private hospitals in Western Australia. Area Areas used in the analysis are the Census Collection District (CD), postcode and Statistical Local Area (SLA). See Glossary for definitions of CD, postal area and SLA. The HMDS includes address details for each separation from a hospital in Western Australia since 1993. These addresses have been linked to a Western Australian street address database to assign northing and easting points (geo-codes). These points were then assigned to the appropriate 1991 or 1996 CD using the ABS CData96 mapping tool. The postcode and SLA of the address were then determined by allocation of CDs to postcode or SLA. The boundaries for CDs and SLAs are consistent. However, boundaries for CDs and postcodes are not, so CDs were allocated to postcodes on a 'best fit' basis (see Glossary). Consequently, comparisons can be made between results for CDs and postcode areas, CDs and SLAs and postcode areas and SLAs. This is particularly important, as much of the area analysis undertaken in the health sector in Australia uses the postcode or the SLA, as a majority of data are only available at these area levels, and it is widely accepted that the larger the area, the less homogenous the population is likely to be. There were 2 297 CDs in Perth at the 1996 Census, with 105 postcodes and 37 SLAs. The average population size at each of these area levels is shown in Table 12 ; these data emphasise the variation in size of the areas at each area level. Measurement of socioeconomic status In the absence of any direct measure of socioeconomic status in the hospital inpatient data, the socioeconomic status of the area of the address of the individual admitted is used as a proxy measure. The Index of Relative Socio-Economic Disadvantage (IRSD) is the measure used to provide the socioeconomic status of the area of the address. The IRSD is one of five Socio-Economic Indexes for Areas (SEIFA) produced by the Australian Bureau of Statistics (ABS) from data collected at the 1996 Population Census. It is calculated at the CD level and can be produced for other area levels. The postcode and SLA level index scores in this report are the population weighted average of the IRSD scores for the CDs in the postcode or SLA. This calculation is undertaken for all CDs in the postcode or SLA, not just those for which hospital episodes were recorded. Each area level (CD, postcode or SLA) was allocated to one of five groups (quintiles). For example, for SLAs, Quintile 1 comprises the SLAs with the highest IRSD scores (most advantaged areas), and Quintile 5 comprises the SLAs with the lowest IRSD score (most disadvantaged areas): each quintile comprises approximately 20% of the Perth population. This process does not provide an exact allocation of population, so the resultant populations are only 'approximately' equal, and the larger the areal units being allocated, the less likely they are to be equal. As shown in Table 13 , when areas were ranked by their IRSD score at the CD level and then grouped to produce quintiles, the resultant populations were relatively close to the ideal population in each quintile of 245 607 (one fifth of 1 288 036). The quintiles based on postcode areas had rather 'lumpier' populations (greater variation around the one fifth figure of 254 859 per quintile – and a higher total of 1 274 297, due to boundary differences between CDs and postcodes. The quintiles based on SLAs were the most variable. For example, the SLA of Wanneroo – South West (with a population of 103 176) had a score marginally below the cut-off score between Quintile 1 and Quintile 2. However, the inclusion of Wanneroo – South West in Quintile 2 resulted in populations in Quintile 1 and 2 of 161 707 and 321 889, respectively. Moving Wanneroo – South West to Quintile 1 left a population of 218 713 in Quintile 2 and increased that in Quintile 1 to 265 883. While these populations are substantially different from the ideal population, they are the best that can be achieved. Table 13 Population of quintiles at various area levels, 1996 Quintile CD Postcode SLA 1 246 131 255 726 265 883 2 245 406 255 942 218 713 3 246 937 259 835 269 879 4 244 072 251 378 234 378 5 245 490 251 416 239 183 Total 1 228 036 1 274 297 1 228 036 Analysis Three (different) IRSD scores were added to each hospital separation record, based on the CD, postcode or SLA that had been previously assigned to the address on that record. It should be noted that these IRSD scores were actually the average score for the particular CD, postcode or SLA as calculated from 1996 Census data. Quintile ranks for each aggregation level were also applied using population weighting as described above. For analyses involving multiple admissions, the IRSD value used was that for the first separation in the five-year period. These 'first' separations were isolated using the internal links between separation records for the same person and the separation date. Of course, many of these 'first' separations could have been preceded by separations occurring before 1994. Rates are crude rates, per 100 000 population. Ideally the data would have been standardised (by the indirect method). However, access to the source data were limited and to requested tables, and standardisation was not an option. As the data were from a complete enumeration (all admissions to hospital), confidence intervals were only calculated for measures of difference (in this case, rate ratios). The Spearman Rank Correlation has been used in the analysis to indicate the degree of correlation between pairs of variables. Glossary CD The Collection District (CD) is the smallest area level in the Australian Bureau of Statistics' statistical geography and is primarily an area used in the five yearly population census. Index of Relative Socio-Economic Disadvantage The Index of Relative Socio-Economic Disadvantage (IRSD) is one of five Socio-Economic Indexes for Areas produced by the Australian Bureau of Statistics at recent population censuses. Produced using Principal Components Analysis, it summarises information available from variables related to education, occupation, income, family structure, race (the proportion of Indigenous people), ethnicity (poor proficiency in use of the English language) and housing. The variables are expressed as percentages of the relevant population. The IRSD is available at the Census Collection District level and was then be calculated for postcodes and SLAs by weighting the CD level scores by their population. The IRSD is calculated to show the relativity of areas to the Australian average for the particular set of variables which comprise it; this average score is set at 1000. Scores below 1000 indicate areas with relative disadvantaged populations under this measure, and scores above 1000 indicate areas with relatively advantaged populations. The IRSD scores at the Census Collection District (CD) level have been grouped to postal area, an area developed by ABS for the presentation of population counts and other Census data from the five-yearly population censuses to approximate postcode areas, as the ABS does not collect the postcode at the Census. Separation The term describing a completed hospital episode is a 'separation'. At the time of admission to hospital, the age, sex, address of usual residence and other personal details of the patient are recorded. At the end of the episode, at the time of separation from hospital, details of the episode itself are recorded, including the date, time and method of separation (discharge, death or transfer of a patient to another care setting eg. hospital, nursing home). Consequently, hospital inpatient data collections are based on separations. Postal area The postal area is an area developed by ABS for the presentation of population counts and other Census data from the five-yearly population censuses. It approximates postcode areas, as the ABS does not collect the Australia Post postcode at the Census. Postal areas comprise Census Collection Districts (CDs) grouped to approximate postcode areas. Where a CD does not fit entirely within a postcode area, it is allocated to the postcode area into which the population largely falls. Where a CD covers more than one postcode area, the total CD population is allocated to one postcode. The IRSD scores at the Census Collection District (CD) level have been allocated to postal areas as described in the section titled 'Methods, Index of Relative Socio-Economic Disadvantage' under 'Methods.'. Similarly, the postal area of each separation was approximated from the CD of the address. The term postcode, rather than postal area, is used in the text, for ease of reading. Postcode See postal area, above. Quintile of socioeconomic disadvantage of area See section titled 'Methods, Measurement of socioeconomic status' under 'Methods.' SLA An SLA in Perth is generally equivalent to a local government area, with additional codes allocated to local government areas split for statistical purposes (mainly local government areas with large populations, split to form SLAs with smaller populations).

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC543455.xml

519029

Glucose-stimulated insulin response in pregnant sheep following acute suppression of plasma non-esterified fatty acid concentrations

Background Elevated non-esterified fatty acids (NEFA) concentrations in non-pregnant animals have been reported to decrease pancreatic responsiveness. As ovine gestation advances, maternal insulin concentrations fall and NEFA concentrations increase. Experiments were designed to examine if the pregnancy-associated rise in NEFA concentration is associated with a reduced pancreatic sensitivity to glucose in vivo. We investigated the possible relationship of NEFA concentrations in regulating maternal insulin concentrations during ovine pregnancy at three physiological states, non-pregnant, non-lactating (NPNL), 105 and 135 days gestational age (dGA, term 147+/- 3 days). Methods The plasma concentrations of insulin, growth hormone (GH) and ovine placental lactogen (oPL) were determined by double antibody radioimmunoassay. Insulin responsiveness to glucose was measured using bolus injection and hyperglycaemic clamp techniques in 15 non-pregnant, non-lactating ewes and in nine pregnant ewes at 105 dGA and near term at 135 dGA. Plasma samples were also collected for hormone determination. In addition to bolus injection glucose and insulin Area Under Curve calculations, the Mean Plasma Glucose Increment, Glucose Infusion Rate and Mean Plasma Insulin Increment and Area Under Curve were determined for the hyperglycaemic clamp procedures. Statistical analysis of data was conducted with Students t -tests, repeated measures ANOVA and 2-way ANOVA. Results Maternal growth hormone, placental lactogen and NEFA concentrations increased, while basal glucose and insulin concentrations declined with advancing gestation. At 135 dGA following bolus glucose injections, peak insulin concentrations and insulin area under curve (AUC) profiles were significantly reduced in pregnant ewes compared with NPNL control ewes (p < 0.001 and P < 0.001, respectively). In hyperglycaemic clamp studies, while maintaining glucose levels not different from NPNL ewes, pregnant ewes displayed significantly reduced insulin responses and a maintained depressed insulin secretion. In NPNL ewes, 105 and 135 dGA ewes, the Glucose Infusion Rate (GIR) was constant at approximately 5.8 mg glucose/kg/min during the last 40 minutes of the hyperglycaemic clamp and the Mean Plasma Insulin Increment (MPII) was only significantly (p < 0.001) greater in NPNL ewes. Following the clamp, NEFA concentrations were reduced by approximately 60% of pre-clamp levels in all groups, though a blunted and suppressed insulin response was maintained in 105 and 135 dGA ewes. Conclusions Results suggest that despite an acute suppression of circulating NEFA concentrations during pregnancy, the associated steroids and hormones of pregnancy and possibly NEFA metabolism, may act to maintain a reduced insulin output, thereby sparing glucose for non-insulin dependent placental uptake and ultimately, fetal requirements.

Background To meet the increasing fetal demands and maternal energy requirements of pregnancy, alterations in the partitioning and utilization of maternal nutrients must occur. These adaptations are regulated by changing blood concentrations of regulatory metabolites and hormones, together with changes in target tissue responsiveness. Of principle interest are alterations in maternal glucose metabolism during pregnancy, as glucose is a major limiting nutrient of fetal growth [ 1 , 2 ]. As sheep pregnancy advances, circulating maternal insulin concentrations decline [ 3 , 4 ] and the insulin response to a glucose load is significantly reduced [ 5 , 6 ]. These decreased insulin concentrations during the last third of gestation and into lactation in ruminants have been postulated to be the result of the decreased response of the pancreas to insulinotropic agents [ 7 ]. Reducing insulin secretion during pregnancy is proposed to be beneficial to fetal well-being, through the creation of an environment which supports minimizing peripheral glucose utilization and maximizing glucose extraction of the gravid uterus [ 8 , 9 ]. Additionally, the sensitivity of peripheral tissues to insulin is reduced [ 10 , 11 ] and increased mobilization of adipose tissue to supply non-esterified fatty acids (NEFA) as an alternative maternal energy source occurs [ 12 - 15 ]. This pregnancy induced mobilization of adipose tissue is accompanied by a decline in lipid synthesis, and occurs as a result of alterations in insulin receptor numbers, a decreased circulating insulin concentration, as well as interactions with specific hormones of pregnancy [ 16 , 17 ]. High plasma NEFA concentrations have been associated with the development of insulin resistance in the peripheral tissues and also the β-cell [ 18 ]. Elevated NEFAs in the muscle inhibit glucose disposal through interactions with the insulin signaling pathways, as certain lipid species act as secondary messengers (ceramide, diacylglycerol and hexosamine), inhibiting insulin signaling [ 19 , 20 ]. Altered serine/threonine phosphorylation of insulin substrate-1 and direct inhibition of components such as protein kinase B, are also sites of action through which NEFAs may give rise to decreased insulin signaling [ 20 , 21 ]. Several of these steps, including the insulin receptor substrate -1 association with phosphatidylinositol, are reduced in liver and muscle during pregnancy [ 22 ]. Furthermore, cardiac function is impaired in situations of elevated ceramide, through apoptotic pathways [ 23 ]. With regards to the pancreas, elevated NEFA concentrations in rats have been reported to decrease pancreatic responsiveness, resulting in a reduction in glucose stimulated insulin secretion [ 24 - 26 ]. In studies with male rat pancreas, following 48 hours of incubation with elevated NEFA concentrations, glucose-induced insulin secretion, as well as proinsulin biosynthetic responses, are reduced [ 24 , 26 ]. In addition, there is a body of evidence which suggests that chronically elevated NEFA concentrations have a lipotoxic effect on the pancreas, through the formation of nitric oxide and β-cell apoptosis [ 18 , 19 , 27 , 28 ]. Studies investigating a possible interaction between increasing NEFA concentrations and maternal pancreatic sensitivity during ovine pregnancy have not yet been reported. The following experiments were designed to examine whether the pregnancy-associated rise in NEFA concentration is associated with a reduced pancreatic sensitivity to glucose in vivo . Reported here are in vivo pregnant sheep studies of insulin responsiveness to glucose, measured using bolus injection and hyperglycemic clamp techniques. Methods Animals and experimental design Twenty-four 3–4 year old, pen-trained Merino ewes of a known gestational age were used. The experiments were conducted adhering to the National Health and Medical Research Council (NH&MRC) guidelines as administered and approved by the University of Western Sydney, Hawkesbury Animal Care and Ethics Committee and the Commonwealth Scientific Research Organization, Division of Animal Production, Animal Care and Experimentation Ethics Committee. All animals displaying estrus following a program of synchronised mating induced by pre-treatment with intravaginal progestogen pessaries were recorded. Pregnancy was later verified by rectal ultrasound scanning at 28 dGA, and then confirmed together with litter size determination at 65 dGA, using abdominal ultrasound. Those ewes displaying estrus, but failing to conceive were used in a non-pregnant, non-lactating (NPNL) control group. A total of 15 NPNL ewes and nine pregnant ewes, three single bearing and six twin bearing, were used in this study. At 90 dGA a trial glucose bolus injection study was conducted in four ewes. At 105 dGA, all animals were studied (total n = 9), in both glucose bolus and hyperglycaemic protocols and then subsequently again at 135 dGA. At each gestational age, NPNL ewes (n = 15) were also studied under both protocols, expect at 90 dGA were only the bolus studies were conducted and only 5 NPNL animals studied. One week prior to the commencement of the experimental period, ewes were weighed. Body weights (BW) were used to determine the glucose bolus doses to be administered at study (0.4 g glucose/kg). Animals were then individually penned under natural lighting conditions and fed 700 g/d per animal, of a 60:40 pelleted ration, consisting of 60% hammer milled lucerne and 40% hammer milled oaten chaff (92.7 ± 0.34 % dry matter (DM), 68.9 ± 1.05% digestible organic matter, and having a calculated metabolizable energy content of 10.33 ± 0.2 MJ/kg DM, crude protein content of 13.8 ± 0.5%). Water was freely available at all times. Glucose Treatments The ewes were subjected to an experiment protocol consisting of two procedures, 1) a bolus glucose study and 2) a hyperglycaemic clamp study. Each experimental period extended over 5 days, with 2 days in between the procedures. On the day before the commencement of the bolus glucose study, each ewe had polyvinyl jugular vein catheters inserted bilaterally under local anaesthesia. Catheters were maintained patent with daily-heparinized saline flushes (35 U heparin/ml). The bolus glucose procedure was conducted as follows. On the day of the experiment, following a 24-h fast, two pre-injection (basal concentration) blood samples were collected 15 min apart. Immediately following the second basal sample, a glucose bolus was administered. Seven two ml blood samples were collected at 5, 10, 20, 35, 55, 155 and 215 min post-injection and stored in an ice bath until centrifuged at 1,200 × g for 15 min at 4°C. A 0.5 ml aliquot of supernatant was removed for glucose determination, while the remaining plasma was frozen at -20°C for later biochemical analysis. Two days after the bolus glucose injection study, animals were also subjected to a 120 min hyperglycaemic clamp procedure, which was imposed after a 24 h fast [ 29 ]. Prior to the commencement of the clamp, three pre-infusion samples were taken 15 min apart (T -30, T -15 and T 0) through the right jugular catheter. At T 0, a bolus glucose injection was administered through the left jugular catheter and then individual peristaltic pumps were activated to begin the glucose infusion through the left jugular catheter. Blood samples (3.0 ml) were then collected every 5 min from the right jugular catheter. To maintain glucose at the bolus injection level, a spot sample of blood was taken from each 5 min sample and glucose concentration determined using a Boehringer Mannheim Accutrend ® blood glucose monitoring kit (Boehringer Germany). The pump speed setting was then adjusted to maintain the required blood glucose concentration. Samples were collected for analysis as described for the bolus protocol. Assays and calculations The plasma concentrations of insulin, growth hormone (GH) and ovine placental lactogen (oPL) were determined by double antibody radioimmunoassay as previously described [ 30 ]. Insulin measurements were made for every collection point, while GH and oPL were determined from the pooled pre-bolus samples. Plasma glucose and NEFA concentrations were determined as previously described [ 31 , 32 ]. While glucose samples were determined at each sample point, samples for NEFA concentration during the hyperglycaemic clamp studies were collected at T -30 and T 0 and at 85 and 90 min relative to the imposition of the clamp. Inter-assay and intra-assay co-efficients of variation for low and high quality control samples for all assays are detailed. Insulin; Inter-assay LQC, 5.2, HQC, 3.9 and intra-assay LQC, 5.7, HQC, 7.3. GH; Inter-assay LQC 10.3, HQC 7.9, and intra-assay LQC 8.6, HQC 7.5. oPL; Inter-assay LQC 9.1, HQC 12.0, and intra-assay LQC 9.2, HQC 10.9. Glucose; Inter-assay LQC 10.1, HQC 9.3, and intra-assay LQC 10.3, HQC 11.9. NEFA; Inter-assay LQC 11.4, HQC 7.8, and intra-assay LQC 10.6, 6.7. Data from the bolus glucose injection studies were expressed using the following parameters; - Area under curve (AUC), for both the glucose ((mg/ml)/min) and insulin ((ng/ml)/min) response profiles, - Glucose clearance rate (CLR; ml/min) - Glucose half life (t1/2, min) AUC for both the glucose and the resulting insulin response profiles were calculated using the trapezoidal rule. Glucose clearance rate (CLR, ml/min) was calculated as Glucose dose /AUC. Glucose and insulin half-life was determined from the interpolated cumulative response curve. Glucose and insulin concentrations at 10 minutes were used as the peak response. In the hyperglycaemic clamp studies, in addition to AUC calculations for the whole study period, the Mean Plasma Glucose Increment (MPGI, mg/dl), Glucose Infusion Rate (GIR; g (glucose)/kg BW/min) and Mean Plasma Insulin Increment (MPII; ng/ml) were determined over the last 40 min of the clamp [ 29 ], during the GIR steady state period. These results were averaged for each animal, and then pooled to give group results. Statistical analyses There was no significant time of experiment effect when comparing NPNL ewe data collected at each of the gestational age studied, and as a consequence NPNL data from studies at 90, 105 and 135 dGA were pooled and analysed as a single control group against each gestational age group. No differences were found between the singleton and twin bearing ewes and litter size data were pooled to give one pregnant animal group at 105 and 135 dGA (n = 9). Glucose data from the 90, 105 and 135 dGA bolus glucose and 105 and 135 dGA hyperglycaemic clamp experiments were analysed through repeated measures ANOVA with Greenhouse Geisser adjustment [ 33 ]. Between ewe variation was included in the ANOVA model. In the hyperglycaemic clamp studies T0 and baseline measurements were used as co-variates. Glucose and insulin AUC, peak concentrations, t1/2 and CLR data from the bolus glucose injection studies were further analysed using the following orthogonal contrasts of a), NPNL versus pregnant ewes; b) day 90 versus day 105 and day 135 ewes and c) day 105 versus day 135. Body weights, glucose doses, insulin, GH, oPL and NEFA concentrations and the derived variables of GIR, MPII and MPGI were all analysed using two sample unpaired Student's t -tests. Changes in NEFA concentrations following hyperglycaemic clamp treatment at three different physiological states, NPNL, 105 dGA and 135 dGA was analysed using 2 way ANOVA. Results Pre experimental body weights, glucose and insulin concentrations There were no significant differences in body weight (Mean weight: 48.4 ± 1.8 kgs) and required glucose bolus injection (19.5 ± 0.7 g) across the physiological states studied. In the bolus glucose experiments, the basal glucose concentration was significantly lower in pregnant ewes measured on day 135 than on day 90 of gestation (p < 0.05). Basal glucose concentrations in the clamp studies, at 105 and 135 dGA, were also significantly reduced compared to NPNL concentrations (p < 0.05, Table 1 ). The basal insulin concentrations as gestation advanced declined, and basal levels measured at 105 and 135 days gestation were significantly reduced compared to NPNL concentrations in both experimental treatments (p < 0.05, Table 1 ). Table 1 Basal glucose (mg/dl) and insulin (ng/ml) concentrations for both bolus glucose injection and hyperglycaemic clamp studies at four gestation age groups. Within columns values with different superscripts are significantly different at p < 0.05 and (n) = number of animals. Glucose Insulin Bolus Clamp Bolus Clamp NPNL (15) 46.9 ± 1.7 a 51.1 ± 1.9 a 0.89 ± 0.13 a 1.09 ± 0.10 a 90 dGA (4) 40.1 ± 4.8 ab 0.64 ± 0.05 ab 105 dGA (9) 29.1 ± 4.0 bc 38.6 ± 3.0 b 0.49 ± 0.11 bc 0.53 ± 0.06 b 135 dGA (9) 21.3± 1.9 c 40.6 ± 4.9 b 0.29 ± 0.07 c 0.36 ± 0.06 b Bolus glucose studies Glucose injection significantly increased glucose concentrations in all groups compared to basal levels (Figure 1a , p < 0.001). After co-variate adjustment for basal glucose concentration, repeated measure ANOVA of glucose concentration after bolus injection revealed no significant differences in maximal glucose concentration obtained between pregnant and NPNL ewes. There was no difference for glucose AUC between NPNL and pregnant ewes, nor did CLR and t1/2 differ between NPNL and pregnant ewes (Table 2 ). Figure 1 (a) Plasma glucose (mg/dl) and (b) insulin concentrations (ng/ml) following a bolus glucose injection (0.4 g/kg) in NPNL ewes (□, n = 15) and pregnant ewes, 90 (×, n = 4), 105 (Δ, n = 9) 135 (■, n = 9) dGA. Table 2 Glucose peaks (mg/dl), AUC (mg/ml)/min), clearance rate (CLR, ml/min) and half life (t1/2, min) for NPNL (n = 15) ewes and pregnant ewes, 90 (n = 4), 105 (n = 9), and 135 (n = 9) dGA following bolus glucose injection. Glucose peak AUC CLR t1/2 NPNL 244.9 ± 6.1 166.46 ± 10.2 129.9 ± 7.8 49.6 ± 2.9 90 dGA 239.3 ± 21.9 162.0 ± 18.5 124.42 ± 11.9 50.9 ± 3.8 105 dGA 192.5 ± 12.1 131.4 ± 9.7 149.1 ± 6.7 48.2 ± 3.2 135 dGA 203.4 ± 9.6 162.4 ± 11.4 123.4 ± 8.4 57.4 ± 2.5 Whereas peak insulin concentrations and the insulin AUC profiles were significantly reduced in pregnant compared with NPNL ewes (Figure 1b , p < 0.001 and Table 3 ). Day 135 ewes had significantly reduced peak insulin responses compared to day 105 ewes (Table 3 , p < 0.04). Ewes at 105 and 135 dGA had significantly reduced AUC when compared to day 90 ewes (Table 3 , p < 0.001). The decline in AUC was linearly related to stage of pregnancy, whilst insulin t1/2 did not vary significantly across the groups (Table 3 ). Table 3 Peak insulin responses (ng/ml), AUC ((ng/ml)/min) and half life (t1/2, min) for NPNL (n = 15) ewes and pregnant ewes 90 (n = 4), 105 (n = 9) and 135 (n = 9) dGA following bolus glucose injection. Peak AUC t1/2 NPNL 13.5 ± 1.4 1095.3 ± 112.8 60.7 ± 4.1 90 dGA 6.9 ± 2.0 458.8 ± 48.9 61.9 ± 12.2 105 dGA 4.7 ± 0.8 277.5 ± 60.3 45.7 ± 7.2 135 dGA 2.5 ± 0.5 169.9 ± 24.9 68.9 ± 10.8 Hyperglycaemic clamp studies The bolus glucose injection, prior to the start of the clamp, significantly elevated glucose concentrations in all groups (Table 1 and Figure 2a , p < 0.001). Following adjustment for basal differences, the glucose concentrations obtained were not statistically different between pregnant and NPNL ewes or between days of gestation. Pregnant ewes while having glucose concentrations not different from NPNL ewes had significantly reduced insulin responses and continued depressed insulin secretion over the 2 hours of the study (Figure 2b , p < 0.001). The measured insulin AUC (ng/ml)/min), were significantly reduced over gestation in 105 dGA and 135 dGA ewes compared to NPNL ewes (NPNL; 1675 ± 110 vs. 105 dGA; 491 ± 67 and 135 dGA; 427 ± 55, p < 0.001). Figure 2 (a) Plasma glucose (mg/dl), (b) insulin concentrations (ng/ml) and (c) glucose infusion rates (mg glucose/kg/min) during a hyperglycaemic clamp conducted on NPNL ewes (□, n = 15) and pregnant ewes at 105 (Δ, n = 9) 135 (■, n = 9) dGA. Mean plasma glucose increment (MPGI), glucose infusion rate (GIR) and Mean plasma insulin increment (MPII) The MPGI did not differ significantly between the three physiological states examined (Table 4 ). Despite a slight rise in glucose concentrations during the clamp, the amount of glucose required to maintain glucose concentrations constant, declined from approximately 14 to 6 mg glucose/kg/min (Figure 2c ). A steady state infusion point was obtained in the last 40 min of the clamp. GIR during this period was constant, at approximately 5.8 mg glucose/kg/min, across all physiological states (Table 4 ). Although MPII during the last 40 min of the glucose hyperglycaemic infusion was significantly greater in NPNL ewes, than in the day 105 and 135 groups (Table 4 , p < 0.001), the glucose infusion rates necessary to maintain hyperglycemia within the groups was not significantly altered (Table 4 ). Table 4 Mean plasma glucose increment (MPGI, mg/dl), glucose infusion rates (GIR, mg glucose/kg/min) and mean plasma insulin increments (MPII, ng/ml) during the last 40 min of hyperglycemic clamp for NPNL (n = 15) ewes and pregnant ewes at 105 (n = 9) and 135 (n = 9) dGA. Within columns values with different superscripts are significantly different at p < 0.05. MPGI GIR MPII NPNL 368.3 ± 12.5 a 5.58 ± 0.39 a 16.44 ± 1.1 a 105 dGA 344.1 ± 24.8 a 6.03 ± 0.51 a 5.1 ± 0.61 b 135 dGA 354.4 ± 16.2 a 5.66 ± 0.71 a 5.03 ± 0.74 b GH, oPL and NEFA concentrations Pre-bolus pooled samples for maternal GH and oPL concentrations displayed increased concentrations with gestational age. Maternal GH concentrations increased with gestation, rising significantly from 0.78 ± 1.1 ng/ml in the NPNL state (p < 0.05), to 2.13 ± 0.35 ng/ml at 105 dGA and 3.5 ± 0.9 ng/ml at 135 dGA. Placental lactogen concentrations were 579 ± 74 ng/ml at 105 dGA and increased to 1211 ± 144 ng/ml (p < 0.05) by 135 dGA. Both pregnant groups displayed higher NEFA concentrations than did the NPNL control ewes following a 24-h fast (Figure 3 ). Pre-fasting, pre-clamp NEFA concentrations at day 105 were significantly greater than those of NPNL ewes (p < 0.05, Figure 3 ). Ewes at 135 dGA displayed significantly elevated NEFA concentrations compared to NPNL NEFA concentrations (P < 0.001, Figure 3 ), and higher concentrations than at 105 dGA (p < 0.06). Following the 2 hour hyperglycaemic clamp, NEFA concentrations were reduced to levels 54% of pre-clamp levels in NPNL ewes, 59% in day 105, and 58% in day 135 ewes (Figure 3 ). Circulating NEFA concentrations in 105 dGA ewes following the hyperglycaemic clamp were not significantly different from NPNL concentrations, nor were they different from post clamp 135 dGA concentrations, though 135 dGA post clamp concentrations remained elevated above NPNL concentrations (p < 0.05, Figure 3 ). Figure 3 Post fast circulating NEFA concentrations (μmol/L), pre (■) and post (□) the imposition of a hyperglycaemic clamp at three physiological states, NPNL (n = 15), day 105 (n = 9) and day 135 (n = 9) ewes. Comparisons are by 2-way ANOVA compared to NPNL post fasting pre-hyperglycaemic clamp NEFA concentrations, * p < 0.05, **p < 0.01 and *** p < 0.001. Discussion There are two important findings of these studies. Firstly, circulating maternal insulin concentration and glucose-stimulated insulin release decrease as gestation advances. Secondly, the imposition of a two-hour hyperglycaemic clamp in pregnant ewes reduces NEFA concentrations, to concentrations not different form the pre clamp concentrations of fasted non-pregnant non-lactating ewes, though despite the reduction, insulin response to glucose remains depressed in 105 and 135 dGA ewes. These results are in agreement with the blunted insulin release during a hyperglycemic clamp treatment observed in lactating cattle, where NEFA concentrations are also elevated late in gestation, prior to a clamp [ 29 , 34 ]. Glucose-stimulated insulin release over gestation in our report was assessed by the capacity of the maternal pancreas to release insulin in response to two forms of glucose challenge. Both of these challenges confirmed that by the end of the second third of gestation (105 dGA, term 147 dGA), a significant and maintained depression in insulin release in the light of elevated glucose levels was observed, which was repeated near term (135 dGA). This suppression of glucose stimulated insulin release was accompanied by increases in maternal NEFA, GH and oPL concentrations. Lipid metabolism during pregnancy has two distinct phases. The first two thirds of pregnancy are characterised by a low lipolytic activity and the majority of maternal energy appears to be derived from dietary carbohydrate. During the later part of pregnancy and into lactation, there is an increased level of lipolytic activity resulting in increased NEFA concentrations [ 17 ], which serve as an alterative energy source, reducing maternal peripheral reliance upon glucose [ 12 , 13 ]. Under these increasing rates of lipolysis, the whole body glycemic response to insulin changes to conserve glucose for uterine uptake and fetal growth [ 13 ] and maternal insulin concentrations decline [ 3 , 4 ]. NEFAs are well documented to promote glucose stimulated insulin secretion [ 35 , 36 , 36 ], however, the chronic effects of elevated NEFA concentrations, specifically as occur during pregnancy, are ill defined, and may be associated with suppressed glucose stimulated insulin release [ 25 ]. The responsiveness of the male pancreas has been reported to be decreased following exposure to elevated NEFA concentrations for periods from 3 up to 48 hours [ 24 - 26 ]. In the pregnancy studies presented here, a two-hour hyperglycaemic clamp at two time points in the later half of gestation, reduced NEFA concentrations by approximately 50%, but maternal insulin release in response to glucose remained suppressed. This is in contrast to the same protocol in NPNL ewes, where NEFA concentrations were also reduced, though pancreatic insulin response was maintained. It is interesting to note that the NEFA concentrations before the clamp in NPNL ewes were similar to pregnant ewes' NEFA concentrations at the end of the clamp procedure, and yet the NPNL insulin response remained unaltered and NPNL NEFA concentrations had declined by approximately 50%. In contrast, in the pregnant ewes, insulin release remained suppressed, despite a comparable depression in NEFA concentrations as generated in the NPNL ewes, similar to NPNL pre-clamp NEFA concentrations. Somewhat similar studies have been conducted in fed and fasted non-pregnant mice [ 36 ]. When a hyperglycaemic clamp was imposed, NEFA concentrations were reduced. While fasted animals had a normal first phase insulin release, similar to the fed animals, this response fell away and significantly reduced insulin secretion rates were observed [ 36 ]. The interesting differences between these mice studies and the pregnant sheep response, is that the initial insulin response is also blunted in pregnancy, suggesting other effects of fasting and or pregnancy are involved in this absolute suppression of a glucose stimulated insulin response. These data demonstrate that the acute suppression of NEFA concentrations per se during pregnancy does not result in the restoration of a NPNL-like insulin response. This suggests that possibly other factors of pregnancy, such as the influence of lactogenic and steroid hormones, and NEFA metabolism may be acting upon the maternal pancreas to suppress the maternal insulin response during pregnancy in sheep. Placental lactogens have been documented in many mammalian species, including the sheep and human, and PL concentrations increase with advancing gestational age, reaching maximal concentrations just prior to term [ 37 , 38 ], similar to what is reported in this study. Homologous studies with human placental lactogen, prolactin and growth hormone have shown significantly elevated in vitro secretion of insulin from pancreatic islets [ 39 ]. Interestingly, more recently it has been suggested that PL and placental GH act in concert to modulate maternal metabolism, resulting in an increase in the available supply of glucose and amino acids to the fetus [ 9 ]. Studies concerning the effect sheep PL may have on pancreatic function however, are not definitive [ 40 - 42 ]. The short-term removal of oPL by immunoneutralisation increased insulin concentration [ 42 ], while an acute oPL infusion failed to demonstrate any significant changes in insulin levels [ 41 ]. However, in the carunculectomy model of fetal growth restriction (FGR), carunculectomised ewes have increased glucose concentration, without any increase in insulin [ 40 ], suggesting that there may be a pregnancy-specific impairment of insulin secretion during sheep pregnancy. Placental progesterone has also been reported to play a major role in modulating insulin release during pregnancy. In rat islet studies, progesterone counteracts lactogenic insulin stimulatory behaviour [ 43 ], and circulating progesterone levels in the sheep increase from approximately 50 dGA [ 38 , 44 ]. Possible differential and interactive effects of ovine PL and progesterone on insulin secretion as gestation advances, as observed in the rat [ 43 , 45 ], remain to be fully explored. Despite this fact, changes in adipose tissue insulin sensitivity suggest that in pregnancy, insulin resistance may develop together with a change in the pattern of substrate utilization, which occur under rising progesterone, prolactin, PL and GH concentrations [ 9 , 13 , 46 ]. Actions of GH on adipose tissue include inhibition of insulin-induced fatty acid synthesis [ 47 ]. Also a down-regulation of GLUT-1 and GLUT-4 in vivo , in rat adipocyte plasma membranes incubated with GH, has resulted in a reduced adipocyte glucose uptake [ 48 ], associated with a down regulation of insulin sensitivity or responsiveness. In addition, despite initial experiments reporting that GH and oPL do not modulate lipolytic activity [ 17 ], there is evidence to suggest PL and prolactin decrease adipocyte glucose transport [ 49 ] and that oPL may act in concert with GH, enhancing GH effects on the pattern of substrate utilization [ 9 , 50 ]. Another candidate that presents itself as possibly regulating maternal insulin response is that of leptin. Leptin concentrations rise during pregnancy [ 51 ], and stimulates lipolysis in muscle tissues [ 52 ], and may also be involved in a unique form of lipolysis where glycerol is released instead of NEFAs [ 53 ]. A direct role of action upon human and rat islet cells has been demonstrated in culture and in vivo, through actions in the regulation of Ca 2+ influx into the β-cell [ 54 , 55 ]. However, when leptin was administered to fasted mice there were no observed changes in plasma insulin or glucose concentrations [ 54 ], suggesting other factors in a fasted state act to regulate pancreatic insulin release in response to glucose. Leptin concentrations were not determined in the study reported here, but the relationship between elevated leptin levels during pregnancy and pancreatic responsiveness is an interesting one, and one that requires further study in vivo . One final possible explanation of the data reported here could be that the chronic elevation in NEFA levels, as occurs in pregnant ewes, does play a role. While the acute suppression of NEFA concentrations during pregnancy did not result in the restoration of a NPNL-like insulin response, a chronic exposure to elevated NEFAs imparts a continuous 'lipotoxic' inhibitory effect. Increased NEFA concentrations have been associated with a lipotoxic effect on the pancreas resulting in suppressed insulin release and β-cell apoptosis, following the formation of nitric oxide [ 18 , 19 , 27 , 28 ], possibly similar to that demonstrated to occur in myocardium [ 23 ]. However no studies have yet been conducted to examine this possible pathophysiological outcome in the pregnant animal. One final aspect of the present studies was the attempt to use the hyperglycemic clamp parameters of MPGI, GIR and MPII [ 29 , 34 ], to define more precisely pregnancy-induced insulin resistance in sheep. The MPGI was constant across all physiological states and MPII, as an index of the ewe's pancreatic responsiveness to elevated glucose concentrations, was reduced in late pregnancy, reflecting the depressed sensitivity of the pancreas throughout the hyperglycemic clamp. During the final 40 minutes of the hyperglycemic clamp, GIR achieved glucose concentrations remained similar between the groups, suggesting both NPNL ewes and pregnant ewes were utilizing and extracting available glucose at the same rate, though to obviously different sinks. This is further supported by the fact that following a fast, both NPNL and pregnant ewes had glucose clearance rates not significantly different from one another. In the present experiments, feed intakes were not different across pregnant and non-pregnant ewes and the constant infusion rates suggest similar total utilization rates, while the difference in MPII reflects different patterns of glucose utilization. Thus as pregnancy advances, the site of utilization changes, from peripheral tissues (muscle and adipose tissues) to the gravid uterus [ 13 , 56 ]. If there were no mechanisms in place to suppress peripheral glucose utilization, be it insulin resistance or reduced insulin output, it should be expected that pregnant animals would show a greater GIR and an increased MPII. Therefore, as GIR was constant across groups and insulin secretion is suppressed, the peripheral glucose utilization, through the induction of peripheral insulin resistance, must be depressed in an effort to support glucose homeorhesis during pregnancy. Conclusions Increasing NEFA concentrations as gestation advances provide the maternal compartment with a double advantage. Firstly they act to an alterative fuel source for maternal metabolism. Secondly, they act to promote the development of an insulin resistant state, which is aided through the suppression of the glucose-stimulated insulin release response. This later action is still not well defined in pregnancy, but could be the result of interactions between hormones of pregnancy and the maternal pancreas, and possibly the result of chronically elevated NEFA concentrations. Through the development of insulin resistance in maternal peripheral tissues and a reduced insulin output, glucose is 'spared' and available for placental uptake and ultimately, fetal demands. List of abbreviations as defined in the text NEFA Non-esterified fatty acids NPNL Non-pregnant, non lactating BW Body weight DM Dry matter GH Growth hormone OPL Ovine placental lactogen AUC Area under curve CLR Glucose clearance rate MPGI Mean plasma glucose increment GIR Glucose infusion rate MPII Mean plasma insulin increment

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC519029.xml

212694

Microarray Analysis

Microarrays can survey genome-wide expression patterns. Not only can these gene expression profiles be used to identify a few genes of interest, they are now being creatively applied for hypothesis generation and testing

Microarrays are used to survey the expression of thousands of genes in a single experiment. Applied creatively, they can be used to test as well as generate new hypotheses. As the technology becomes more accessible, microarray analysis is finding applications in diverse areas of biology. Microarrays are simply a method for visualizing which genes are likely to be used in a particular tissue at a particular time under a particular set of conditions. The output of a microarray experiment is called a “gene expression profile.” Gene expression profiling has moved well beyond the simple goal of identifying a few genes of interest. The notion that this is the major objective of microarray studies has engendered the oft-repeated criticism that the approach only amounts to “fishing expeditions.” The sophistication of microarray analysis very much blurs the distinction between hypothesis testing and data gathering. Hypothesis generation is just as important as testing, and very often expression profiling provides the necessary shift in perspective that will fuel a new round of progress. In many gene expression profiling experiments, the hypotheses being addressed are genome-wide integrative ones rather than single-gene reductionist queries. In general, without a hypothesis only the most obvious features of a complex dataset will be seen, while clear formulation of the scientific question undoubtedly fuels better experimental design. And in some cases, the results of a microarray screen that was initially designed as an effort at cataloguing expression differences are so unexpected that they immediately suggest novel conclusions and areas of enquiry. Fundamental Microarray Technology All microarray experiments rely on the core principle that transcript abundance can be deduced by measuring the amount of hybridization of labeled RNA to a complementary probe. The idea of a microarray is simply to lay down a field of thousands of these probes in perhaps a 5 sq cm area, where each probe represents the complement of at least a part of a transcript that might be expressed in a tissue. Once the microarray is constructed, the target mRNA population is labeled, typically with a fluorescent dye, so that hybridization to the probe spot can be detected when scanned with a laser. The intensity of the signal produced by 1,000 molecules of a particular labeled transcript should be twice as bright as the signal produced by 500 molecules and, similarly, that produced by 10,000 molecules half as bright as one produced by 20,000 molecules. So a microarray is a massively parallel way to survey the expression of thousands of genes from different populations of cells. Trivially, if fluorescence is observed for a gene in one population but not another, the gene can be inferred to be on or off, respectively. With appropriate replication, normalization, and statistics, though, quantitative differences in abundance as small as 1.2-fold can readily be detected. The output of all microarray hybridizations is ultimately a series of numbers, which covers a range of almost four orders of magnitude, from perhaps one transcript per ten cells to a few thousand transcripts per cell ( Velculescu 1999 ). It is the comparison of gene expression profiles that is usually of most interest. This is because the visualization is done at the level of transcript abundance, but just seeing a transcript does not guarantee that the protein is produced or functional. If, however, a difference in transcript abundance is observed between two or more conditions, it is natural to infer that the difference might point to an interesting biological phenomenon. A general approach to performing gene expression profiling experiments is indicated as a flow diagram in Figure 1 . Having performed the experiment, quality control checks, statistical analysis, and data-mining are performed. More and more, investigators are interested not just in asking how large the magnitude of an expression difference is, but whether it is significant, given the other sources of variation in the experiment. Similarly, we might want to evaluate whether some subset of genes show similar expression profiles and so form natural clusters of functionally related genes. Or we may combine expression studies with genotyping and surveys of regulatory sequences to investigate the mechanisms that are responsible for similar profiles of gene expression. Finally, all of the expression inferences must be integrated with everything else that is known about the genes, culled from text databases and proteomic experiments and from the investigator's own stores of biological insight. Figure 1 Flow Diagram of Gene Expression Profiling Fishing for Hypotheses The ability to survey transcript abundance across an ever-increasing range of conditions gives geneticists a fresh look at their cellular systems, in many cases providing a more holistic view of the biology, but at the same time feeding back into the classical hypothetico-deductive scientific framework. The technology has rapidly advanced beyond the simple application of fishing for candidate genes and now sees applications as diverse as clinical prediction, ecosystem monitoring, quantitative mapping, and dissection of evolutionary mechanisms. Two of the better-known examples of the interplay between microarray profiling and hypothesis testing are provided by the studies of Ideker et al. (2001) and Toma et al. (2002) . The latter authors profiled the difference in expression between strains of flies that had been divergently selected for positive and negative geotaxis, a supposedly complex behavior relating to whether flies prefer to climb or stay close to the ground. They identified two dozen differentially expressed genes, several of which were represented by mutant or transgenic stocks that allowed tests of the effect of gene dosage on behavior. At least four of the candidate genes indeed quantitatively affect geotaxis. Ideker et al. (2001) took this approach a step further in arguing for a four-step iterative feedback between profiling, identifying candidate genes, knocking them out, and then profiling once more. They showed how thoughtful experimentation can considerably enhance our understanding of genetic regulatory pathways such as the yeast galactose response. Much excitement has been generated recently by the potential for clinical applications of gene expression profiling in relation to complex diseases such as cancer, diabetes, aging, and response to toxins. An early foray into this realm was provided by Alizadeh et al. (2000) , who demonstrated that diffuse large B-cell lymphomas have two major subtypes defined by molecular profiles. Whereas it is difficult to predict clinical outcome on the basis of histology, these profiles define a set of genes that provide quite a strong indicator of long-term survival. Similarly, van't Veer et al. (2002) have described a “poor prognosis” signature in breast cancer biopsies from young women prior to the appearance of metastases in the lymph nodes. Much statistical and empirical work remains to be done before these tools see clinical application, but the idea that gene expression integrates signals from the genotype and environment provides potent motivation for studying disease with microarrays. A good example of the ability of microarray analyses to simply surprise us is provided by the study reported in this issue of PLoS Biology by DeRisi and colleagues ( Bozdech et al. 2003 ). They reasoned that profiling transcript abundance throughout the erythrocyte phase of the lifecycle of the malaria parasite Plasmodium falciparum might identify a handful of genes that are induced at critical times and hence might be novel drug targets. Employing very careful staging, a platform with low experimental noise, and appropriate statistical procedures, they discovered an extremely tight molecular lifecycle within the organism. Families of functionally related genes are induced as a unit, one after another, in a tightly orchestrated rhythm that testifies to incredible integration of the physiology of the parasite. They show that with microarray analysis it is possible to model the physiology and biochemistry of the pathways instead of just targeting a few genes. In the coming years, expect to see microarrays developed for an extremely diverse range of organisms and applied to an even wider range of questions, from parasitology to nutritional genomics. Consensus on a core set of statistical options will likely emerge, as will agreement on data quality standards. Applications will encompass defining gene function; inferring functional networks and pathways; understanding how variation is distributed among individuals, populations, and species; and developing clinical protocols relating to cancer prognosis and detection of toxin exposure. Similar profiling methods for proteins and metabolites will attract just as much attention as functional genomics, building on the foundations laid by genome sequencing.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC212694.xml

535702

Sleep Duration Affects Appetite-Regulating Hormones

Some of us, when awake in the middle of the night, feel an urge to visit the kitchen. This could explain results of previous studies that have shown a link between short sleep duration and high body mass index (BMI). But a study by Emmanuel Mignot and colleagues suggests that it's not just the additional snacking opportunities that make short sleepers more likely to be overweight. Intrigued by the connection between sleep and BMI, and by recent studies showing that sleep deprivation in laboratory settings can cause a decrease in serum levels of leptin, a hormone known to control appetite, Emmanuel Mignot and colleagues set out to study the levels of various hormones known to regulate appetite and energy expenditure under “real life” conditions. They took advantage of the Wisconsin Sleep Cohort Study, an ongoing longitudinal study of sleep habits and disorders in the general population. The study began in 1989, when researchers mailed state employees aged 30–60 years a survey on sleep habits, health, and demographics. Mail surveys were repeated at 5-year intervals, and some of the respondents were recruited to sleep a night in the laboratory and undergo various tests. A number of participants were also asked to keep a sleep diary for 6 days. The study has already shown connections between sleep apnea and hypertension, and between menopause and sleep-disordered breathing. For their study, Mignot and colleagues measured sleep duration (habitual and immediately prior to blood sampling), BMI, and pre-breakfast blood hormone levels in 1,024 participants. Consistent with previous studies, they found that in individuals who sleep less than 8 hours (74% of all participants), BMI was inversely proportional to sleep duration. In addition, short sleep was associated with low leptin and high ghrelin levels (ghrelin is a hormone thought to stimulate food intake).These hormonal differences are likely to increase appetite, which could be responsible for the increased BMI in short sleepers. Peaceful sleep (Photo: Sharad Taheri) These findings could explain, at least in part, why societies in which excess calories are much easier to come by than a good night's sleep are more prone to obesity. Mignot and colleagues plan to test this in intervention studies where they make people sleep more and measure the effects on body mass. “Good sleep, healthy eating habits, and regular exercise each may have important roles in fighting obesity in modern society,” suggests Mignot.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535702.xml

514531

The AMC Linear Disability Score project in a population requiring residential care: psychometric properties

Background Currently there is a lot of interest in the flexible framework offered by item banks for measuring patient relevant outcomes, including functional status. However, there are few item banks, which have been developed to quantify functional status, as expressed by the ability to perform activities of daily life. Method This paper examines the psychometric properties of the AMC Linear Disability Score (ALDS) project item bank using an item response theory model and full information factor analysis. Data were collected from 555 respondents on a total of 160 items. Results Following the analysis, 79 items remained in the item bank. The remaining 81 items were excluded because of: difficulties in presentation (1 item); low levels of variation in response pattern (28 items); significant differences in measurement characteristics for males and females or for respondents under or over 85 years old (26 items); or lack of model fit to the data at item level (26 items). Conclusions It is conceivable that the item bank will have different measurement characteristics for other patient or demographic populations. However, these results indicate that the ALDS item bank has sound psychometric properties for respondents in residential care settings and could form a stable base for measuring functional status in a range of situations, including the implementation of computerised adaptive testing of functional status.

Background It is now widely accepted that examining quality of life is an important aspect in the treatment and evaluation of many conditions. Functional status is seen as an important determinant of quality of life. A wide variety of instruments have been developed to quantify functional status [ 1 ]. These instruments tend to have a fixed length and all items are administered to the whole group of patients under scrutiny. However, currently interest is moving towards the more flexible framework offered by item banks. An item bank is a collection of items, for which the measurement properties of each item are known [ 2 , 3 ]. When using an item bank, it is not essential for all respondents to be examined using all items. This enables the burden of testing to be considerably reduced for both patients and researchers. It is even possible to select the 'best' items for individual patients using computerised adaptive testing algorithms [ 4 ]. Furthermore, results from studies using different selections of items from an item bank can be directly compared. Item banks, measuring concepts such as quality of life [ 2 , 5 ], the impact of headaches [ 6 ] or functional status [ 7 , 8 ], have been developed. The AMC Linear Disability Score (ALDS) project item bank was developed to quantify functional status [ 7 , 9 ]. The ALDS item bank covers a large number of activities, which are suitable for assessing respondents with a very wide range of functional status and many types of chronic condition. The item bank is particularly suitable for use in the Netherlands. The ALDS items were obtained from a systematic review of generic and disease specific functional health instruments [ 1 ]. Five psychometric aspects of the ALDS item bank need to be considered before it can be implemented. These are: (a) there needs to be enough variation in the response categories used for each item [ 9 ]; (b) estimates of the item response theory model parameters should not depend on patient characteristics such as age or gender [ 10 , 11 ]; (c) estimates of the item response theory model parameters, which are stable across different subsets of items from the instrument and based on a sufficiently large sample [ 12 ] of respondents, should be available [ 9 ]; (d) an examination of the extent to which the ALDS items represent a single construct; and (e) testing whether a simpler item response theory model is suitable for the set of items. This paper examines these five aspects of the ALDS item bank using the responses given by residents of supported housing schemes, residential care and nursing homes in and around Amsterdam, the Netherlands. This, mainly elderly, population has been chosen because they generally experience some level of functional restriction and consume a large amount of health care services. Methods Data collection This paper considers 160 items, which were considered to be applicable in a residential care setting. Each item has two response categories: 'I could carry out the activity' and 'I could not carry out the activity'. If a respondent had never had the opportunity to experience an activity 'not applicable' was recorded. In the analysis, responses in the category 'not applicable' were treated as if the individual items had not been presented to the individual respondents [ 13 ]. It was felt that presenting all 160 items to each respondent would place an unnecessary and unacceptable burden on those responding to the items. Therefore, the data described in this paper were collected using an incomplete, anchored calibration design [ 7 , 9 , 14 , 15 ] with four sets of 80 items. Item sets A and B have half their items in common, as do item sets B and C, item sets C and D and item sets A and D . The items in common between two sets of items are known as 'anchors' and allow all items and patients to be calibrated on the same scale. The patterns of missing data in this type of design are, in statistical terms, ignorable [ 16 ]. The item sets were administered randomly to 150 respondents (item set A ), 143 respondents (item set B ), 138 respondents (item set C ) and 124 respondents (item set D ). Respondents A total of 555 residents of supported housing, residential care and nursing homes were interviewed. The median age was 84 years (range 37 to 101 years), while 444 (80%) were female. Since the respondents were interviewed 'at home', accurate data on medical conditions were not available. All respondents gave informed consent. The study was approved by the medical ethics committee in our hospital. The item response theory models In this paper the data were analysed using the two-parameter logistic item response theory model [ 7 , 9 , 17 , 18 ]. In this model, the probability, P ik , that patient k responds to item i in the category 'can' is modelled using where θ k denotes the ability of patient k to perform activities of daily life. The discrimination parameter ( α i ) and difficulty parameter ( β i ) describe the measurement characteristics of item i . The larger the value of β i , the more difficult item i is. In addition, the larger the value of α i , the better an item is a distinguishing between abilities above and below β i . If the values of α i are constrained to be equal for all items, the model in equation 1 becomes the one-parameter logistic item response theory model [ 19 ]. The model in equation 1 can be extended to test whether the values of β i for, say males and females, are significantly different. If the values of β i for different groups of respondents are significantly different, then there is evidence of differential item functioning. Full-information factor analysis also uses an extension of the model in equation 1. These approaches are described in mathematical terms in the Appendix. In this paper, estimates of α i and β i were obtained using a marginal maximum likelihood based procedure [ 20 ]. This method assumes that the ability parameters ( θ k ) follow a Normal distribution and can account for incomplete designs, as described in the Appendix. Expected a posteriori methods were used to estimate θ k [ 21 ]. Statistical analysis To achieve the objectives of this study, there were five steps in the statistical analysis. In step (a), the amount of variation in the response categories used for each item [ 9 ] was considered and items demonstrating too little variation were removed. Items were excluded from further analysis if fewer than 10% or more than 90% of the patients responded in the category 'cannot'. In step (b), the items were examined to investigate whether the value of the item difficulty parameter ( β i ) was similar for male and female patients and for patients younger than 85 years and those aged 85 or older. The model is described in depth in the Appendix. Items were excluded from further analysis if the value of the item difficulty parameter was significantly different (1% level) between gender or aged based groups. In this step, the fit of the model to the data from each item was not assessed. In step (c), estimates of the item parameters ( α i and β i ) were obtained. The fit of the model to the data from each item was assessed using G 2 statistics [ 22 ]. Items, for which the fit statistic had a p -value of less than 0.01, were excluded from the item bank. In addition, the stability of the estimates of the item parameters over different sets of items was examined using the model from step (b). Items were excluded from further analysis if the value of the item difficulty parameter was significantly different (1% level) between item sets A and B, B and C, C and D or A and D . Furthermore, a Kolmogornov-Smirnov test was carried out to examine whether the ability parameters ( θ k ) were Normally distributed. In step (d), the dimensionality of the item bank was examined using item response theory based full information factor analysis [ 18 , 22 , 23 ]. The number of latent roots greater than 1 is regarded as an indicator of the number of factors in the data set. This method is described in more depth in the Appendix. Four exploratory factor analyses were carried out, one on each of the anchors between item sets A and B (293 respondents), B and C (281 respondents), C and D (262 respondents) or A and D (274 respondents). A fifth, confirmatory, factor analysis was carried out on the whole data set (555 respondents). In addition, Cronbach's coefficient alpha was calculated for each anchor and the whole data set [ 24 , 25 ]. In step (e) the one-parameter logistic item response theory model was fitted to the remaining items. The differences between the -2log likelihoods of this model and the two-parameter model fitted in step (c) was tested using a χ 2 test. The analysis in steps (a), (b), (c) and (e) was carried out in Bilog, version 3.0 [ 22 ]. The analysis in step (d) was carried out using TESTFACT, version 4.0 [ 22 ]. Results Of the 160 items included in the item bank, one was removed because it was worded differently in two different item sets. Of the 159 remaining items, 77 were removed from the item bank. This process is described in Table 1 . In step (a), 28 items were excluded from further analysis because fewer than 10% or more than 90% of responses were in the category 'cannot'. In step (b), 26 items were removed because they had significantly different estimates of the item difficulty parameter ( β i ) for for males and females and/or for younger and older respondents. Of these 26 items, 19 had different measurement characteristics for females and for males, 5 items had different measurement characteristics for those aged under 85 and for those aged 85 or over, and 2 items had different measurement characteristics for both males and females and for older and younger respondents. In step (c), 23 items had an item fit statistic p -value of less than 0.01. In addition, 3 items were excluded from further analysis because the value of the item difficulty parameter ( β i ) was significantly different between two item sets of items. Hence, 79 psychometrically sound items remained in the item bank. A short description of the content of the 79 items in the final version of the calibrated item bank, together with estimates of the dispersion ( α ) and difficulty ( β ) parameters and their standard errors, are given in Tables 2a and 2b . Following step (c) of the analysis, the anchors between the sets of items contained between 13 and 23 items. In addition, there was no evidence to suggest that estimates of θ do not follow a Normal distribution (Kolmogorov-Smirnov test, p -value = 0.637). In step (d), the full information factor analysis indicated that, for three of the four anchors between the item sets, there was only one latent root of the correlation matrix larger than 1. In the fourth item set, a second latent root was marginally above 1. The percentage of the variance explained by the first factor varied between 67% and 72%. The values of Cronbach's alpha coefficient for the four anchors were between 0.86 and 0.93. The confirmatory factor analysis carried out on the whole data set indicated that 70% of the variance was explained by the first factor. Cronbach's alpha coefficient for the whole data set equalled 0.98. In step (e), the one-parameter logistic item response theory model was fitted to the 79 items remaining after step (c). This model fitted the data significantly less well than the two-parameter model ( p -value < 0.0001). For 3 items, the item fit statistic had p -value < 0.01. After removal of these items, the two-parameter model was still significantly better than the one-parameter model ( p -value < 0.0001). Table 1 The number of items proceeding to each step of the analysis The number of and examples of items removed at each stage of the psychometric analysis. Stage of analysis Number of items removed Reason for removal Examples 1 Concerns about the way the item was presented (a) 28 < 10% or > 90% of responses in 'cannot' Reaching for a cup and taking a sip of water Combing hair at a sink Cycling on a heavily laden bicycle (b) 26 Significant difference between M and F and/or under and over 85 years Washing up (easier for older respondents) Crossing the street (easier for younger respondents Preparing a warm meal (easier for female respondents) (c) 26 Item fit p -value < 0.01 or estimates of β i not stable Taking oral medication Cycling Getting money out of the bank using an ATM In item bank 79 See Table 2 Total 160 Table 2 The 79 items remaining in the calibrated item bank. The items remaining in the calibrated item bank. The number of respondents, to whom the item was offered (Offered to), the number responding in the category 'not applicable' (NA), the number responding in the category 'can' (can) and the number responding in the category 'cannot' (cannot) are given. The discrimination ( β ) and difficulty ( β ) parameters are given along with their standard errors in parentheses. Description of item content Offered to Item response category Location parameter ( β ) Discrimination parameter ( α ) NA can cannot Walking up stairs with a bag 262 0 19 243 -3.607 (2.404) 1.122 (0.892) Mopping a flight of stairs 262 5 16 241 -2.830 (1.708) 0.447 (0.411) Cleaning the top of a high cupboard 281 2 27 252 -2.816 (1.946) 0.480 (0.393) Cleaning a bathroom 293 1 37 255 -2.621 (2.061) 0.323 (0.338) Vacuuming 274 0 33 241 -2.408 (1.844) 0.287 (0.280) Going for a walk in the woods 281 0 31 250 -2.343 (1.636) 0.362 (0.340) Fetching groceries for 3–4 days 293 0 36 257 -2.262 (1.623) 0.353 (0.343) Mopping the floor 281 2 41 238 -2.225 (1.902) 0.339 (0.374) Caring for plants on a balcony 262 1 32 229 -2.108 (1.616) 0.314 (0.325) Travelling by bus or tram 281 0 44 237 -2.093 (1.835) 0.308 (0.370) Walking up two flights of stairs 274 0 39 235 -1.921 (1.532) 0.277 (0.298) Cleaning a fridge 293 2 64 227 -1.406 (1.464) 0.171 (0.236) Going to a restaurant 293 3 60 230 -1.335 (1.238) 0.159 (0.188) Carrying a tray 281 1 75 205 -1.304 (1.774) 0.222 (0.326) Going for a long walk (15+ minutes) 281 0 72 209 -1.290 (1.629) 0.178 (0.277) Going to the dentist 293 18 75 200 -1.283 (1.881) 0.205 (0.299) Sweeping the floor 262 1 70 191 -1.239 (1.902) 0.196 (0.313) Cutting toe nails 262 0 45 217 -1.175 (0.786) 0.136 (0.144) Walking up a hill or bridge 281 3 73 205 -1.133 (1.425) 0.137 (0.198) Walking up one flight of stairs 274 2 67 205 -1.127 (1.382) 0.155 (0.222) Going to a concert 262 0 57 205 -0.996 (0.860) 0.113 (0.128) Going to the pharmacist 262 2 75 185 -0.976 (1.572) 0.131 (0.201) Hanging a load of washing out 293 10 84 199 -0.960 (1.514) 0.144 (0.240) Going to the post office or bank 274 0 88 186 -0.948 (1.881) 0.151 (0.248) Going to a party 281 1 69 211 -0.924 (0.878) 0.109 (0.129) Filling an official form in 281 1 68 212 -0.896 (0.790) 0.106 (0.119) Using a washing machine 281 6 95 180 -0.851 (1.743) 0.153 (0.244) Visiting an outpatients' clinic 293 0 95 198 -0.815 (1.317) 0.112 (0.161) Taking bottles to the bottle bank 281 5 108 168 -0.675 (1.840) 0.153 (0.312) Short walk (less than 15 minutes) 274 0 95 179 -0.645 (1.358) 0.108 (0.166) Putting a rubbish bag outside 293 5 108 180 -0.573 (1.338) 0.116 (0.198) Reaching into a high cupboard 274 0 95 179 -0.569 (1.097) 0.099 (0.172) Using a dustpan and brush 262 2 103 157 -0.537 (1.865) 0.135 (0.335) Opening and closing a high window 281 0 140 141 -0.078 (1.290) 0.096 (0.166) Fetching groceries for one day 262 0 128 134 -0.043 (1.373) 0.099 (0.184) Using a public toilet 293 5 159 129 0.139 (1.835) 0.110 (0.241) Putting flowers in a vase 293 2 162 129 0.169 (1.787) 0.107 (0.240) Frying an egg 281 3 154 124 0.178 (1.982) 0.115 (0.261) Warming up a tin of soup 293 1 164 128 0.203 (1.919) 0.113 (0.232) Cleaning a toilet 262 0 149 113 0.308 (1.528) 0.105 (0.203) Putting socks and lace up shoes on 281 1 167 113 0.314 (1.165) 0.093 (0.157) Changing the bulb in a table light 281 1 177 103 0.533 (1.541) 0.116 (0.174) Cleaning a bathroom sink 281 5 170 106 0.564 (2.089) 0.126 (0.302) Cutting finger nails 262 0 168 94 0.605 (1.337) 0.114 (0.173) Rubbing lotion into whole body 262 4 164 94 0.627 (1.469) 0.115 (0.184) Reaching into a low cupboard 274 0 184 90 0.672 (1.131) 0.106 (0.152) Picking something up off the floor 262 0 172 90 0.712 (1.466) 0.129 (0.198) Making porridge 293 2 191 100 0.714 (1.704) 0.119 (0.216) Getting in and out of a car 281 3 185 93 0.738 (1.656) 0.132 (0.209) Shaking a tablecloth out 274 2 190 82 0.906 (1.438) 0.125 (0.204) Making a bed 281 0 193 88 1.003 (2.028) 0.152 (0.292) Preparing breakfast or lunch 262 1 186 75 1.117 (1.729) 0.169 (0.279) Using the lift in a public building 262 2 199 61 1.208 (1.299) 0.158 (0.186) Putting an alarm clock right 281 4 216 61 1.319 (1.431) 0.165 (0.199) Pulling a blanket up 293 0 253 40 1.485 (0.898) 0.167 (0.149) Visiting the neighbours 293 1 231 61 1.548 (1.685) 0.221 (0.252) Travelling as a passenger in a car 274 3 230 41 1.592 (1.126) 0.222 (0.192) Shaving face or applying make up 274 1 233 40 1.593 (1.075) 0.180 (0.164) Watering a house plant 262 3 204 55 1.600 (1.681) 0.226 (0.259) Opening and closing a window 262 0 201 61 1.735 (2.137) 0.246 (0.343) Putting trousers on 293 2 224 67 1.821 (2.372) 0.295 (0.406) Making coffee or tea 293 0 235 58 1.832 (1.936) 0.237 (0.290) Peeling an apple 281 1 233 47 1.859 (1.631) 0.226 (0.219) Making a bowl of cereal 281 1 225 55 1.860 (1.921) 0.222 (0.256) Eating a meal at the table 293 0 255 38 2.081 (1.509) 0.253 (0.225) Hanging clothes up in a cupboard 262 0 203 59 2.105 (2.595) 0.344 (0.481) Opening and closing curtains 262 0 216 46 2.129 (1.958) 0.366 (0.357) Moving between two dining chairs 281 0 237 44 2.214 (1.905) 0.389 (0.375) Putting a scarf and gloves on 293 1 259 33 2.364 (1.617) 0.306 (0.246) Making a cheese sandwich 281 1 243 37 2.416 (1.856) 0.392 (0.333) Moving to sit on the edge of a bed 262 1 231 30 2.457 (1.658) 0.309 (0.244) Putting a coat on 274 0 227 47 2.463 (2.323) 0.425 (0.395) Putting a shirt or blouse on 262 0 228 34 2.495 (1.842) 0.360 (0.287) Washing upper body at a sink 274 0 243 31 2.705 (1.875) 0.420 (0.327) Answering the front door 274 1 233 40 2.792 (2.373) 0.481 (0.449) Getting out of bed into a chair 262 0 232 30 3.019 (2.132) 0.581 (0.448) Washing lower body at sink 293 1 241 51 3.037 (3.098) 0.722 (0.761) Putting a T-shirt on 293 2 257 34 3.440 (2.664) 0.718 (0.630) Locking a door 262 0 230 32 3.366 (2.512) 0.970 (0.749) NA denotes that the category 'not applicable' was chosen Discussion In this study, the psychometric properties of the item bank have been examined using a sample of 555 respondents and an incomplete calibration design. Each item was presented to between 262 and 293 respondents. These figures are above the minimum, of 200 respondents, regarded as necessary to implement the models used in this paper [ 12 ]. It could be argued that it would have been desirable for all respondents to be presented with all items, but this would have placed an unacceptable burden on the, often frail, population in this study. Incomplete calibration designs are regularly implemented in the development and maintenance of item banks used in educational testing [ 4 , 14 ] and have gained some recognition in health related applications [ 15 ]. Developments in psychometric theory mean that it is now possible to perform the same types of analysis on data resulting from incomplete designs, as is performed on data from complete calibration designs [ 22 , 23 , 25 ]. The number of items in the anchors following the analysis, indicate that the design was still amply linked [ 9 ]. One of the major assumptions underlying the use of the item response theory models described in this paper is that the items reflect a single latent trait ( θ ). This has been examined using item response theory based full-information factor analysis. Part of the full-information factor analysis was performed on sub-sets of the data, as exploratory analyses on incomplete designs may lead to instable results. However, the confirmatory factor analysis was performed on all data. The results, together with the high level of internal consistency, as measured by Cronbach's alpha, and the acceptable fit of the two-parameter logistic item response theory model to the data indicate that the items presented in this paper probably represent a unidimensional construct in a population of respondents requiring residential care. Another important assumption when using item response theory models in conjunction with marginal maximum likelihood estimation procedures is that the values of the latent trait ( θ ) follow a pre-specified, usually Normal, distribution. In this study, there was no evidence that these values did not follow a Normal distribution. This is in contrast to many previously published studies into health and quality of life outcomes, where a strongly skewed distribution was found. The authors feel that there are two reasons for this contrast. Firstly, in this study, the respondents all had some level of restriction in their ability to perform activities of daily life. Secondly, the item bank includes items well above and well below the level of functional status enjoyed by the respondents. This means that the item bank did not have a ceiling or floor effect with respect this this population. In this study, 81 (51%) of 160 items were removed from the item bank because they did not conform to the psychometric standards required of the item bank. This is a much higher level than would be expected in the calibration of an item bank for use in educational measurement. However, when the results are examined more carefully, 28 items were removed because they were too difficult or too easy for the population in this study. In addition, 26 items were removed because they had different item parameters for different groups of respondents. These problems would have been identified much earlier in an educational item bank. Hence, only 26 (25%) of 106 items were removed due to item misfit. The number of items retained in the item bank may have been higher if a more flexible model, based on, for example, non-parametric smoothing techniques had been used [ 26 ]. However, this type of model is less suitable as a base for implementing modern testing algorithms, such as computerised adaptive testing. In addition, it is possible that more items could be made available if the items demonstrating differential item functioning were included in the item bank with different item location parameters ( β i ) for males and females or for younger and older respondents. This may seem complicated, but is straightforward in the framework of a computerised item bank. This paper has concentrated on the two-parameter logistic item response theory model. However, the one-parameter logistic item response theory model was also fitted to the 79 items remaining in the item bank. This model fitted the data significantly less well than the two-parameter model, even after 3 items demonstrating misfit at the item level were removed. This confirms the choice of the two-parameter model. This model was chosen because it allows the probability of responding in the category 'can' to be modelled more flexibly than when the one-parameter logistic model is used. This enables a more realistic model for the data to be built than when the more restrictive approach associated with the one-parameter model is chosen [ 18 ]. This paper has examined the calibration of the ALDS item bank in a population requiring residential care. It has been shown that the item bank has sound psychometric properties and could form a stable base for a wide range of applications. However, it is possible that the items will have different measurement characteristics for patients requiring treatment for specific chronic conditions or in other countries. Hence, it is important that the ALDS item bank is tested carefully before it is used to assess the functional status in other groups of respondents or in other countries. Conclusions Now that the measurement properties of the ALDS item bank have been examined carefully, the item bank can be used as a foundation for quantifying functional status. If modern algorithms, such as computerised adaptive testing, are implemented, it will be possible to obtain accurate measurements, whilst keeping the burden of testing on respondents and interviewers to a minimum. Items can be selected for use in further research, for allocation individuals to appropriate care settings and for calculating institutional funding based on the actual care load. It is hoped that the ALDS item bank will play an important part in the implementation of computerised adaptive testing of functional status. Appendix Differential item functioning The model in equation 1 can be extended to test whether different groups of respondents to have different values of β i . This is known as differential item functioning. For instance, if interest is in possible differences in β i between males and females, then the probability, P ik , that patient k responds to item i in the category 'can' is written as where β iM is the item difficulty for male respondents, β iF - M is the difference between the item difficulty for males and for females and I k is an indicator variable taking the value 0 if respondent k is male and the value 1 if respondent k is female. The hypothesis H 0 : β iF - M = 0 can be tested to examine whether item i has the same measurement characteristics for males and for females. Item parameter estimation in incomplete designs In this study, the item parameters ( α i and β i ) were estimated using marginal maximum likelihood methods. The likelihood, L , over n items and K ( K = 555) respondents can be written as where I ik is an indicator variable taking the value 1 if respondent k was offered item i and the value 0 otherwise and where J ik is an indicator variable taking the value 1 if respondent k responded to item i in the category 'can' and the value 0 otherwise. Furthermore, the probability, P ik , that respondent k responded to item i in the category 'can' is as in equation 1, or, where appropriate, as in equation 2 or 4. In the estimation process, the values of θ k or θ km were assumed to follow a Normal distribution with mean equal to 0 and unknown variance, σ 2 , and were integrated out of the likelihood to obtain the marginal likelihood. The marginal likelihood was maximised using an EM algorithm [ 20 ]. Full information factor analysis Full information factor analysis is a technique based on multidimensional item response theory models where the ability is represented by M variables, denoted θ km where m = 1, 2,..., M [ 22 , 23 ]. The model, in equation 1, for the probability, P ik , that person k responds to item i in the category 'can' can be extended to where θ km denotes the value of the latent variable θ m associated with person k and α im denotes the discrimination parameter for item i with respect to the latent variable θ m . Furthermore, δ i is a difficulty type parameter. The loading, a im of item i on factor m can be calculated using The value of the standard difficulty parameter, ( β i ), can be calculated using Generally, the parameters α im and δ i are estimated using marginal maximum likelihood methods. Abbreviations ALDS = AMC Linear Disability Score Competing interests None declared. Funding RH and RL were supported by a grant from the Anton Meelmeijer fonds, a charity supporting innovative research in the Academic Medical Center, Amsterdam, The Netherlands. Authors contributions RL conceived the study and supervised the data collection. RH prepared the first draft and carried out the analyses. RL, MV and RJdH critically reviewed the manuscript. RH prepared the final version.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514531.xml

554101

Central neuropeptide Y receptors are involved in 3rd ventricular ghrelin induced alteration of colonic transit time in conscious fed rats

Background Feeding related peptides have been shown to be additionally involved in the central autonomic control of gastrointestinal functions. Recent studies have shown that ghrelin, a stomach-derived orexigenic peptide, is involved in the autonomic regulation of GI function besides feeding behavior. Pharmacological evidence indicates that ghrelin effects on food intake are mediated by neuropeptide Y in the central nervous system. Methods In the present study we examine the role of ghrelin in the central autonomic control of GI motility using intracerobroventricular and IP microinjections in a freely moving conscious rat model. Further the hypothesis that a functional relationship between NPY and ghrelin within the CNS exists was addressed. Results ICV injections of ghrelin (0.03 nmol, 0.3 nmol and 3.0 nmol/5 μl and saline controls) decreased the colonic transit time up to 43%. IP injections of ghrelin (0.3 nmol – 3.0 nmol kg -1 BW and saline controls) decreased colonic transit time dose related. Central administration of the NPY 1 receptor antagonist, BIBP-3226, prior to centrally or peripherally administration of ghrelin antagonized the ghrelin induced stimulation of colonic transit. On the contrary ICV-pretreatment with the NPY 2 receptor antagonist, BIIE-0246, failed to modulate the ghrelin induced stimulation of colonic motility. Conclusion The results suggest that ghrelin acts in the central nervous system to modulate gastrointestinal motor function utilizing NPY 1 receptor dependent mechanisms.

Background The presence or absence of food in the gut stimulates the release of several regulatory peptides. These orexigenic (NPY, AGRP, ghrelin, MCH, Orexin-A, ...) and anorexigenic (CRF, CCK, CART, GLP-1, leptin, insulin, ...) peptides participating in the hypothalamic control of feeding behavior and satiety have been shown to be additionally involved in the autonomic control of gastointestinal (GI) functions like secretion and motility. For example fasted motor activity of the GI tract, e.g. the colon, is observed after intracerebroventricular (ICV) injection of neuropeptide Y whereas CRF ICV-treatment cause the disruption of fasted colonic motor activity [ 1 ]. Stomach-derived ghrelin is the first peripheral orexigenic peptide identified [ 2 - 6 ]. There is convincing evidence from several groups of investigators that ghrelin acts in the CNS and the periphery to simulate not only feeding but also GI function such as gastric acid secretion and gastric motility in rodents [ 7 , 9 - 11 ]. However, it is still unknown whether ghrelin is involved in the CNS control of other digestive functions besides gastric acid secretion and motility. Recent studies suggest that CNS-signaling by circulating ghrelin is mediated downstream by neurons of arcuate nucleus and the paraventricular nucleus of the hypothalamus, in particular, neurons expressing neuropeptide Y and agouti-related protein (AGRP) [ 12 - 14 ]. Furthermore it has been demonstrated that there is an anatomical interaction and functional relationship between ghrelin and neuropeptide Y. Using electrophysiological recordings Cowley et al have found that ghrelin stimulated the activity of arcuate NPYergic neurons and mimicked the effect of NPY in the paraventricular nucleus of the hypothalamus [ 15 ]. In addition ghrelin simulates food intake through hypothalamic NPY 1 receptors [ 1 , 16 , 17 ]. Thus, the question came up "are NPY receptors involved in the ghrelin effect on GI function"? Among others, neuropeptide Y plays a role in the CNS control of gastrointestinal function [ 1 , 18 ]. NPY activates at least six receptor subtypes, NPY 1 to NPY 6 . NPY binds preferentially with high affinity to Y 1 and Y 2 receptors, and there is evidence suggesting that these two receptor subtypes are involved in CNS regulation of digestive function by NPY action in arcuate nucleus and the paraventricular nucleus of the hypothalamus [ 18 ]. Taken together there is overwhelming evidence that ghrelin, beside its satiety modulatory capacity, is involved in the CNS control of digestive function of the upper gastrointestinal tract. In the CNS ghrelin and NPY, the most potent orexigenic neuropeptides known, are anatomical associated and functionally related. Moreover hypothalamic NPYergic neurons are downstream mediators of feeding related ghrelin action. In the present study we scrutinize the hypothesis that central neuropeptide Y receptor activation is involved in the ghrelin induced modulation of gastrointestinal motility using a microinjection-model in conscious fed and freely moving rats. Methods Animals All experimental components described were performed in accordance with the requirements of German legislation for the protection of animals and were licensed and supervised by the appropriate government body. Male Sprague-Dawley rats with a mean body weight of 350 ± 50 g were maintained on a 12 : 12 h photoperiod. They were housed in colony cages under conditions of controlled humidity and temperature (22 ± 2°C) for at least 7 days prior to the surgical procedure. The animals were fed a standard rat diet (Altromin ® , Lage, Germany) an tap water ad libitum . After surgical procedures, rats were housed individually. During experimental procedure the animals had continuous access to food and water. Drugs Ghrelin (Bachem, Heidelberg, Germany) doses of 0.03 nmol (100 ng), 0.3 nmol (1 μg) and 3 nmol (10 μg)/5 μl were dissolved in 0.15 M sterile saline (B. Braun, Melsungen, Germany). The NPY 1 receptor antagonist, BIBP-3226 (200 nmol/5 μl; Sigma-RBI, Natrick, MA, USA) [see Ref.: [ 19 ]] and the NPY 2 receptor antagonist, BIIE-0246 (120 nmol/5 μl; Boehringer-Ingelheim, Biberach, Germany) [see Ref.: [ 21 ]] were dissolved in sterile 0.15 M saline. The NPY receptor antagonists were used in similar equipotent nanomolar concentrations. The used intracerebroventricular concentrations of the receptor antagonists were comparable with the ICV-dosages used by other groups in rodents [ 18 , 20 ]. Probes were aliquoted and frozen (-80°C). Fresh aliquots were thawed on each experimental day before injections. Any excess was discarded. In our hands nanomolar concentrations of BIBP-3226 and BIIE-0246 were effective in antagonization of NPY receptor subtypes without any side effects. In particular no central depressive effects or conspicuous behavior was observed after BIBP-3226 treatment [ 18 ]. Cerebral cannulas For surgical procedures, rats were anesthetized with a mixture of ketamine (75 mg kg -1 i.p., Parke-Davis, Freiburg, Germany) and xylazine (5 mg kg -1 i.p., Bayer AG, Leverkusen, Germany). Animals were positioned in a stereotactic apparatus (David Kopf Instruments, Tujunga, CA). The head was fixed in a nose-down-position (-3 mm) and the skull exposed. Then trepanation of the skullcap was performed according to coordinates obtained from Paxinos and Watson [ 22 ] (mm from bregma: anterior-posterior = -3.30; lateral = ± 0.0; dorsoventral = -3.8). According to these coordinates a 22-gauge guide cannula (Bilarney / Plastic one, Düsseldorf, Germany) was implanted into the third ventricle. The cannula was anchored by dental cement and stainless steel screws affixed to the skull. Dummy cannulas (28 G), extending 2 mm beyond the guide cannula tips, were inserted to prevent blockage. After cerebral surgery, animals were individually housed. The animals were allowed 4 days recovery after guide cannula surgeries before the abdominal surgical procedures were performed. Colonic catheter This method was performed as described elsewhere [ 1 ]. Prior to all abdominal surgeries, the animals were food deprived overnight. Four days after cerebral surgery, rats were anaesthetized with a mixture of ketamine (75 mg kg -1 ) and xylazine (5 mg kg -1 ). After laparatomy a polyethylene microcatheter (inside diameter, 1.2 mm; outside diameter 1.7 mm; Becton Dickinson, New Jersy, USA) was chronically implanted into the proximal colon 1 cm distal from the caecocolonic junction. The catheter was fixed at the colonic wall by a purse-string suture and routed subcutaneously to the interscapular region, where it was exteriorized through the skin and secured. The animals were allowed 7 days recovery after abdominal surgeries before the beginning of habituation training sessions. Experiments were performed in fed, conscious rats. Intraperitoneal (IP) and intracerebroventricular (ICV) microinjection The doses of ghrelin were calculated according to the lowest effective doses to stimulate food intake [see Ref.: [ 23 ]]. For IP injection a 1 ml syringe (Hamilton, Reno, NV, USA) was used. Ghrelin and vehicle were injected IP after central administration of vehicle or NPY receptor antagonist. For IP injections the low dose of ghrelin administered peripherally was 0.3 nmol kg -1 /1 ml 0.15 M saline and the high dose of ghrelin was 3 nmol kg -1 /1 ml saline. NPY receptor antagonists were injected ICV 15 min before ghrelin was given peripherally at doses of 200 nmol/rat (BIBP-3226) and 120 nmol/rat (BIIE-0246) respectively. For ICV injections a 1 μl micro syringe (Hamilton, Reno, NV, USA), attached to a 32 G injection needle via a PE-50 tube-catheter was used. The stainless steel injection cannulas (32 G) were cut to protrude 2 mm beyond the tips of the guide cannulas. The conscious animals were gently restrained by hand, the injection needle was inserted through the guide cannula, and vehicle (5 μl 0.15 M saline) or NPY receptor antagonists (BIBP-3226 200 nmol/5 μl; BIIE-0246 120 nmol/5 μl) and ghrelin (0.03 nmol, 0.3 nmol or 3 nmol/5 μl), were consecutively injected ICV slowly over a 60 s period. We used a 15 min time interval between ICV injection of receptor antagonist or vehicle and ICV ghrelin administration. The injection needle was left in place for 2 min after injection to allow diffusion of the solution and to prevent back flow. Then dummy cannulas were reinserted into the guide cannulas. After the last experimental testing session, the rats were anesthetized and 5 μl of alcian blue 8GX were injected ICV. Colonic transit time measurement Colonic transit time was calculated by using an enteral dye-marker. Trypan blue, a non-absorbable dye, was injected in 0.2 ml volume through the catheter positioned in the proximal colon and followed by a 0.2 ml saline flush immediately after the ICV or IP microinjection. Colonic transit time was evaluated as the time interval between dye injection and the discharge of the first blue pellet. Faecal pellet output was monitored continuously by a self-developed, automated observation system that mechanically registers the time of all bowel movements for 24 h. The device consists of a conveyor belt placed under the mesh bottom cage, which transports faecal pellets with defined velocity to a collector. Brain histology The methods were performed as described in previous studies [ 17 ]. When experiments were completed, rats were anaesthetized with ketamine (75 mg kg -1 i.p.) and xylazine (5 mg kg -1 i.p.), and 0.05% alcian blue 8GX was microinjected intracerebroventricular under the same conditions as vehicle or peptide. The anaesthetized animals were transcardially perfused with phosphate buffered saline (PBS) buffer (0.1 M, pH 7.4) followed by Zamboni's fixative (2% formaldehyde and 2% picric acid in 0.1 M PBS buffer, pH 7.4). The brains were removed and cryoprotected in 25% sucrose. The site of injection was confirmed by inspection of intracerebroventricular dye distribution. Animals that received injections outside of the 3 rd ventricle were excluded from data analysis. Experimental design Experiment I: Effect on colonic motorfunction of peripheral (IP) and central (ICV) ghrelin administration The aim of the first experiment was to determine whether exogenous ghrelin would alter colonic motor function. Thus in experiment I, dose response effects of ghrelin in the cerebrospinal fluid (ICV) and the periphery (IP) on colonic transit time were examined. Ghrelin or saline as a vehicle was administered ICV or IP in conscious lightly restrained rats as previously described. For IP injection the low dose of ghrelin administered peripherally was 0.3 nmol kg -1 BW and the high dose of ghrelin was 3.0 nmol kg -1 BW. After injections, rats were subsequently returned to their home cages and maintained in a non-stressful environment to monitor colonic transit time. In order to minimize interindividual variation, and to reduce the number of animals needed to perform this study, animals were tested twice in this study. In randomized order, each rat received vehicle and a single dose of ghrelin or vehicle ICV or IP. The time interval between the experiments performed on the same animal was at least 4 days. Experiment II: Effect of BIBP-3226 and BIIE-0246 pretreatment on centrally and peripherally injected ghrelin induced modulation of colonic transit In experiment II the hypothesis that ghrelin acts at the CNS to modulate colonic motor function via a NPY receptor dependent pathway was addressed. Therefore, we determined if a pretreatment with selective NPY 1 - (BIBP-3226) and NPY 2 (BIIE-0246) receptor antagonists administered into the cerebrospinal fluid would block the alterations of colonic motor activity induced by centrally and peripherally administered ghrelin. The animals were pretreated with the NPY-Y 1 and -Y 2 receptor antagonists, injected ICV or vehicle (0.15 M saline), 15 min prior to ICV or IP ghrelin injections. Thereafter colonic transit time was assessed as described above. Data analysis The criterion used to include results in the data analysis of the ICV-injected group was the correct placement of the ICV cannulas. Results are expressed as mean ± SEM. The data of all studies were analyzed by ANOVA and differences between groups were evaluated by the Student-Newmann-Keuls test. P < 0.05 was considered significant. Results Effect of peripherally (IP) and centrally (ICV) administered ghrelin on colonic motor function In experiment I, dose response effect of peripherally and centrally administered ghrelin on colon transit time in fed and freely moving rats were examined. As demonstrated in Fig. 1 , ghrelin injected into the cerebrospinal fluid (CSF) stimulates colonic transit time dose dependently. In rats microinjected with vehicle into the CSF or IP, the average colonic transit time was 322 ± 8 min. As demonstrated in Fig. 1 , 0.03 nmol, 0.3 nmol and 3 nmol/5 μl ghrelin injected ICV dose-dependently decreased transit time by 24%, 34% and 43% respectively in conscious fed rats. Peripherally administered ghrelin accelerated transit time up to 36% (Fig. 1 ). Effect of ICV NPY receptor antagonist pretreatment on 3 rd ventricular and peripherally ghrelin induced stimulation of colonic transit The hypothesis that ghrelin acts in the brain to stimulate colonic transit via NPY receptor dependent mechanisms was addressed. As shown in Fig. 2 , pre-treatment with BIBP-3226 (200 nmol) which is a selective NPY 1 receptor antagonist 15 min prior to ICV application of 0.3 nmol ghrelin totally blocked the ghrelin induced effect on colonic transit. Application of BIBP-3226 into the CSF of the control group that was microinjected with vehicle, had no effect. Changes in colonic transit time induced by IP injection of ghrelin (3.0 nmol kg -1 BW) were canceled by ICV injection of NPY 1 receptor antagonist, BIBP-3226. (Fig.: 2 ). Pretreatment with the selective NPY 2 receptor antagonist, BIIE-0246, failed to affect the ghrelin induced alteration of colonic transit time. (Fig.: 2 ) Discussion The present experiments using a freely moving conscious rat model permit the measurement of colonic motility in rats in the physiological fed status. The results demonstrated that ghrelin given ICV and IP stimulates gastrointestinal motility indicated by shortened colonic transit time. In addition we found that the NPY type 1 receptor is primarily involved in the ghrelin induced modulation of fasted motor activity of the colon. There is convincing evidence that the most effective appetite stimulating peptides, NPY and ghrelin, act in the CNS and the periphery to simulate not only feeding but also GI function such as gastric acid secretion and gastric motility [ 1 , 8 - 10 ]. Stomach derived ghrelin, first described in 1999 by Kojima et al., is the first peripheral orexigenic peptide identified [ 4 ]. Ghrelin was identified as endogenous ligand for the GH secretagogue receptor (GHS R) and a peripheral metabolic signal informing the brain about stomach nutrient load [ 17 , 24 ]. Physiological studies suggest a functional relationship of ghrelin and NPY within the brain. It has been demonstrated that peripherally (i.v.) and central (ICV) administered ghrelin increases the expression of the immediate early gene c- fos , a marker of neuronal activity, in the arcuate nucleus and the PVN in awake fed rats [ 25 ]. Furthermore, exogenous ghrelin increases mRNA levels for NPY into the arcuate nucleus and simulates food intake through hypothalamic NPY 1 receptors [ 14 , 16 , 26 ]. Further Fujino et al. have demonstrated that ICV pretreatment with neuropeptide Y antiserum completely blocked ghrelin induced gastric and duodenal motoractivity [ 9 ]. Taken together these data suggest that there is an anatomical interaction and functional relationship between ghrelin and NPY within the brain. Six recognized subtypes of neuropeptide Y receptors have been described (NPY 1 to NPY 6 ). Two of these, NPY 1 - and NPY 2 receptors, are found in high density in the hypothalamus. There is compelling evidence that, in particular, NPY 1 and NPY 2 receptors are involved in the CNS regulation of gastrointestinal function. [ 1 , 8 , 27 ] For this reason, we focused on neuropeptide Y 1 and Y 2 receptor pathways in the present study and did not investigate the role of neuropeptide Y receptor subtypes Y 4 and Y 5 which are also expressed in the hypothalamus and are also involved in the autonomic control of feeding behavior and GI function. In the present study pretreatment with the NPY 1 receptor antagonist, BIBP-3226, blocked stimulation of colonic motility induced by systemic microinjection of exogenous ghrelin (ICV and IP). In our hands BIIE-2046, which is a selective antagonist of the NPY Y 2 receptor, failed to affect the ghrelin induced induction of fasted motor activity. It was previously described that knocking out NPY significantly decreases ghrelin stimulated feeding [ 17 , 28 ]. In this context Fujino et al. have recently demonstrated that the ghrelin induced fasted gastroduodenal motor activity in rats is blocked by ICV injection of GHS-R antagonist as well as NPY antiserum [ 9 ]. The results presented by Fujino et al. also suggest that the vagal pathway may mediate the action of centrally administered ghrelin on gastroduodenal motility [ 9 ]. Thus we can speculate that central NPY pathways, e.g. centrally NPY receptor activation, are the primary downstream mediator of circulating ghrelin. This interpretation is consistent with neuroanatomical and physiological facts: Neuropeptide Y works at two sites, locally within the arcuate nucleus to inhibit POMC neuronal activity and at afferent-terminal sites, in particular the paraventricular nucleus of the hypothalamus. Guan et al. have shown that neuropeptide Y- and ghrelin like immunoreactive (LI) neurons within the arcuate nucleus could influence each other by complex synaptic transmissions [ 29 ]. Furthermore Cowley et al. have demonstrated that ghrelin stimulated the activity of arcuate neuropeptide Y-LI neurons and mimicked the effect of neuropeptide Y in the PVN [ 15 ]. Compelling evidence showed that NPY projections from the arcuate nucleus (ARC) to the PVN are involved in the CNS regulation of food intake and other physiological functions of the organism, e.g. digestive function, by neuroendocrine and autonomic pathways [ 17 , 18 ]. For example NPY released from ARC neurons activates NPY-Y 1 receptors in the hypothalamus, e.g. the PVN, and results in the stimulation of GI motor function [ 18 ]. Furthermore arcuate NPYergic neurons have been thought to regulate feeding behavior by NPY receptor subtypes Y1 and Y5 in the PVN and adjacent areas [ 17 ]. Pretreatment with a Y1, but not other receptor antagonist markedly inhibited ghrelin-induced feeding, pointing to NPY receptor Y1 as one of the downstream pathways [ 9 , 17 ]. With regard to the characteristic physiological feature that peripheral ghrelin does not cross the blood-brain barrier in rodents it is important to note that the arcuate nucleus is the only hypothalamic structure located outside the brain-blood barrier [ 30 ]. Thus we can speculate that circulating ghrelin modulates gastrointestinal motility via activation of hypothalamic, in particular by using NPYergic pathways via activation of NPY-Y1 receptors, in the arcuate nucleus. This hypothesis is in good agreement with our observation that the effect of peripherally (IP) administered ghrelin on colonic motility is blocked by ICV pretreatment with the specific NPY 1 receptor antagonist, BIBP-3226. The NPY 2 receptor antagonist BIIE-0246 injected in the 3 rd ventricle at the equipotent dose as BIBP-3226 was not effective to antagonize the ghrelin effect on GI motility significantly. This data suggests that ghrelin unfolds a stimulatory effect on colonic motility primarily by acting on central NPY 1 and not via NPY 2 receptors. This interpretation is confirmed by the observation that Y 1 receptors acts rather postsynaptically and the Y 2 receptor rather presynaptically [ 31 , 32 ] The question of whether neuropeptide Y 4 or Y 5 receptors in the CNS are involved in the CNS control of gastrointestinal function should be examined in future studies. Conclusion We hypothesize that circulating ghrelin exhibits its effect by activating hypthalamic neurons, in particular neurons in the arcuate nucleus bearing GHS- and NPY-Y1 receptors. Further this ghrelin induced neuronal activation leads to stimulation of GI motor function by activation of higher hypothalamic brain sites, e.g. activation of neuronal projections within the paraventricular nucleus of the hypothalamus. On the other hand it is possible that the site of action of circulating ghrelin is not the hypothalamus but other brain sites. In our model using 3 rd ventricular injection of ghrelin this could simply mean that the peptide gained access to the 4 th ventricle and reached further caudal brain sites, e.g. NTS, DVC and medulla oblongata. With respect to the distribution of GHS- and NPY receptors in the CNS this hypothesis is possible, but how ghrelin action on any of these brain sites would modulate digestive function is not known. This question should be examined in future studies In summary, we presented evidence that ghrelin is involved in the CNS control of GI function. Apart from humoral pathways ghrelin acts into the CNS to control GI function by a mechanism of action involving neuropeptide Y1 receptor pathways. Further this study support the hypothesis giving by Chen et al. that ghrelin has an absolute requirement for neuropeptide Y pathways to unfold its physiological effects [ 28 ]. List of abbreviations AGRP agouti-related peptide ARC arcuate nucleus CART cocain- and amphetamine-regulated transcript CCK cholecystokinin CNS central nervous system CSF cerebrospinal fluid CRF corticotropin releasing factor DVC dorsal motor nucleus of vagus GHS-R growth hormone secretagogue receptor GLP-1 glucagon like peptide-1 ICV intracerebroventricular MCH melanocortin hormone MI microinjection NPY neuropeptide Y NTS nucleus of the solitary tract POMC proopiomelanocortin PVN paraventricular nucleus of the hypothalamus Competing interests The author(s) declare that they have no competing interests. Authors' contributions JJT participated in the design and coordination of the study, performed the microinjection studies and drafted the manuscript. CGT was the surgeon in charge and participated in the animal experiments. M-KHS and SM were involved in the design and coordination of the study. MR participated in the analysis and interpretation of data and revised the manuscript critically. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554101.xml

554115

Oral vitamin B12 therapy in the primary care setting: a qualitative and quantitative study of patient perspectives

Background Although oral replacement with high doses of vitamin B 12 is both effective and safe for the treatment of B 12 deficiency, little is known about patients' views concerning the acceptability and effectiveness of oral B 12 . We investigated patient perspectives on switching from injection to oral B 12 therapy. Methods This study involved a quantitative arm using questionnaires and a qualitative arm using semi-structured interviews, both to assess patient views on injection and oral therapy. Patients were also offered a six-month trial of oral B 12 therapy. One hundred and thirty-three patients who receive regular B 12 injections were included from three family practice units (two hospital-based academic clinics and one community health centre clinic) in Toronto. Results Seventy-three percent (63/86) of respondents were willing to try oral B 12 . In a multivariate analysis, patient factors associated with a "willingness to switch" to oral B 12 included being able to get to the clinic in less than 30 minutes (OR 9.3, 95% CI 2.2–40.0), and believing that frequent visits to the health care provider (OR 5.4, 95% CI 1.1–26.6) or the increased costs to the health care system (OR 16.7, 95% CI 1.5–184.2) were disadvantages of injection B 12 . Fifty-five patients attempted oral therapy and 52 patients returned the final questionnaire. Of those who tried oral therapy, 76% (39/51) were satisfied and 71% (39/55) wished to permanently switch. Factors associated with permanently switching to oral therapy included believing that the frequent visits to the health care provider (OR 35.4, 95% CI 2.9–432.7) and travel/parking costs (OR 8.7, 95% CI 1.2–65.3) were disadvantages of injection B 12 . Interview participants consistently cited convenience as an advantage of oral therapy. Conclusion Switching patients from injection to oral B 12 is both feasible and acceptable to patients. Oral B 12 supplementation is well received largely due to increased convenience. Clinicians should offer oral B 12 therapy to their patients who are currently receiving injections, and newly diagnosed B 12 -deficient patients who can tolerate and are compliant with oral medications should be offered oral supplementation.

Background Intramuscular injections of vitamin B 12 (cobalamin) have been the mainstay of B 12 deficiency treatment for decades. However, because approximately 1% of orally ingested B 12 is absorbed via simple diffusion throughout the gastrointestinal tract (i.e., independently of intrinsic factor)[ 1 , 2 ], oral replacement with high doses of cobalamin is both effective and safe, regardless of the etiology of B 12 deficiency [ 1 , 3 - 12 ]. Oral B 12 therapy would decrease physician burden, increase patient control over therapy, and avoid patient discomfort and inconvenience. Switching patients from B 12 injections to oral therapy would also result in savings to the health care system [ 13 ]. While some commentators have argued that clinicians should switch to oral B 12 therapy [ 13 - 15 ], little is known about patients' views concerning the acceptability and effectiveness of oral B 12 . Efforts to switch patients who are well established on parenteral therapy (and who have previously been told they require lifelong injections) may fail without an understanding of what factors influence patient acceptability of oral therapy. Therefore, we combined qualitative and quantitative methods to test the hypothesis that patients offered oral B 12 are willing to switch to oral therapy and to explore the reasons for their choice. Methods We administered a questionnaire to patients with B 12 deficiency, offered the option of a six-month trial of oral replacement, and administered a follow-up questionnaire. Throughout the study, semi-structured interviews were conducted for the qualitative arm. We received ethics approval from the Sunnybrook and Women's College Health Sciences Centre Research Ethics Board. All patients provided informed consent. Quantitative arm Data collection The study took place at two academic family practice units and a community health centre with diverse practice profiles in Toronto. We included all patients who received regular injections (i.e., every 1 to 3 months), regardless of age and etiology of B 12 deficiency. We excluded patients if their last injection occurred more than three months before the start of patient recruitment, if they had left the practice, if their B 12 therapy had been discontinued or if they had already switched to oral therapy, if they did not speak English and did not have access to a translator, or if they had been recently diagnosed with a serious illness. After applying these criteria, 133 patients were included in the study (Figure 1 ). Figure 1 Flow diagram of patients in study. The recruitment cover letter included one sentence describing the equivalency of oral B 12 therapy to injections: "Studies have shown that vitamin B 12 pills are just as effective and safe as B 12 injections." Educational sessions on the effectiveness of oral therapy were held with the nursing staff (since they administer the injections) at all sites, and they were encouraged to pass this information on to their patients. As well, contact information for the study investigators was provided to the patients in case they had any questions. We placed the initial questionnaire [see Additional file 1 ] in patients' charts to be completed at their next visit for a B 12 injection. The initial questionnaire elicited demographic data, medication history, logistical aspects of B 12 related visits, history of past B 12 therapy, and attitudes about injection and oral therapy. Non-responders received up to two telephone reminders. Patients willing to try oral therapy were given a six-month supply of B 12 tablets (one 1000 μg tablet daily), and were offered testing of serum B 12 levels at baseline (one month after their last injection) and after the six-month trial. Patients had the option to withdraw from the trial and return to injections at any time for any reason. A follow-up questionnaire was placed in the charts of patients who tried the pills. This questionnaire re-assessed patients' attitudes about the different forms of B 12 therapy and asked whether they would continue on oral supplementation or return to injection therapy (see Additional file 2 ). Statistical analyses Where variables were not already dichotomous (e.g., for satisfaction with injections, the questionnaire listed "very satisfied," "satisfied," "neutral," "unsatisfied," and "very unsatisfied" as choices), we dichotomized the variable of interest (e.g., "satisfied" vs. "neutral/unsatisfied"). To ascertain the relationship between questionnaire responses and preference for oral therapy (as "willingness to switch" and "permanently switching") we performed bivariate and multivariate analyses. We used a p-value of 0.20 (for the continuity-adjusted χ 2 and Fisher's exact tests) as the cut-off for inclusion of individual patient factors into multivariate logistic regression models. We then performed backwards stepwise regression to determine statistically significant relationships after adjustment. We generated both crude and adjusted odds ratios with 95% confidence intervals. Qualitative arm Participants and setting One investigator (D.C.) conducted 17 semi-structured interviews in a private meeting room at the family practice unit. A purposive sample was selected amongst those willing to be interviewed to reflect diversity in terms of sex, age, willingness to switch, and final choice of therapy after the trial of B 12 pills. These interviews were performed at various stages of the study (before, during, and after the trial of oral therapy). Amongst the 17 participants, four were selected from those not willing to switch, while of the 13 who were willing to try oral therapy, six were interviewed prior to, two during, and five after the trial of oral therapy. Data collection and analysis The interviews were audiotaped and transcribed. They lasted 20–30 minutes, with questions about the patient's knowledge and history of their B 12 therapy, his or her relationship with health care providers, perceived advantages and disadvantages of both pills and injections, and attitudes about past and current B 12 therapy. Analyses of the transcripts were performed independently by two investigators (J.K. and D.T-K.), using a three-step content analysis approach to identify and collate relevant themes [ 16 ]. Through a process of clarification, confrontation, and consensus, we reached agreement regarding the themes and sub-themes. Results Quantitative arm For our initial questionnaire, we received responses from 86 out of the sample of 133 patients, for a response rate of 64.7%. Selected characteristics of the study population are presented in Table 1 . Non-responders were younger and more likely to be female. Table 1 Selected characteristics of study population and a comparison of responders vs. non-responders Responders n = 86 Non-responders n = 47 p-value Mean age – Yr ± SD 72 ± 15 65 ± 20 0.03 Female sex – No. (%) 48 (56) 37 (79) 0.009 Level of education completed High school or less 28 (33) Some post-secondary education 21 (25) Bachelors degree or more 35 (42) Annual household income Less than $40K 40 (53) $40–79K 20 (26) $80K or greater 16 (21) Self-reported perception of health Above average 25 (29) Average 40 (47) Below average 20 (24) Prescription medications taken 0 8 (10) 1 to 3 46 (55) 4 to 6 16 (19) 7 + 14 (17) Monthly episodes of forgetting medications 0 45 (54) 1–2 26 (31) 3+ 12 (14) Years on B12 therapy 0 to 2 19 (24) 3 to 5 25 (31) 6 to 10 17 (21) 11 to 19 13 (16) 20+ 6 (8) Frequency of B12 injections Less than once monthly 9 (10) Once monthly 74 (86) More than once monthly 3 (3) Satisfaction with B12 injections Satisfied 22 (26) Neutral 59 (69) Unsatisfied 5 (6) Monthly visits to doctor for other reasons 0 49 (60) 1 23 (28) 2 + 10 (12) Mode of travel to visit doctor Personal vehicle 49 (58) Public transit 23 (27) Walk 9 (11) Taxi 4 (5) Travel time to visit doctor < 15 min 27 (32) 15–29 min 29 (35) 30–44 min 20 (24) 45–59 min 5 (6) 60+ min 3 (3) Patients at each study site Sunnybrook Campus, SWCHSC* 51 (59) Flemingdon Health Centre 19 (22) Women's College Campus, SWCHSC* 16 (19) * Sunnybrook and Women's College Health Sciences Centre Willingness to switch to oral B 12 Sixty-three of our 86 respondents reported a willingness to switch to oral therapy. A large number of patient factors had no clear statistical relationship to a "willingness to switch" to oral therapy (i.e., p > 0.20) and were excluded from the multivariate model: age, education, income, drug insurance coverage, study site, number of prescription medications, mode of travel to the clinic, years of past B 12 therapy, number of non-B 12 related visits per month, improvement in perceived well-being since initiating B 12 injections, several perceived disadvantages of injections (risk of complications, travel/parking costs), and several perceived disadvantages of pills (take too many pills already, would have to pay for them, won't work as well as injections). The following factors were included in the multivariate model: gender, time to clinic, satisfaction with past B 12 injections, several perceived disadvantages of injections (shots are painful, frequent visits to see MD/nurse, cost to health care system), and one perceived disadvantage of pills (won't see MD/nurse as often). Factors associated with willingness to switch after multivariate adjustment were: being able to get to the clinic more quickly (i.e., in less than 30 minutes) (OR 9.29, 95% CI 2.16–39.97), and believing that injection therapy is disadvantageous due to the need for frequent visits to health care provider (OR 5.41, 95% CI 1.10–26.56) and the increased costs to the health care system (OR 16.68, 95% CI 1.51–184.22) (Table 2 ). Table 2 Patient factors associated with willingness to switch to oral therapy on initial questionnaire. No. (%) of subjects Unadjusted Adjusted* Patient factor Willing to switch n = 63 Not willing to switch n = 23 OR (95% CI) p-value OR (95% CI) p-value Time to clinic 0–29 minutes 49 (87) 7 (13) 9.34 <0.001 9.29 0.003 30+ minutes 12 (43) 16 (57) (3.14–27.78) (2.16–39.97) Perceived disadvantages of injections Frequent visits to see MD/nurse Agree 38 (86) 6 (14) 4.31 0.010 5.41 0.038 Disagree 25 (60) 17 (40) (1.49–12.42) (1.10–26.56) Cost to health care system Agree 26 (93) 2 (7) 7.38 0.010 16.68 0.023 Disagree 37 (64) 21 (36) (1.59–34.23) (1.51–184.22) * Adjusted for patient factors at least weakly associated (i.e., p < 0.20) with a willingness to switch to oral B 12 therapy in bivariate analyses: gender, time to clinic, satisfaction with past B 12 injections, perceived disadvantages of injections (shots are painful, frequent visits to see MD/nurse, cost to health care system), perceived disadvantage of pills (won't see MD/nurse as often) Trial of oral therapy Of the 63 patients who were willing to switch to oral B 12 therapy, eight changed their minds before the trial started. Therefore, 55 patients were started on oral B 12 . Five dropped out and three were lost to follow-up, leaving 47 patients who completed the six-month trial. Reasons cited for discontinuing oral therapy included fatigue, neurological symptoms, and gastrointestinal intolerance. Fifty-two patients returned the follow-up questionnaire. Over three-quarters (39 of 51 respondents) reported being satisfied or very satisfied with oral therapy. Only 8% of patients (4/51) perceived that they felt worse with pills, while 23.5% (12/51) felt better and the rest felt the same. Self-reported compliance was good; 48 patients or 92% reported forgetting to take the pills two times or less per month. Thirty-nine patients (71% of the 55 who actually switched) stated that they wished to permanently switch to B 12 pills. Of the 35 patients who reported feeling the same with the pills as with injections, 28 (80%) chose pills. Again, many patient factors had no association with a desire to permanently switch to oral B 12. Patient factors that on bivariate analyses had a p-value of less than 0.20 and were consequently included in the multivariate model were: patient beliefs that frequent visits to see the doctor or nurse and the associated travel/parking costs are disadvantages of injections, and the belief oral B 12 would add unnecessarily to an already large number of prescribed oral medications. Table 3 presents the factors significantly associated, after multivariate adjustment, with permanently switching to oral therapy: agreeing that disadvantages of injections included frequent visits to see the doctor/nurse (OR 35.41, 95% CI 2.90–432.70) and travel/parking costs (OR 8.66, 95% CI 1.15–65.30). Table 3 Patient factors associated with permanently switching to oral therapy on follow-up questionnaire No. (%) of subjects Unadjusted Adjusted* Patient factor Choosing oral therapy n = 39 Choosing injection therapy n = 13 OR (95% CI) p-value OR (95% CI) p-value Perceived disadvantages of injections Frequent visits to see MD/nurse Agree 27 (96) 1 (4) 27.00 <0.001 35.41 0.005 Disagree 12 (50) 12 (50) (3.14–231.87) (2.90–432.70) Travel/parking costs Agree 23 (92) 2 (8) 7.91 0.016 8.66 0.036 Disagree 16 (59) 11 (41) (1.54–40.60) (1.15–65.30) * Adjusted for patient factors at least weakly associated (i.e., p < 0.20) with permanently switching to oral B 12 therapy in bivariate analyses: perceived disadvantages of injections (frequent visits to see MD/nurse, travel/parking costs), perceived disadvantage of pills (take too many pills already) Serum B 12 levels Baseline and post-intervention serum B 12 levels were obtained for 39 of the 55 patients who switched to pills (Figure 2 ). The mean serum level increased from 387 to 698 pmol/L (p < 0.0001) (reference range: ≥ 180 pmol/L, unlikely to have B 12 deficiency). Only one patient's serum level decreased but even she remained within the normal range. Figure 2 Serum B12 concentrations before and after six months of oral B12 therapy for forty patients. Qualitative arm The mean age of interview participants was 71 (range 54–92) years, with 7 men and 10 women. Perceived efficacy of B 12 therapy Some patients felt that the B 12 injections they had been receiving in the past were effective: "I find it gives me more energy." Patients who expressed this view were often not willing to switch to oral therapy, or if they did try the pills, they eventually went back to injections. Others felt that the injections hadn't helped much: "There is a little feeling that, 'Oh, I'm taking this. I should feel better,' but I'm not sure if I do." These patients tended to switch permanently to oral therapy. When asked about the anticipated efficacy of oral therapy, some patients, particularly those who were not willing to switch, were sceptical: "Well, I don't think it will work. If the doctors suggested that I needed to go on the injections, I'm certainly pretty sure that the pills are not going to work. Otherwise he'd have put me on pills, would he not?" Many others were unsure about the efficacy of oral therapy. Reasons for switching to oral therapy Nearly every patient cited convenience as an advantage of switching to B 12 pill. Other advantages of oral therapy included savings to the health care system and ease of travel. Cited disadvantages of injections included decreased compliance due to the need for frequent visits and potential complications associated with injections. A number of patients decided to switch because they were interested in participating in a research study. Some patients acknowledged that oral therapy would benefit those who are averse to needles, however, needles did not bother most of the patients interviewed: "...needles don't bother me, because – well, I don't know. They just don't seem to bother me. I know some people are very concerned about it." Satisfaction with oral therapy For both patients interviewed during their trial of oral B 12 and for three of the five patients interviewed afterward, all of whom permanently switched to oral therapy, there was a high level of satisfaction with oral B 12 . One patient even reported feeling better on oral B 12 : "Well, I'm more level. I don't feel at the end of the month that I'm running out of energy. I'm quite well aware of that." (see Additional file 3 ). Reasons for staying with/switching back to injection therapy Disadvantages of switching to oral therapy included having to take an additional oral medication, fear of side effects, concern about swallowing difficulties, the inconvenience of taking a medication daily, and the potential for losing contact with health care providers and the opportunity for minor drop-in consultations. Some patients were under the impression that injections would work more quickly and directly. Two patients who switched back to injections after trying the pills were interviewed. These patients switched back due to side effects – one cited neurological and gastrointestinal symptoms (her post-trial serum B 12 level was significantly increased from baseline) and the other cited decreased energy (his post-trial serum B 12 level was not available). (see Additional file 4 ). Discussion We found that switching patients from injection to oral B 12 is both feasible and acceptable to patients. Of those who responded to our initial questionnaire, nearly three-quarters were willing to try oral B 12 , and of those who did switch, most were satisfied and the majority wished to remain permanently on oral therapy. These patients believed that injections are disadvantageous because they are associated with too many visits to their health care provider and with higher patient costs (in terms of travel and parking expenses). In the qualitative arm of our study, most patients cited convenience as the reason for wanting to switch from injections to pills. We were also able to confirm findings from other studies indicating that oral B 12 is biochemically equivalent to, if not better than, parenteral therapy. One non-intuitive finding from the initial questionnaire assessing the factors associated with trying oral therapy (Table 2 ) was that those who were able to get to the clinic more quickly (i.e., in fewer than 30 minutes) were more likely to try switching. While one might expect that those who live closer to their clinic would benefit less from the convenience of not having to visit their health care provider as frequently, we speculate that perhaps these patients are more comfortable with trying oral therapy because they know that it is easy for them to access care if they need it, rather than having to rely on regular visits for their injections to see their primary care provider. Another interesting finding – that the perception of travel expenses being a disadvantage of the injections was not associated with being willing to try injections but was associated with permanently switching after the trial – suggests that perhaps patients did not realize the benefit of the saved travel costs until several visits had been averted. By monitoring trough serum B 12 levels (i.e., 1 month after their last injection) and again 6 months after oral therapy, we observed that one patient was a non-responder (i.e., serum level decreased), and 4 patients had serum levels that increased but remained below 295 pmol/L (a level considered by some to be borderline B 12 deficient)[ 17 ]. This represents 13% out of the 39 patients who underwent testing, suggesting that follow-up testing may be warranted with serum B 12 levels and, if available, functional tests such as serum homocysteine and methlymalonic acid. The sub-optimal response in these patients may be due to patient non-compliance or sub-optimal dosing (we used a 1-mg daily dose, rather than the 2-mg daily dose used by Kuzminski et al .) [ 8 ]. Further studies may be required to clarify the optimal dosing regimen for oral therapy. The primary strength of our study was that we achieved triangulation through the use of both quantitative and qualitative methods. By using questionnaires, we were able to provide estimates such as switch rates and levels of patient satisfaction; with the qualitative arm, we were able to more thoroughly explore patient perspectives. The main limitation of our study was the relatively small sample size for the quantitative arm, which led to large confidence intervals for the estimates of the relationship between patient factors and switching to oral B 12 . However, the sample was large enough to identify a number of patient factors associated with switching from injection to oral therapy and to determine that patient attitudes toward oral B 12 therapy are greatly influenced by its convenience. A second limitation was that we did not assess the knowledge and attitudes of our patients' physicians and nurses, as they may have influenced patient's expectations and perceptions regarding the effective of oral B 12 therapy. Finally, another limitation was the potential participant bias, evidenced by the differences between responders and non-responders in terms of age and sex. However, among the responders, neither age nor sex was significantly associated with switching to oral therapy, suggesting that any differences between participants and non-participants were likely irrelevant. While there have been several studies examining physicians' perspectives on switching from intramuscular to oral therapy [ 18 - 21 ], we know of few studies that have examined patients' views. In a recent study that involved switching forty patients to oral therapy, the authors reported that 83% preferred the oral form [ 15 ]. However, no further details on patient perspectives were reported. Another study found that patients become very attached to receiving their injections; after identifying 48 patients who did not meet diagnostic criteria for B 12 deficiency and providing educational sessions to encourage discontinuing injections, 38% stated they would leave the practice if denied their injections [ 22 ]. Explanations suggested by the authors include the reluctance to discontinue a therapy initiated by a trusted and respected physician, and the belief that intramuscular injections are more potent and more effective than medications taken orally. While the perception of increased potency of injections was reported by some patients in the qualitative arm of our study, this belief was supported by only 20% of respondents in the initial questionnaire (data not shown). While oral B 12 therapy may be acceptable for most patients, it may not be appropriate for those who will not be compliant with oral medications, such as patients with significant memory impairment or cognitive dysfunction, unless they have caregivers who could ensure compliance. Oral therapy also may not be appropriate for those with swallowing difficulties; for such patients, sublingual B 12 therapy, which has been shown to be equally effective, may more appropriate [ 23 ]. Conclusion In summary, the results of our study suggest that clinicians should offer oral B 12 therapy to their patients who are currently receiving injections if they can tolerate and are compliant with oral medications. Most patients who switch from parenteral to oral therapy are satisfied and wish to stay permanently on oral therapy. They often cite the inconvenience and the travel-related expenses associated with frequent visits to the doctor as disadvantages of parenteral therapy. Therefore, oral B 12 therapy is generally well-received by patients, and should be considered by clinicians as a superior alternative to the traditional injections for most patients. Competing interests The author(s) declare that they have no competing interests. Authors' contributions JK designed and coordinated the study, analyzed the qualitative and quantitative data, and drafted the manuscript. DC conducted the in-depth interviews and assisted in data collection. ID helped to develop the questionnaire and performed statistical analyses of the quantitative data. DTK analyzed the qualitative data. RU conceived the study and provided guidance on all aspects of the project. All authors participated in the development of the manuscript and gave approval to its final submission for publication. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 Initial questionnaire Click here for file Additional File 2 Follow-up questionnaire Click here for file Additional File 3 Reasons for switching to oral therapy Click here for file Additional File 4 Reasons for not switching to oral therapy Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554115.xml

544851

An empirical analysis of training protocols for probabilistic gene finders

Background Generalized hidden Markov models (GHMMs) appear to be approaching acceptance as a de facto standard for state-of-the-art ab initio gene finding, as evidenced by the recent proliferation of GHMM implementations. While prevailing methods for modeling and parsing genes using GHMMs have been described in the literature, little attention has been paid as of yet to their proper training. The few hints available in the literature together with anecdotal observations suggest that most practitioners perform maximum likelihood parameter estimation only at the local submodel level, and then attend to the optimization of global parameter structure using some form of ad hoc manual tuning of individual parameters. Results We decided to investigate the utility of applying a more systematic optimization approach to the tuning of global parameter structure by implementing a global discriminative training procedure for our GHMM-based gene finder. Our results show that significant improvement in prediction accuracy can be achieved by this method. Conclusions We conclude that training of GHMM-based gene finders is best performed using some form of discriminative training rather than simple maximum likelihood estimation at the submodel level, and that generalized gradient ascent methods are suitable for this task. We also conclude that partitioning of training data for the twin purposes of maximum likelihood initialization and gradient ascent optimization appears to be unnecessary, but that strict segregation of test data must be enforced during final gene finder evaluation to avoid artificially inflated accuracy measurements.

Background The number of generalized hidden Markov model (GHMM) gene finders reported in the literature has increased fairly dramatically of late [ 1 - 8 ], and the community is now contemplating various ways to extend this attractive framework in order to incorporate homology information, with a handful of such systems having already been built (e.g., [ 9 - 12 ]). GHMMs offer a number of clear advantages which would seem to explain this growth in popularity. Chief among these is the fact that the GHMM framework, being (in theory) purely probabilistic, allows for principled approaches to constructing, utilizing, and extending models for accurate prediction of gene structures. While the decoding problem for GHMM gene finders is arguably well understood, being a relatively straightforward extension of the same problem for traditional HMMs and amenable to a Viterbi-like solution (albeit a more complex one), methods for optimally training a GHMM gene finder have received scant attention in the gene-finding literature to date. What information is available (e.g., [ 2 , 4 ]) seems to indicate that the common practice is to optimize the submodels of the GHMM independently, without regard for the optimality of the composite model. The training of HMMs and GHMMs has traditionally been carried out using some form of maximum likelihood estimation (MLE). Baum-Welch training [ 13 ], which is an instance of the well-known expectation maximization (EM) procedure, is itself a form of MLE [ 14 ]. In the case of GHMM gene finders, one typically applies some form of MLE to each of the submodels (states) in the GHMM so as to render training features of each type (e.g., exon, intron, donor site) maximally likely under the induced (sub)model; i.e., maximizing: for state q and for S i a feature of length d i from the state- q -specific training set T . The submodels are then merged into a composite model (i.e., the full GHMM) by observing transition probabilities between features in the training data corresponding to each of the GHMM states. For example, an exon state in a GHMM can be trained by collecting n -gram statistics (i.e., counts of n -letter substrings) from known exon sequences and normalizing these into transition probabilities for an ( n -1) th -order Markov chain [ 15 ]. Similarly, intron, intergenic, and untranslated region (UTR) states can be modeled by collecting appropriate statistics from corresponding sample features and using these to train individual content-scoring models, such as Markov chains, neural networks, decision trees, etc. Signal sensors for donor and acceptor splice sites and start and stop codons can be trained by aligning known signals of the appropriate type and counting nucleotide frequencies at each position within a fixed window around the signal; converting these counts to relative frequencies produces probability estimates for use in a weight matrix or similar type of model. Transition and duration probabilities can likewise be estimated by observing appropriate frequencies in training data. All of these estimation activities can be performed independently, resulting in a GHMM consisting of distinct subsets of maximum likelihood parameters. Such an approach does not, however, attend to the global optimality of the GHMM as a whole. Ideally, one would like to maximize the expected accuracy of the gene finder on unseen data. A reasonable approximation to this ideal would be to maximize the average probability of the gene parses in the training set: where the collection of model parameters making up the GHMM is denoted θ and the elements ( S , φ) of the training set T comprise pairs of sequences S and their known parses φ. This argmax gives us the parameterization under which the full gene parses (rather than the sequences ) in the training set will be maximally likely (on average). Decomposing each parse φ into a series of ( q i , d i ) pairs, for state q i and state duration (i.e., feature length) d i , we get: where P e , P t , and P d represent the emission, transition, and duration probabilities of the GHMM, respectively. Whereas the common MLE training procedure for GHMMs as described above optimizes the individual terms in the numerator of Equation 3 independently, the argmax above calls instead for these terms to be jointly tuned so as to optimize the entire ratio in parentheses. Intuitively, one can think of this alternate formulation as attempting to account for the process in the Viterbi algorithm (during later decoding) whereby the individual submodels "compete" for nucleotides (in the sense that each nucleotide can be emitted by only one submodel in any given parse, and the Viterbi algorithm chooses the final, predicted parse based on the values of the model parameters). Our hope is that by addressing the issue of submodel competition explicitly during parameter estimation, we will thereby empower the gene finder to better discriminate at a global sequence level between the features modeled by individual submodels in the GHMM, thereby producing more accurate gene predictions. A similar optimization problem occurs in the field of speech recognition, in which systems of interacting acoustic models and language models are employed to optimally parse an audio stream into a series of discrete words. Interestingly, the trend in that field, starting with Bahl et al . in 1986 [ 16 ], has increasingly been away from the sole use of MLE and toward an alternative approach very similar to that prescribed by Equation 2 known as global discriminative training [ 17 - 19 ] or conditional maximum likelihood [ 20 ]. The problem also appears in a slightly different form in the related field of statistical natural language parsing, in which it has been suggested that global methods for optimizing competing stochastic grammar models may improve the accuracy of systems at the level of whole-sentence parses [ 21 ]. Maximum discrimination HMMs have already been applied successfully to problems in the realm of biological sequence analysis [ 22 ], though their use in gene finding has apparently not yet seen widespread adoption. To our knowledge, the only gene finder reported to use discriminative training is HMMgene [ 23 ], a gene finder based on a non-generalized HMM. In light of these considerations, it is worth contemplating the possible gains in gene finder accuracy that might be obtained through the use of some form of discriminative training applied to a GHMM – that is, training aimed more directly at optimizing the ability of the gene finder to discriminate between exons and non-exons, thereby improving the expected accuracy of the gene finder's predictions. Anecdotal evidence already suggests that investigation of such methods may indeed be fruitful, as the process of manual tuning of GHMM parameters (i.e., "tweaking") after MLE training is commonly acknowledged by those with experience training GHMM-based gene finders (including our own systems). The practice of performing such tuning on the training set, especially when done iteratively, can be viewed as a manual form of gradient ascent optimization using the percentages of correctly predicted nucleotides, exons, and whole genes as surrogates for the Σ (S,φ)∈T P(φ|S,θ) term in Equation 2. We therefore decided to investigate the use of a simple form of global discriminative training for gene-finding. We did this by building a rudimentary gradient ascent optimizer and applying it to a subset of the model parameters for our GHMM-based gene finder, TigrScan, as described in the Methods. Results Maximum likelihood versus discriminative training Results for Arabidopsis thaliana are shown in Table 1 and those for Aspergillus fumigatus are shown in Table 2 . The two methods being compared are maximum likelihood estimation (MLE) versus maximum likelihood followed by gradient ascent parameter estimation (GRAPE). Table 1 Results on Arabidopsis thaliana method train test nucAcc exonF geneSn GRAPE CV CV 95 ± 1% 82 ± 2% 49 ± 3% GRAPE CV H 93 ± 1% 80 ± 2% 44 ± 3% GRAPE T T 95% 86% 57% GRAPE T H 94% 81% 48% MLE CV CV 90 ± 1% 72 ± 2% 33 ± 4% MLE T T 91% 75% 36% MLE T H 90% 71% 33% GRAPE = GRadient Ascent Parameter Estimation, MLE = Maximum Likelihood Estimation only. CV=cross validation, T = training set, H = 1000-gene hold-out ("test") set. CV in the train column means training on 800 genes from T. CV in test column means testing on 200 genes from T. In rows with a CV in either column, numbers are averages from 5 runs. nucAcc = nucleotide accuracy, exonF = exon F score, geneSn = gene sensitivity. F = 2SnSp/(Sn+Sp) for Sn = sensitivity and Sp = specificity. CV averages are reported ± SD. Table 2 Results on Aspergillus fumigatus method train test nucAcc exonF geneSn GRAPE CV CV 88 ± 1% 54 ± 4% 35 ± 4% GRAPE CV H 88 ± 1% 51 ± 2% 29 ± 1% GRAPE T T 92% 65% 48% GRAPE T H 87% 51% 31% MLE CV CV 81 ± 3% 27 ± 8% 16 ± 5% MLE T T 88% 42% 28% MLE T H 83% 30% 18% See Table 1 for legend. The train column indicates whether training (i.e., parameter estimation) was performed on the entire training set (T) or on separate 800-gene cross-validation partitions (CV). The test column indicates whether accuracy was measured on the full training set (T), on one-fifth of the training set (CV), or on the unseen data (H). We will consider the evaluation on H to be the most reliable measure of gene finder accuracy. For any row containing a CV, we report the average of five runs, where each run used a different 800-gene subset of the training data for parameter estimation. Both tables give compelling evidence for the value of gradient ascent training, as shown in Figure 1 . In Arabidopsis , gradient ascent applied to the full training set improved over the MLE method from 71% to 81% at the level of exons and 33% to 48% at the level of whole genes. In Aspergillus the improvement was even more dramatic: 30% to 51% at the exon level and 18% to 31% for whole genes. A gain of 4% nucleotide accuracy was measured for both organisms. Figure 1 Maximum likelihood versus gradient ascent Gradient ascent parameter estimation (GRAPE) improves accuracy over MLE at the nucleotide, exon, and whole gene levels. arab = Arabidopsis thaliana , asp = Aspergillus fumigatus . Data partitioning and cross validation A tangible improvement was still seen when a cross-validation design was used to split the training set so as to separate the data used for maximum likelihood estimation (800 genes) and subsequent gradient ascent (200 genes). However, results from both organisms suggest that this separation did not improve the accuracy of the gene finder, as shown in Figure 2 . Indeed, on Arabidopsis , gradient ascent training produced greater gains in accuracy when performed on the entire training set rather than using the cross-validation structure, while on Aspergillus the improvement due to using a cross-validation structure was either small (nucleotide level: 1%), zero (exon level), or negative (gene level: -2%). Thus, the recommended training protocol would be to apply MLE to the entire training set followed by gradient ascent on the full training set as well. Figure 2 Data partitioning for gradient ascent Separating the training set into an 800-gene MLE set and a 200-gene gradient ascent set provides no improvement over simply performing MLE and GRAPE on the full training set. Although use of a cross-validation structure to split the training set for the twin purposes of maximum likelihood estimation of ~90,000 parameters and gradient ascent refinement of 29 parameters is therefore not justified (according to the above results), cross-validation does seem to have some value in terms of predicting how well the gene finder will perform on unseen data, as suggested by Figure 3 . Figure 3 Cross-validation versus testing on unseen data Cross-validation scores provide a reasonably accurate prediction of performance on unseen data. Results shown for A. thaliana only; results for A. fumigatus are given in Table 2. On both genomes and at all levels (nucleotide, exon, gene), accuracy measurements obtained through cross-validation were closer to the accuracy measured on unseen data than were the measurements taken from the full training set, as we expected. This was true both with and without gradient ascent, though when gradient ascent was applied, even the cross-validation results were slightly inflated. The latter observation is presumably attributable to the "peeking" that was permitted (see Methods), whereby the gradient ascent procedure received feedback from the 200 evaluation genes held out from the training set, T. This suggests that estimating even small numbers of parameters (in this case 29) from the test set can artificially inflate accuracy measurements on that set. Figure 4 illustrates the effects of testing the gene finder on the training set. As can be seen from the figure, the accuracy measurements taken from the training set can be substantially inflated relative to the more objective measurements taken from the hold-out set, thereby promoting overly optimistic expectations for how the gene finder will perform on unseen data. Figure 4 Evaluation on the training set Accuracy measurements taken from the training set were artificially inflated, as expected. Results are shown only for A. thaliana ; results for A. fumigatus were even more extreme. Discussion The results presented above provide a clear demonstration that independent maximum likelihood estimation of submodel parameters is sufficiently neglectful of global GHMM behavior as to compromise gene finder accuracy. Even such a crude method as our 29-parameter gradient ascent procedure proved to be effective at significantly improving accuracy over that achievable by simple MLE training. The potential for more sophisticated global discriminative training methods to produce even greater improvements is surely worthy of investigation. It is interesting to observe that the natural language processing and speech recognition communities, from whom HMM-based methods were originally borrowed for use in bioinformatics, have been moving toward global discriminative training methods for some time. The two most popular forms of discriminative training for speech recognition are Maximum Mutual Information (MMI) and Minimum Classification Error (MCE). Both methods can be implemented using an iterative gradient ascent/descent algorithm. Our approach is most similar in spirit to that of MCE. In the case of "pure" (i.e., non-generalized) HMMs, expectation-maximization (EM) update formulas have been derived for both MMI and MCE. These formulas allow model parameters to be updated in an axis-oblique (rather than axis-parallel ) manner; i.e., multiple parameters can be adjusted simultaneously, so that the optimizer is less constrained in following the direction of steepest gradient in parameter space. This may reduce the number of steps required for convergence. Indeed, more rapid convergence (in terms of numbers of re-evaluation steps) has been cited as a concrete advantage of these EM-style formulations over more generalized gradient ascent methods [ 23 ]. However, EM-style approaches to the discriminative training problem for HMMs have typically involved a number of simplifying assumptions and/or heuristics, thereby voiding formal assurances of optimality (e.g., [ 17 , 24 , 18 , 26 ]). Furthermore, as with more generalized gradient ascent procedures, EM often tends to find only a local optimum rather than a global one [ 13 ]. In the case of GHMM-based gene finders, the advantages of EM over a generalized gradient ascent procedure may indeed be rather slim. The very flexibility which we find attractive in GHMMs can be expected to complicate the derivation of such EM-like update formulas for arbitrary GHMM-based gene finders, likely requiring additional assumptions and approximations that would further compromise the optimality of the EM procedure. It was for this reason that we decided to employ a more generalized gradient ascent method for the present study. A rudimentary gradient ascent optimizer is simple to implement, and the use of prediction accuracy as an objective function affords great convenience in approximating Σ (S,φ)∈T P(φ|S,θ). Although P(φ|S,θ) can be more directly computed using a modified Forward algorithm [ 23 ], to do so would in theory be no more efficient than running the full gene finder, since the asymptotic run times of the Forward and Viterbi algorithms for GHMMs are equivalent. Nevertheless, inasmuch as the Forward algorithm provides a more direct approximation of P(φ|S,θ), its use for this purpose is worthy of investigation. There are a number of other variations and enhancements which we are at present contemplating for our discriminative trainer. One of these involves the joint training of pairs of submodels in the GHMM using a maximum discrimination criterion rather than the usual one based on maximum likelihood. Although such an approach would not in itself directly attend to the global optimality of the GHMM (indeed, we already apply such an approach to our signal sensors during our so-called "MLE" training regime, as remarked earlier), it would at least seem to offer a promising direction for improving our existing optimizer and may be feasible without increasing the computational cost beyond what is practical. For the present, we feel confident in making the recommendation that others tasked with the training of GHMM gene finders consider applying an automated gradient ascent procedure like that described here as a more systematic alternative to manual tuning of parameters following maximum likelihood training of individual submodels. Beyond the obvious advantage of likely improving gene finder accuracy, such an automated method may offer some degree of reproducibility (notwithstanding the typically stochastic nature of such methods) and uniformity for the purposes of comparing gene finders and gene finding algorithms. In addition, we urge those practicing manual tuning on their final "test" set to consider that their reported accuracy results may well be inflated as a result of "peeking" at the test set before the final evaluation – a practice that has been criticized in the field of machine learning (eg., [ 27 ]). That significant inflation was seen in our studies as a result of tuning only 29 of the ~90,000 GHMM parameters on the 200-gene "test" set suggests that the phenomenon may conceivably occur to some degree even when an automated procedure is not employed. Finally, we would like to make note of an unfortunate consequence of discriminative training of HMMs for biological sequence analysis, namely, that while the resulting models may possess improved ability for discrimination and therefore greater utility for specific tasks such as gene prediction, their suitability as representative models of biological knowledge (especially probabilistic knowledge) may well be reduced relative to models induced with simple MLE techniques. Indeed, some authors in the field of speech recognition (e.g., [ 20 ]) have noted that more accurate discrimination can sometimes be obtained by relaxing sum-to-one constraints for probability distributions, thereby permitting the gradient ascent procedure to automatically discover appropriate weightings between states or inputs. This is reminiscent of the exon "optimism" parameter which we employ and which seems to have no principled justification (and indeed, we might speculate that this extraneous parameter proved useful precisely because it enabled a primitive form of discriminative training by providing an explicit "correction factor" or weighting between submodels). Thus, despite the apparent value of discriminative training in improving gene finder accuracy, our ability to extract biological knowledge by inspecting the parameters of a gene finder trained in this way may be somewhat hindered. For the present, this does not seem to be of great practical significance, but it is a consideration worthy at least of mention. Conclusions We have shown that discriminative training for GHMM-based gene finders is feasible using a rudimentary gradient ascent approach, and have briefly explored the relation between this method and the EM-like techniques which have been proposed in the field of speech recognition. Our experiments show that the gradient ascent method can result in a gene finder with substantially greater prediction accuracy. It is our hope that even greater gains in accuracy will result from extension and refinement of discriminative training techniques applied to GHMM-based gene finders. Methods Description of the GHMM The gene finder TigrScan [ 8 ] is a GHMM-based program similar to Genie [ 1 ] and Genscan [ 2 , 28 ]. The forward-strand model contains six signal states (donor and acceptor sites; start and stop codons; promoter; poly-A signal) and eight content states (intron; intergenic; 5' and 3' UTR; initial, internal, final, and single exons). The reverse-strand model mirrors that of the forward strand. Four relative frequency histograms are used to estimate the duration probabilities of the four exon types; the four noncoding states are assumed to have geometric duration distributions and are therefore each parameterized by a single value representing the mean duration. Each content state is scored using a separate fifth-order Interpolated Markov Model (IMM) [ 29 ]. TigrScan offers a number of signal sensors, including WMMs, WAMs, WWAMs, and MDD trees [ 28 ] having any of the foregoing signal sensors as leaf models; for this study we used only (non-MDD) WAMs, though the order of the Markov chains within the WAMs was allowed to vary. Putative signals scoring below a given signal threshold are ignored by TigrScan. This threshold is chosen separately for each signal sensor so as to achieve a desired sensitivity Sn ( Sn = TP /( TP + FN ), TP = true positive count, FN = false negative count) on a training set of true and "decoy" signals. "Boosting" of signal sensors was performed by iteratively retraining each signal sensor on sets of training features in which the lowest scoring features were duplicated so as to focus the training procedure on the most difficult examples. Boosting has been found to improve signal detection in other application areas [ 30 ]. Most transitions in the GHMM are obligatory (such as "donor site → acceptor site"); of the non-obligatory transitions, sum-to-one constraints and the forward/reverse strand equivalence reduce the number which can be independently varied to just four. Transitions into exon states are modified by an exon "optimism" multiplier (similar to that described in [ 6 ]) which has been seen anecdotally to be useful in improving prediction accuracy (unpublished data). Parameters to be optimized The total number of parameters which need to be estimated when training TigrScan is roughly 90,000; the large bulk of these are the n-gram statistics comprising the IMMs used for the content sensors. As an initial attempt at applying discriminative training to TigrScan, we selected 29 of these ~90,000 parameters to subject to gradient ascent optimization. Although this is a miniscule proportion of the available parameters, our previous experiences with hand-tuning our GHMM on other data sets suggested that these 29 parameters exert a disproportionately large influence on the accuracy of the gene predictions. By limiting the number of parameters to be optimized we hoped to both accelerate the training procedure and also reduce the risk of overtraining. The selected parameters were: • mean intron, intergenic, and UTR lengths (3) • transition probabilities (4) • exon optimism (1) • WAM size and relative positioning (8) • WAM order (4) • signal sensitivity (1) • number of signal boosting iterations (8) • skew and kurtosis of exon length distributions Modifications to skew and kurtosis of exon length distributions were found during early exploration to produce no improvements; these parameters were therefore left unchanged in all further experiments. All remaining parameters were estimated using standard MLE techniques. For those runs in which gradient ascent was disabled (see below), the following methods were used to estimate the above 29 parameters: mean intron and UTR lengths as well as transition probabilities were estimated using MLE from training data; mean intergenic length was set to a fixed value based on the known intergenic lengths in the test set; exon optimism was set to zero; remaining parameters were selected so as to minimize the misclassification rate on a set of true and "decoy" signals selected from the training set. Objective function and optimization procedure As an objective function for use by the gradient ascent procedure, we decided to measure the accuracy of the current parameterization by running the gene finder on a subset of the training genes. Our hope was that this accuracy measure would provide a reasonable approximation of Σ (S,φ)∈T P(φ|S,θ) by indicating roughly how often the current model θ would cause the correct parse φ to be predicted for training sequence S. We defined the nucleotide accuracy A nuc as the percentage of nucleotides correctly classified as coding vs. noncoding; A exon was defined as an average of exon sensitivity and specificity (where a predicted exon is considered correct only if both boundary coordinates were predicted correctly); and A gene was defined as the percentage of training genes which were predicted exactly correctly. These were all rounded to integral percentages between 0 and 100%. The objective function was then defined as: f (θ) = 100 A nuc + A exon + A gene .    (4) The A nuc and A exon terms were included in an effort to smooth the function, which would otherwise have been insensitive to changes not reflected in the number of genes predicted exactly correctly – i.e., a step function. Though the A nuc term was given much greater weight for this study, additional work needs to be undertaken to determine the most suitable set of weights for our objective function. Parameters were optimized using an iterative gradient ascent procedure operating in the selected 29-dimensional parameter space, as illustrated schematically in Figure 5 . Steps were taken in an axis-parallel manner (one step per axis per iteration), with the step size for each axis decreasing by half whenever a local maximum was reached on that axis. Figure 5 Gradient ascent training Schematic diagram of gradient ascent training procedure. Of 29 parameters modified by gradient ascent, some (e.g., WAM size) were used to control the MLE estimation procedure, while others (e.g., mean intron length) were used directly as parameters to the GHMM. Testing of the gradient direction was performed on the 200-gene cross-validation set, which was part of the 1000-gene training set, T. Data and experimental design The quality of a given parameterization θ was measured by evaluating the objective function f (θ) on a held-out subset of the training set. The training set was limited to 1000 genes, and all experiments were repeated separately on two highly divergent species, the model plant Arabidopsis thaliana and the pathogenic fungus Aspergillus fumigatus . Five-fold cross-validation was employed, so that the entire optimization procedure was carried out five times on four-fifths of the data (800 genes) and each time evaluated on the remaining one-fifth (200 genes); accuracy results reported here were obtained by averaging the five sets of accuracy numbers obtained from the cross-validation. The held-out one-fifth was also used by the gradient ascent procedure to tune the selected 29 parameters. The practice of using a held-out set for smoothing or to estimate a small number of additional parameters is common in the natural language processing field [ 31 ], where it is recognized that such "peeking" at the test set (by which we mean iterative re-estimation of model parameters from the training set after receiving accuracy feedback on the test set) by the training procedure can (unfortunately) artificially inflate reported accuracy numbers. For this reason, an additional 1000 genes were used for testing the gene finder after each cross-validation run. The results of this final testing were not made available to the optimizer, but are instead reported here as a more objective assessment of final model accuracy. We will refer to the training set as T and the additional 1000 genes for testing as H. BLAST [ 32 ] was used to ensure that no two genes in T∪H were more than 80% similar over 80% of their lengths at the nucleotide level. This training protocol is illustrated in Figure 6 . Figure 6 Cross-validation experiments Five-fold cross-validation was used both in the gradient ascent and in the MLE-only experiments. For gradient ascent training, MLE was performed on four-fifths of the training set (T) and then gradient ascent was performed on the other one-fifth. A separate hold-out set (H) of 1000 genes was used to obtain an unbiased evaluation of all final models. Several variations of this experiment were also performed. To evaluate the utility of splitting the training set and performing MLE and gradient ascent parameter estimation on separate subsets (as described above), we also performed MLE followed by gradient ascent training on the full training set T and again evaluated the induced models on H. To assess whether gradient ascent provided any improvement in accuracy we also trained a model on T using only MLE and evaluated that model on H. Although the virtues of cross-validation have been well explored in the context of many other applications, we decided to use the above experimental design as a convenient opportunity to verify our expectation that it would also prove useful for objective analysis of gene finder accuracy. Authors' contributions Software implementation and computational experiments were performed by WHM. The manuscript was written by WHM with assistance from SLS.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544851.xml

543469

Validity of the AusTOM scales: A comparison of the AusTOMs and EuroQol-5D

Background Clinicians require brief outcome measures in their busy daily practice to document global client outcomes. Based on the UK Therapy Outcome Measure, the Australian Therapy Outcome Measures were designed to capture global therapy outcomes of occupational therapy, physiotherapy and speech pathology in the Australian clinical context. The aim of this study was to investigate the construct (convergent) validity of the Australian Therapy Outcome Measures (AusTOMs) by comparing it with the EuroQuol-5D (EQ-5D). Methods The research was a prospective, longitudinal cohort study, with data collected over a seven month time period. The study was conducted at a total of 13 metropolitan and rural health-care sites including acute, sub-acute and community facilities. Two-hundred and five clients were asked to score themselves on the EQ-5D, and the same clients were scored by approximately 115 therapists (physiotherapists, speech pathologists and occupational therapists) using the AusTOMs at admission and discharge. Clients were consecutive admissions who agreed to participate in the study. Clients of all diagnoses, aged 18 years and over (a criteria of the EQ-5D), and able to give informed consent were scored on the measures. Spearman rank order correlation coefficients were used to analyze the relationships between scores from the two tools. The clients were scored on the AusTOMs and EQ-5D. Results There were many health care areas where correlations were expected and found between scores on the AusTOMs and the EQ-5D. Conclusion In the quest to measure the effectiveness of therapy services, managers, health care founders and clinicians are urgently seeking to undertake the first step by identifying tools that can measure therapy outcome. AusTOMs is one tool that can measure global client outcomes following therapy. In this study, it was found that on the whole, the AusTOMs and the EQ-5D measure similar constructs. Hence, although the validity of a tool is never 'proven', this study offers preliminary support for the construct validity of AusTOMs.

Background The costs of operating public health services in Australia are rapidly rising. Health administrators and practitioners are under pressure to document client outcomes and demonstrate the effectiveness of therapy interventions [ 1 - 3 ]. Increasingly, the allied health professions have come to see the need for quick, easy to use measures that describe the result of interventions in terms of client outcomes, and provide evaluative data for benchmarking between health service providers [ 4 ]. An outcome measure is a tool for documenting change in client status following therapist intervention. This involves the therapist administering a standardized measure at two time points (for example, at admission and at discharge) or at designated time points throughout therapy and then calculating how much change has occurred. The effectiveness of a therapy is shown when the therapist is able to demonstrate that the change in client status was attributable to treatment and not to other factors such as spontaneous recovery [ 3 , 5 ]. In response to the need for outcome measures, a study titled Australian Therapy Outcome Measures (AusTOMs) was funded by the Australian Department of Health and Ageing from 2001–2003. The goal of the study was to develop a reliable and valid measure of therapy outcome for the three largest allied health professions in Australia; occupational therapy (OT), physiotherapy (PT) and speech pathology (SP). The AusTOMs was based on the UK Therapy Outcome Measure (TOM) and adapted and developed to suit the current practices of therapists in Australia [ 6 ]. Clinicians can use AusTOMs data which show client change over time in a variety of ways. Clinicians can benchmark their service against other similar facilities which may prompt changes in the type or duration of therapy services offered [ 4 ]. Tools such as AusTOMs can also be used in research (such as randomised controlled trials) to evaluate the effectiveness of therapy services. The original TOM was developed for use by speech and language therapists in the UK for therapists to measure client outcomes in a clinical setting [ 1 ]. Later, scales were developed to measure the effects of interventions by occupational therapists, physiotherapists, and rehabilitation nurses [ 7 , 8 ]. Both sets of tools were used to provide benchmarks for therapist practice between service providers [ 4 , 8 - 12 ]. The development of the TOM was considerably influenced by the International Classification of Impairments, Disabilities and Handicaps 1 and 2 (ICIDH 1&2) [ 13 ]. The TOM draws on the ICIDH domains and allows therapists to monitor client status over time in relation to Impairment, Disability, and Handicap. In addition, the developers of TOM added a domain to measure therapist perception of client Wellbeing or Distress (now referred to in this article as Wellbeing). The inter-rater reliability for the four domains of the TOM have been reported for Occupational Therapy as .84 for impairment, .85 for disability, .74 for handicap and .58 for Wellbeing. The reliability for physiotherapists was .66 for impairment, .74 for disability, .77 for handicap and .57 for Wellbeing and for Speech Pathologists the reliability was .89 for impairment, .90 for disability, .84 for handicap and .57 for Wellbeing [ 1 , 14 ]. The AusTOMs was designed to measure client therapy outcome separately for occupational therapists, speech pathologists and physiotherapists. Similar to TOM, AusTOMs provides a 'snapshot' rating that is determined by the clinical judgment of the therapist, which broadly reflects the client's status. The development of the scales and content validity of AusTOMs has been published [ 6 ], as has preliminary data concerning the reliability of the scales [ 15 ]. Attention has now turned to whether the instrument performs in a manner consistent with the theoretically derived hypotheses underpinning the constructs being measured [ 16 ]. The purpose of this paper is to continue the process of validating the AusTOMs, by establishing construct (convergent) validity. Construct validity refers in part to the ability of an instrument to measure an abstract concept or construct. Because constructs are not directly observable and are usually multidimensional, it is important to ascertain that the constructs adequately define and represent the variables that the instrument purports to measure [ 16 ]. In particular, convergent validity indicates the degree to which two instruments are measuring similar constructs. Therefore, examination of the construct validity of the AusTOM scales concerns whether the scales actually measure the intended underlying construct of global health-related outcomes. The researchers attempted to find a 'gold standard' tool to investigate the concurrent validity of the AusTOMs. Health-related quality of life (HRQoL) tools have increasingly been used to assess multiple aspects of health-related quality of life in clinical trials [ 17 ]. Tools such as the General Sickness Impact Profile (SIP) [ 18 ] measures or infers aspects of activity and participation. The Medical Outcomes Study (MOS) Short Form Health Survey (SF-36) [ 19 ] and the Nottingham Health Profile (NHP) [ 20 ] measure or infers aspects of impairment, activity and participation. The widely used Functional Independence Measure [ 21 ] records the single health domain of activity limitation. However, no tools could be found that measure all four health domains as provided by the AusTOMs, and the tools that were reviewed required too much therapist administration time to be included in the present study. Since there is no gold standard global health status and therapist administered tools with which to compare AusTOMs, it was decided to compare the constructs of AusTOMs with those of EuroQoL-5D (EQ-5D) [ 22 ] in order to investigate the convergent validity of the tool [ 16 ]. The EQ-5D was chosen for this study since it is widely used in European [ 22 ] and Australian studies [ 23 , 24 ], has been used in number of clinical trials [ 17 ], is simple and quick to use [ 25 ] and similar to the AusTOM, purports to measure global health-related outcomes. However, the potential advantage of using the AusTOMs over the EQ-5D is that while the EQ-5D measures health related outcomes globally, the AusTOMS measures global outcomes in relation to the four specific domains of impairment, activity limitation, participation restriction and wellbeing/ distress. EQ-5D is a short and simple to administer generic HRQoL measure of health status [ 25 ]. EQ-5D provides a simple descriptive profile of client problems on five dimensions, an overall score for client self-rated health, and generates a single index value that can be used in the clinical and economic evaluation of health care and in population health surveys [ 17 ]. EQ-5D was initially developed in Dutch, English, Finnish, Norwegian and Swedish and is now available in 42 official translations and adaptations [ 22 ]. While in principle, health professionals support the notion of measuring health status, there is no consensus regarding the method of measurement [ 26 , 27 ]. While the AusTOMs is rated by therapists, the EQ-5D is rated by the client's themselves. This may be viewed as the main limitation in selecting the the EQ-5D for comparison with the AusTOMS. Nonetheless, it was expected that scores on the AusTOMs scale would vary in relation to scores generated on the EQ-5D since both seek to measure global health-related outcomes. Some researchers prefer the objectivity offered by therapist ratings from observation of client performance [ 28 ]. Others support client self-report [ 29 , 30 ] as an accurate reflection of the client's perception of their status, which is becoming increasingly important in consumer-driven heath services. Self -report tools are also considerably cheaper than therapist administered ones, hence, self-report assessments are typically used in a climate requiring cost containment [ 31 ]. However, it is also becoming increasingly clear that therapist and client ratings of client performance may not be related [ 27 , 32 ]. In view of the lack of therapist-administered tool suitable to validate the AusTOMs against, and given the time administration advantages of the use of the EQ-5D which were significant to the success of this research program, the EQ-5D was selected for inclusion in the present study. The purpose of this study was to examine the measurement properties of the AusTOMs and to compare them with the EQ-5D in 'real practice'. The main question being; does AusTOMs perform in a similar manner to the EQ-5D? The study sought to investigate the following hypotheses: 1. There will be a clear pattern of correlations for the admission, discharge and change scores between the AusTOMs domains and the EQ-5D Health Status and Thermometer. Several scale-specific correlations are expected. For example : a. There will be a moderate negative correlation of the admission, discharge and change scores between the PT AusTOMs Scale 'Pain', Impairment domain and the EQ-5D Health Status Subscale 'Pain'. b. There will be a moderate negative correlation of the admission, discharge and change scores between OT AusTOMs Scale 'Functional Mobility and Walking', Activity Limitation Domain and the EQ-5D Health Status Subscale 'Mobility'. c. There will be a moderate negative correlation of the admission, discharge and change scores between OT AusTOMs Scale 'Self-care', Activity Limitation domain, and the EQ-5D Health Status Subscale 'Self-care'. 2. There will be a moderate positive correlation of the admission, discharge and change scores between all the Physiotherapy, Occupational Therapy and Speech Pathology AusTOMs Scales for the Wellbeing /Distress scores and the EQ-5D Thermometer. Methods The research was designed as a prospective, longitudinal cohort study, with data collected over a seven month time period. Participants Thirty-eight occupational therapists, 30 physiotherapists and 47 speech pathologists were trained at 13 participating facilities to collect AusTOMs data, and to present the EQ-5D for clients to complete. However, it is possible that not all these therapists collected data (data collection forms did not require therapists to record their identity). The facilities included acute hospitals, rehabilitation hospitals, and community care facilities. Therapists recorded AusTOMs data and obtained client EQ-5D ratings from 205 clients (110 from Physiotherapy, 67 from Occupational Therapy and 28 from Speech Pathology). These clients were from a larger group of 1007 clients who participated in the study (284 from Physiotherapy, 466 from Occupational Therapy and 257 from Speech Pathology). While some of these participants refused to complete the EQ-5D, or the therapists chose not to burden the client with completing this form, many were children or non-cognizant adults and the EQ-5D is not validated for these groups. Otherwise, the sample was sequential admissions to therapist caseloads over a seven month period. Instruments AusTOMs is comprised of three separate sets of scales for Occupational Therapy (12 scales), Speech Pathology (6 scales) and Physiotherapy (9 scales). The title of each scale is provided in Table 1 . Table 1 AusTOMs scales for occupational therapists, speech pathologists and physiotherapists Scale Occupational Therapy Speech Pathology Physiotherapy 1 Learning & Applying Knowledge Speech Balance & Postural Control 2 Functional Walking & Mobility Cognitive-Communication Cardiovascular System Related Functions 3 Upper Limb Use Language Musculoskeletal Movement Related Functions 4 Carrying Out Daily Life Tasks & Routines Voice Neurological Movement related Functions 5 Transfers, Swallowing Pain 6 Using Transport Fluency Respiratory Related Functions 7 Self-care Sensory functions 8 Domestic Life – Home Skin functions 9 Domestic Life – Managing Resources Urinary and bowel continence 10 Interpersonal Interactions & Relationships 11 Work, employment and Education and Community Life 12 Recreation, Leisure and Play. Each scale requires a rating for four domains of client function, that is, Impairment, Activity Limitation, Participation Restriction and Wellbeing/Distress. An additional optional rating can be made of a caregiver's level of Wellbeing/Distress if the clinician has had contact with a caregiver, and feels that therapy is directed toward the caregiver in some way. Each of the domains are rated by therapists on an 11-point ordinal scale (6 defined points from 0 [most severe] to 5 [normal], and 5 undefined half points). Although clinicians are only required to use the 6 defined scale points, clinicians overwhelmingly chose to include the half points in the AusTOM scoring system to increase scale sensitivity. The use of the half points also facilitates international benchmarking of data against the UK TOM. A generic description of each of the domains of client function is presented in Table 2 . Three of the AusTOM's four domains were drawn from the World Health Organisation (WHO)'s International Classification of Function (ICF) [ 33 ]. Based on TOM, the AusTOMs were developed by focus groups of expert clinicians in the state of Victoria in Australia who determined both the scale headings, and scalar descriptions for each of the 6 levels for each of the four domains. These scales were then sent out to clinicians across Australia for further refinement. More information on scale development was reported in an earlier publication [ 6 ]. In addition, a publication in press [ 15 ] reports the reliability of the AusTOM's domains for the majority of scales as ranging from 60–100% agreement, within .5 scalar points for most domains. Table 2 Generic AusTOMs scales (Perry et al, 2004) Impairment of either Structure or Function (as appropriate to age): Impairments are problems in body structure (anatomical) or function (physiological) as a significant deviation or loss . 0 The most severe presentation of impairment (either structure or function) 1 Severe presentation of this impairment 2 Moderate/severe presentation 3 Moderate presentation 4 Mild presentation 5 No impairment of structure or function Activity Limitations (as appropriate to age): Activity limitation results from the difficulty in the performanceof an activity. Activity is the execution of a task by the individual . 0 Complete difficulty 1 Severe difficulty 2 Moderate/severe difficulty 3 Moderate difficulty 4 Mild difficulty 5 No difficulty Participation Restrictions (as appropriate to age): Participation restrictions are difficulties the individual may have in the manner or extent of involvement in their life situation. Clinicians should ask themselves: "given their problem, is this individual experiencing disadvantage?" 0 Unable to fulfill social, work, educational or family roles. No social integration. No involvement in decision-making. No control over environment. Unable to reach potential in any situation. 1 Severe difficulties in fulfilling social, work, educational or family roles. Very limited social integration. Very limited involvement in decision-making. Very little control over environment. Can only rarely reach potential with maximum assistance. 2 Moderately severe difficulties in fulfilling social, work, educational or family roles. Limited social integration. Limited involvement in decision-making. Control over environment in one setting only. Usually reaches potential with maximum assistance. 3 Moderate difficulties in fulfilling social, work, educational or family roles. Relies on moderate assistance for social integration. Limited involvement in decision-making. Control over environment in more than one setting. Always reaches potential with maximum assistance and sometimes reaches potential without assistance. 4 Mild difficulties in fulfilling social, work, educational or family roles. Needs little assistance for social integration and decision-making. Control over environment in more than one setting. Reaches potential with little assistance. 5 No difficulties in fulfilling social, work, educational or family roles. No assistance required for social integration or decision-making. Control over environment in all settings. Reaches potential with no assistance. Wellbeing/Distress (as appropriate to age): The level of concern experienced by the individual. Concern may be evidenced by anxiety, anger, frustration etc . 0. High and consistent levels of distress or concern. 1. Severe concern, becomes distressed or concerned easily. Requires constant reassurance. Loses emotional control easily. 2. Moderately severe concern. Frequent emotional encouragement and reassurance required. 3. Moderate concern. May be able to manage emotions at times, although may require some encouragement. 4. Mild concern. Able to manage emotions in most situations. Occasional emotional support or encouragement needed. 5. Able to cope with most situations. Accepts and understands own limitations. The EQ-5D consists of two parts; the self -classifier or questionnaire, and the EQ-Vas or Thermometer. The EQ-5D self-classifier is a one-page questionnaire, which captures respondent descriptions of health problems on a 5-dimensional classification of mobility, self-care, usual activities, pain and discomfort and anxiety and depression. Each dimension is rated by respondents on a three-level scale from 1 (no problem) to 3 (unable or extreme problem) [ 22 ]. The EQ-Vas is a 20-centimeter visual analogue scale, portrayed as similar to a thermometer, on which the respondent rates his/her health state today between 0 (worst imaginable) to 100 (best imaginable). Overall, respondent's health status is either expressed as a score on the visual analogue scale (EQ-Vas), as a profile of their scores on each of the five dimensions (self-classifier), or by combining the scores on the five dimensions. This research utilised the combined scores from the 5 dimensions. The combined dimensions describe 243 theoretically possible health states, that can be converted into a weighted health index score (EQ-Index) for use in cost-effective analysis [ 26 ]. The EQ-5D has been shown to be both reliable and valid when used with adult clients with a wide variety of health-related conditions [ 17 , 22 , 25 , 26 ]. Procedure Approval from the Human Ethics Committee at La Trobe University and the participating facilities was obtained. Study packs were collated for the collection of data. Each pack contained AusTOMs Scale Manual, AusTOMs and EQ-5D data collection forms, informed consent information and consent forms (if these were required by the facility ethics committee). The packs were sent to a contact person in occupational therapy, speech pathology and physiotherapy departments at each site participating in the project. The role of the contact at each site was to receive the packs, disseminate the packs to therapists, check the packs after completion and return them by postage paid envelope. On admission, the therapists (who had each been previously trained in the use of the scales) briefed each client about the study and after verbal agreement, clients were given a statement of informed consent to read and sign. Clinicians then recorded relevant demographic information and established with the client a specific goal or set of goals for the first episode of care. The therapist then chose the AusTOMs scale/s that best described the main areas targeted for therapy intervention. An admission rating was made by the therapist for each of the four domains of AusTOMs (impairment, activity limitation, participation and wellbeing/distress) on a scale from 0 (most severe) to 5 (least severe). A rating for Wellbeing/ distress was also made for the client's carer if this was applicable to the client's situation. Therapists report that the AusTOMs takes approximately 5 minutes to complete. The therapist then asked the client to complete the self -classifier section of the EQ-5D and the EQ-Vas (Thermometer). Clients were instructed to indicate which statements best described their own health state today, by placing a tick in one box for each of the dimension of mobility, personal care, usual activities, pain/discomfort and anxiety depression. Finally, clients completed the EQ-Vas. Information on the form stated, 'to help people say how good or bad a health state is, we have drawn a scale (rather like a thermometer) on which the best state you can imagine is marked 100 and the worst state you can imagine is marked 0. We would like you to indicate on this scale how good or bad your own health is today, in your opinion. Please do this by drawing a line from the box below to whichever point on the scale indicates how good or bad your health state is today' [ 22 ]. Clients completed the EQ-5D in approximately 5 to 20 minutes. The therapist rating for AusTOMs was repeated at client discharge, and clients were asked to again complete both sections of the EQ-5D. Data Analysis The data were analyzed separately for each profession given the differences in the AusTOMs scales. Correlational analyses were performed to investigate the relationship between AusTOMs and EQ-5D. Given the ordinal nature of the scales, a non-parametric approach was adopted, hence all analyses use Spearman's rank-order correlation coefficients (Spearman's Rho). Given the number of correlations performed, alpha (to determine statistical significance) was set at .01, and magnitude of the relationship was considered using the guidelines from Colton [ 34 ] where .00 – .25 = little or no relationship, .25 – .50 = a weak to fair relationship, .50 – .75 moderate to good relationship and .76 and above considered good to excellent. In this paper, only relationships that are .5 – .75 (moderate to good), and .76 and above (good to excellent) are reported. In addition, only expected correlations are reported. The optional AusTOMs domain of 'Caregiver Wellbeing' was not included in the analyses since the sample sizes were generally too small to enable computations. Analyses were undertaken across the scales for each profession, and since sample sizes permitted, for the physiotherapy scales: Balance and Postural Control, Musculoskeletal and Neurological, and for the occupational therapy scales: Functional Walking and Mobility, Upper Limb Use, and Self-care. Sample sizes were not sufficient to enable individual scale analysis for speech pathology scales. The analyses were conducted using only the first scaled selected by the therapist to rate the client. It is also important to note the directions of relationships reported. The EQ-5D Health Status subscales are 1 = no problem -> 3 = unable or extreme problem and the AusTOMs scores are 5 = Normal -> 0 = unable or extreme problem, hence, we expect to see negative correlations. However, the EQ-5D Thermometer scores 0 as the worst state and 100 as the best state and the overall EQ-5D Health Status self classifier score also indicates a better outcome as the score increases, and the AusTOMs scores are 5 = Normal -> 0 = unable or extreme problem. Hence, we expect to see positive correlations between these scores. In the 'Results' the statement is made that the results are in the 'expected direction'. In line with the research aims and hypotheses, the following analyses were undertaken across each profession's data set. First, a correlation considering all the AusTOMs scales for each domain with EQ-5D Health Status (self classifier score and the 5 dimensions) and Thermometer at admission was performed. Next, AusTOMs scores for each domain for a subset of the most frequently used OT and PT scales with EQ-5D Health Status (self classifier score and the 5 dimensions) and Thermometer at admission were obtained. Then, considering all the AusTOMs scales for each domain were correlated with EQ-5D Health Status (self classifier score and the 5 dimensions) and Thermometer at discharge. Following this, AusTOMs scores for each domain for a subset of the most frequently used occupational therapy and physiotherapy scales were correlated with EQ-5D Health Status (self classifier score and the 5 dimensions) and the Thermometer at discharge. Finally, correlations were obtained for change from admission to discharge scores for AusTOMs (considering all the scales overall and for individual scales) with change from admission to discharge scores for EQ-5D Health Status (self classifier score and the 5 dimensions) and the Thermometer. Results A brief summary of demographic data from the sample is provided in Table 3 . The results are presented in relation to the five analyses performed with the data set from each profession. The moderate to good, statistically significant correlations are reported in Table 4 (physiotherapy), Table 5 (occupational therapy), and Table 6 (speech pathology). Rather than present all correlations, only those that would be theoretically expected are presented. In Tables 4 , 5 , 6 , an asterisk is also marked where correlations were expected that were not found, and the sample sizes the analyses were performed on are included since in many cases there is an inadequate sample to detect a relationship. Table 3 Summary of client demographic data Variable Occupational Therapy Clients (n = 67) Speech Pathology Clients (n = 28) Physiotherapy Clients (n = 110) Mean Age 67.24 (SD 16.65) 64.44 (SD 13.43) 65.44 (SD 20.84) SEX No. Females 41 (61.2%) 11 (39.3.1%) 67 (60.9%) No. Males 25 (37.3%) 16 (57.1%) 42 (38.2%) Missing 1 (1.5%) 1 (3.6%) 1 (0.9%) 3 most frequently recorded aetiologies Acquired neurological 24 (35.8%) Orthopaedic 12 (17.9%) Spinal 6 (9%) Acquired neurological 14 (50%) Oncology 7 (25%) Neurosurgery 3 (10.7%) Orthopaedic 44 (40%) Acquired neurological 19 (17.3%) Spinal 9 (8.2%) Musculoskeletal 9 (8.2%) 3 most frequently recorded disorders Inadequate muscle power 16 (23.9%) Decreased general mobility 11 (16.4%) Multifactorial 11 (16.4%) Pain 9 (13.4%) Dysphagia (feeding) 9 (32.1%) Acquired language disorder 5 (17.9%) Disorders of voice 5 (17.9%) Dysarthria 4 (14.3%) Cognitive impairment 4 (14.3%) Abnormal joint mobility 29 (26.4%) Decreased general mobility 24 (21.8%) Inadequate muscle power 15 (13.6%) SETTING No. inpatient 44 (65.7%) 17 (60.7%) 78 (70.9%) No. outpatient 21 (31.3%) 8 (28.6%) 32 (29.1%) Missing 2 (3.0%) 3 (10.7%) SERVICE TYPE Acute 7 (10.4%) 1 (3.6%) 17 (15.5%) Subacute 49 (73.1%) 24 (85.7%) 68 (61.8%) Community 9 (13.4%) 2 (7.1%) 15 (13.6%) Home 0 (0%) 0 (0%) 10 (9.1%) Missing 2 (3.0%) 1 (3.6%) Mean No. of occasions of service 9.05 (SD7.50) 23.28 (SD40.97) 12.36 (SD11.84) Table 4 Summary of Physiotherapy Results: Moderate to strong, statistically significant Spearman's Rho correlations between AusTOMs and EQ-5D EQ-5D Therm. EQ-5D Health status EQ-5D Mobility Subscale EQ-5D Self-care Subscale EQ-5D Usual activities subscale EQ-5D Pain/ Discom-fort EQ-5D Anxiety/ depression AusTOM Over all Impairment AusTOM Over all Activity Limitation * * AusTOM Over all Participation AusTOM Over all Wellbeing/ Distress 0.508 0.537 * AusTOM n = 16 Balance & Pos control Impairment -0.691 -0.677 AusTOM Balance & Pos control Activity Limitation * * AusTOM Balance & Pos control Participation AusTOM Balance & Pos control Wellbeing/ Distress 0.655 -0.739 AusTOM n = 66 Musculoskeletal Impairment AusTOM Musculoskeletal Activity Limitation -0.546 * AusTOM Musculoskeletal Participation AusTOM Musculoskeletal Wellbeing/ Distress 0.597 0.614 -0.539 AusTOM n = 18 Neurological Impairment AusTOM Neurological Activity Limitation -0.801 -0.746 * AusTOM Neurological Participation AusTOM Neurological Wellbeing/ Distress 0.770 * Key: Admission correlation coefficients in normal font Discharge correlation coefficients in bold font Change from admission to discharge correlation coefficients in italic Correlations expected but not obtained marked with * Table 5 Summary of Occupational Therapy Results: Moderate to strong, statistically significant Spearman's Rho correlations between AusTOMs and EQ-5D EQ-5D Therm. EQ-5D Health status EQ-5D Mobility Subscale EQ-5D Self-care Subscale EQ-5D Usual activities subscale EQ-5D Pain/ Discom-fort EQ-5D Anxiety/ depression AusTOM Over all Impairment AusTOM Over all Activity Limitation * * AusTOM Over all Participation AusTOM Over all Wellbeing/ Distress * -0.612 AusTOM n = 13 Walk & Mobility Impairment AusTOM Walk & Mobility Activity Limitation * * AusTOM Walk & Mobility Participation AusTOM Walk & Mobility Wellbeing/ Distress * * AusTOM n = 18 Upper limb use Impairment AusTOM Upper limb use Activity Limitation 0.707 * AusTOM Upper limb use Participation AusTOM Upper limb use Wellbeing/ Distress AusTOM n = 16 Self-care Impairment AusTOM Self-care Activity Limitation 0.748 -0.645 -0.623 -0.683 AusTOM Self-care Participation AusTOM Self-care Wellbeing/ Distress * * Key: Admission correlation coefficients in normal font Discharge correlation coefficients in bold font Change from admission to discharge correlation coefficients in italic Correlations expected but not obtained marked with * Table 6 Summary of Speech Pathology Results: Moderate to strong, statistically significant Spearman's Rho correlations between AusTOMs and EQ-5D EQ-5D Therm. EQ-5D Health status EQ-5D Mobility Subscale EQ-5D Self-care Subscale EQ-5D Usual activities subscale EQ-5D Pain/ Discom-fort EQ-5D Anxiety/ depression AusTOM Over all Impairment AusTOM Over all Activity Limitation * * AusTOM Over all Participation AusTOM Over all Wellbeing/ Distress * * Key: Admission correlation coefficients in normal font Discharge correlation coefficients in bold font Change from admission to discharge correlation coefficients in italic Correlations expected but not obtained marked with * Over all AusTOMs scales for each domain with EQ-5D Health Status (self-classifier score and the 5 dimensions) correlated with the Thermometer at admission (in other words, over all AusTOMs scales for each domain correlated with the EQ-5D Thermometer at admission). These results are reported in the first 4 rows of Tables 4 , 5 , 6 , normal font. The correlations found, that were expected are all in the expected direction. AusTOMs scores for each domain for a subset of the most frequently used OT and PT scales with EQ-5D Health Status (self-classifier score and the 5 dimensions) correlated with the Thermometer at admission. Several moderate to strong correlations were found that were expected and these are presented in Table 4 , rows 5–16 for physiotherapy, and Table 5 , rows 5–16 for occupational therapy, all in normal font. Again, all correlations were in the expected direction. Over all AusTOMs scales for each domain with EQ-5D Health Status (self-classifier score and the 5 dimensions) correlated with the Thermometer at discharge. In relation to this correlation, the results are reported in the first 4 rows of Tables 4 , 5 , 6 , bold font. The correlations found (that were expected) for physiotherapy and occupational therapy were all in the expected direction. AusTOMs scores for each domain for a subset of the most frequently used OT and PT scales (only) with EQ-5D Health Status (self-classifier score and the 5 dimensions) correlated with the Thermometer at discharge. The correlations expected that were found are presented in Table 4 , rows 5–16 for physiotherapy, and Table 5 , rows 5–16 for occupational therapy, all in bold font. Again, all correlations were in the expected direction. Change from admission to discharge scores for AusTOMs (overall and for individual scales) with change from admission to discharge scores for EQ-5D Health Status (self-classifier score and the 5 dimensions) correlated with the Thermometer. The moderate to good, statistically significant correlations expected between change on the EQ-5D and AusTOMs overall, or in relation to the six AusTOMs scales where sample size permitted are presented as follows: for physiotherapy (see Table 4 , rows 1–16, italic font), occupational therapy (see Table 5 , rows 1–16, italic font), and speech pathology (see Table 6 , rows 1–4, italic font). Discussion There was some support for the first hypothesis; '...There will be a clear pattern of correlations for the admission, discharge and change scores between the AusTOMs domains and the EQ-5D Health Status and Thermometer (except in relation to the EQ-5D Thermometer and the AusTOMs Wellbeing/ Distress domain as presented in the final hypothesis)'. There were several areas where relationships between constructs measured on AusTOMs and EQ-5D were expected (as described below), and it generally appeared that these two tools are measuring similar constructs. This lends some support to the construct (convergent) validity of AusTOMs. However, not all expected correlations were found and while AusTOMs seems to be measuring global change from the therapist's perspective in relation to four distinct domains (Impairment, Activity Limitation, Participation Restriction and Wellbeing), EQ-5D (as expected), is measuring client perceptions of how they feel about their health status. Hence, while both assessments attempt to capture global health-related outcomes, the differing perceptions of the raters (clinicians versus clients) does seem to impact on the establishment of construct validity. Suggestions for overcoming this problem are described below. The next sub-hypotheses dealt with specific correlations that were expected in these data. Unfortunately, there were insufficient data to determine if a moderate negative correlation between admission, discharge and change scores between the PT AusTOMs Scale 'Pain', Impairment domain and the EQ-5D Health Status Subscale 'Pain' existed. Similarly, there were insufficient data (n = 13) to explore the hypothesis that '...there will be a moderate negative correlation between admission, discharge and change scores between the OT AusTOMs Scale 'Functional Mobility and Walking', Activity Limitation Domain and the EQ-5D Health Status Subscale Mobility'. The final scale-specific hypothesis predicted a moderate negative correlation between admission, discharge and change scores between the OT AusTOMs Scale 'Self-care', Activity Limitation domain and the EQ-5D Health Status Subscale Self-care. This hypothesis was supported. The results indicate that therapist and client perceptions of client self-care ability status on admission, discharge (and in relation to the change scores) were moderately correlated. Since many occupational therapists spend considerable time working on self-care with clients, and talking about progress in this area, it is reasonable that clients and therapists would rate client status in this area in a similar manner. Finally, it was hypothesised that there would be a moderate positive correlation between admission, discharge and change scores across all the PT, OT and SP AusTOMs Scales for the Wellbeing domains and the EQ-5D Thermometer. There was only limited support for this hypothesis. However, in some cases the sample sizes were on the small side. There were no moderate, statistically significant correlations when analyzing across all combined OT and SP AusTOMs Scales for the Wellbeing domains and the EQ-5D Thermometer. However, there were moderate to good correlations at both admission and at discharge across all PT AusTOMs Scales for the Wellbeing domains and the EQ-5D Thermometer. Study limitations and directions for further research Given the number of correlations performed for this study, it is important not to over-interpret the relatively small number of moderate and good correlations found. When considering these findings it is also important to note the relatively small sample size since a larger EQ-5D data set may have produced more, significant correlations. The low EQ-5D return rate from speech pathology is not surprising considering that clients seen by speech pathologists often have communication/ cognitive difficulties, and this increases the difficulty in using a self-administered tool such as the EQ-5D. Clinicians also reported that it was difficult to ask quite acutely unwell clients to complete the EQ-5D although they were able to score the client using the AusTOMs. The validity of a tool is never confirmed. Rather, many studies are required over time to demonstrate that a tool is operating in the manner which developers intended. Future validity studies could investigate the ability of AusTOMs to predict client discharge data from admission status, and to determine the capacity of the tool to discriminate between clients with differing severity levels of impairments and activity limitations. This has already been reported for the physiotherapy profession in relation to the UK TOM [ 4 ]. In addition, it would be interesting to compare therapist ratings of clients on the EQ-5D with client ratings on this tool. Such research would provide greater insights to the issue of how similar client and therapist views of clients' health status are. Future validity studies could also compare client data from the AusTOMs with data from other global measures such as the Medical Outcomes Study (MOS) Short Form Health Survey (SF-36) [ 19 ] or the Nottingham Health Profile (NHP) [ 20 ] measure. Conclusion The EQ-5D is used extensively in cost effectiveness analysis [ 22 ]. It is based on client's self report and is thus consistent with the theoretical basis of economic evaluation as the summation of individual utilities. In contrast, AusTOMs are based in therapists' assessment of clinical progress. In the introduction, it was stated that it might be expected that client scores for these two assessments could be different. However, this study revealed that client and therapist assessment appear to be somewhat similar on some domains, thus lending some support for the construct (convergent) validity of AusTOMs. Yet the fact that more, stronger correlations were not found helps to explain some of the differences in perceptions between policy makers, clients, and therapists. Therapists see a range of clients with a given condition and because of their training and experience, have an understanding of what might be achievable in therapy. Clients, on the other hand, make their assessment based on their own experiences and expectations. The differences between client and therapist expectations could perhaps be minimised with better and clearer communication between client and therapist, although neither party in that dyad may be able to accept the inherent limitations of the rehabilitation process. Nonetheless, client perceptions of the success of therapy are vitally important, and more research is required to investigate reasons for the different perceptions of 'therapy success' of these two groups. The use of different tools across different disciplines to measure improvement can lead to different conclusions about benefits. If outcome measures of cost effectiveness are based on client perceptions it could well be the case that therapy interventions which are seen by therapists to lead to statistically significant improvements in outcome may not be so valued by clients. As a result, those interventions may not be found to be cost effective in an economic sense, if such an evaluation is based on measures of client perception, such as EQ-5D. These differences in perceptions may then contribute to conflict between policy makers, therapists and clients. Alternative economic measures of client outcome, such as return to work, may not be suitable in environments where a significant proportion of clients are beyond working age. Although the tools appear to be measuring somewhat similar constructs, the results of this study suggest that therapy outcome measures such as AusTOMs may need to be supplemented by client-based measures. As part of the treatment process, differences between responses should be discussed to improve understanding between client and therapist about expectations and achievable outcomes from therapy. This may in turn assist goal setting for the therapy process. Authors' contributions CU and SD have made substantial contributions to conception and design of the study, analysis and interpretation of data and have been involved in drafting the article and revising it critically for important intellectual content. DD has made substantial contributions in the acquisition of data, and has been involved in drafting the article and revising it critically for important intellectual content. APand JS have made substantial contributions to conception and design of the study, acquisition of data, and have been involved in revising the article critically for important intellectual content. NT has made substantial contributions in the acquisition of data and has been involved in revising the article critically for important intellectual content. All authors have given final approval of the version to be published.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC543469.xml

529459

Multimodal pressure-flow method to assess dynamics of cerebral autoregulation in stroke and hypertension

Background This study evaluated the effects of stroke on regulation of cerebral blood flow in response to fluctuations in systemic blood pressure (BP). The autoregulatory dynamics are difficult to assess because of the nonstationarity and nonlinearity of the component signals. Methods We studied 15 normotensive, 20 hypertensive and 15 minor stroke subjects (48.0 ± 1.3 years). BP and blood flow velocities (BFV) from middle cerebral arteries (MCA) were measured during the Valsalva maneuver (VM) using transcranial Doppler ultrasound. Results A new technique, multimodal pressure-flow analysis (MMPF), was implemented to analyze these short, nonstationary signals. MMPF analysis decomposes complex BP and BFV signals into multiple empirical modes, representing their instantaneous frequency-amplitude modulation. The empirical mode corresponding to the VM BP profile was used to construct the continuous phase diagram and to identify the minimum and maximum values from the residual BP (BP R ) and BFV (BFV R ) signals. The BP-BFV phase shift was calculated as the difference between the phase corresponding to the BP R and BFV R minimum (maximum) values. BP-BFV phase shifts were significantly different between groups. In the normotensive group, the BFV R minimum and maximum preceded the BP R minimum and maximum, respectively, leading to large positive values of BP-BFV shifts. Conclusion In the stroke and hypertensive groups, the resulting BP-BFV phase shift was significantly smaller compared to the normotensive group. A standard autoregulation index did not differentiate the groups. The MMPF method enables evaluation of autoregulatory dynamics based on instantaneous BP-BFV phase analysis. Regulation of BP-BFV dynamics is altered with hypertension and after stroke, rendering blood flow dependent on blood pressure.

Background Noninvasive assessment of cerebral vasoregulation is a major challenge in medical diagnostics and post-stroke care. Dynamic autoregulatory mechanisms adapt cerebral perfusion to spontaneous variations in intracranial and systemic pressure within a few heartbeats. Decline of cerebral blood flow that occurs with normal aging is further potentiated by presence of risk factors for cerebrovascular disease such as hypertension. Cerebral autoregulation is damaged by acute stroke, rendering cerebral blood flow dependent on blood pressure (BP) [ 1 - 3 ]. The duration of post-stroke autoregulatory impairment and the degree of recovery after stroke are not known. Stroke is more common in older subjects, but the consequences of stroke in younger subjects may last for decades. It is not known if cerebral autoregulation is impaired in subjects with minor chronic infarcts and good neurological outcome. This study, employing a new signal analysis method, addresses the unresolved issue of whether the dynamics of cerebral autoregulation are altered in younger subjects with minor chronic stroke. Continuous monitoring of BFV using transcranial Doppler ultrasound enables assessment of dynamic autoregulation from spontaneous BP and BFV fluctuations [ 4 ], and during interventions inducing a sudden BP reduction, such as the VM, thigh cuff deflation, and the sit-to-stand test [ 5 - 7 ]. The VM induces a sudden increase in intrathoracic and cerebrospinal fluid pressure that is associated with rapid declines in BP and BFV. The BFV response to intracranial pressure changes precedes peripheral BP responses. The end of straining is followed by a BP increase ≈30 mm Hg above baseline, associated with an increase in BFV and cerebrovascular resistance [ 8 , 9 ]. BP and BFV fluctuations evoked by the VM are transient and highly nonstationary. Rapid changes in vascular tone that act to adjust perfusion pressure during these fluctuations may yield a nonlinear pressure/flow relationship. These short and nonstationary time series present a methodological challenge since they are not suitable for analysis using traditional analytic techniques based on Fourier transform analysis [ 4 , 10 ] or Volterra-Wiener moving average modeling [ 11 ] which assume signal linearity and stationarity. Accordingly, we: 1) introduce a new method, multimodal pressure flow analysis (MMPF), based on the Hilbert-Huang transformation [ 12 ], to quantify the relationships between two nonstationary signals; 2) apply the MMPF method to the assessment of dynamic autoregulation using the instantaneous BP-BFV phase relationships during the VM; 3) compare pressure/flow dynamics in younger subjects with a minor chronic stroke to that of normotensive and hypertensive subjects without stroke, and 4) compare the MMPF method with standard indices of autoregulation in the stroke and non-stroke groups. Methods Subjects Studies were conducted at the Autonomic Nervous System Laboratory at the Department of Neurology at The Ohio State University and at the SAFE (Syncope and Falls in the Elderly) Laboratory at the Beth Israel Deaconess Medical Center at Harvard Medical School. All subjects signed informed consent, approved by the Institutional Review Boards. Subjects in the stroke and non-stroke groups were recruited from the Neurology Stroke Service and through advertisement. Demographic characteristics are summarized in Table 1 . Table 1 Demographic characteristics and baseline blood pressure and blood flow velocities in MCAs Group Normotensive Hypertensive Stroke Men/Women 9/6 8/12 5/10 Age (yrs) 40.2 ± 2.0 49.9 ± 2.0 53.1 ± 1.6** Race W/AA 14/1 15/5 14/1 Stroke side, R/L -- -- 4/11 Baseline values mean BP (mm Hg) 84.2 ± 2.2 102.1 ± 3.0 97.2 ± 2.4*** mean BFV MCAR (cm/s) 53.1 ± 3.4 60.7 ± 3.7 59.5 ± 4.3 mean BFV MCAL (cm/s) 53.0 ± 4.3 58.5 ± 3.5 57.2 ± 3.8 EtCO 2 (mm Hg) 37.5 ± 1.9 35.4 ± 1.3 36.2 ± 1.1 Demographic characteristics, age and baseline values (mean ± SE, ** p < 0.003, p < 0.0001), gender and race (W = white, AA = African American) and the stroke side (R/L = right/left), BP = mean blood pressure, BFV = mean blood flow velocity in the right (MCAR) and left (MCAL) middle cerebral artery, EtCO 2 = end tidal carbon dioxide Normotensive group 15 healthy normotensive subjects (age 40.2 ± 2.0 years, mean ± SE). Hypertensive group 20 patients with essential hypertension, controlled by antihypertensive medications (age 49.9 ± 2.0 years). Stroke group 15 subjects with the first minor ischemic stroke (18.3 ± 4.5 months after acute onset) (age 53.1 ± 1.6 years) including 9 subjects treated for hypertension (stroke-hypertensive) and 6 subjects who were normotensive (stroke-normotensive) with no BP treatment. Stroke subjects had a documented infarct on MRI or CT affecting <1/3 of the vascular territory with a minor neurological deficit (Modified Rankin Score scale <3). The side of the lesion was determined by neurological evaluation and confirmed with MRI and CT. The lesion was in the right hemisphere in 4 and in the left hemisphere in 11 subjects. The infarct types and locations were as follows: small vessel infarcts in 13 subjects and large vessel infarcts in 2 subjects. Cortical infarcts were found in 10 subjects and subcortical infarcts in 5 subjects. Normal carotid Doppler ultrasound study was required for participation. Patients with hemorrhagic strokes, clinically important cardiac disease including major arrhythmias, diabetes and any other systemic illness were excluded. All subjects were carefully screened with a medical history, physical and laboratory examination. Subjects with hypertension (with or without stroke) were treated with antihypertensive agents from the following categories (diuretics, beta-adrenergic blockers and angiotensin-converting enzyme inhibitors). Antihypertensive medications were gradually tapered over 3 days and discontinued for 3 days prior to the study. Anticoagulants and other medications that did not affect cardiovascular or autonomic nervous system function were allowed. Experimental protocol VM After instructions and several practice sessions, each subject rested for 5 minutes in the supine position. The subject was then asked to take a breath and expire forcefully through a mouthpiece with a small air-leak, maintaining for 15 seconds a pressure of 40 mm Hg monitored on a pressure gauge connected to the mouthpiece. All data were continuously acquired over the 5 minute period during which BP returned to baseline. The VM was repeated twice and the signal showing the more prominent VM oscillation by visual inspection was selected. Data acquisition, processing and analysis The experiments were done in the morning or > 2 hours after the last meal. The electrocardiogram was measured from a modified standard lead II or III using a Spacelab Monitor (SpaceLab Medical Inc., Issaquah, WA). Beat-to-beat BP was recorded from a finger with a Finapres device (Ohmeda Monitoring Systems, Englewood CO), which is based on a photoplethysmographic volume clamp method. During the study protocol, BP was verified by arterial tonometry. With finger position at the heart level and temperature kept constant, the Finapres device can reliably track intraarterial BP changes over prolonged periods of time [ 13 ]. Respiratory waveforms were measured with a nasal thermistor. CO 2 was measured from a mask using an infrared end tidal volume CO 2 monitor (Datex Ohmeda, Madison WI). The right and left MCAs were insonated from the temporal windows, by placing the 2-MHz probe in the temporal area above the zygomatic arch using a transcranial Doppler ultrasonography system (MultiDop X4, DWL Neuroscan Inc, Sterling, VA). Each probe was positioned to record the maximal BFV and fixed at a desired angle using a three-dimensional positioning system attached to the light-metal probe holder. Special attention was given to stabilize the probes, since their steady position is crucial for reliable, continuous BFV recordings. BFV and all cardiovascular analog signals were continuously acquired at 200 Hz and exported at 50 Hz for off-line post-processing. Data were visually inspected and occasional extrasystoles and outlier data points were removed using linear interpolation. Fourier transform of the Doppler shift (the difference between the frequency of the emitted signal and its echo (frequency of reflected signal) was used to calculate BFV. BFVs in the MCA correlate with invasive measurements of blood flow with xenon clearance [ 14 ], laser Doppler flux [ 15 ] and positron emission tomography [ 16 ]. Since MCA diameter is relatively constant under physiological conditions, BFV can be used for blood flow estimates [ 17 ]. Multimodal pressure-flow method To quantify the dependency between cerebral blood flow and systemic pressure, we developed a novel computational procedure, called multimodal pressure-flow (MMPF) analysis. The MMPF analysis implemented the Hilbert-Huang transformation [ 12 ] technique to measure the coupling between two nonstationary signals. This method was motivated by the fact that the original BP and BFV signals are recorded over time. However, different subjects vary in the amount of time spent in each stage of the VM. Therefore, it is essential to find an alternative coordinate system for both BP and BFV signals that allows for a meaningful, non-time dependent cross referencing of these two signals. Since the complete VM cycle can be treated as a full cycle of BP oscillation, the oscillatory phase of the BP modulation during the VM can serve as such a useful coordinate system. To implement this approach, we first need to calculate how BP phase changes as a function of time. Then we can map the original time-varying BP and BFV signals to the new axis of reference, namely BP oscillatory phase. To precisely calculate the BP phase, we need to extract the characteristic (dominant) BP oscillation induced by the VM. We applied the empirical mode decomposition (EMD) technique developed by Huang et al. [ 12 ] The EMD algorithm decomposes complex signals such as BP and BFV into multiple empirical modes. Each mode represents the frequency-amplitude modulation at a specific time scale. Figure 1A shows the original BP waveform, which is modulated by multiple frequencies corresponding to the systolic peak, dicrotic notch, heart rate, respiratory frequency and BP fluctuations induced by the VM. Figure 1B shows the decomposed empirical modes (1–10) of the original BP signal. Empirical modes 1–5, corresponding to the faster frequencies, were removed from the signal. The remaining lower frequency BP signal, termed the "residual BP" or BP R (shown as thick curve in Figure 1A ), was used to identify the maximum and minimum values during the VM. The empirical mode best representing the dominant BP profile during the VM, denoted as BP VM (mode 6 in this example, plotted as the thick line in Figure 1B ), was visually identified and used for subsequent phase analysis. We followed the same procedure to obtain the residual BFV signal, denoted as BFV R . Figure 1 Schematic diagram showing Hilbert-Huang decomposition of the original blood pressure (BP) signal into the empirical modes corresponding to amplitude-frequency modulation for different time scales. Panel A shows the profile of the BP waveform over the course of the VM: I- indicates the beginning of the maneuver, II- the duration of straining, III-the end of straining and IV- the BP overshoot above baseline. (Note that the transient BP decrease in phase III is due to inspiration.) Panel B shows the empirical modes for each component frequencies and their corresponding amplitudes were detected from the signal (mode 1–10). Empirical modes corresponding to the faster frequencies (modes 1–5) were removed from the original BP and BFV signals. The empirical mode corresponding to the characteristic BP profile induced by the VM (BP VM – mode 6 in this example) was used to obtain phase information. Modes 7–10 reflect BP modulations at slow frequencies. Similarly, the empirical mode corresponding to the characteristic BFV profile was extracted from the raw BFV waveform (not shown). In the next step of the MMPF analysis, we applied the Hilbert transform to the BP VM signal to calculate its instantaneous phases. This phase was then used as a reference coordinate both for BP and BFV signals. From this point on, the term phase refers to the phase of the BP VM oscillation during the VM. Unlike the Fourier transform, the Hilbert transform does not assume that signals are composed of superimposed sinusoidal oscillations of constant amplitude and frequency. Real-world biological fluctuations, such as BP and BFV, are not stationary and are better described by analytical methods that can quantify variations of amplitude and frequency. Mathematically, the first two steps of the MMPF algorithm can be summarized in the following way: Any complex signal s(t) can be represented as the superimposition of more basic (simpler) components: S ( t ) = ∑ k S k ( t ), where S k are empirical modes that fulfill certain criteria of the original signal [ 12 ]. For each empirical mode, its Hilbert transform is defined as: where the Cauchy principal value is taken in the integral. Instantaneous amplitude, A k ( t ), and instantaneous phase, φ k ( t ), can be calculated by The BP VM profile from the first peak at the beginning of the VM to the subsequent BP VM maximum forms a complete phase cycle from 0–360° over 30–40 seconds. We assigned the first peak at the beginning of the VM to phase 0°, the BP VM minimum during the VM to phase 180° and the subsequent BP VM maximum to phase 360°. To quantify the relationships between BP and BFV signals, we measured the BP-BFV phase shift, defined as the difference between the phase at the BP R minimum (maximum) value and the phase at the BFV R minimum (maximum) value. We have calculated phase values for all data points in this interval and, in principle, we can measure the phase shift at all point of the VM cycle. However, for simplicity, we only calculated the phase shift at the minimum and maximum of these two signals for statistical analysis. Since these BP-BFV phase shifts reflect dynamical changes in peripheral and cerebral vascular tone over the course of the VM, they can be used as a sensitive index of cerebral autoregulation dynamics in normal and pathological conditions. Autoregulation indices We also assessed autoregulation using a standard index, calculated using the second-order differential equation model proposed by Tiecks et al[ 5 ] This model assumes a linear flow-pressure relationship and a constant cerebral perfusion pressure over the course of a sudden BP reduction, such as occurs during thigh cuff deflation. This technique can be also applied to the sudden BP decline during the VM. The autoregulation index ranges from 0 = "no autoregulation" to 9 = "the fastest autoregulation;" value 5 reflects "normal autoregulation." We also calculated the "rate of autoregulation" (RoR) using the slope of the linear regression fitted to the original BP and BFV waveforms signals during the period between the baseline and the BP minimum (descending slope) and between BP minimum and maximum (ascending slope). Statistical analysis We used one-way analysis of variance for between-group comparisons of baseline, minimum and maximum BP R and BFV R values and phases. Two-way analysis of variance was used for side-to-side comparisons of BFV between groups (JMP-5.0 SAS Institute, Cary, NC). For the group comparisons, we used the BFV R in the right and left MCAs for the normotensive and hypertensive groups compared to the BFV R in the stroke-side and non-stroke side MCA in the stroke group. Age was different between the groups (p < 0.003). However, age and stroke subtypes had no significant effects when included as co-variants in the analysis. Data are presented as mean ± SE. Results Figure 2A shows representative raw BP and BFV waveforms (right and left MCAs) during the VM for a normotensive 31 year old man. The BP R and BFV R signals are superimposed on the raw waveforms. For comparison, we show similar signals (the right, non-stroke side, MCA and the left, stroke side, MCA) for a 48 year old woman with the left temporal stroke in Figure 2B . The BP R and BFV R for the right (non-stroke side) MCA and for the left (stroke side) MCA are shown in the bottom three panels as a function of the phase. With normal autoregulation, the BFV R changes precede BP R changes over the course of the VM. Therefore, the phase at the BFV R minimum is smaller than the phase at the BP R minimum. (Phase at the BFV R minimum for the right MCA = 129° and the left MCA = 112° vs. the phase at BP R minimum = 178°.) In contrast, with post-stroke cerebral autoregulation, the phase at the BFV R minimum is similar to the phase at the BP R minimum, suggesting that BFV is dependent on blood pressure. (Phase at the BFV R minimum for the non-stroke MCA = 173° and for the stroke MCA = 180° vs. the phase at BP R minimum = 185°). Figure 2 Panel A shows blood pressure (BP) and blood flow velocity (BFV) waveforms from the right and left MCAs (MCAR and MCAL respectively) during the VM for a normotensive subject (top 3 panels). The duration of the VM straining is indicated by a horizontal line. The thick black line indicates the BP R and BFV R that reflect the characteristic VM oscillation. Bottom 3 panels show BP R and BFV R in the MCAR and MCAL. Arrows indicate phases at the BP R and BFV R minima. With normal autoregulation, BFV R minimum preceded BP R minimum. Panel B shows BP and BFV waveforms for a subject with a left temporal infarct (MCAR = non stroke-side MCA, MCAL = stroke side MCA) (top 3 panels). Horizontal line indicates duration of the VM. Black thick line indicates the BP R and BFV R obtained from the BP and BFV raw waveforms. Bottom 3 panels show BP R and BFV R in the non-stroke side MCA and in the stroke-side MCA expressed as a function of BP VM phase. Arrows indicate that the phase at BFV minimum was similar to the phase at BP minimum. Pressure flow relationship at the BFV R minimum Figure 3A shows the group averages of BFV R values and the phases at the BFV R minimum and maximum values for the right MCA in the normotensive and hypertensive groups and for the non-stroke side MCA in the stroke group. Figure 3B shows the BFV R values and the phases for the left MCA in the normotensive and hypertensive group, and the stroke side MCA in the stroke group. Mean BP R values and corresponding phases are shown in panel C. The phase at BFV R minimum was different between groups for the right (non-stroke side) (p = 0.0005) and left (stroke side) (p = 0.004) MCAs. In the normotensive group, the phase at BFV R minimum was shorter than the phase at BP R minimum. In the stroke group, the phase at BFV R minimum was similar to the phase at BP R minimum. The phase at BFV R minimum was greater in the non-stroke (p = 0.002) and stroke (p = 0.03) MCAs, compared to the normotensive group. In the hypertensive group, the phase at BFV R minimum was also greater for the right (p = 0.0007) and left (p = 0.006) MCAs compared to the normotensive group. No significant differences were found between the stroke and hypertensive groups. The phase at BP R minimum was similar between groups (normotensive group = 172.7 ± 3.9°, hypertensive group = 178.6 ± 2.0°, stroke group = 178 ± 1.9°). Average BP R values were higher in the stroke and hypertensive groups compared to the normotensive group at baseline (p = 0.001), at BP R minimum (p = 0.054) and at BP R maximum (p = 0.0001) (Figure 3C ). Average BFV R values at baseline, at BFV R minimum and maximum, and BFV R change from baseline were not different among groups. Average BP R change and percent change from baseline to BP R minimum and maximum were not different. Average BFV R change and percent change from baseline to BFV minimum and maximum were also not different. Figure 3 Panel A shows the phase and corresponding residual blood flow velocity (BFV R ) values at baseline, BFV R minimum and BFV R maximum for the right MCA for the normotensive -●- and - -▼- hypertensive groups and for - O- the non-stroke side MCA in the stroke group. Panel B shows the phase and corresponding BFV R values for the left MCA in the normotensive and hypertensive groups and for the stroke side MCA in the stroke group. BFV R phase was significantly greater in the stroke and hypertensive groups compared to the normotensive group for BFV R minimum and maximum in both MCAs (between groups phase comparisons *** p < 0.005, ** p < 0.01). Panel C shows the phase and corresponding residual blood pressure (BP R ) values for the BP R minimum and maximum (between groups BP R values comparisons +++ p < 0.001, mean ± SE). Pressure flow relationship at the BFV R maximum In the normotensive group, the phase at BFV R maximum preceded the phase at BP R maximum (Figure 3 ). The phase at BFV R maximum was different between groups for the right (non-stroke side) (p = 0.009) and left (stroke side) (p = 0.003) MCAs. In the stroke group, the phase at BFV R maximum was similar to the phase at BP R maximum. The phase was greater for the non-stroke side (p = 0.008) and stroke side (p = 0.03) MCAs, compared to the normotensive group. In the hypertensive group, the phase at BFV R maximum was also greater for the right (p = 0.005) and left (p = 0.009) MCAs compared to the normotensive group. The phase at BP R maximum was similar among groups (normotensive group = 354 ± 4.7°, hypertensive group = 360.1 ± 2.2° stroke group = 364.1 ± 4.3°). BP-BFV phase shifts Figure 4A summarizes the differences between the phases at BP R and BFV R minimum and between the phases at BP R and BFV R maximum for the right MCA in normotensive and hypertensive groups and for the non-stroke side MCA in the stroke group. The BP-BFV phase shifts at the minimum points were smaller in the stroke and hypertensive groups compared to the normotensive group (p = 0.009). The BP-BFV phase shifts at the maximum points were smaller in stroke and hypertensive groups compared to the normotensive group (p = 0.03). Figure 4B shows the phase shift for the left MCA in normotensive and hypertensive groups and for the stroke-side MCA in the stroke group. The BP-BFV phase shifts at the minima and maxima were also smaller in the stroke and hypertensive groups compared to the normotensive group (p = 0.0002). The BP-BFV phase shifts at the maxima were smaller in the stroke and hypertensive groups compared to the normotensive group (p = 0.008). In the stroke group, the BP-BFV phase shift at the maxima was greater compared to the phase shift at the minima for the non-stroke MCA (p = 0.02). The BP-BFV phase shifts at the minima and maxima for the stroke MCA did not reach statistical significance (p = 0.08). The BP-BFV phase shifts between the stroke and non-stroke MCA at the minima and maxima were not different. In the normotensive group, the phase difference between BP R and BFV R minima was about 60 degrees corresponding to a time difference of about 3.6 seconds. In contrast, in the stroke and hypertensive groups, the time difference between BFV and BP phases was < 0.5 second. Figure 4 Panel A shows the phase shift between BP R minimum and BFV R minimum and the phase shift between BP R maximum and BFV R maximum for the right MCA for the □ normotensive and hypertensive groups and ■ for the non-stroke side MCA in the stroke group. Figure 4B shows the phase shift between BP R minimum and BFV R minimum and the phase shift between BP R maximum and BFV R maximum for the left MCA for the normotensive, and hypertensive groups and for the stroke side MCA in the stroke group. Phase shift was greater in the normotensive compared to other groups (between group comparisons *** p < 0.005, ** p < 0.01 *p < 0.05, mean ± SE). Autoregulation indices The standard autoregulation index was not different between groups and between MCAs in both hemispheres (Table 2 ). The rate of autoregulation (RoR) of BFV responses to BP reduction and increases during the VM were also not different (Table 2 ). Table 2 Autoregulation Indices Variable Normotensive Hypertensive Stroke MCA side Right Right Non-stroke-side ARI 5.7 ± 3.3 6.1 ± 2.6 6.1 ± 2.6 RoR – descending slope 0.7 ± 0.1 1.1 ± 0.6 0.6 ± 0.2 RoR – ascending slope 0.5 ± 0.1 1.6 ± 1.0 0.6 ± 0.2 MCA side Left Left Stroke-side ARI 5.9 ± 3.2 5.9 ± 2.6 5.1 ± 1.9 RoR-descending slope 0.6 ± 0.1 0.8 ± 0.4 0.5 ± 0.2 RoR-ascending slope 0.5 ± 0.1 1.5 ± 1.0 0.6 ± 0.2 Autoregulation index (ARI) and the rate of autoregulation (RoR) for the right and left MCAs in the normotensive and hypertensive groups and for the non-stroke and stroke side MCA in the stroke group. RoR – descending slope of the linear portion of the BP and BFV reduction between the baseline and BP minimum. RoR – ascending slope of the linear portion of the BP and BFV increase between BP minimum and maximum during the VM. ARI and RoR were not significantly different between the groups. Data are presented as mean ± SE Stroke-normotensive and stroke-hypertensive subjects In a subset analysis, we separated stroke-normotensive (N = 6) and stroke-hypertensive subjects (N = 9) and compared them to the non-stroke normotensive and hypertensive groups. For the non-stroke side MCA, the phases at BFV R minima (p = 0.01) and maxima (p = 0.04) were greater in the stroke-normotensive group compared to the normotensive group. For the stroke side MCA, the phases at BFV R minima (p = 0.04) and maxima (NS, p = 0.08) were greater in the stroke-normotensive group compared to the normotensive group. The phases at BFV R minima and maxima were not different between the stroke-hypertensive and hypertensive groups for both MCAs. No significant differences were found between the stroke and non-stroke side MCA. Discussion This study introduces a new technique, based on the Hilbert-Huang transformation [ 12 , 18 ], termed multimodal pressure-flow analysis, for assessing the relationships between systemic blood pressure and cerebral blood flow changes associated with provocative maneuvers. Development of this method was motivated by the facts that 1) that the duration of the VM stages and resulting BP and BFV responses vary over time and among subjects, and 2) these types of time series are short and nonstationary, and therefore, not suitable for analysis using standard Fourier transform and autoregressive type approaches. We implemented the MMPF method to evaluate the dynamics of cerebral autoregulation using the instantaneous systemic BP and MCA BFV phase relationships during the VM. The frequency and corresponding BP R and BFV R amplitudes were computed for each data sample to construct a continuous phase diagram. The BP R and BFV R profiles were similar over the course of the VM, but the phase relationships were different. The autoregulation indices, calculated using the standard methods, did not differentiate the groups. Cerebral vasoregulation compensates for rapid BP and BFV transitions over the course of the VM. A sudden and parallel increase in intrathoracic [ 19 ] and cerebrospinal fluid [ 8 ] pressures is associated with a rapid decline in BP and an initial increase of cerebrovascular resistance, resulting in a rapid decline in BFV. With active autoregulation, cerebrovascular resistance diminishes, enabling BFV to recover in the face of falling perfusion pressure, and BFV responses in the MCA precede the systemic BP changes. Therefore, in healthy controls, the phases corresponding to the BFV R minimum and maximum were smaller than BP R phases, reflecting an active vasoregulatory process. With delayed or impaired autoregulation, BFV becomes synchronized with blood pressure. In the stroke group, the BFV R and BP R phase diagrams were similar, suggesting that cerebral blood flow was entrained by systemic BP. In the stroke and hypertensive groups, BFV phases at the BFV R minimum and maximum were greater compared to the normotensive group. The BP-BFV phase shifts were smaller in the stroke and hypertensive group, compared to healthy controls. Methods evaluating dynamic cerebral autoregulation that use the Fourier transform-based coherence and transfer functions assume a linear relationship between stationary signals. Coherence, phase and gain derived from the transfer function of spontaneous BP and BFV fluctuations have been used to assess autoregulation. These analyses have shown a significant phase lead of cerebral BFV with respect to systemic BP [ 4 , 20 , 21 ]. However, assumptions about signal stationarity and a linear flow/pressure relationship are not met for the short nonstationary time series from the VM, and therefore transfer function gain was not evaluated. The joint time-frequency distributions [ 22 ], such as the new MMPF developed here, that make no assumptions about signal characteristics and can reliably track simultaneous changes of spectral powers and frequencies, are better suited for these short nonstationary signals. The variable time delay between BP and BFV suggests that the relationship between the cerebral blood flow and systemic pressure is not linear. Previous studies indicated bilateral impairment of the dynamics of pressure autoregulation after an acute [ 2 , 3 ] and subacute ischemic stroke [ 23 ]. Dynamic indices of autoregulation that were calculated from spontaneous BP and BFV fluctuations were altered, with no significant difference between autoregulatory indices in the affected and unaffected hemispheres. No significant differences were found in autoregulatory indices between large vessel anterior and posterior circulation infarcts and lacunar infarcts [ 2 ]. In patients with large hemispheric strokes, head elevations from 0° to 45° induced reductions in arterial BP, BFV, intracranial and perfusion pressures [ 1 ]. We have reported [ 24 ] that cerebral vasomotor responses to hypocapnia and hypercapnia were diminished following minor chronic infarctions. Baseline BFVs in the MCAs were similar between the stroke and the non-stroke groups, but differed during head-up tilt. BFV declined in the stroke side MCA during the head-up tilt. Side-to-side BFV differences were the most prominent in stroke-normotensive subjects with lower BP during head-up tilt compared to stroke-hypertensive subjects and non-stroke groups. The present study has confirmed bilateral impairment of autoregulation dynamics in stroke-normotensive and stroke-hypertensive subjects and also in a non-stroke hypertensive group. The BP-BFV phase relationships suggest that the autoregulatory responses were delayed in both hemispheres and that the BFV responses were dependent on perfusion pressure. About one third of stroke patients are hypertensive upon hospital admission. Hypertension and stroke may exert similar pathophysiological effects on vascular compliance, sympatho-vagal interactions [ 25 ] and blood pressure regulation [ 26 , 27 ]. Increased vascular stiffness, impaired vasodilatation and shift of the autoregulatory responses toward higher BP values may also affect the timing of autoregulatory responses in both hemispheres. There are several limitations of this study: 1) The MMPF method was implemented for the VM, which is widely used for clinical autonomic testing. The VM allows noninvasive evaluation of pressure autoregulation, and testing can be completed in less than 5 minutes. However, the VM requires active patient cooperation, and may not be advisable in acute stroke settings where change in intracranial pressure should be avoided. 2) This study evaluated a population of younger subjects with minor stroke. A larger cohort is needed to determine the effects of ischemic stroke subtypes on the dynamics of autoregulation. 3) Age was different between groups; however, it had no significant effect on BP-BFV relationship in our analysis. Aging and cardiovascular risk factors exert significant but distinct effects on regulation of cerebral blood flow. The vasomotor reactivity to hypercarbia declines with aging and hypertension, while the dynamics of pressure regulation can be preserved [ 7 ]. This effect may be in part due to a shift of the autoregulatory range toward higher blood pressure values. Conclusions Multimodal pressure-flow analysis is a new method that enables evaluation of short nonstationary time-series not suitable for Fourier–based techniques. The MMPF method provides high time and frequency resolution and permits construction of instantaneous phase diagrams on a beat-to-beat basis. This method may be particularly useful as a complementary measure of cerebral autoregulation for the short and nonstationary time series acquired during provocative interventions such as the VM. Application of this method reveals that the regulation of BP-BFV dynamics is altered in both hemispheres after minor stroke, rendering blood flow dependent on blood pressure. Hypertension without stroke is also associated with delayed BP-BFV dynamics. List of abbreviations BP = blood pressure BP R = residual BP calculated by MMPF method BP VM = empirical mode of BP corresponding to the dominant VM oscillation BFV = blood flow velocity in the MCA BFV R = residual BFV calculated by MMPF method EMD = empirical mode decomposition HHT = Hilbert-Huang Transform MCA = middle cerebral artery MMPF = multimodal pressure-flow analysis VM = Valsalva maneuver Competing interests The author(s) declare that they have no competing interests. Authors' contributions VN – designed the study, conducted the experiments, participated in the analysis and wrote the first draft of the manuscript. ACY – contributed to the MMPF development and analysis; LL – conducted the data and statistical analysis. ALG – contributed to the method application, data interpretation and manuscript preparation. LAL – contributed to data interpretation and manuscript preparation. CKP – designed and developed MMPF method, contributed to data analysis and manuscript preparation.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529459.xml

555948

Time trends in the impact factor of Public Health journals

Background Journal impact factor (IF) is linked to the probability of a paper being cited and is progressively becoming incorporated into researchers' curricula vitae . Furthermore, the decision as to which journal a given study should be submitted, may well be based on the trend in the journal's overall quality. This study sought to assess time trends in journal IF in the field of public, environmental and occupational health. Methods We used the IFs of 80 public health journals that were registered by the Science Citation Index from 1992 through 2003 and had been listed for a minimum period of the previous 3 years. Impact factor time trends were assessed using a linear regression model, in which the dependent variable was IF and the independent variable, the year. The slope of the model and its statistical significance were taken as the indicator of annual change. Results The IF range for the journals covered went from 0.18 to 5.2 in 2003. Although there was no statistical association between annual change and mean IF, most of the fastest growing journals registered mean IFs in excess of 1.5, and some represented emerging areas of public health research. Graphs displaying IF trends are shown. Conclusion In view of the delay between the publication of IFs and that of any given paper, knowing the trend in IF is essential in order to make a correct choice of journal.

Background Scientific journal impact factor (IF) is directly linked to the probability of a paper being cited. It is currently accepted that a higher IF indicates a better journal quality, existing a correlation among IF and quality indicators at least in some health disciplines [ 1 ]. As a result these indices are progressively becoming incorporated into researchers' curricula vitae . In Spain, publication in top-IF journals has occupational implications in terms of academic careers and obtaining research grants [ 2 ]. The most widespread and important bibliometric indicators are those referring to the repercussion of scientific activity, and among these, IF has a leading role [ 3 ]. This pressure means that when it comes to having to select a journal in which to publish their studies, researchers turn to journals with IF, and assess the possibility of publishing in those that have the highest IF possible. Journals having the top IF within each medical specialty tend to be those with the greatest international prestige and highest profile, i.e., the most widely read by researchers and most in demand to publish their studies. However, the publication of extraordinary relevant scientific findings are concentrated in a small list of core journals, most of them not directly related with the specialty of the author [ 1 , 4 ] Positive evaluation of IF which failed to take account of its trend over time would tend to favor publications of recent, but not necessarily lasting, interest. In contrast, publications of steadily growing interest and stable impact would be undervalued, though there may be differences of opinion about this [ 5 ]. Hence, in addition to journal quality, the decision as to which journal a manuscript should be submitted may be based on the trend in its IF over time. The aim of this study was to analyze IF time trends of journals in the "Public, Environmental and Occupational Health" category, thereby furnishing a new criterion on which to base the choice of journal for publication. Methods For study purposes, we selected 80 journals that: were included in the "Public, Environmental and Occupational Health" category of the hard copy version of the Journal Citation Reports (JCR) from 1992 through 2003 [ 6 ]; were listed for a minimum period of 3 consecutive years; and had IFs in the JCR for 2003. We consulted the 2003 JCR-IFs via the ISI Web of Knowledge [ 7 ]. The impact factor is one of the quantitative tools provided by JCR for ranking, evaluating, categorizing, and comparing journals. The annual impact factor of a journal is calculated by dividing the number of current year citations to the source items published in that journal during the previous two years [ 6 ]. Impact factor time trends were assessed using a linear regression model, in which the dependent variable was IF and the independent variable, the year. The slope of the model (index of annual change-IAC) and its statistical significance were taken as the indicator of year-to-year variation. Results Shown in Table 1 is a list of all journals included, ranked by their impact factor in 2003. This table also shows the index of annual change (slope of the model) and its statistical significance. Table 2 shows the same list, but ranked this time according to the IAC. This method of ranking can prove useful, since, by comparing the IF trends between two journals with similar IFs, the better choice would be the journal with the better index of annual change. Table 1 Journals ranked by Impact Factor (IF) in 2003 Title IAC IF(2003) mean p-value ANNU. REV. PUBL. HEALTH 0.225 5.179 3.158 0.010 CANCER. EPIDEMI. BIOMAR 0.214 4.720 3.475 0.001 AM. J. EPIDEMIOL 0.099 4.486 3.788 0.000 EPIDEMIOLOGY 0.250 4.220 3.093 0.000 AM. J. PUBLIC. HEALTH 0.076 3.363 3.057 0.010 EPIDEMIOL. REV -0.175 3.306 3.203 0.076 INT. J. EPIDEMIOL 0.125 3.289 1.820 0.001 AM. J. PREV. MED 0.232 3.256 1.440 0.000 TOBACC. CONTROL 0.509 3.164 2.052 0.181 MED. CARE 0.105 3.152 2.379 0.004 ENVIRON. HEALTH. PERSP 0.220 3.038 2.192 0.000 DRUG. SAFETY 0.238 2.971 2.059 0.000 CANCER. CAUSE. CONTROL 0.087 2.726 2.623 0.061 B. WORLD. HEALTH. ORGAN 0.115 2.442 1.838 0.001 ANN. EPIDEMIOL 0.141 2.345 1.995 0.001 J. EPIDEMIOL. COMMUN. H 0.075 2.332 1.679 0.003 PSYCHIATR. SERV 0.138 2.274 1.658 0.004 GENET. EPIDEMIOL 0.018 2.265 1.681 0.544 J. CLIN. EPIDEMIOL 0.065 2.227 1.872 0.001 TROP. MED. INT. HEALTH 0.181 2.156 1.477 0.003 TR. ROY. SOC. TROP. MED. H 0.064 2.114 1.553 0.001 AM. J. TROP. MED. HYG 0.019 2.105 1.950 0.058 QUAL. LIFE. RES -0.171 2.000 2.089 0.149 INFECT. CONT. HOSP. EP 0.081 1.951 2.074 0.035 PREV. MED 0.013 1.889 1.540 0.568 J. MED. SCREEN -0.033 1.867 1.815 0.696 OCCUP. ENVIRON. MED 0.100 1.847 1.755 0.013 SCAN. J. WORK. ENV. HEA 0.075 1.816 1.433 0.001 NEUROEPIDEMIOLOGY 0.095 1.762 1.411 0.001 J. ADOLESCENT. HEALTH 0.082 1.674 1.361 0.000 PAEDIATR. PERINAT. EP 0.132 1.673 1.176 0.005 J. WOMEN. HEALTH. GEN. B 0.388 1.561 0.928 0.007 AM. J. IND. MED 0.049 1.542 1.256 0.001 EPIDEMIOL. INFECT 0.023 1.509 1.594 0.199 J. OCCUP. ENVIRON. MED 0.081 1.472 1.349 0.121 QUAL. HEALTH. CARE 0.056 1.466 1.232 0.221 J. AEROSOL. MED 0.056 1.459 0.818 0.006 ENVIRON. RES 0.058 1.452 1.390 0.068 INT. ARCH. OCC. ENV. HEA 0.026 1.388 1.086 0.072 ANN. OCCUP. HYG 0.070 1.357 1.041 0.002 J. URBAN. HEALTH 0.316 1.286 0.723 0.002 EUR. J. PUBLIC. HEALTH 0.002 1.281 1.044 0.983 J. EXPO. ANAL. ENV. EPID 0.124 1.263 1.033 0.001 PALLIATIVE. MED -0.103 1.185 1.627 0.060 PUBLIC. HEALTH. REP 0.025 1.139 1.012 0.192 STAT. MED 0.030 1.134 1.238 0.094 PATIENT. EDUC. COUNS 0.103 1.130 0.774 0.001 COMUNITY. DENT. ORAL 0.069 1.100 0.976 0.002 INT. J. HYG. ENVIR. HEAL 0.302 1.085 0.822 0.142 J. OCCUP. HEALTH -0.040 1.047 1.049 0.445 SCAN. J. PUBLIC. HEALT 0.207 1.018 0.714 0.044 ANN. TROP. MED. PARASIT 0.052 1.010 0.837 0.000 J. PUBLIC. HEALTH. DENT -0.010 1.000 0.787 0.568 J. PUBLIC. HEALTH. MED 0.032 0.973 0.805 0.014 EUR. J. EPIDEMIOL 0.022 0.972 0.676 0.080 AVIAT. SPACE. ENVIR. MD 0.075 0.946 0.681 0.007 FLUORIDE 0.034 0.907 0.560 0.018 ANN. HUM. BIOL 0.026 0.885 0.787 0.001 ARCH. ENVIRON. HEALTH -0.054 0.878 1.391 0.028 J. SCHOOL. HEALTH 0.039 0.868 0.688 0.185 ANN. AGR. ENV. MED 0.216 0.827 0.590 0.065 HEALTH. PHYS 0.012 0.777 0.865 0.385 J. ENVIRON. SCI. HEAL. B -0.034 0.758 0.718 0.131 INT. J. TECHNOL. ASSESS 0.013 0.754 0.922 0.686 SOZ. PREVENTIV. MED 0.170 0.750 0.525 0.013 PUBLIC. HEALTH 0.025 0.697 0.522 0.010 OCCUP. MED. OXFORD 0.041 0.693 0.464 0.010 BIOMED. ENVIRON. SCI -0.036 0.609 0.557 0.596 AIHAJ 0.180 0.601 0.449 0.166 INT. J. ENVIRON. HEAL. R 0.094 0.588 0.419 0.088 ENVIRON. GEOCHEM. HLTH 0.027 0.565 0.369 0.082 INDOOR. BUILT. ENVIRON 0.085 0.525 0.496 0.542 TOXICOL. IND. HEALTH 0.106 0.508 1.051 0.206 REV. EPIDEMIOL. SANTE 0.024 0.485 0.401 0.003 IND. HEALTH 0.015 0.474 0.497 0.262 TROP. DOCT 0.022 0.347 0.326 0.089 J. ENVIRON. HEALTH 0.021 0.341 0.228 0.007 J. PUBLIC. HEALTH. POL -0.023 0.314 0.615 0.675 WILD. ENVIRON. MED -0.019 0.280 0.339 0.822 B. SOC. PATHOL. EXOT -0.092 0.183 0.262 0.154 IAC = index of annual change Table 2 Journals ranked in descending order, by index of annual change (IAC). Title IAC IF(2003) mean p-value TOBACC. CONTROL 0.509 3.164 2.052 0.181 J. WOMEN. HEALTH. GEN. B 0.388 1.561 0.928 0.007 J. URBAN. HEALTH 0.316 1.286 0.723 0.002 INT. J. HYG. ENVIR. HEAL 0.302 1.085 0.822 0.142 EPIDEMIOLOGY 0.250 4.220 3.093 0.000 DRUG. SAFETY 0.238 2.971 2.059 0.000 AM. J. PREV. MED 0.232 3.256 1.440 0.000 ANNU. REV. PUBL. HEALTH 0.225 5.179 3.158 0.010 ENVIRON. HEALTH. PERSP 0.220 3.038 2.192 0.000 ANN. AGR. ENV. MED 0.216 0.827 0.590 0.065 CANCER. EPIDEMI. BIOMAR 0.214 4.720 3.475 0.001 SCAN. J. PUBLIC. HEALT 0.207 1.018 0.714 0.044 TROP. MED. INT. HEALTH 0.181 2.156 1.477 0.003 AIHAJ 0.180 0.601 0.449 0.166 SOZ. PREVENTIV. MED 0.170 0.750 0.525 0.013 ANN. EPIDEMIOL 0.141 2.345 1.995 0.001 PSYCHIATR. SERV 0.138 2.274 1.658 0.004 PAEDIATR. PERINAT. EP 0.132 1.673 1.176 0.005 INT. J. EPIDEMIOL 0.125 3.289 1.820 0.001 J. EXPO. ANAL. ENV. EPID 0.124 1.263 1.033 0.001 B. WORLD. HEALTH. ORGAN 0.115 2.442 1.838 0.001 TOXICOL. IND. HEALTH 0.106 0.508 1.051 0.206 MED. CARE 0.105 3.152 2.379 0.004 PATIENT. EDUC. COUNS 0.103 1.130 0.774 0.001 OCCUP. ENVIRON. MED 0.100 1.847 1.755 0.013 AM. J. EPIDEMIOL 0.099 4.486 3.788 0.000 NEUROEPIDEMIOLOGY 0.095 1.762 1.411 0.001 INT. J. ENVIRON. HEAL. R 0.094 0.588 0.419 0.088 CANCER. CAUSE. CONTROL 0.087 2.726 2.623 0.061 INDOOR. BUILT. ENVIRON 0.085 0.525 0.496 0.542 J. ADOLESCENT. HEALTH 0.082 1.674 1.361 0.000 INFECT. CONT. HOSP. EP 0.081 1.951 2.074 0.035 J. OCCUP. ENVIRON. MED 0.081 1.472 1.349 0.121 AM. J. PUBLIC. HEALTH 0.076 3.363 3.057 0.010 AVIAT. SPACE. ENVIR. MD 0.075 0.946 0.681 0.007 J. EPIDEMIOL. COMMUN. H 0.075 2.332 1.679 0.003 SCAN. J. WORK. ENV. HEA 0.075 1.816 1.433 0.001 ANN. OCCUP. HYG 0.070 1.357 1.041 0.002 COMUNITY. DENT. ORAL 0.069 1.100 0.976 0.002 J. CLIN. EPIDEMIOL 0.065 2.227 1.872 0.001 TR. ROY. SOC. TROP. MED. H 0.064 2.114 1.553 0.001 ENVIRON. RES 0.058 1.452 1.390 0.068 J. AEROSOL. MED 0.056 1.459 0.818 0.006 QUAL. HEALTH. CARE 0.056 1.466 1.232 0.221 ANN. TROP. MED. PARASIT 0.052 1.010 0.837 0.000 AM. J. IND. MED 0.049 1.542 1.256 0.001 OCCUP. MED. OXFORD 0.041 0.693 0.464 0.010 J. SCHOOL. HEALTH 0.039 0.868 0.688 0.185 FLUORIDE 0.034 0.907 0.560 0.018 J. PUBLIC. HEALTH. MED 0.032 0.973 0.805 0.014 STAT. MED 0.030 1.134 1.238 0.094 ENVIRON. GEOCHEM. HLTH 0.027 0.565 0.369 0.082 ANN. HUM. BIOL 0.026 0.885 0.787 0.001 INT. ARCH. OCC. ENV. HEA 0.026 1.388 1.086 0.072 PUBLIC. HEALTH 0.025 0.697 0.522 0.010 PUBLIC. HEALTH. REP 0.025 1.139 1.012 0.192 REV. EPIDEMIOL. SANTE 0.024 0.485 0.401 0.003 EPIDEMIOL. INFECT 0.023 1.509 1.594 0.199 EUR. J. EPIDEMIOL 0.022 0.972 0.676 0.080 TROP. DOCT 0.022 0.347 0.326 0.089 J. ENVIRON. HEALTH 0.021 0.341 0.228 0.007 AM. J. TROP. MED. HYG 0.019 2.105 1.950 0.058 GENET. EPIDEMIOL 0.018 2.265 1.681 0.544 IND. HEALTH 0.015 0.474 0.497 0.262 INT. J. TECHNOL. ASSESS 0.013 0.754 0.922 0.686 PREV. MED 0.013 1.889 1.540 0.568 HEALTH. PHYS 0.012 0.777 0.865 0.385 EUR. J. PUBLIC. HEALTH 0.002 1.281 1.044 0.983 J. PUBLIC. HEALTH. DENT -0.010 1.000 0.787 0.568 WILD. ENVIRON. MED -0.019 0.280 0.339 0.822 J. PUBLIC. HEALTH. POL -0.023 0.314 0.615 0.675 J. MED. SCREEN -0.033 1.867 1.815 0.696 J. ENVIRON. SCI. HEAL. B -0.034 0.758 0.718 0.131 BIOMED. ENVIRON. SCI -0.036 0.609 0.557 0.596 J. OCCUP. HEALTH -0.040 1.047 1.049 0.445 ARCH. ENVIRON. HEALTH -0.054 0.878 1.391 0.028 B. SOC. PATHOL. EXOT -0.092 0.183 0.262 0.154 PALLIATIVE. MED -0.103 1.185 1.627 0.060 QUAL. LIFE. RES -0.171 2.000 2.089 0.149 EPIDEMIOL. REV -0.175 3.306 3.203 0.076 The IF range for the journals covered went from 0.18 to 5.2 in 2003. Although there was no statistical association between annual change and mean IF, most of the fastest growing journals registered mean IFs in excess of 1.5 (Journal of Womens Health and Gender Based Medicine IAC = 0.388 p = 0.007, Journal of Urban Health IAC = 0.316 p = 0.002, Epidemiology IAC = 0.250 p < 0.001, Drug Safety IAC = 0.238 p < 0.001), and some represented emerging areas of public health research (Table 2 ). Figure 1 depicts IF trends on a multiple graph, thereby allowing a quick idea to be formed of the evolution of the indicator in all the journals included. Figure 1 Time trends in the impact factor of Public Health journals by alphabetical order. If journals that publish review papers are excluded, then the "American Journal of Epidemiology" ranked first in 1991, a position occupied in 1992 by a recently created publication with a fairly specific content matter, the "Cancer Epidemiology Biomarkers & Prevention", in tandem with a journal of similar orientation and seniority, viz., "Cancer Causes and Control". The rise of both these journals may be indicative of current priorities in epidemiologic research [ 8 , 9 ]. This upward trend pattern among journals addressing cancer epidemiology remained in evidence throughout the study period. In 2003 the epidemiology journals with the highest IF ranged from 1.5 through 4.5, though it should be stressed that special caution is called for when dealing with the individual journal lists for the respective medical specialties. Discussion The use of IF as an indicator of a journal's profile or prestige has become widespread among researchers, editors, libraries, and even among the agencies that fund research. Nevertheless, this indicator has a number of limitations that have been extensively debated in the literature [ 5 , 10 ]. Thus, for instance, journals that publish review papers receive a high number of citations and their impact factors are particularly high. It should be pointed out here that the ISI seeks to offer an overview of international science, with the result that journals covering topics or disciplines of more local interest are scarcely covered. Within each category, therefore, it is frequent for journals that are more basic -and thus of universal interest- to be associated with a higher impact factor than those that are more applied -and so of more local interest- given that the latter's circulation is more restricted. The publication of extraordinary relevant scientific findings are concentrated in a small list of core journals, most of them not directly related with the specialty of the author [ 1 , 4 ]. The mere quality of the documents published by a journal, albeit essential, will not suffice for it to be cited. The number of citations can be enhanced using management techniques, such as expanding a journal's international circulation, raising its profile in databases and on web pages, and increasing the number of papers. It becomes necessary for journals to be known among the international community and attain sufficient prestige to be subsequently cited. The policy pursued in Spain in recent years aimed at fostering high-quality, competitive science has induced Spanish scientists to bypass Spanish journals and send their publications instead to journals enjoying a wide international circulation, something that is often associated with journals having the greatest IF. Our results suggest that it would be of interest to add the index of annual change to the criteria used for selecting a journal for publication. Studies conducted in another JCR area have shown the importance of analyzing IF trends by category [ 11 ]. IF trend might, however, be determined by certain factors that should be discussed. The journals analyzed may be assigned to more than two categories in the SCI-JCR. The Public Health category changed names in 1996, and from 1997 onwards was called "Public, Environmental and Occupational Health", thus explaining why the number of titles jumped from 61 to 73 from one year to the next (Table 3 ). Table 3 Trend in the number of journals classified in the Public, Environmental and Occupational Health category. Year 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 No. of journals 58 60 58 60 61 73 80 84 87 88 90 89 Similarly, until the year 2000, the "Epidemiology" category came within the umbrella of "Public Health", thereafter disappearing and falling within the category of study without any specific entry. The categories covered by the JCR have changed in name and number over the years and, logically, this is equally true of the journals that comprise them. In addition to the comments regarding the list of journals included in the category of study, it would be a wise decision on the ISI's part if allocation of journals to categories were made in agreement with the researchers [ 9 ]. However, it is not the intention of this paper to dwell on the use and abuse of the IF or the arbitrariness of the study-category journal list, an aspect previously analyzed in connection with the public health sphere [ 10 ]. This study solely included journals listed for 3 years in the JCR. Recently launched journals with a policy geared to novel publication, such as the BMC group, were not taken into consideration. The "BMC Public Health" journal has been listed for two years, with IFs of 0.29 in 2002 and 0.93 in 2003, and plots a growth pattern similar to that of journals addressing emerging issues. In general terms and in view of Figure 1 , the linear model seems to be adequate, inasmuch as there are very few journals with trends that display evident turning points. Many of the journals maintain their trend over the years. Most of them (60%) add 0.03 points or more to their IFs every year and there are very few that register a decline with time; indeed this is statistically significant in only one instance. A number of categories can be drawn up based on the trend pattern, namely: long-standing journals with the top IF, which maintain their trend; new journals focused on emerging issues, which seem to enjoy good acceptance; journals that maintain a very stable intermediate ranking; and a small group that has witnessed a decrease in their respective IFs. The reasons why a journal changes its IF trend has been commented before and some of them does not have any relationship with a better quality of its papers. Probably a better or worst management of the journal also is related with the changes in the citations trend and could deserve some study. In the publication of a scientific paper, a long time elapses between deciding upon a journal and the date of publication: on average, more than one year can go by. In view of the stability of the indicators in the area of study targeted, relying upon IF time trends in order to choose a particular journal might perhaps not be very relevant. Yet it may be a critical aspect in other areas of science where there are increases of around one point per year. For researchers/authors who know only too well how costly it can prove to see their paper published and are, moreover, aware that it is going to be evaluated on the basis of concepts as abstract as the impact and number of their publications, this decision is important. Conclusion Leaving aside the speculative components of the choice, and given the delay that accumulates between the publication of IFs and that of any given paper, knowing the trend in the IF is yet another factor that will help authors make the correct choice of journal. Competing interests The author(s) declare that they have no competing interests. Authors' contributions CMT was responsible for the development of intellectual content and the study design, collected the data, data coding and entry, interpretation of the results, manuscript drafting and the critical revisions of manuscript. GLA was responsible for the development of intellectual content, statistical analyses, interpretation of the results and manuscript drafting. 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/PMC555948.xml

529465

Results of a planned interim toxicity analysis with trimodality therapy, including carboplatin AUC = 4, paclitaxel, 5-fluorouracil, amifostine, and radiation for locally advanced esophageal cancer: preliminary analyses and treatment recommendations from the North Central Cancer Treatment Group

Purpose An aggressive trimodality approach from the Minnie Pearl Cancer Research Network [carboplatin AUC = 6, days 1 and 22; 5-fluorouracil 225 mg/m2 continuous infusion, days 1–42, paclitaxel 200 mg/m2, days 1 and 22; 45 Gy] has resulted in remarkable pathologic response rates but notable toxicity. This trial was designed to mitigate this toxicity by starting with a lower carboplatin dose, AUC = 4, and by adding subcutaneous amifostine. Methods This phase II trial included patients with locally advanced, potentially resectable esophageal cancer. All were to receive the above regimen with modifications of carboplatin AUC = 4 and amifostine 500 mg subcutaneously before radiation. All were then to undergo an esophagectomy. A planned interim toxicity analysis after the first 10 patients was to determine whether the carboplatin dose should escalate to AUC = 6. Results Ten patients were enrolled, and all required dose reductions/omissions during neoadjuvant therapy. One patient died from paclitaxel anaphylaxis. Six patients manifested a complete pathologic response. Conclusion With this regimen, carboplatin AUC = 4 for patients with locally advanced esophageal cancer is appropriate.

Combined modality treatment is often administered for locally advanced esophageal cancer [ 1 , 2 ]. Particularly noteworthy are data from Meluch and others from the Minnie Pearl Cancer Research Network [ 3 ]. These investigators published an innovative trial that examined neoadjuvant therapy that consisted of the following: carboplatin area under the curve (AUC) of 6 on days 1 and 22, paclitaxel 200 mg/m2 days 1 and 22, 5-fluorouracil 225 mg/m2 per day by continuous infusion days 1–42 along with radiation 45 Gy in 25 fractions. Yielding some of the most remarkable response rates observed in the treatment of esophageal cancer, this regimen resulted in a pathologic complete response rate of 46% within an initial cohort of 37 patients. Although long-term survival had not been reported, these data appear promising when one acknowledges that pathologic complete response predicts a longer survival and better prospect for cure for patients with this malignancy [ 4 ]. Moreover, recent preliminary data suggest that these response rates are reproducible [ 5 , 6 ]. However, this high complete response rate occurred at the expense of substantial treatment-related toxicity. During neoadjuvant therapy, grade 3 or worse leukopenia occurred in 65% of patients. Twenty-two percent of patients required hospitalization for neutropenic fevers. Grade 3 or worse esophagitis occurred in 31%. Over half of patients suffered weight loss of 10 pounds or more during neoadjuvant therapy. Finally, although there were no deaths during the administration of neoadjuvant therapy, the postoperative death rate was 9%. These toxicity data, coupled with these high response rates, underscore the need to test such novel, aggressive treatments in conjunction with methods to mitigate toxicity. This report describes the planned interim toxicity analysis from an ongoing North Central Cancer Treatment Group (NCCTG) trial that was undertaken with this goal in mind. The NCCTG trial employed two modifications to the Minnie Pearl Regimen in an attempt to mitigate toxicity: 1) carboplatin dosing was reduced from AUC = 6 to AUC = 4 in the first 10 patients with the possibility of dose escalation thereafter; 2) amifostine 500 mg subcutaneously to be given before radiation, as prompted by data from Koukourakis and others [ 7 ] as well as by other subsequent reports [ 8 ]. An interim analysis of toxicity in this NCCTG trial was planned after enrollment of the first 10 patients to decide whether to continue this trial and, if so, whether to escalate the carboplatin to AUC = 6. The results of this interim toxicity analysis are presented below. Methods Overview This trial was initiated and conducted within the NCCTG. The Institutional Review Boards at each site approved the protocol before patient enrollment, and all patients provided signed informed consent at the time of enrollment. Eligibility and Exclusion Criteria Eligibility criteria consisted of the following: 1) age ≥ 18 years; 2) histologic or cytologic evidence of squamous cell carcinoma or adenocarcinoma of the esophagus; 3) surgically resectable tumor, as deemed by a surgeon; 4) Eastern Cooperative Oncology Group (ECOG) performance score of 0–2; 5) anticipated life expectancy of ≥ 12 weeks; 7) the laboratory parameters < 14 days prior to registration of absolute neutrophil count ≥ 1.5 × 10 9 /L, platelet count ≥ 100 × 10 9 /L, total bilirubin ≤ 1.5 times the upper limit of normal, serum creatinine ≤ 1.5 times the upper limit of normal, asparatate aminotransferase ≤ three times the upper limit of normal. Exclusion criteria consisted of the following: 1) uncontrolled infection; 2) prior chemotherapy for esophageal cancer; 3) pregnancy or unwillingness to utilize contraception if pregnancy was a possibility; 4) New York Heart Association classification III or IV; 5) other severe underlying illness that, in the opinion of the treating oncologist, would make the patient inappropriate for study entry; 6) prior radiation that would overlap anticipated radiation fields; 7) antihypertensive or diuretic medications that could not be safely discontinued, if necessary, for several days during study treatment. Treatment Regimen The neoadjuvant treatment regimen consisted of both chemotherapy and radiation. Chemotherapy consisted of carboplatin with an area under the curve (AUC) of 4 to be given intravenously on days 1 and 22, paclitaxel 200 mg/m2 to be given intravenously on days 1 and 22, and 5-fluorouracil 225 mg/m2/day to be given by continuous intravenous infusion on days 1 through 42. In addition, amifostine was to be administered as a 500 mg flat dose subcutaneously immediately before each radiation treatment. For evaluation purposes, a treatment "cycle" is defined by the administration of carboplatin and paclitaxel, where the initiation of these agents heralded the start of a new chemotherapy "cycle." In effect, cycle 1 occurred during days 1–21 of the treatment, and cycle 2 between days 22–42. The protocol was written to allow for a dose increase of carboplatin to an AUC of 6 in the event that a planned interim toxicity analysis after the first 10 patients deemed this increase could be undertaken safely. Radiation was prescribed at a total dose of 4500 centigray. Each fraction size was 180 centigray, and a total of 25 fractions were to be given. Chemotherapy dose reductions were initiated based on toxicity. As determined by the National Cancer Institute's Common Toxicity Criteria (NCI CTC), version 2, Grade 3 or worse stomatitis, esophagitis, or diarrhea prompted holding 5-fluorouracil until symptoms were grade 2 or better. If treatment was held for diarrhea, the protocol called for resuming the continuous infusion 5-fluorouracil at 80% of the prior dose. 5-Fluorouracil that was held was not to be made up. Severe myelosuppression on weekly blood counts (absolute neutrophil count ≥ 0.5 × 10 9 /L for greater than 2 days and/or platelet count ≤ 25 × 10 9 /L) prompted a 25% dose reduction of both paclitaxel and carboplatin on day 22. Similarly, on day 22, the protocol mandated that paclitaxel and carboplatin be held until the absolute neutrophil count was ≥ 1.5 × 10 9 /L and the platelet count ≥ 100 × 10 9 /L. For grade 3 or worse esophagitis, the paclitaxel and carboplatin were held until the esophagitis resolved to grade 1 or less. Paclitaxel was also held for grade 3 or worse neuropathy. Re-treatment was permitted with a 30% dose reduction in the event the neuropathy resolved to grade 2 or better. For any grade 3 or 4 event attributable to amifostine, the amifostine was to be decreased to 300 mg and subsequently discontinued if the event recurred at the lower dose. Finally, radiation was to be held for myelosuppression (absolute neutrophil count < 1.0 × 10 9 /L and/or the platelet count < 50 × 10 9 /L) or for grade 4 esophagitis. Aggressive supportive care measures were recommended throughout the neoadjuvant portion of the regimen. These measures included, but were not limited to, premedication with corticosteroids prior to paclitaxel, use of antiemetics before and during chemotherapy and before amifostine, and nutrition and hydration support, as needed. An esophagectomy was to be performed within 4–8 weeks after completion of radiation for all patients still deemed to be operative candidates. The protocol also included an optional, additional two cycles of post-operative chemotherapy with paclitaxel and carboplatin with dosing, for the most part, left to the discretion of the treating oncologist. Pretreatment and Follow Up Evaluations All patients underwent a history and physical examination within 21 days of trial registration. Other testing was performed within this time frame as well and included a complete hemogram, chemistry profile, computed tomography scan or magnetic resonance imaging of the chest and abdomen, chest radiograph, electrocardiogram, and an esophagoscopy. A bone scan was required if the alkaline phosphatase was two times the institution's upper limit of normal, and a bronchoscopy was required if the primary tumor was adjacent to the trachea or left main stem bronchus or if the patient was experiencing respiratory symptoms. Other testing was optional and included endoscopic ultrasound of the upper gastrointestinal tract and positron emission tomography scanning. All patients were monitored throughout the period of radiation and chemotherapy administration with a weekly history and physical examination and weekly hemograms. History, physical examination, hemogram and chemistry profiles were mandated before days 1 and 22 of chemotherapy. Tumor assessments were performed two weeks prior to surgery, and RECIST criteria, as recommended by the NCI , were used to determine tumor response [ 9 ]. Additionally pathological response was assessed post-operatively. Other testing, such as quality of life assessment and genotyping, were also obtained but will be reported at a later date. Statistical Analyses The purpose of this study was to assess the safety of this treatment regimen for patients with locally advanced esophageal cancer. The protocol called for a three-stage phase II study design with a safety analysis in the first 10 patients. The first 10 patients received treatment, as described above, with a carboplatin AUC = 4. It was decided a priori that if more than two of the initial patients developed grade 4 or 5 non-hematologic adverse events during neoadjuvant therapy or died within 15 days of surgery, then the study would not permit a carboplatin dose increase to an AUC = 6. All data on toxicity and response are presented descriptively in this initial 10-patient report on the first phase of this trial. Summary statistics, including means and median values, frequency tables, and graphical methods were used to describe the distributions of drug administration, clinical characteristics, and adverse event rates. Unless otherwise specified, all adverse event data are reported regardless of relationship to study treatment. Results Demographics Ten patients were recruited from late April 2001 to early March 2002. Patient characteristics are listed in Table 1 , which shows that the median age of the cohort was 63 years (range 49–81) and that nine patients had an Eastern Cooperative Performance Score of 0 with one having a score of 1. Table 1 Characteristic N = 10 Age Median, in years (range) 63 (49–81) Male/Female 9/1 Performance Score (ECOG) 0 9 1 1 Drug Administration During administration of neoadjuvant therapy, all 10 patients required a dose reduction or an omission of drug administration because of treatment-related toxicity (Table 2 ). Overall, nine patients received both cycles of neoadjuvant chemotherapy. Table 2 Patient Amifostine dose received (%)* Neoadjuvant chemotherapy dose received (%) 1 st cycle/2 nd cycle* Disease Status 1 31 81/75 CR** 2 7 71/57 CR 3 16 64/11 PR*** 4 27 100/26 PR 5 73 100/100 CR 6 24 52/3 CR 7 63 62**** PR 8 87 86**** dead 9 7 91/28 CR 10 67 72/79 CR *denotes the lowest percentage of administered drug for a cycle ** complete response (CR) ***partial response (PR) ****second cycle not give Toxicity Severe adverse events during neoadjuvant therapy consisted of 1 death (paclitaxel-related allergic reaction); 15 grade 4 events (myelosuppression (11); mucositis (1); dyspnea (1); neutropenia (1); and cerebral vascular accident (1); and 55 grade 3 events of various types (Table 3 ). Six of eight patients experienced grade 3 or greater adverse events within 15 days of surgery, and four of these were grade 4 events. Some patients experienced more than one event. Specifically, two patients suffered acute respiratory distress syndrome; one patient thrombophlebitis; two patients dyspnea; one patient non-specific constitutional symptoms; one patient neutropenia; and one patient a dysrhythmia. Of the two patients with acute respiratory distress syndrome, one remains alive 6 months after surgery as of this report, and the other died roughly 3 months after surgery. Table 3 Salient Grade 3 or Worse Adverse Events Attributed to Neoadjuvant Study Treatment ADVERSE EVENT ABSOLUTE INCIDENCE NUMBER OF PATIENTS WITH EVENT death (paclitaxel reaction) 1 1 neutropenia 11 9 leukopenia 12 10 febrile neutropenia 5 5 hypotension 4 3 vomiting 4 3 nausea 2 2 dehydration 2 2 diarrhea 2 2 dysphagia 6 6 electrolyte abnormalities 6 4 dyspnea 1 1 pain (non-specific) 1 1 rash 1 1 mucositis 1 1 fatigue 2 2 syncope 1 1 hypersensitivity 1 1 neuropathy 1 1 infection 1 1 melena 1 1 Preliminary Response Data A total of 9 patients underwent an esophagectomy, demonstrating 6 complete pathologic responses. Discussion This report describes a planned interim toxicity analysis of the first 10 patients who were enrolled on a multi-institutional trial for patients with locally advanced esophageal cancer. This NCCTG trial tested a version of the Minnie Pearl Regimen with two main modifications: a drop in the carboplatin AUC dose from 4 to 6 and an addition of amifostine 500 mg subcutaneously (flat dose) prior to each dose of radiation. Preliminary analyses show that this modified regimen resulted in a substantial rate of severe toxicity: one death from a paclitaxel allergic reaction, several episodes of grade 4 neutropenia/leukopenia, one episode of grade 4 mucositis as well as several grade 3 events – all of which were attributable to the study regimen. At the same time, however, this regimen resulted in 6 of 10 patients manifesting a complete pathologic response. The results of this planned analysis led to the NCCTG's decision to reopen this trial with a carboplatin dose of AUC = 4. In effect, the results of this interim analysis suggest that oncologists who might choose to treat patients with the Minnie Pearl Regimen should consider treating with this lower carboplatin dose, especially should they choose to include amifostine for its purported cytoprotective effects. In fact, a major modification of the Minnie Pearl Regimen in the NCCTG trial described here is the addition of subcutaneous amifostine. It remains unclear whether the addition of amifostine compounded or mitigated the toxicity observed in this trial. Certainly, the inclusion of amifostine into this regimen did not permit a dose escalation of carboplatin to an AUC of 6, as tested in the original Minnie Pearl Regimen. Although amifostine is considered cytoprotective, it also carries toxicity in its own right including nausea, vomiting, and hypotension. Preliminary data suggest that subcutaneous administration of amifostine might be better tolerated than intravenous administration, but these conclusions are not based on large, comparative studies. The fact that the amifostine, carboplatin, 5-fluorouracil, paclitaxel, and radiation were all given in combination as part of the phase II study presented here makes it impossible to sort out attribution as it pertains to an individual study agent, such as amifostine. Should the results of this trial continue to appear promising, a larger phase III trial would be necessary to provide a definitive answer on the contribution of subcutaneous amifostine to the efficacy and toxicity profile of this regimen. In short, the preliminary toxicity assessment of this trial suggests that the Minnie Pearl Regimen should include a carboplatin dose with an AUC of 4 rather than 6, especially if subcutaneous amifostine is included as part of the regimen.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529465.xml

423148

Deconstructing Brain Waves: Background, Cue, and Response

Light waves from an awaited signal—a white circle—arrive at the subject's eye; within a fraction of a second, the subject's thumb presses a button. Between eye and thumb lies the central nervous system, its feats of perception, integration, and response largely opaque to scientific scrutiny. Imaging techniques like magnetic resonance imaging can detail brain anatomy but can only broadly show changes in activity levels occurring over seconds—indirect echoes of brain function. Electrodes stuck to the scalp record coordinated neuronal symphonies, and wires inserted among neurons can capture the single-cell firing patterns of the individual instruments of the neural orchestra. But how these electrical signals map to information processing within and across neural circuits remains blurry. A new analysis sharpens the focus by separating individual brain wave patterns, measured from multiple sites across the scalp, into nine distinct process classes, each centered in an anatomically relevant brain area and producing predictable patterns as human subjects receive visual cues and produce responses. Schematic representation of the source and strength of task-related EEG signals. The animation can be accessed online at http://dx.doi.org/10.1371/journal.pbio.0020176.v001 Scalp electroencephalograms (EEGs) are dominated by waves of synchronized neuronal activity at specific frequencies. Decades of research have associated wave patterns recorded at different scalp regions with different states of alertness—attending, drowsy, sleeping, or comatose; eyes open or closed—and gross abnormalities, such as seizure, brain damage, and tumor. In order to separate EEG responses to specific events from background, state-related activity, researchers repeat an experiment like the button-press exercise tens or hundreds of times and average the EEG across trials. By averaging out background activity, this technique reveals a characteristic waveform, called an event-related potential (ERP). It differs by electrode location, but often contains a large positive wave that peaks 300 milliseconds or more after an awaited visual cue. In the current paper, Scott Makeig et al. argue that ERP averaging removes important information about ongoing processes and their interactions with event-related responses. Instead of averaging multiple recordings from each of 31 electrode sites, the authors applied an algorithm that seeks independent signal sources contributing to the individual tracings. The researchers measured signal source activities by the frequency and phase of wave patterns and source locations by comparing signal strength and polarity at different electrodes. Altogether, the researchers identified nine classes of maximally independent sources, each having similar locations and activities across subjects. The results dovetail neatly with prior anatomical and functional observations. This analysis demonstrates that average waveforms identified in ERP studies probably sum multiple, separate processes from several brain regions. In particular, the large positive ERP seen 300 milliseconds or more after a visual cue reflects different waveforms from frontal, parietal, and occipital cortex—areas involved in task planning, spatial relationships and movement, and visual processing, respectively. In addition, this study showed a two-cycle burst of activity in the 4–8 (theta) frequency band after button presses—another common ERP feature. The theta activity was coordinated across several signal sources, and localized to areas associated with planning and motor control. Notably, the planning component seemed to lead the motor signal. Suppression or resynchronization of several EEG processes followed the visual cue or button press. The authors theorize that such coordination might influence the speed or impact of communication between brain areas and help retune attention after significant events. Using this approach in more subjects, and under differing conditions, could provide an unprecedented glimpse of how the brain translates perception and planning into action. The results suggest that EEG data contain an untapped richness of information that could give researchers and clinicians a new window into thought in action.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC423148.xml

519001

EGF Signal Propagation during C. elegans Vulval Development Mediated by ROM-1 Rhomboid

During Caenorhabditis elegans vulval development, the anchor cell (AC) in the somatic gonad secretes an epidermal growth factor (EGF) to activate the EGF receptor (EGFR) signaling pathway in the adjacent vulval precursor cells (VPCs). The inductive AC signal specifies the vulval fates of the three proximal VPCs P5.p, P6.p, and P7.p. The C. elegans Rhomboid homolog ROM-1 increases the range of EGF, allowing the inductive signal to reach the distal VPCs P3.p, P4.p and P8.p, which are further away from the AC. Surprisingly, ROM-1 functions in the signal-receiving VPCs rather than the signal-sending AC. This observation led to the discovery of an AC–independent activity of EGF in the VPCs that promotes vulval cell fate specification and depends on ROM-1. Of the two previously reported EGF splice variants, the longer one requires ROM-1 for its activity, while the shorter form acts independently of ROM-1. We present a model in which ROM-1 relays the inductive AC signal from the proximal to the distal VPCs by allowing the secretion of the LIN-3L splice variant. These results indicate that, in spite of their structural diversity, Rhomboid proteins play a conserved role in activating EGFR signaling in C. elegans, Drosophila, and possibly also in mammals.

Introduction Intercellular signaling pathways control many diverse processes, such as cell proliferation, differentiation, survival, migration, shape changes, and responses to the environment. In most instances, the release of the signaling molecules by the signal-sending cells constitutes a rate-limiting step that determines the spatial distribution and temporal duration of the response ( Freeman and Gurdon 2002 ). On the other hand, the binding sites on the receptors that are presented by the signal-receiving cells are usually in excess of the available ligands ( Freeman and Gurdon 2002 ). Specificity is achieved by the tissue-specific expression of the signal or by the regulated activation of an inactive precursor molecule. For example, growth factor peptides are often produced as inactive precursors that need to be processed before they can be released and activate their cognate receptors on the signal-receiving cells ( Arribas et al. 1996 ). The epidermal growth factor (EGF) receptor (EGFR) acts in a highly conserved signal transduction pathway that controls various cell fate decisions in metazoans ( Bogdan and Klambt 2001 ). EGFR ligands of the transforming growth factor-α (TGF-α) family are produced as membrane-tethered precursor proteins with a single extracellular EGF repeat that is cleaved off the membrane anchor ( Pandiella and Massague 1991 ; Bosenberg et al. 1993 ). The best-studied example of EGF processing is probably the Drosophila growth factor Spitz, which activates the EGFR in multiple developmental processes ( Rutledge et al. 1992 ; Golembo et al. 1996 ). Genetic analysis of Drosophila EGFR signaling has identified Rhomboid-1 as a protein necessary for Spitz activation in the signal-sending cell ( Bier et al. 1990 ; Golembo et al. 1996 ; Guichard et al. 2000 ; Wasserman et al. 2000 ). Drosophila Rhomboid-1 is the founding member of a family of seven-pass transmembrane proteins that function as intramembrane serine proteases ( Urban et al. 2001 ). Site-specific cleavage of Spitz by Rhomboid-1 in the Golgi apparatus allows the secretion of the extracellular portion of Spitz by the signal-sending cell ( Lee et al. 2001 ; Urban and Freeman 2003 ). The Drosophila genome encodes a total of seven rhomboid genes with partially overlapping functions in different tissues that utilize the EGFR pathway ( Wasserman et al. 2000 ; Urban et al. 2002 ). There are four predicted Rhomboid homologs in humans and five in C. elegans ( Wasserman et al. 2000 ). Rhomboid-like proteins are even found in yeast and bacteria ( Gallio et al. 2002 ; McQuibban et al. 2003 ). Rhomboids are part of the larger family of I-Clip proteases that includes the aspartyl protease Presenilin ( Wolfe et al. 1999 ) and the Zn 2+ metalloprotease S2P ( Urban and Freeman 2002 ). On the other hand, the secretion of vertebrate TGF-α involves processing by a disintegrin metalloprotease (ADAM), TNF-α-converting enzyme ( Peschon et al. 1998 ). Whether a Rhomboid protease is also involved in the processing of TGF-α is currently unknown. The development of the C. elegans hermaphrodite vulva serves as a simple model by which to study signal transduction and cell fate determination during organogenesis ( Kornfeld 1997 ; Sternberg and Han 1998 ). During C. elegans postembryonic development, the anchor cell (AC) in the somatic gonad induces three out of six equivalent vulval precursor cells (the VPCs, termed P3.p through P8.p) in the ventral hypodermis to adopt vulval cell fates ( Sulston and White 1980 ; Kimble 1981 ). The AC produces the LIN-3 growth factor, which is similar to Drosophila Spitz and mammalian TGF-α ( Hill and Sternberg 1992 ). The VPCs express the EGFR homolog LET-23 on the basolateral surface that faces the AC ( Whitfield et al. 1999 ), and they are all competent to activate the RAS/mitogen-activated protein kinase (MAPK) signaling pathway in response to the inductive LIN-3 EGF signal. The VPC closest to the AC, P6.p, adopts the primary (1°) cell fate characterized by a symmetrical cell lineage leading to eight 1° vulval cells ( Sternberg and Horvitz 1986 ). The neighbors of P6.p, P5.p and P7.p, adopt the secondary (2°) cell fate, which is characterized by an asymmetrical lineage leading to seven 2° vulval cells. The more distally located VPCs, P3.p, P4.p, and P8.p, adopt the uninduced tertiary (3°) cell fate. After dividing once, they fuse with the surrounding hypodermal syncytium (hyp7). LIN-3 EGF dosage experiments have suggested that the inductive signal acts in a graded manner ( Katz et al. 1995 ). According to this model, P6.p receives the highest amount of the inductive signal and thus adopts the 1° fate, while an intermediate level of the LIN-3 signal specifies the 2° fate in P5.p and P7.p. The distal VPCs, P3.p, P4.p, and P8.p, receive too little signal to adopt vulval cell fates. However, in response to the inductive signal, P6.p produces a lateral signal that activates the LIN-12 NOTCH signaling pathway in the neighboring VPCs, P5.p and P7.p ( Greenwald et al. 1983 ). The LIN-12 NOTCH signal is both necessary and sufficient to induce the 2° cell fate ( Sternberg 1988 ; Simske and Kim 1995 ). Thus, the graded LIN-3 EGF signal may act redundantly with the lateral LIN-12 NOTCH signal to specify the 2° vulval cell fate in the neighbors of P6.p ( Kenyon 1995 ). Like its Drosophila and vertebrate homologs, LIN-3 EGF is synthesized as a transmembrane precursor protein ( Hill and Sternberg 1992 ). Experiments with dig-1 mutants in which the AC is dorsally displaced indicate that the AC is capable of inducing vulval cell fates from a distance, suggesting that proteolytic cleavage of membrane-bound LIN-3 occurs in the AC ( Thomas et al. 1990 ). Here, we report the identification of the C. elegans Rhomboid homolog ROM-1 as a positive regulator of vulval induction. Surprisingly, we find that ROM-1 acts in the signal-receiving VPCs rather than in the signal-sending AC. Furthermore, we uncover an AC-independent function of LIN-3 EGF that depends on ROM-1 activity in the VPCs. Two LIN-3 splice variants, termed LIN-3S and LIN-3L, that differ by an insertion of 15 amino acids in the region in the juxtamembrane domain critical for processing, have been described ( Hill and Sternberg 1992 ). Genetic epistasis experiments indicate that LIN-3L activity in the VPCs depends on ROM-1, while LIN-3S or a truncated form of LIN-3 lacking the transmembrane domain act independently of ROM-1. We propose a relay model in which ROM-1 is required for the activation of LIN-3L in the proximal VPCs to transmit the inductive AC signal to the distal VPCs. Results Five Rhomboid-Like Proteins in C. elegans Since the LIN-3 EGF growth factor is produced as a transmembrane precursor protein ( Hill and Sternberg 1992 ), we asked whether an intramembrane serine protease of the Rhomboid family is involved in the proteolytic processing of LIN-3 EGF. Rhomboid proteins in metazoans share a characteristic secondary structure consisting of seven transmembrane domains ( Bier et al. 1990 ; Urban et al. 2001 ). We searched the complete C. elegans genome sequence for genes with similarity to Drosophila rhomboid-1 and identified five rhomboid- like genes termed rom-1 (F26F4.3), rom-2 (C48B4.2), rom-3 (Y116A8C.14), rom-4 (Y116A8C.16), and rom-5 (Y54E10A.14) ( Figure 1 A). All five C. elegans ROM proteins display the typical secondary structure of Rhomboids ( Wasserman et al. 2000 ). The transmembrane domains show the highest degree of sequence conservation, while the hydrophilic N termini are more divergent ( Figure 1 B). ROM-1 is most similar to Drosophila Rhomboid-1 (35% identity), followed by ROM-2 (29% identity) and the more diverged ROM-3 (24% identity), ROM-4 (26% identity), and ROM-5 (29% identity). Mutagenesis experiments with Drosophila Rhomboid-1 have identified a catalytic triad formed by conserved asparagine, serine, and histidine residues that are necessary for the serine protease activity ( Urban et al. 2001 ). This catalytic triad is conserved only in ROM-1 (black triangles in Figure 1 B), suggesting that the other four Rhomboid-like proteins do not function as serine proteases . Figure 1 The C. elegans Rhomboid Genes (A) Dendogram showing the relation between the seven-pass transmembrane domains of Rhomboids from C. elegans (C.e.), Drosophila melanogaster (D.m.), and Homo sapiens (H.s.) calculated with the neighbor joining method using CLUSTAL X ( Thompson et al. 1997 ). (B) Alignment of C. elegans (C.e.) ROM-1 and ROM-2 and Homo sapiens (H.s.) Rho-1 relative to Drosophila melanogaster (D.m.) Rho-1. Residues identical to those of Drosophila Rho-1 are highlighted in black, and similar residues are highlighted in grey. The thick black lines indicate the predicted seven-pass transmembrane domains. The three black triangles point at the residues forming a catalytic triad that forms a charge-relay system to activate the essential serine residue during peptide bond cleavage, and the three open triangles indicate other conserved residues necessary for the enzymatic activity as identified in D.m. Rho-1 ( Urban et al. 2001 ). The region underlined with a dotted line indicates the extent of deletion in the rom-1 ( zh18 ) allele. (C) Intron-exon structure of the rom-1 locus and extent of the deletion in the rom-1(zh18) strain. The numbers indicate the position of the deletion break-points relative to the A in the ATG start codon. In order to confirm the predicted intron-exon structure of rom-1, we isolated rom-1 cDNA by RT-PCR. An SL1 trans-spliced leader sequence was identified at the 5′ end of the message that was spliced to the second of the six exons predicted by the C. elegans genome project ( Figure 1 C) (see http://www.wormbase.org ). The remaining intron-exon boundaries were confirmed experimentally and corresponded exactly to the predicted boundaries. The conceptual translation of the 1,071-bp open reading frame (ORF) predicts a protein of 356 amino acids, with very short stretches of hydrophilic amino acids between the seven-pass transmembrane domains, except for a longer loop consisting of 43 amino acids between the first and second transmembrane domains (see Figure 1 B). ROM-1 and ROM-2 Are Not Essential for Normal Vulval Development As a first step to examine the biological function of the rom genes, we used RNA interference (RNAi) to transiently knock down their expression ( Fire et al. 1998 ; Fraser et al. 2000 ; Kamath et al. 2001 ). Double-stranded RNA derived from a 352-bp rom-1 or a 718-bp rom-2 cDNA fragment of the divergent N-terminal portion was injected into the hermaphrodite gonads, and vulval development was examined in the F1 progeny under Nomarski optics. No obvious vulval phenotype was observed when rom-1 or rom-2 RNAi was performed in a wild-type background. Also, feeding wild-type or let-60(n1046gf) animals with bacteria producing rom-3 dsRNA had no effect on vulval development (unpublished data). Due to the high degree of sequence similarity between rom-3 and rom-4 (69.8% identity), rom-3 RNAi is likely to simultaneously reduce rom-4 function. Using a PCR-based assay to screen a library of mutagenized worms, we isolated a 1,556-bp deletion in the rom-1 gene (see Figure 1 C) ( Jansen et al. 1997 ; Berset et al. 2001 ). The zh18 deletion removes 206 amino acids from the N terminus, including the first three transmembrane domains and 384 bp of promoter sequences. Thus, the zh18 deletion probably results in a complete loss of rom-1 function and will be referred to as rom-1(0) . The rom-1(0) single mutants exhibited no obvious phenotype; they were healthy and fertile. In addition, we obtained the rom-2(ok966) allele from the C. elegans Gene Knockout Consortium. The rom-2(ok966) animals carry a 530-bp deletion that removes the fifth exon, which contains the predicted catalytic center with the essential histidine residue (see Figure 1 B) ( Urban et al. 2001 ). Since this allele is predicted to inactivate any potential protease activity of ROM-2, we refer to it as rom-2(rf) . Consistent with the RNAi experiments, both rom-1(0) and rom-2(rf) single mutants exhibited normal vulval development ( Table 1 , rows 2 and 3). Also, in rom-1(0) rom-2(rf) double mutants, no defects in vulval development were observed, ruling out a possible redundant function of the two genes ( Table 1 , row 4). Thus, neither ROM-1 nor ROM-2 are required for vulval induction under normal conditions. Table 1 Suppression of Multivulva Mutants by rom-1(0) Vulval induction was scored using Nomarski optics as described in Materials and Methods . % Vul indicates the fraction of animals with fewer than three induced VPCs, % Muv indicates the fraction of animals with more than three induced VPCs, and the induction index indicates the average number of VPCs per animal that had adopted 1° or 2° vulval fates. Number of animals scored is designated by n . Alleles used: pry-1(mu38), rom-1(zh18), rom-2(ok996), lin-12(n137gf), dpy-19(e1259), huIs7[hs::bar-1 Δ NT], let-60(n1046gf), gaIs36[hs::mpk-1, D-mek-2(gf)], lin-15(n309), zhEx22[lin-3(+), sur-5::gfp, unc-119(+)], and syIs12[hs::lin-3extra]. Statistical analysis was done as described in Materials and Methods a Five independent transgenic lines generated with this construct displayed induction indices ranging from 4.1 to 5.4, and one of the lines displaying a penetrant Muv phenotype ( zhEx22 ) was used for the further analysis in the different backgrounds b dsRNA was injected into the syncytial gonad of the parents, and vulval induction in the F1 progeny was scored for rows 4 and 11 by inspection under a dissecting microscope c To provide a low dose of LIN-3extra, L1 larvae were heat-shocked for 5 min at 33 °C and grown at 25 °C until L4 d To provide a high dose of LIN-3extra, or MPK-1 or BAR-1ΔNT, respectively, early L21 larvae were heat-shocked for 30 min at 33 °C and grown at 25 °C until L4 e These strains carried the dpy-19(e1259) mutation in cis to lin-12(n137gf) ** p ≤ 0.001; *** p ≤ 0.0001; numbers in brackets next to asterisks indicate the row to which a dataset was compared n.d., no data ROM-1 Positively Regulates the EGFR/RAS/MAPK Pathway in Distal VPCs Next, we examined whether loss of rom-1 or rom-2 function affects vulval induction in a sensitized genetic background by using mutations that hyperactivate the EGFR/RAS/MAPK pathway. The rom-1(0) mutation as well as rom-1 RNAi partially suppressed the multivulva (Muv) phenotype caused by overexpression of the LIN-3 EGF growth factor [lin-3(+)] ( Hill and Sternberg 1992 ) or by the n1046 gain-of-function (gf) mutation in the let-60 ras gene, which renders vulval development partially independent of upstream signaling ( Beitel et al. 1990 ; Chang et al. 2000 ) ( Table 1 , rows 5–7 and 12–14). In addition, the rom-1(0) mutation suppressed the Muv phenotype of hs::mpk-1 animals that overexpress the wild-type MAPK MPK-1 under control of a heat-shock promoter together with Drosophila MEK-2 under control of the interferon-1α promoter ( Lackner and Kim 1998 ) ( Table 1 , rows 16 and 17). In contrast to rom-1, neither the rom-2(rf) mutation nor rom-2 RNAi affected the Muv phenotype of let-60(gf) animals ( Table 1 , row 15; unpublished data). The rom-1(0) mutation did not significantly enhance the vulvaless (Vul) phenotype caused by the lin-3(e1417), lin-2(n397), sem-5(n2019), or let-60(n2021) mutations that reduce the activity of the receptor tyrosine kinase (RTK)/RAS/MAPK pathway (unpublished data). Since these Vul mutants affect the cell fates of only the proximal VPCs (P5.p, P6.p, and P7.p), ROM-1 plays no role in the induction of the proximal VPCs by the AC. Thus, ROM-1 enhances the activity of the EGFR/RAS/MAPK pathway to allow the induction of the distal VPCs P3.p, P4.p, and P8.p. ROM-1 Regulates LIN-3 EGF Activity during Vulval Induction A soluble form of LIN-3 that consists of the extracellular domain with the EGF repeat but lacks the transmembrane and intracellular domains is biologically active and causes a Muv phenotype when overexpressed under control of a heat-shock promoter (hs::lin-3extra) ( Katz et al. 1995 ). Unlike full-length LIN-3, the Muv phenotype induced by a low or high dosage of LIN-3extra was not suppressed by rom-1(0) ( Table 1 , rows 8–11). In lin-15(rf) mutants, all VPCs adopt vulval cell fates independently of the LIN-3 signal, though induction in lin-15(rf) mutants depends on the activity of LET-23 and the other components of the EGFR/RAS/MAPK pathway ( Clark et al. 1994 ; Huang et al. 1994 ). The rom-1(0) mutation did not suppress the Muv phenotype of lin-15(rf) animals, suggesting that loss of rom-1 function affects the LIN-3-dependent induction of vulval cell fates rather than the LIN-3-independent activity of the EGFR/RAS/MAPK pathway ( Table 1 , rows 18 and 19). Finally, we examined the genetic interaction between rom-1 and the Notch and Wnt pathways, since both pathways control vulval cell fate specification in parallel with the RTK/RAS/MAPK pathway ( Wang and Sternberg 2001 ). In lin-12 notch(gf) animals, no AC is formed, and all VPCs adopt the 2° cell fate ( Sternberg and Horvitz 1989 ). The same phenotype was observed in rom-1(0) lin-12(gf) double mutants ( Table 1 , rows 20 and 21). In addition, the Muv phenotype caused by hyperactivation of the Wnt pathway through a reduction-of-function mutation in pry-1 axin or by overexpression of a N-terminally truncated BAR-1 β-catenin protein was not suppressed by the rom-1(0) mutation ( Table 1 , rows 22–25) ( Gleason et al. 2002 ). In summary, these experiments suggest that ROM-1 promotes the LIN-3-dependent activation of the EGFR/RAS/MAPK signaling pathway. ROM-1 likely acts at the level or upstream of LIN-3 since full-length but not a soluble form of LIN-3 was sensitive to loss of rom-1 function. ROM-1 Is Required to Transmit the Inductive Signal to the Distal VPCs To assess how much inductive signal each VPC receives, we examined the expression pattern of the egl-17::cfp reporter, which is a transcriptional target of the EGFR/RAS/MAPK pathway ( Yoo et al. 2004 ). In mid L2 larvae, before the LIN-12 NOTCH-mediated lateral inhibition becomes effective, egl-17::cfp is expressed in a graded manner with highest levels in P6.p, intermediate levels in P5.p and P7.p ( Yoo et al. 2004 ), and lower levels in P3.p, P4.p, and P8.p ( Figure 2 A and 2 B). We therefore examined the effect of the rom-1(0) mutation on the egl-17::cfp expression pattern in mid L2 larvae. For this purpose, larvae were synchronized in the mid L1 stage at 13 h of development by letting them hatch in the absence of food, and then development was allowed to proceed by adding food for another 24 h until they reached the mid L2 stage (approximately 37 h of development). Loss of ROM-1 function had no effect on egl-17::cfp expression in the proximal VPCs (P5.p, P6.p, and P7.p), but significantly reduced egl-17::cfp expression in the distal VPCs when compared to wild-type animals ( Figure 2 C and 2 D). Thus, ROM-1 increases the range of the inductive LIN-3 signal, allowing the distal VPCs to activate the EGFR/RAS/MAPK pathway. The altered egl-17::cfp expression pattern in rom-1(0) animals is consistent with the epistasis data, which showed that loss of rom-1 function affects the induction of only the distal VPCs (see above). Figure 2 Expression of the egl-17::cfp Reporter in rom-1(0) and lin-3(rf) Mutants Photographic images on the left (A, C, E, G, and I) show the expression of the arIs92[egl-17::cfp] reporter in the VPCs of mid-L2 larvae of the different genotypes indicated. Pie graphs on the right (B, D, F, H, and J) show semi-quantitative representations of the expression levels observed in individual VPCs in the different backgrounds. A solid black color indicates the strongest expression of EGL-17::CFP as it was observed in P6.p of many (59%) wild-type animals; dark grey indicates intermediate, light grey weak, and white undetectable expression. The numbers inside the pie charts are the corresponding percentage values, and n refers to the number of animals examined for each case. EGL-17::CFP expression in each VPC of rom-1(0) or lin-3(e1417rf) animals was compared against the same VPC in wild-type animals (considered as expected value) with a Chi 2 test for its independence; *** p ≤ 0.0001, ** p ≤ 0.001. The row to which a dataset was compared is indicated on the right. All photographs were taken with identical exposure and contrast settings. The scale bar in (I) is 20 μm. ROM-1 Is Expressed in the VPCs but Not in the AC during Vulval Induction To analyze the expression pattern of ROM-1, we generated a transcriptional rom-1 reporter by fusing 6.9 kb of the 5′ rom-1 promoter/enhancer region to the green fluorescent protein (gfp) ORF carrying a nuclear localizing signal (zhIs5[rom-1::nls::gfp]) . With a translational full-length rom-1::gfp fusion construct, we failed to obtain transgenic lines that consistently expressed ROM-1::GFP. Moreover, a genomic DNA fragment encompassing the entire rom-1 locus failed to produce stable transgenic lines even when injected at relatively low concentrations (1–10 ng/μl), suggesting that elevated levels of ROM-1 are toxic to the animals. The transcriptional rom-1::nls::gfp reporter was widely expressed in somatic cells throughout development. Surprisingly, we did not detect any rom-1::nls::gfp expression in the gonadal AC before the L4 stage, while consistent expression was observed in the Pn.p cells and the Pn.a-derived neurons from the L1 stage on. In early L2 zhIs5 larvae, the six VPCs expressed rom-1::nls::gfp at equal levels ( Figure 3 A and 3 B). The Pn.p cells that are not part of the vulval equivalence group and had fused to hyp7 at the end of the L1 stage showed relatively higher rom-1::nls::gfp expression than the VPCs (for example, P1.p, P2.p, and P9.p in Figure 3 C and 3 D). Toward the end of the L2 stage, rom-1::nls::gfp expression decreased in distal VPCs adopting the 3° uninduced fate and persisted in the proximal VPCs adopting induced vulval fates ( Figure 3 C and 3 D). In 60% of zhIs5 animals, we observed an up-regulation of rom-1::nls::gfp in P6.p, and in 35% and 45% of the cases, rom-1::nls::gfp expression was higher in P5.p and P7.p, respectively ( n = 20). After vulval induction, rom-1::nls::gfp was down-regulated in the 1° and 2° descendants of P5.p, P6.p and P7.p, while the 3° descendants of P3.p, P4.p, and P8.p again expressed high levels of rom-1::nls::gfp after they had fused with hyp7 ( Figure 3 E and 3 F). Expression of rom-1::nls::gfp was observed in the AC and other cells of the somatic gonad only beginning in the L4 stage, before the AC fused with the uterine seam cell and persisting after fusion ( Figure 3 G and 3 H) ( Sulston and Horvitz 1977 ; Newman et al. 1996 ). To test whether the inductive AC signal is required for the elevated rom-1::nls::gfp expression in the proximal VPCs, we ablated in zhIs5 animals the precursors of the somatic gonad Z1 and Z4 ( Kimble 1981 ). Uniformly low rom-1::nls::gfp expression was found in all six VPCs of gonad-ablated zhIs5 animals at the late L2 to early L3 stage, before the descendants of the 3° VPCs had fused to hyp7 ( Figure 3 J and 3 K, n = 20). To test whether rom-1::nls::gfp expression depends on RTK/RAS/MAPK signaling in the VPCs, we introduced the zhIs5 transgene into lin-7(e1413) mutants that exhibit a penetrant Vul phenotype due to reduced LET-23 EGFR activity ( Simske et al. 1996 ). In lin-7(e1413); zhIs5 animals, the up-regulation of rom-1::nls::gfp occurred less frequently (in 13%, 33%, and 7% of the cases in P5.p, P6.p, and P7.p, respectively, n = 15). Thus, the AC signal up-regulates rom-1::nls::gfp expression in the VPCs that adopt vulval cell fates. Figure 3 Expression Pattern of rom-1::nls::gfp Expression pattern of the zhIs5[rom-1::nls::gfprom-1::] transcriptional reporter during vulval development. Images on the left (A, C, E, G, and I) show the corresponding Nomarski pictures with the arrows pointing at the Pn.p cell nuclei and the arrowhead indicating the position of the AC nucleus. (B) A mid L2 larva before vulval induction with uniform rom-1::nls::gfp expression in all the Pn.p cells. (D) An early L3 larva in which rom-1::nls::gfp expression was decreased in all VPCs except P6.p (see text for a quantification of the expression pattern). Note that the nuclei of hyp7 and the Pn.p cells that had fused to hyp7 displayed strong rom-1::nls::gfp expression (P1.p, P2.p, P3.p and P9.p in the example shown). (F) A mid to late L3 larva in which P6.p had generated four descendants. Expression of rom-1::nls::gfp occurred only in the 3° descendants of P.4.p and P8.p after they fused to hyp7. (H) An L4 larva during vulval invagination. No rom-1::nls::gfp was detectable in the 1° and 2° descendants of P5.p, P6.p, and P7.p, but the AC and the surrounding uterine cells displayed strong rom-1::nls::gfp expression. (K) A late L2 to early L3 larva following the ablation of the precursors of the somatic gonad. No up-regulation of rom-1::nls::gfp in P5.p, P6.p, or P7.p was observed. The scale bar in (K) is 10 μm. ROM-1 Acts in an AC-Independent Pathway that Promotes Vulval Induction To examine whether ROM-1 acts in cells other than the AC (which is part of the somatic gonad), we tested the effect of loss of rom-1(+) function on vulval induction in gonad-ablated animals. If ROM-1 acts exclusively in the AC, then the rom-1(0) mutation should not affect vulval induction in gonad-ablated animals. On the other hand, if ROM-1 acts in cells other than the AC, then the rom-1(0) mutation should suppress vulval induction even in the absence of the AC. Since the inductive AC signal is absolutely required to initiate vulval development ( Kimble 1981 ), we performed the gonad ablation experiments in let-60(gf) or hs::mpk-1 animals that exhibit a hyperactive EGFR/RAS/MAPK signaling pathway causing AC-independent vulval induction ( Table 2 , rows 5 and 8) ( Beitel et al. 1990 ; Lackner and Kim 1998 ; Chang et al. 2000 ). In addition, we examined lin-3(+) animals because, as reported previously by Hill and Sternberg (1992) , in animals that overexpress wild-type lin-3 under control of its own promoter, some vulval differentiation could still be observed in the absence of the AC, pointing at an additional source of LIN-3 from the transgene in cells outside of the gonad ( Table 2 , row 2). Loss of rom-1 function in gonad-ablated lin-3(+), let-60(gf), or hs::mpk-1 animals caused a strong further reduction in vulval induction ( Table 2 , compare rows 2 with 3, 5 with 6, and 8 with 9). In contrast, vulval induction in gonad-ablated lin-15(rf) animals that exhibit lin-3 independent vulval differentiation was not affected by the rom-1(0) mutation (Table2, rows 13 and 14). Table 2 Gonad-Independent Function of rom-1 and lin-3 Vulval induction was scored as described in the legend to Table 1 . See Table 1 legend for key to abbreviations and terminology. Alleles used: rom-1(zh18), let-60(n1046gf), gaIs36[hs::mpk-1, D-mek-2(gf)] , lin-3(n1049null) [the sterile Dpy nonUnc progeny segregated by dpy-20(e1282) lin-3(n1049)/ unc-44(e362) unc24(e138); gaIs36 mothers was examined], lin-15(n309), zhEx22[lin-3(+), and sur-5::gfp, unc-119(+)]. a The gonad precursors Z1 through Z4 were ablated in L1 larvae where indicated b L2 larvae were heat-shocked for 30 min at 33 °C and grown at 25 °C until L4 Analogous results were obtained by examining the egl-17::cfp expression pattern after removal of the AC. In gonad-ablated animals, residual egl-17::cfp expression was observed in all VPCs ( Figure 2 E and 2 F). In many cases, P6.p expressed higher levels of the reporter than did the other VPCs despite the absence of the AC. Loss of rom-1 function in gonad-ablated animals caused a further decrease in egl-17::cfp expression in all VPCs ( Figure 2 G and 2 H). Thus, ROM-1 acts in cells outside of the somatic gonad to promote vulval induction. An AC-Independent Activity of LIN-3 EGF Next, we used an analogous strategy to test whether endogenous LIN-3 acts with ROM-1 in an AC-independent pathway. The decrease in vulval induction in hs::mpk-1 animals that was caused by the lin-3(n1049) loss-of-function mutation [lin-3(0)] was much stronger than the decrease observed in gonad-ablated lin-3(+); hs::mpk-1 animals ( Table 2 , compare rows 8 and 10; the L1 larval lethal phenotype caused by the lin-3(0) mutation was suppressed by the hs::mpk-1 transgene). Vulval induction in lin-3(0); hs::mpk-1 animals was not affected by gonad ablation since the lin-3(0) allele eliminated lin-3 function in the AC ( Table 2 , compare rows 10 and 11). Thus, a complete loss of lin-3 function had a more severe effect on vulval induction than did just the removal of the AC. Likewise, the lin-3(e1417) reduction-of-function mutation almost completely abolished the expression of the egl-17::cfp marker ( Figure 2 I and 2 J). Thus, LIN-3 is also necessary for the AC-independent egl-17::cfp expression in the VPCs. Loss of rom-1 function in a lin-3(0); hs::mpk-1 background caused no further decrease in vulval induction, suggesting that ROM-1 does not affect vulval development in the absence of LIN-3 ( Table 2 , compare rows 10 and 12). Taken together, these experiments indicate that not only the AC but also cells outside of the gonad produce LIN-3 to promote vulval fate specification. This AC-independent activity of LIN-3 requires ROM-1 function. ROM-1 Can Act in the Pn.p Cells The absence of detectable rom-1::nls::gfp expression in the AC around the time of vulval induction and the AC-independent function of rom-1 and lin-3 suggested that rom-1 may act cell-autonomously in the VPCs. To test this hypothesis, we expressed rom-1 under control of the Pn.p cell-specific lin-31 promoter (lin-31::rom-1) ( Tan et al. 1998 ). The lin-31::rom-1 transgene restored vulval induction in rom-1(0); let-60(gf) and rom-1(0); hs::mpk-1 double mutants to levels comparable to those found in let-60(gf) and hs::mpk-1 single mutants ( Table 3 , rows 1–3 and 8–10). A transgene encoding bacterial Cre recombinase under control of the lin-31 promoter (lin-31::cre) that was used as a negative control had no effect on vulval induction ( Table 3 , rows 4 and 11) ( Hoier et al. 2000 ). Consistent with a function of rom-1 in an AC-independent pathway, the lin-31::rom-1 transgene also increased induction in gonad-ablated rom-1(0); let-60(gf) animals ( Table 3 , rows 5–7). Finally, we expressed rom-1 in the AC under control of the AC-specific enhancer (ACEL) (ACEL::rom-1), which is located in the third intron of the lin-3 locus ( Hwang and Sternberg 2003 ). In contrast to lin-31::rom-1, the ACEL::rom-1 transgene did not rescue the suppression of the let-60(gf) Muv phenotype by rom-1(0) ( Table 3 , rows 12 and 13). Thus, the tissue-specific expression of ROM-1 in the Pn.p cells efficiently rescues a loss of rom-1 function. Table 3 Expression of rom-1 in the Pn.p Cells but Not in the AC Rescues the rom-1(0) Phenotype Vulval induction was scored as described in the legend to Table 1 . See Table 1 legend for key to abbreviations and terminology. Alleles used: rom-1(zh18), let-60(n1046gf), gaIs36[hs::mpk-1, D-mek-2(gf)], zhEx66[lin-31::rom-1, unc-119(+), sur-5::gfp], zhEx81[lin- zhEx81[lin-31::cre, unc-119(+), myo-3::gfp], and zhEx89[ACEL::rom-1, sur-5::gfp]. a The gonad precursor cells Z1 through Z4 were ablated in L1 larvae where indicated b All three independent transgenic lines examined increased the induction index of rom-1(0); let-60(gf) animals to 4.2–4.5 c L2 larvae were heat-shocked for 30 min at 33 °C and grown at 25 °C until L4 d Two independent transgenic lines examined displayed an induction index in rom-1(0); let-60(gf) animals of 3.3 and 3.5 LIN-3 EGF from the Pn.p Cells Amplifies the AC Signal To examine whether the VPCs or their descendants are the source of the AC-independent LIN-3 signal, we expressed lin-3 dsRNA in the Pn.p cells in order to down-regulate by RNAi any possible lin-3 expression in the VPCs ( Timmons et al. 2003 ). For this purpose, a vector consisting of an inverted repeat of a 921-bp lin-3 cDNA fragment under control of the same Pn.p cell-specific lin-31 promoter used above (lin-31::lin-3i) was introduced into wild-type animals ( Tan et al. 1998 ). Vulval induction occurred normally in lin-31::lin-3i animals ( Table 4 , row 1), although the adult animals displayed an 80% penetrant egg-laying defective (Egl) phenotype due to a defect in vulval morphogenesis ( n = 122). In wild-type L4 larvae, the 1° descendants of P6.p in the vulF toroid ring (P6.papl/r and P6.ppal/r,), secrete LIN-3 to specify the ventral uterine (uv1) cell fate in the somatic gonad ( Chang et al. 1999 ). If LIN-3 expression in the F cells is blocked through a mutation in the egl-38 pax transcription factor, then the uv1 cell adopts a uterine seam fate, resulting in an Egl phenotype. Thus, lin-31::lin-3i appeared to efficiently reduce LIN-3 expression in the vulval F cells without reducing the activity of LIN-3 in the AC. To further authenticate the efficiency of this approach, we crossed animals carrying the lin-31::lin-3i transgene to animals expressing the short splice variant of lin-3 cDNA in the Pn.p cells under control of the lin-31 promoter ( lin-31::lin-3S, see below). The lin-31::lin-3i transgene almost completely suppressed the Muv phenotype caused by the lin-31::lin-3S transgene, while the lin-31::cre transgene that was used as negative control had no effect ( Table 4 , rows 2–4). Furthermore, the lin-31::lin-3i transgene significantly reduced vulval induction in lin-3S animals that carry a lin-3 minigene encoding the short splice variant ( Figure 4 A), as well as in let-60 ras(gf) and hs::mpk-1 animals ( Table 4 , rows 5–14) . Consistent with an AC-independent function of LIN-3 in the VPCs, the lin-31::lin-3i transgene also affected vulval induction in let-60(gf) animals lacking a gonad ( Table 4 , rows 11 and 12). Figure 4 Alternative Splicing of lin-3 mRNA (A) RT-PCR amplification of lin-3 mRNA from mixed-stage N2 cDNA before (left) and after (right) size fractionation by preparative agarose gel electrophoresis. The lowest band corresponding to LIN-3S is most prominent, and the two upper bands correspond to LIN-3L and LIN-3XL. (B) Intron-exon structure of the lin-3 locus. The lin-3L splice variant is generated by the usage of an alternative (more 3′ located) splice donor in exon 6a. The lin-3XL variant contains the additional exon 6b inserted between exons 6a and 7. The regions encoding the EGF repeat in exon 5 and part of 6a and the transmembrane domain in exon 7 are outlined, and the positions of the PstI sites used for the construction of the minigenes are indicated (see Materials and Methods ). The structure of the lin-3S and lin-3L minigenes is shown in the lower part of the graphic. (C) Sequence alignment of the alternatively spliced region in LIN-3 with the corresponding region in Drosophila Spitz. The 15 and 41 amino acids in LIN-3L and LIN-3XL, respectively, in the juxtamembrane region break the alignment of LIN-3 with Spitz. The C-terminal end of the EGF domain is underlined with a horizontally hatched bar, and the beginning of the transmembrane domain is underlined by a diagonally hatched line. Table 4 Pn.p Cell-Specific Function of lin-3 Vulval induction was scored as described in the legend to Table 1 . See Table 1 legend for key to abbreviations and terminology. Alleles used: let-60(n1046gf), egl-38(n578), gaIs36[hs::mpk-1, D-mek-2(gf)], zhEx72[lin-31::lin-3S, unc-119(+), sur-5::gfp], zhEx68[lin-3S, unc-119(+), sur-5::gfp], zhEx88[lin-31::lin-3i, unc-119(+), myo-3::gfp], and zhEx81[lin-31::cre, unc-119(+), myo-3::gfp]. a The gonad precursor cells Z1 through Z4 were ablated in L1 larvae where indicated b Three independent transgenic lines displayed a penetrant Egl phenotype and suppressed the let-60(gf) phenotype to induction indices L1 ranging from 3.1 to 3.3, and one line was used for further analysis. Two lines displayed no Egl phenotype and did not suppress the let-60(gf) phenotype (induction index 4.2 and 4.3). All five lines exhibited normal vulval induction in a wild-type background c L1 and L2 larvae were heat-shocked for 30 min at 33 °C and grown at 25 °C until L4 As an independent test to determine if a reduction of LIN-3 expression in the vulval cell lineage affects induction, we used the n578 reduction-of-function mutation in the egl-38 pax transcription factor, because this egl-38 allele has been shown to eliminate LIN-3 expression in the 1° cell lineage ( Chang et al. 1999 ). Although egl-38(rf) single mutants exhibited wild-type levels of vulval induction, the egl-38(rf) mutation reduced the Muv phenotype of hs::mpk-1 animals to a similar degree as the rom-1(0) mutation or the lin-31::lin-3i transgene ( Table 4 , rows 15 and 16). Thus, EGL-38 is necessary for the AC-independent function of LIN-3. In summary, these experiments indicated that, during the process of vulval cell fate specification, some of the Pn.p cells (probably the VPCs) produce LIN-3 to amplify the inductive signal. Three LIN-3 EGF Splice Variants that Differ in the Juxtamembrane Domain The lin-3 locus encodes two splice variants termed LIN-3S (short) and LIN-3L (long) that are generated by the differential choice of the splice donor of exon 6 ( Figure 4 B) ( Hill and Sternberg 1992 ). While performing RT-PCR experiments using a primer pair flanking the differentially spliced exons, we discovered a third splice variant, termed LIN-3XL, that is generated by the insertion of an additional exon (6b) between exons 6 and 7 ( Figure 4 A and 4 B). The LIN-3XL splice variant was independently isolated from the yk1053b07EST clone. LIN-3XL contains a 41 amino acid insert, and LIN-3L contains a 15 amino acid insert, in the region between the EGF repeat and the transmembrane domain, when compared to LIN-3S ( Figure 4 C). Since the analogous region in Drosophila Spitz EGF is required for the proteolytic processing of Spitz by Rhomboid ( Bang and Kintner 2000 ; Lee et al. 2001 ; Urban and Freeman 2003 ), we sought to determine which of the LIN-3 splice variants depend on ROM-1 activity. To address this question, we first constructed lin-3 minigenes by replacing the differentially spliced exons with cDNA fragments encoding either of the splice forms (see Figure 4 B). Both lin-3L and lin-3S minigenes were capable of inducing a Muv phenotype, but we observed a marked difference in the dosages required to elicit this phenotype. All (12 out of 12) transgenic lines generated by injection of a relatively low (1 ng/μl) or high (100 ng/μl) concentration of the lin-3S minigene exhibited a strong Muv phenotype (with induction indices ranging from 4.1 to 5.6). In contrast, the lin-3L construct caused a Muv phenotype only when injected at a high concentration. (None of the seven lin-3L lines obtained by injecting 1 ng/μl exhibited a Muv phenotype, while all nine lines obtained by injecting 100 ng/μl exhibited a Muv phenotype, with induction indices ranging from 4.2 to 5.0). For the lin-3XL construct, we obtained variable results; some lines exhibited a weak Muv and others no or even a Vul phenotype (unpublished data). Since we failed to observe a consistent phenotype with this minigene construct, we did not further pursue the analysis of the lin-3XL minigene. ROM-1 Is Necessary for the Activation of the Long LIN-3 Splice Variant To investigate the genetic interactions between rom-1 and the lin-3 splice variants, we compared one line for each of the lin-3S and lin-3L minigenes that displayed a similar degree of vulval induction ( Table 5 , rows 1 and 4). Since the presence of endogenous LIN-3 might mask a specific requirement of either LIN-3 splice variant, we introduced the two minigenes into a lin-3(0) background. The lin-3S and lin-3L transgenes both rescued the larval lethality of lin-3(0) mutants, yielding adult Muv animals ( Table 5 , rows 2 and 5 and Table 6 , rows 1 and 3). Since we used multicopy arrays that are silenced in the germ cells and LIN-3 is required in the oocytes to induce ovulation ( Clandinin et al. 1998 ), the rescued lin-3(0); lin-3S and lin-3(0); lin-3L animals were sterile. Loss of rom-1 function did not affect the viability or the Muv phenotype of lin-3(0); lin-3S animals ( Table 5 , rows 2 and 3 and Table 6 , row 2). In contrast, the efficiency of the lin-3L transgene in rescuing the larval lethality of lin-3(0) mutants was reduced by loss of rom-1 function ( Table 6 , row 4). Moreover, the rare rom-1(0), lin-3(0); lin-3L animals that escaped the larval lethality exhibited a weaker Muv phenotype than lin-3(0); lin-3L animals, suggesting that the function of LIN-3L during vulval induction partially depends on ROM-1 activity ( Table 5 , rows 5 and 6). Table 5 The lin-3L Splice Form Depends on rom-1 Activity Vulval induction was scored as described in the legend to Table 1 . See Table 1 legend for key to abbreviations and terminology. Alleles used: rom-1(zh18), lin-3(n1049null) [the sterile Dpy nonUnc progeny segregated by dpy-20(e1282) lin-3(n1049)/ unc24(e138) unc-unc-44(e362) mothers were examined], zhEx68[lin-3S, sur-5::gfp, unc-119(+)], zhEx69[lin-3L, sur-5::gfp, unc-119(+)], zhEx72[lin-31::lin-3S, unc-119(+), sur-5::gfp], and zhEx73[lin-31::lin-3L, unc-119(+) sur-5::gfp] a The gonad precursor cells Z1 through Z4 were ablated in L1 larvae where indicated Table 6 Rescue of the Larval Lethality in lin-3(0) Mutants by the lin-3S and lin-3L Minigenes See Table 1 legend for key to abbreviations and terminology. To confirm that the viable Dpy nonUnc animals were the rescued lin-3(0) homozygotes and not recombinants between dpy-20 and lin-3, they were scored for fertility 4–5 d later. All animals counted for this table developed into sterile adults due to the silencing of the rescuing transgenes in the germline. In contrast, all viable Dpy nonUnc progeny segregated by dpy-20(e1282) lin-3(n1049)/ unc-44(e362) unc24(e138) mothers lacking a lin-3 minigene developed into fertile adults, indicating recombination between dpy-20 and lin-3 . (Seven recombinants were found among 1,000 F1 progeny animals.) Alleles used: rom-1(zh18), lin-3(n1049) , dpy-20(1282), unc24(e138), unc-44(e362), zhEx68[lin-3S, sur-5::gfp, unc-119(+)], and zhEx69[lin-3L,sur- sur-5::gfp, unc-119(+)]. a To score viability, 100–200 GFP-positive embryos segregated by dpy-20(e1282) lin-3(n1049)/ unc24(e138) unc-44(e262) mothers carrying the indicated minigenes were placed on NGM plates. After 24 and 48 h, the viable Dpy nonUnc larvae representing the rescued lin-3(0) animals and the dead larvae were counted, and the % viability was calculated as 100× [Dpy nonUncs/(Dpy nonUncs+ dead larvae)] To specifically test the function of the LIN-3 splice variants in the Pn.p cells, we cloned full-length cDNAs encoding the LIN-3S and LIN-3L splice variants under control of the Pn.p cell-specific lin-31 promoter (lin-31::lin-3S and lin-31::lin-3L) . The lin-31::lin-3S and lin-31::lin-3L transgenes both caused a strong Muv phenotype in the presence and absence of the AC ( Table 5 , rows 7, 9, 11, and 13). Loss of rom-1 function did not change the phenotype of lin-31::lin-3S animals ( Table 5 , rows 8–10), but it strongly suppressed the Muv phenotype of lin-31::lin-3L animals ( Table 5 , rows 12 and 14). Thus, ROM-1 is required for the activity of the LIN-3L splice variant in the Pn.p cells, while LIN-3S functions independently of ROM-1. Discussion ROM-1 Positively Regulates LIN-3 EGF Mediated Vulval Induction Of the five Rhomboid proteins predicted by the complete C. elegans genome sequence, only ROM-1, the closest homolog of Drosophila Rhomboid-1, possesses the hallmarks of a serine protease with an intact catalytic center ( Urban et al. 2001 ). Here, we show that ROM-1 acts as a positive regulator of the EGFR/RAS/MAPK signaling pathway during vulval induction, as loss of rom-1 function partially suppresses ectopic vulval induction caused by hyperactivation of the EGFR/RAS/MAPK pathway. Our epistasis analysis points at a role of ROM-1 in activating LIN-3 EGF. The activity of a soluble form of LIN-3 lacking the transmembrane and intracellular domains is completely independent of ROM-1 activity, but the activity of full-length LIN-3 EGF is sensitive to loss of ROM-1 function. Moreover, a mutation in lin-15, which renders vulval induction independent of LIN-3 activity, is not suppressed by loss of rom-1 function, but mutations in LET-23 or downstream components of the EGFR/RAS/MAPK pathway efficiently suppress the lin-15 Muv phenotype ( Clark et al. 1994 ; Huang et al. 1994 ). Although ROM-1 enhances the activity of the inductive LIN-3 EGF signal, ROM-1 is not required for vulval induction under normal growth conditions. Loss of rom-1 function does not enhance the Vul phenotype caused by mutations that reduce RTK/RAS/MAPK signaling in the proximal VPCs, indicating that ROM-1 is not required for the induction of the proximal VPCs by the AC. These observations point at the existence of another, yet unidentified protease that mediates the release of LIN-3 from the AC. Like vertebrate TGF-α, and unlike Drosophila Spitz, which absolutely depends on Rhomboid function, membrane-bound LIN-3 might be cleaved at the surface of the AC by an ADAM family metalloprotease ( Peschon et al. 1998 ). Another possibility our experiments have not ruled out is that in rom-1(0) mutants an unprocessed, membrane-bound form of LIN-3 that is retained on the plasma membrane of the AC induces the 1° fate in the adjacent VPC P6.p through juxtacrine signaling ( Anklesaria et al. 1990 ). Once the AC has induced the 1° cell fate in P6.p, the 2° cell fate specification in the neighboring VPCs P5.p and P7.p can occur exclusively through lateral LIN-12 NOTCH signaling, resulting in a wild-type vulva ( Greenwald et al. 1983 ; Kenyon 1995 ; Simske and Kim 1995 ). However, the AC is separated from the VPCs by two adjacent basal laminas that dissolve only after the vulval cell fates have been induced ( Sherwood and Sternberg 2003 ). It is therefore difficult to predict whether LIN-3 anchored in the plasma membrane of the AC could reach its receptor LET-23 EGFR on the basolateral surface of P6.p. ROM-1 Is Required for LIN-3 EGF Activity in the Pn.p Cells Three lines of evidence indicate that ROM-1 functions in the signal-receiving VPCs rather than the signal-sending AC. First, a rom-1::nls::gfp transcriptional reporter is expressed in the VPCs but not in the AC around the time of vulval induction. The rom-1 gene appears to be a transcriptional target of the EGFR/RAS/MAPK pathway, as rom-1::nls::gfp expression is up-regulated in the induced VPCs in response to AC signaling. Second, the expression of rom-1 in the Pn.p cells rescues loss of rom-1 function. Third, loss of rom-1 function in animals lacking an AC results in a further suppression of vulval induction and reduction of egl-17::cfp expression, indicating that the main, if not the only, focus of rom-1 action is outside of the gonad. On the other hand, our epistasis analysis and the biochemical experiments done with Drosophila Rhomboid-1 ( Bang and Kintner 2000 ; Lee et al. 2001 ; Urban and Freeman 2003 ) suggest that ROM-1 is required cell-autonomously for the activation of a membrane-bound LIN-3 precursor. This apparent discrepancy can be explained by the previously published observation of residual lin-3 activity in the absence of the gonad ( Hill and Sternberg 1992 ) and by the additional experiments presented in this paper that uncovered an AC-independent function of LIN-3 during vulval induction. The most likely source of LIN-3 besides the AC are the VPCs, as reducing lin-3 function in the Pn.p cells by tissue-specific RNAi or a mutation in the egl-38 gene, which is required for lin-3 expression in vulval cells ( Chang et al. 1999 ), had essentially the same effect as loss of rom-1 function. Using transcriptional reporter constructs, lin-3 expression has been observed in vulval cells of the 1° lineage beginning in the early L4 stage ( Chang et al. 1999 ), and occasionally we observed weak lin-3 expression in the VPCs or their daughter cells (unpublished data). It is possible that the reporter constructs used were lacking some of the regulatory sequences necessary to drive strong lin-3 expression in the VPC lineage. Other potential sources of LIN-3 may be the posterior ectoderm or the excretory system in the head. However, it seems unlikely that LIN-3 secreted from cells at the anterior or posterior end of the animal influences vulval induction, since we did not observe a bias favoring the induction of anterior or posterior VPCs in the absence of the AC. A Relay Model for Vulval Induction Expression levels of rom-1::nls::gfp are highest in the proximal VPCs (P5.p, P6.p, and P7.p) that adopt 1° and 2° vulval cell fates, suggesting that the proximal VPCs are competent to secrete LIN-3 in response to the inductive AC signal. LIN-3 from the proximal VPCs may facilitate the induction of the more distally located VPCs by paracrine signaling ( Figure 5 ). Such a relay model is reminiscent of the EGF signaling during Drosophila oogenesis ( Freeman et al. 1992 ; Wasserman and Freeman 1998 ). The Gurken growth factor produced by the Drosophila oocyte initially activates the EGFR/RAS/MAPK pathway in the adjacent epithelial follicle cells on the dorsal side of the oocyte independently of Rhomboid. In response to the Gurken signal, the dorsal follicle cells secrete Spitz in a Rhomboid-dependent manner and activate the EGFR in the neighboring follicle cells by paracrine signaling, allowing the signal to spread along the dorsal follicle cell layer. In contrast to Drosophila oogenesis, signal spreading is not necessary for the development of a wild-type C. elegans vulva. The AC needs to induce only the nearest VPC, P6.p, since a 1° cell can specify the 2° fate in the neighboring cells exclusively through lateral LIN-12 NOTCH signaling ( Greenwald et al. 1983 ; Simske and Kim 1995 ). On the other hand, dosage experiments have indicated that low levels of inductive LIN-3 signal can directly specify the 2° cell fate in the absence of lateral signaling ( Katz et al. 1995 ). It is therefore possible that the relay signal generated by ROM-1 and LIN-3 in P6.p contributes to the specification of the 2° cell fate in the neighboring VPCs in combination with the lateral LIN-12 NOTCH signal. LIN-3 secreted from the proximal VPCs could initially serve to maintain the competence of all VPCs, while at a later phase of induction the AC and lateral signals would seal the 1° fate of P6.p and the 2° fate of P5.p and P7.p, respectively. A similar two-step model of vulval induction has been proposed for other rhabditid nematode species such as Oscheius sp. ( Felix et al. 2000 ). In C. elegans, ROM-1 is dispensable for the induction of the proximal VPCs, and the relay mechanism mediated by LIN-3 and ROM-1 only becomes apparent in a sensitized genetic background in which distal VPCs adopt induced cell fates. It is interesting to note in this context that in Mesorhabditis and Teratorhabditis, in which the vulva develops in the posterior body region, induction occurs without a signal from an AC or any other gonad cell ( Sommer and Sternberg 1994 ). In these posterior-vulva Rhabditidae, the VPCs are not equivalent, because only P5.p and P6.p are competent to adopt the 1° fate. The mechanism that generates this intrinsic difference among the VPCs is unknown. It is possible that in these nematodes the specification of the VPC cell fates occurs in a cell-autonomous manner that could involve EGF signaling between the VPCs. The AC in C. elegans serves to position the vulva in the central body region, and the function of an AC appears to be absent in the posterior-vulva Rhabditidae . Figure 5 A Relay Model for Vulval Induction The AC initiates vulval development by secreting the LIN-3 growth factor independently of ROM-1. In response to the AC signal, the proximal VPCs up-regulate ROM-1 expression and start secreting LIN-3 in a ROM-1-dependent manner to relay the AC signal. Splice Variant-Specific Action of ROM-1 The two previously identified LIN-3 splice forms (LIN-3S and LIN-3L) as well as the newly identified longer variant (LIN-3XL) differ by 15 and 41 amino acid insertions in the juxtamembrane region just prior to the predicted Rhomboid cleavage site at the start of the transmembrane domain ( Hill and Sternberg 1992 ). Our experiments with LIN-3 minigenes indicate that the activity of the shortest splice form (LIN-3S) is completely independent of ROM-1 function. LIN-3S is expressed at all stages of development, and the LIN-3S minigene rescued all phenotypes caused by the lin-3(0) mutation, including the ovulation defects (unpublished data). This may explain why loss of rom-1 function causes neither the larval lethality nor the sterility observed in lin-3 mutants. Furthermore, our data indicate that LIN-3L function in the VPCs almost completely depends on ROM-1 activity. It seems improbable that the 15 amino acid insertion in LIN-3L could change the substrate specificity toward the ROM-1 protease. A more likely explanation for the inherent difference in the dependence of the LIN-3 splice variants on ROM-1 is suggested by the experiments performed with Drosophila Spitz ( Lee et al. 2001 ; Tsruya et al. 2002 ). When expressed in mammalian cells that lack Rhomboid activity, Spitz is retained in the Golgi apparatus. Introducing functional Rhomboid into these cells allows the cleavage and release of Spitz from the Golgi apparatus, resulting in the secretion of the extracellular portion of Spitz. In analogy to Spitz, the small insert in LIN-3L could cause the retention of this LIN-3 isoform in the Golgi apparatus, thus rendering LIN-3L dependent on ROM-1 mediated processing. The tissue distribution of the LIN-3 splice variants is unknown, although all three forms can be detected by RT-PCR in L2 and L3 larvae around the time of vulval induction (unpublished data). In view of the relay model discussed above ( Figure 5 ), tissue-specific splicing may account for the distinct functions of LIN-3. The roles of the two Drosophila EGF-like growth factors Gurken and Spitz may be fulfilled in C. elegans by the splice variants LIN-3S and LIN-3L, respectively. In this model, the AC uses the ROM-1-independent isoform LIN-3S to induce vulval development, and the proximal VPCs relay the AC signal to the distal VPCs by secreting LIN-3L. Alternative splicing has also been reported for the Neuregulin family of EGF-like ligands in vertebrates ( Chang et al. 1997 ). Different Neuregulin isoforms elicit distinct responses by activating different EGFRs ( Meyer et al. 1997 ). The tissue-specific expression of Rhomboid family proteases could determine which isoforms a particular cell type can secrete, thus adding another level of regulation. Materials and Methods General methods and strains used Standard methods were used for maintaining and manipulating Caenorhabditis elegans ( Brenner 1974 ). The C. elegans Bristol strain, variety N2, was used as the wild-type reference strain in all experiments. Unless noted otherwise, the mutations used have been described in Riddle and National Center for Biotechnology Information (U.S.) (2001) and are listed below by their linkage group: LGI: pry-1(mu38) ( Gleason et al. 2002 ); LGIII: dpy-19(e1259), lin-12(n137gf), rom-1(zh18) (this study), rom-2(ok966) ( C. elegans Gene Knockout Consortium), and unc-119(e2498); LGIV: let-60(n1046) ( Beitel et al. 1990 ), lin-3(n1049), unc-5(e53), unc-44(362), lin-45(sy96), unc-24(e138), mec-3(e1338), dpy-20(e1282), egl-38(n578), and mec-3(n3197); LGX: sem-5(n2019) and lin-15(n309); extrachromosomal and integrated arrays: zhEx22[lin-3(+), sur-5::gfp, unc-119(+)], zhE66[lin-31::rom-1, unc-119(+), sur-5::gfp], zhEx72[lin-31::lin-3S, unc-119(+), sur-5::gfp], zhEx68[lin-3S, unc-119(+), sur-5::gfp], zhEx69[lin-3L, sur-5::gfp, unc-119(+)],zhEx73[lin-31::lin- zhEx73[lin-31::lin-3L, unc-119(+) sur-5::gfp], zhEx78[ACEL:: Δ pes-10::nls::gfp, unc-119(+)], zhEx81[lin-31::cre, unc-119(+), myo-3::gfp], zhE88[lin-31::lin-3i, unc-119(+), myo-3::gfp], zhEx89[ACEL::rom-1, sur-5::gfp], syIs12[hs::lin-3extra] ( Katz et al. 1995 ), zhIs5[rom-1::nls::gfp, unc-119(+)] , huIs7[hs::bar-1 Δ NT, dpy-20(+)] ( Gleason et al. 2002 ), gaIs36[hs-mpk-1, IF1alpha-Dmek-2] ( Lackner and Kim 1998 ), and arIs92[egl-17::cfp] ( Yoo et al. 2004 ). Unless noted in the table legends, all experiments were conducted at 20 °C. Transgenic lines were generated by injecting the experimental DNA at a concentration of 100 ng/μl or at the concentrations indicated in the text into both arms of the syncytial gonad as described ( Mello et al. 1991 ).The constructs pUnc-119 (20 ng/μl), pPD93.97 ( myo-3::gfp , 40 ng/μl), and pTG96 ( sur-5::gfp , 100 ng/μl) were used as a cotransformation markers ( Maduro and Pilgrim 1995 ; Yochem et al. 1997 ). The extrachromosomal array zhEx[rom-1::nls::gfprom-1::gfp; unc-119(+)] was integrated in the genome animals following γ-irradiation with 3,000 Rad to generate the array zhIs5 and backcrossed six times before analysis. Double and triple mutants were constructed using standard genetic methods. Where cis-linked markers were used they are indicated in the table legends. Plasmid constructs The transcriptional rom-1::nls::gfp reporter construct (pRH2) was generated by ligating a HindIII-NheI–restricted 6,998-bp genomic fragment spanning the entire 5′ upstream region of F26F4.3 isolated by PCR amplification with the primers OAD49 (5′-GGAAGCTTGCATGCCCAACGAAATCGATA-3′) and OAD59 (5′-GGGCTAGCCATGTTGTGGAGAAGGAGAAC-3′) into the HindIII-XbaI site of pPD96.04. The rom-1::gfp translational reporter construct (pAD31) was generated by PCR amplification of a 3,146-bp genomic fragment containing 1,849 bp of 5′ sequences and the entire rom-1 ORF using the primers OAD47 (5′-GACTCTAGAGTTGTCAAAAGGTCACGGG-3′) and OAD51 (5′-ATCCTCTAGAGTTGAGCAATTTTCGTTGTTCCAC-3′') followed by XbaI restriction and ligation to XbaI-digested vector pPD95.75. The upstream promoter region of this construct was further extended by replacing a 420-bp PstI fragment with a 2,099-bp PstI genomic fragment corresponding to positions –1,432 and –3,531 relative to the predicted translation start codon of F26F4.3. The lin-31::rom-1 construct (pAD16) was generated by ligating a 1,601-bp SalI-NotI fragment spanning the entire rom-1 coding sequence amplified with the primers OAD44 (5′-TTTTGGTCGACCTCCTTCTCCACAAC-3′) and OAD45 (5′-TTTGGCGGCCGCCTATGAGCAATTTTCG-3′) into the SalI-NotI site of the pB253 vector ( Tan et al. 1998 ). To generate the ACEL::rom-1 construct, a 2.3-kb SalI genomic lin-3 fragment encompassing the third intron, which contains the ACEL ( Hwang and Sternberg 2003 ), was cloned into the SalI site of the pTB11 plasmid, which consists of an nls::gfp reporter cassette under control of the truncated pes-10 minimal promoter ( Berset et al. 2001 ). Transgenic animals carrying the resulting ACEL:: Δ pes-10::nls::gfp construct (pAH67) showed strong and specific GFP expression in the AC beginning in the mid L2 stage as reported ( Hwang and Sternberg 2003 ). The KpnI-EcoRI fragment encoding the nls::gfp cassette was then replaced with a 1.6-kb KpnI-NotI rom-1 fragment isolated from the lin-31::rom-1 plasmid described above to yield the ACEL::rom-1 construct. For the lin-31::lin-3 splice variant constructs, partial cDNAs covering the differentially spliced region in the lin-3 mRNA were amplified with the primers OAD31 (5′-CCCTTCGTGGTTTCGTCAAGAACGTAGTGC-3′) and OAD32 (5′-CGTATCTGCAGAATCCAACTCGATATTAATTAC-3′) using first-strand cDNA synthesized from mixed-stage total RNA as template. The PCR-amplified products were size-fractionated by agarose gel electrophoresis, cloned into the pGEMT vector (Promega), and sequenced to identify clones encoding individual splice variants. To obtain full-length lin-3 cDNA construct (pAD27), a 1,996-bp XhoI fragment from the EST clone yk1053b07 (confirmed to encode full-length lin-3XL cDNA by DNA sequencing) was first subcloned into the XhoI site of a modified pBluescriptSK (Stratagene, La Jolla, California, United States) vector (pAD23) in which the PstI site had been destroyed by restriction with EcoRV and SmaI and religation of the resulting blunt ends. To generate full-length lin-3S and lin-3L cDNA constructs (pAD25 and pAD26, respectively), 1,065-bp and 1,110-bp PstI fragments specific for each splice variant isolated from the partial cDNA clones described above were used to replace the 1,188-bp PstI fragment in the full-length lin-3XL cDNA construct (pAD27). The lin-31::lin-3S and lin-31::lin-3L constructs were generated by cloning the 1,133-bp and 1,178-bp XhoI cDNA fragments of the S and L splice variants into the SalI site of pB253 ( Tan et al. 1998 ). The lin-3 splice variant minigene constructs (pAH63, pAH64, and pAH65 for lin-3S, lin-3L, and lin-3XL, respectively) were generated by cloning a 6.1-kb genomic fragment spanning the entire ORF of lin-3 and 574 bp of 5′ and 236 bp of 3′ sequences amplified with the primers OAH137 (5′-CCAGAAAGTTCATGTGAATCAT-3′) and OAH138 (5′-TCACAGGAACTGAGAGGGAGAGTG-3′) into the pGEMT vector. From this construct, a 6,206-bp ApaI-SacI fragment was subcloned into pAD23 to obtain pAH62. The minigenes encoding each of the splice variants were obtained by replacing the 2,728-bp lin-3 genomic PstI fragment with 1,065-, 1,110-, and 1,188-bp PstI fragments isolated from cDNAs of the different splice variants. To construct the lin-31::lin-3 hairpin plasmid (pAD35), a 964-bp NdeI-HindIII lin-3S cDNA fragment from pAD25 was cloned into NdeI-HindIII–digested pAD27 using the recA – E. coli SURE strain as host to obtain pAD32. The resulting 1,918-bp lin-3 hairpin fragment was excised with XhoI from pAD32 and subcloned in E. coli SURE into the SalI site of the pB253 vector to obtain pAD35. RNA interference The dsRNA to interfere with rom-1 function was generated by in vitro transcription using a 350-bp rom-1 cDNA fragment corresponding to the nucleotides –17 to 725 relative to the predicted start codon of the ORF inserted into pGEM-T (Stratagene) as template. Transcripts were prepared using T7 and Sp6 RNA polymerase and annealed prior to injection as described ( Fire et al. 1998 ). Progeny of the injected animals were assayed at 20 °C. RNAi of rom-2 and rom-3 was done by feeding the animals dsRNA-producing E. coli at 20 °C as described ( Kamath et al. 2001 ). Isolation of the rom-1(zh18) deletion allele The rom-1(zh18) deletion mutant was isolated from an ethyl methanesulfonate–mutagenized library consisting of approximately 10 6 haploid genomes as previously described ( Jansen et al. 1997 ; Berset et al. 2001 ). DNA pools were screened by nested PCR with primers Rho13 (5′-GAGACCGGGGACCGTATTCTGGCAC-3′) and Rho10 (5′-GAGAGCATAAACTCCTGCGGAAGCACC-3′) in a first PCR reaction, and Rho35 (5′-GGGAATCCGACGGTGGTAGAAGC-3′) and Rho10 (5′-GAGAGCATAAACTCCTGCGGAAGCACC-3′) in a second PCR reaction. The zh18 deletion removes 1,556 bp including 384 bp of 5′ upstream sequences and 618 bp of the rom-1 ORF (positions 4906804–4908360 in the cosmid F26F4). The mutant strain was backcrossed six times against N2 before further experiments were done. Vulval induction assay Vulval induction was scored by examining worms at the L4 stage under Nomarski optics as described ( Sternberg and Horvitz 1986 ). The number of VPCs that adopted a 1° or 2° vulval fate was counted for each animal as described ( Sternberg and Horvitz 1986 ), and the induction index was calculated by dividing the number of 1° or 2° induced cells by the number of animals scored. Statistical analysis was performed using a t-test for independent samples. To remove the AC, the nuclei of the Z1 to Z4 gonadal precursor cells were ablated in early L1 larvae with a laser microbeam as described ( Sulston and White 1980 ; Kimble 1981 ). The operated animals were allowed to develop until the L4 stage. Only those animals in which neither gonad arm developed and no residual gonadal cells survived were scored. Supporting Information Accession Numbers The GenBank ( http://www.ncbi.nlm.nih.gov/ ) accession numbers of the Rhomboid genes discussed in this paper are C.e. ROM-1 (AAA91218), C.e. ROM-2 (CAA82377), C.e. ROM-3 (CAB55154), C.e. ROM-4 (CAB55122), C.e. ROM-5 (AAF60768), D.m. Rho-1 (CAA36692), D.m. Rho-2 (AAK06752), D.m. Rho-3 (AAK06753), D.m. Rho-4 (AAK06754), D.m. Rho-6 (NP_523557), H.s. Rho-1 (CAA76629).

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC519001.xml

529303

Early detection of Pseudomonas aeruginosa – comparison of conventional versus molecular (PCR) detection directly from adult patients with cystic fibrosis (CF)

Background Pseudomonas aeruginosa (PA) is the most important bacterial pathogen in patients with cystic fibrosis (CF) patients. Currently, routine bacteriological culture on selective/non- selective culture media is the cornerstone of microbiological detection. The aim of this study was to compare isolation rates of PA by conventional culture and molecular (PCR) detection directly from sputum. Methods Adult patients (n = 57) attending the regional adult CF centre in Northern Ireland, provided fresh sputum following airways clearance exercise. Following processing of the specimen with sputasol (1:1 vol), the specimen was examined for the presence of PA by plating onto a combination of culture media ( Pseudomonas isolation agar, Blood agar & McConkey agar). In addition, from the same specimen, genomic bacterial DNA was extracted (1 ml) and was amplified employing two sequence-specific targets, namely (i) the outer membrane protein ( opr L) gene locus and (ii) the exotoxin A (ETA) gene locus. Results By sputum culture, there were 30 patients positive for PA, whereas by molecular techniques, there were 35 positive patients. In 39 patients (22 PA +ve & 17 PA -ve), there was complete agreement between molecular and conventional detection and with both PCR gene loci. The opr L locus was more sensitive than the ETA locus, as the former was positive in 10 more patients and there were no patients where the ETA was positive and the opr L target negative. Where a PCR +ve/culture -ve result was recorded (10 patients), we followed these patients and recorded that 5 of these patients converted to being culture-positive at times ranging from 4–17 months later, with a mean lag time of 4.5 months. Conclusions This study indicates that molecular detection of PA in sputum employing the opr L gene target, is a useful technique in the early detection of PA, gaining on average 4.5 months over conventional culture. It now remains to be established whether aggressive antibiotic intervention at this earlier stage, based on PCR detection, has any significant benefits on clinical outcome.

Introduction Cystic fibrosis [CF] is the most commonly inherited fatal disease in persons originating from a white and European background, currently affecting approximately 30,000 adults and children in the US [ 1 ]. The defective gene carrying the mutation is carried in one in every 31 Americans [one in 28 Caucasians], equating to more than 10 million people being a symptomless carrier of the defective gene [ 1 ]. It is an autosomal recessive condition whereby two alleles carrying a polymorphism in the cystic fibrosis transmembrane conductance regulator [CFTR] gene phenotypically manifest the disease state through a variety of multiorgan problems, associated with a pharmacological disfunction to regulate sodium and chloride secretion across cell membranes. The most common complication of CF is the recurrence of chronic chest infections usually caused by bacterial pathogens [ 2 ]. CF patients continue to suffer from recurrent and chronic respiratory tract infections and most of their morbidity and mortality is due to such infections throughout their life [ 3 ]. These infections are usually dominated by Gram-negative organisms, especially by the pseudomonads, including Pseudomonas aeruginosa , Burkholderia cepacia and Stenotrophomonas maltophilia . However, with modern antibiotic management with improved antimicrobial agents, such as the aminoglycosides and carbapenems, CF patients have an improved survival, resulting in more adults in employment. Previous studies have described alternative laboratory markers to conventional bacteriological culture for the detection of Pseudomonas aeruginosa in CF patients.[ 4 , 5 ] These markers have included serological testing of CF patients' sera for the presence of antibody to P. aeruginosa ,[ 4 ] as well as molecular detection of DNA signature sequences of P. aeruginosa in CF patients [ 5 ]. Presently, the routine employment of molecular (PCR) detection of P. aeruginosa directly from the sputum of CF patients in the UK and Ireland is uncommon. Most clinical microbiology laboratories in the UK and Ireland, which support a CF centre, have developed standard operating procedures (SOPs) for the isolation of this organism from patients' sputum. Hence, it was the aim of this study to compare the conventional and molecular detection of P. aeruginosa from the sputum of adult patients attending the adult centre in Northern Ireland, as well as to estimate the lag time of conventional culture to detection, compared with molecular (PCR) detection. Materials & Methods Qualitative conventional detection of Pseudomonas aeruginosa from sputum Duplicate sputa (1 ml minimum) specimens were collected from 57 adult patients with a well characterized history of CF in sterile (100 ml) plastic disposable containers. Sputum was collected immediately after a standardized session of physiotherapy and was stored at ambient temperature and was processed within 4 h from collection. Fresh sputum (1 ml min) was mixed with an equal mass (1:1) of Sputasol (Oxoid SR089A, Oxoid Ltd., Poole, England) and was incubated in a water bath at 37°C for 15 min, before further qualitative processing for the detection of Pseudomonas aeruginosa . Processed sputa (10 μ l) were inoculated and incubated, onto several selective media for the isolation of Pseudomonas aeruginosa , including:- Columbia Blood Agar (Oxoid CM0331) supplemented with 5% (v/v) defribinated horse blood, MacConkey Agar (Oxoid CM0007) and Pseudomonas Isolation Agar (PIA) (Oxoid CM0559 + SR0102). All media were incubated aerobically at 37°C for 48 h, unless otherwise stated. The PIA plates were incubated at room temperature for a further three days following initial 48 hrs incubation. In addition, all different phenotypes from the sputum of each patient were identified phenotypically employing a combination of conventional identification methods (e.g. oxidase), as well as the API Identification schemes (API 20NE, API 20E) (Biomérieux, Les Halles, France). Molecular (PCR) detection of Pseudomonas aeruginosa from sputum DNA extraction All DNA isolation procedures were carried out in a Class II Biological Safety Cabinet (MicroFlow, England) in a room physically separated from that used to set up nucleic acid amplification reaction mixes and also from the "post-PCR" room in accordance with the Good Molecular Diagnostic Procedures (GMDP) guidelines of Millar et al [ 6 ], in order to minimise contamination and hence the possibility of false positive results. Bacterial genomic DNA was extracted directly from the patients' sputum, as well as from the reference strain Pseudomonas aeruginosa (Schroeter; Migula) ATCC 27853, by employment of the Roche High Purity PCR Template Preparation Kit (Roche, England), in accordance with the manufacturer's instructions. Extracted DNA was stored at -80°C prior to PCR amplification. For each batch of extractions, a negative extraction control containing all reagents minus sputum, was performed, as well as an extraction positive control with P. aeruginosa . PCR amplification All reaction mixes were set up in a PCR hood in a room separate from that used to extract DNA and the amplification and post-PCR room in order to minimise contamination. Initially PCR amplification conditions were optimised by separately varying magnesium chloride concentration, annealing temperature, primer concentration and DNA template concentration. Following optimisation, reaction mixes (100 μl) were set up as follows:-10 mM Tris-HCl, pH 8.3, 50 mM KCl, 2.5 mM MgCl 2 , 200 μM (each) dATP, dCTP, dGTP and dTTP; 1.25 U of Taq DNA polymerase (Amplitaq; Perkin Elmer), 0.1 μM (each) of the each set of primers (exoA & opr L) (Table 1 ) and 4 μl of DNA template. The reaction mixtures following a "hot start" were subjected to the following empirically optimized thermal cycling parameters in a Perkin Elmer 2400 thermocycler: 96°C for 5 min followed by 40 cycles of 96°C for 1 min, 55°C for 1 min, 72°C for 1 min, followed by a final extension at 72°C for 10 min. Positive ( P. aeruginosa ATCC 27853 DNA) and multiple negative (water) amplification controls were included in every set of PCR reactions. In addition, a broad-range/universal PCR was employed with each sputum to demonstrate successful extraction of bacterial DNA from the specimen, as well as lack of PCR inhibition, by employing the highly conserved 16S rDNA primers, PSL/PSR (Table 1 ), as previously described [ 7 ]. Any sputum which failed to amplify this locus was re-extracted and amplified until a positive signal was obtained. Table 1 Oligonucletoides and optimised experimental PCR amplification conditions Target Gene locus Primer 5'-----------------------3' Optimum MgCl 2 concentration (mM) Annealing temperature Size of Amplicon (bp) Eubacteria [7] 16S rRNA f: 5'-AGG ATT AGA TAC CCT GGT AGT CCA-3' 2.5 55°C 312 r: 5'-ACT TAA CCC AAC ATC TCA CGA CAC-3' Pseudomonas aeruginosa [8] opr L 1 f: 5'-ATG GAA ATG CTG AAA TTC GGC-3' 2.5 55°C 504 r: 5'-CTT CTT CAG CTC GAC GCG ACG-3' Pseudomonas aeruginosa [9] exo A 2 f: 5'-GAC AAC GCC CTC AGC ATC ACC AGC-3' 2.5 72°C 396 r: 5'-CGC TGG CCC ATT CGC TCC AGC GCT-3' 1 , outer membrane lipoprotein; 2 , exotoxin A. Detection of amplicons Following amplification, aliquots (10 μl) were removed from each reaction mixture and examined by electrophoresis (80 V, 45 min) in gels composed of 2% (w/v) agarose (Gibco, UK) in TAE buffer (40 mM Tris, 20 mM acetic acid, 1 mM EDTA, pH 8.3), stained with ethidium bromide (5 μg/100 ml). Gels were visualised under UV illumination using a gel image analysis system (UVP Products, England) and all images archived as digital graphic files (*.bmp). Where a band was visualized at the correct expected size for either exo A or opr L loci, the specimen was considered positive for P. aeruginosa . Results and Discussion A comparison of the occurrence of P. aeruginosa in patients' sputum by conventional and molecular techniques is shown (Table 2 ). By culture, there were 30 patients positive for P. aeruginosa , whereas by molecular techniques, there were 35 positive patients. In 39 patients (22 P aeruginosa [PA] positive [+ve] & 17 P. aeruginosa [PA] negative [-ve]), there was complete agreement between molecular and conventional detection techniques. The opr L locus was more sensitive than the ETA locus, as the former was positive in 10 more patients and there were no patients where the ETA was positive and the opr L target negative, However, there were five patients who were culture positive but PCR -ve by both gene targets. Given that the universal 16S rDNA PCR was positive for these sputum specimens and that the appropriate molecular controls were working optimally, this may accounted for by potential phenotypic misidentification of P. aeruginosa , which has been recently described.[ 10 ] Alternatively, discrepant results (PCR+ve/culture-ve) could reflect true P. aeruginosa colonization with a false-negative culture result due to sample overgrowth by other bacteria or to the presence of non-cultivable organisms or auxotrophic mutations in the organism. Where a PCR +ve/culture -ve result was recorded (10 patients), we followed these patients and recorded that five of these patients converted to being culture-positive at times ranging from 4–17 months later, with a mean lag time of 4.5 months, whereas the remaining five patients remained negative for P. aeruginosa . Table 2 Comparison of conventional culture with molecular detection of Pseudomonas aeruginosa from the sputum of adult CF patients PCR (OprL) PCR (exotoxin A) Culture Frequency of patients (%) + + + 22 (38.6%) + - + 3 (5.3%) - - + 5 (8.8%) + + - 3 (5.3%) + - - 7 (12.3%) - - - 17 (29.7%) In this study, two PCR assays were performed individually for the molecular detection of P. aeruginosa directly from the sputum of patients with CF. Two gene loci were targeted, as specific markers of P. aeruginosa , namely the exotoxin A (ETA) gene and the outer membrane lipoprotein ( opr L) gene. ETA is produced by the majority of P. aeruginosa strains and can inhibit eucaryotic protein biosynthesis at the level of polypeptide chain elongation factor 2, similarly to diphtheria toxin.[ 9 ] OprL is an outer membrane lipoprotein which has been implicated in efflux transport systems, as well as affecting cell permeability.[ 8 ] Pseudomonas aeruginosa is the most important bacterial pathogen in patients with CF [ 11 ], as demonstrated with high prevalence data in most of the National CF Registries. Chronic Pseudomonas colonisation of the major airways leading to delibating exacerbations of pulmonary infection, is the major cause of morbidity and mortality in patients with CF, hence it is important to be able to reliably detect P. aeruginosa from patients' sputum. Emerson et al . [ 12 ] recently published their findings of a US Cystic Fibrosis Foundation (CFF) registry-based study, which showed that infection related to P. aeruginosa was a major predictor of morbidity and mortality, whereby the 8-year risk of death parameter was 2.6 times higher in patients who had positive sputum cultures for this organism, as well as having a significantly lower percent predicted forced expiratory volume (FEV1). These workers suggested that early interventions may help decrease associated morbidity and mortality of young patients with CF. More recently, Rosenfeld et al . [ 13 ] described the pathophysiology and risk factors for early P. aeruginosa infection in CF. These workers suggested that chronic lower airway infection with P. aeruginosa is associated with significant morbidity and mortality among CF patients [ 13 ]. However, they suggested that first acquisition of P. aeruginosa does not appear to cause an immediate and rapid decline in lung function, as early isolates are generally non-mucoid, antibiotic-sensitive and present at low densities, suggesting a possible "window of opportunity" for early intervention, as they describe [ 13 ]. This study also concluded that there are no controlled trials demonstrating clinical benefit in young children following early intervention, particularly for anti-pseudomonal therapy, as well as the long-term safety profile and optimal drug regimen.[ 13 ] It is therefore important that primary diagnostic bacteriology laboratories have the ability to detect transient and early Pseudomonas colonization as early as possible, so that (i) aggressive antibiotic regimes may be considered,(ii). the patient is managed optimally, in an attempt to avoid early biofilm formation and chronic colonization with Pseudomonas and (iii) appr opr iate infection control precautions are considered. More recently, West et al . showed by a combination of serum IgG, IgA and IgM anti P. aeruginosa antibodies, in conjunction with these authors' Wisconsis Cystic Fibrosis Radiograph score, that P. aeruginosa infections occurred approximately 6–12 months before the organism was recovered from respiratory secretions.[ 4 ] In addition, this study demonstrated that mixing of young child with older chronically colonised children was associated with a significantly increased risk of P. aeruginosa acquisition. Given that not all laboratories employ molecular detection methods for Pseudomonas aeruginosa , either from culture plates or patients' sputum, small numbers of Pseudomonas colonies (n = 1–2) may therefore be missed when present in the early stages of colonization preceding infection of the patient's airways, particularly where such single colonies are mixed alongside other phenotypically similar genera on the primary culture plate.[ 14 ] Pragmatic, practical and cost implications leave it impossible to qualitatively identify the total bacterial microflora present on non-selective primary plates from sputum. Therefore, any rapid molecular screening method should be encouraged to detect low copy numbers of organisms in the early stages of colonization/infection, where the main value of this diagnostic assay is in the rapid screening of patients with no or an intermittent history of Pseudomonas colonization. Although such assays are not generally available in most clinical diagnostic laboratories, these laboratories generally do have access to such technology at regional specialist microbiology centres and it may therefore be prudent to establish routine analysis of children's sputum, particularly from patients with no or an intermittent history by culture of Pseudomonas colonization, at annual review. Furthermore, it is very important to follow up any such patients, which demonstrate a PCR +ve/culture -ve finding from their sputum, in order to establish whether there is transient infection in patients, which never result in established colonization leading to chronic infection. Additionally, it is important to monitor such PCR +ve/culture -ve patients in terms of optimal antibiotic management and infection control. Overall, our study demonstrated that molecular assays for the detection of P. aeruginosa were able to detect P. aeruginosa at an early stage, on average 4.5 months sooner than culture detection. However, it remains unclear what this "window of opportunity" potentially offers in terms of a decrease in P. aeruginosa -associated morbidity and mortality. Authors' contributions JEM, PGM and JSE jointly conceived the study and designed the experiments. JX executed and analyzed all practical aspects of the study, as well as analyzing the data. JEM executed certain molecular components of experimentation, analyzed the data and prepared the manuscript. BCM helped guide the molecular aspects of experimentation and was involved in drafting of the manuscript. PGM provided expert microbiological analysis and interpretation of data. JSE provided clinical expertise, interacted with the CF patients and was involved in interpretation of data, as well as critically reviewing the final manuscript. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529303.xml

524169

Is autoinducer-2 a universal signal for interspecies communication: a comparative genomic and phylogenetic analysis of the synthesis and signal transduction pathways

Background Quorum sensing is a process of bacterial cell-to-cell communication involving the production and detection of extracellular signaling molecules called autoinducers. Recently, it has been proposed that autoinducer-2 (AI-2), a furanosyl borate diester derived from the recycling of S-adenosyl-homocysteine (SAH) to homocysteine, serves as a universal signal for interspecies communication. Results In this study, 138 completed genomes were examined for the genes involved in the synthesis and detection of AI-2. Except for some symbionts and parasites, all organisms have a pathway to recycle SAH, either using a two-step enzymatic conversion by the Pfs and LuxS enzymes or a one-step conversion using SAH-hydrolase (SahH). 51 organisms including most Gamma-, Beta-, and Epsilonproteobacteria, and Firmicutes possess the Pfs-LuxS pathway, while Archaea, Eukarya, Alphaproteobacteria, Actinobacteria and Cyanobacteria prefer the SahH pathway. In all 138 organisms, only the three Vibrio strains had strong, bidirectional matches to the periplasmic AI-2 binding protein LuxP and the central signal relay protein LuxU. The initial two-component sensor kinase protein LuxQ, and the terminal response regulator luxO are found in most Proteobacteria, as well as in some Firmicutes, often in several copies. Conclusions The genomic analysis indicates that the LuxS enzyme required for AI-2 synthesis is widespread in bacteria, while the periplasmic binding protein LuxP is only present in Vibrio strains. Thus, other organisms may either use components different from the AI-2 signal transduction system of Vibrio strains to sense the signal of AI-2, or they do not have such a quorum sensing system at all.

Background Quorum sensing through small signal molecules called autoinducers is an important process for the regulation of population density dependent cellular processes in bacteria, including the production of antibiotics and virulence factors, conjugation, transformation, swarming behaviour and biofilm formation [ 1 , 2 ]. Recently it was discovered that two different density dependent signal transduction cascades are present in Vibrio which converge to trigger luminescence in V. harveyi [ 3 ] and expression of virulence factors in V. cholerae [ 4 , 5 ]. Two chemically different autoinducers are involved in this regulation. While autoinducer-1 is an acylated homoserine lactone (AHL) (N-butanoyl-homoserine lactone) in V. harveyi , the structure of autoinducer-2 (AI-2) has been determined in a complex with the sensor protein LuxP [ 6 ] and shown to be a furanosyl-borate-diester. It is synthesized in two enzymatic steps (by the Pfs enzyme and the LuxS enzyme) from S-adenosyl-homocysteine (SAH), resulting in 4,5-dihydroxy 2,3 pentanedione (DPD) which undergoes spontaneous cyclization and is then complexed with borate to form AI-2 (Fig. 1 ). Figure 1 Enzymes involved in the detoxification of SAH and synthesis of AI-2 and AI-3. Abbreviations: SAM, S-adenosyl-methionine; SAH, S-adenosyl-homocysteine; SRH, S-ribosyl-homocysteine; DPD, 4,5-dihydroxyl-2,3-pentanedione; Pro-AI-2, Ai-2 precursor; AI-2, autoinducer 2; AI-3, autoinducer 3. The numbers in brackets show the numbers of analyzed organisms that have that enzyme (reciprocal best hit). The LuxS enzyme responsible for the last enzymatic step of AI-2 synthesis is present in a wide phylogenetic range of bacterial genera and AI-2 produced by heterologous organisms triggers luminescence in the V. harveyi reporter strains [ 7 , 8 ]. Thus it was hypothesized that AI-2 might be a universal signal molecule. In addition, a large fraction of E. coli genes is transcribed differently with culture supernatants containing AI-2 compared to culture supernatants from luxS - mutants [ 9 , 10 ]. Recently it could be shown that the expression of important genes of the virulence islands in E. coli serotype O157:H7 (EHEC) is controlled by a LuxS dependent molecule, which was later shown to be not AI-2 but AI-3 whose structure is not known yet and which does not activate the V. harveyi bioreporter strain [ 11 ]. It is also produced by the gut microflora. Moreover, the host hormone epinephrine activates virulence gene transcription through the same signalling pathway as AI-3, resulting in cross-talk between host and bacterium [ 11 ]. Knock-out mutants for luxS have been investigated for modifications of infectious phenotypes in some of the currently sequenced pathogens. In V. cholerae [ 4 , 5 ], Streptococcus pyogenes [ 12 , 13 ], Streptococcus pneumoniae [ 14 ], Neisseeria meningitidis [ 15 ] and Clostridiuim perfringens [ 16 ] luxS - mutants showed severe defects in the expression of virulence factors. In some other pathogens luxS - mutants showed none or very subtle changes in virulence related traits (e.g. Borrelia burgdorferi , [ 17 ]; Porphyromonas gingivalis , [ 18 ]; Shigella flexneri , [ 19 ]). In Salmonella typhimurium AI-2 controls the expression of an ABC transporter which is responsible for the back transport of AI-2 into the cell, presumably to conserve metabolic energy or to interfere with quorum sensing mechanisms of the gut microflora [ 20 ]. The signal detection system for AI-2 from Vibrio strains is well experimentally proven [ 4 ] (Fig. 2 ). In V. harveyi it is composed of a soluble periplasmic AI-2 binding protein LuxP, and a phosphorelay cascade resulting in density dependent activation of the lux operon (Fig. 2 ). The first step in this cascade is formed by the hybrid sensor kinase LuxQ, which contains both a N-terminal periplasmic membrane bound sensory domain and a C-terminal intracellular response regulator domain [ 21 ]. The signal is then transferred to the phosphorelay protein LuxU [ 22 ]. This phosphotransferase receives phosphorylation signals both from LuxQ and from LuxN, the parallel, homoserine lactone based quorum sensing circuit of Vibrio strains. It phosporylates the final response regulator, LuxO, which belongs to a large, highly conserved family of sigma54 dependent transcriptional regulators. LuxO has three conserved domains, e.g. the response regulator domain, the sigma54 activation domain, and a HTH (helix turn helix) motif for direct DNA binding [ 23 ]. At low cell density and in the absence of autoinducers, LuxQ autophosphorylates. The signal is transferred from its conserved aspartate residue to the histidine residue of LuxU, which phorphorylates the aspartate residue of the response regulator LuxO. In its phosphorylated (activated) form, and together with sigma54, LuxO activates the expression of small regulatory RNAs (sRNAs). The complexes of these sRNAs and the sRNA chaperone protein Hfq destabilize the mRNA of the quorum-sensing master regulator LuxR, resulting in the indirect repression of the lux operon transcription [ 24 ]. At high cell density, AI-2 present in the periplasmic space binds to the protein LuxP, which converts LuxQ from kinase to phosphatase. This reverses the flow of phosphate through the pathway, from LuxO to LuxU and then to LuxQ. In this case, sRNAs are not expressed. Without destabilization of the sRNA-Hfq complex, LuxR is translated and consequentially the transcription of the Lux operon is switched on. Figure 2 Genes involved in the signalling cascade for detection of AI-2 in Vibrio . Abbreviations: OM outer membrane; IM inner membrane; H histidine; D aspartate; HTH helix turn helix motif; sRNA small regulatory RNAs; Hfq chaperone protein. Numbers indicate analyzed organisms that have an orthologous gene for that protein using a reciprocal best hit search strategy, numbers in brackets indicate the number of analysed organisms having a similar gene based on standard blast search. The LuxS enzyme responsible for the last enzymatic step of AI-2 synthesis has at the same time an important function in the activated methyl cycle of the cell, since it is necessary for recycling of the toxic intermediate SAH [ 25 ]. Two pathways are known to be able to degrade and recycle SAH (Fig. 1 ). One pathway is composed of two successive enzymatic reactions catalysed by LuxS and Pfs. This pathway produces adenine, homocysteine and DPD which can be complexed with borate and converted to AI-2 by two non-enzymatic spontaneous reactions. The other pathway contains only one enzymatic step catalysed by SAH hydrolase (SahH) and produces adenosine and homocysteine. There is no AI-2 production through this pathway. Homocysteine can be further recycled to methionine by MetE or MetH and then activated to SAM by SAM synthetase. Despite intensive research on AI-2 in the last years the available data do in many cases not allow to clearly separate the metabolic function of the LuxS gene product from its possible signalling activity in interspecies communication. Winzer and his colleagues [ 25 , 26 ] analysed the available genomes (by April 2003) for the presence of LuxS and related genes and critically reviewed the available experimental data with respect to the potential signalling function of AI-2. The emphases of these studies were on the genes pfs, luxS and sahH. A large-scale and more detailed comparative genomic analysis of other genes involved in the AI-2 related metabolic and signal transduction pathways is missing. Therefore, we present here a comprehensive investigation of the phylogenetic distribution of all the genes involved in the synthesis of AI-2, the detoxification of SAH, as well as the signalling cascade necessary for the detection of AI-2 by analysing 138 completely sequenced genomes from the KEGG database [ 42 ] and the EMBL database [ 43 ]. While LuxS is the enzyme necessary for AI-2 production, it is not required for AI-2 signal transduction. Theoretically, an organism may not be able to produce AI-2 but have the ability to detect the presence of coexisting or competing bacterial species by sensing the environmental concentration of AI-2. This is the case in Pseudomonas aeruginosa [ 27 ]. Therefore, not only the LuxS-containing organisms but also all other sequenced genomes have been analysed for the existence of the AI-2 signal transduction cascade. Results Metabolic pathways involved in SAH degradation and recycling SahH and Pfs/LuxS are alternative pathways for recycling of SAH The distribution of the orthologs of the proteins involved in AI-2 production, SAH degradation and recycling, and AI-2 signalling is listed in Supplementary Table s1 and s2 [ additional file 1 and 2 ]. As shown in Supplementary Table s1 [ additional file 1 ], 80% of the 138 completely sequenced genomes have at least one pathway to degrade SAH. 51 organisms have only the two-step pathway using Pfs and LuxS, while 60 have only the one-step pathway using SahH (Fig. 1 ). The remaining one-fifth having neither pathway mainly belong to symbionts, intracellular parasites, Mollicutes or Chlamydiates. They probably rely on their host to recycle the toxic intermediate. With the exception of Bifidobacterium longum NCCC2705 and Escherichia blattae , no organism has both the sahH and luxS gene. Interestingly, each organism has only a single copy of the highly conserved luxS gene. These results are consistent with the studies of Winzer and his colleagues [ 26 ]. Phylogenetic distribution of SAH detoxification pathways The presence of either a one-step or a two-step detoxification pathway for SAH follows a phylogenetic pattern. Eukarya and Archaea use exclusively the one-step pathway to degrade SAH, while Bacteria use either the one-step or the two-step pathway depending on their phylogenetic position (Fig. 3 ). The two-step Pfs/LuxS pathway is consistenly present in all Firmicutes and absent in Actinobacteria (with the exception of Bifidobacterium longum ). Within the Proteobacteria, Alphaproteobacteria clearly use the one-step detoxification pathway, while there is a dividing line going across the Betaproteobacteria and the Gammaproteobacteria. For the Betaproteobacteria, the pathogen Neisseria meningitidis uses the two step pathway, while Ralstonia solanacearum and Nitrosomonas europaea use the one-step pathway. With the exception of the Xanthomonadales and Pseudomonadales, all Gammaproteobacteria presently sequenced use the Pfs/LuxS pathway. Figure 3 Distribution of one-step (red) and two-step (green) detoxification pathways of SAH within the three domains of life. All sequences shown in the tree could be found in the ARB database (ssujun02.arb), hence were already aligned, and for the desired presentation were all transferred to the rudimentary tree (tree_demo) by the ARB function "quick add by parsimony". After marking all species (sequences) of interest only these were kept by removing all unmarked species with an integrated ARB function. This leaves the original topology of the tree intact while eliminating all species and branches which are unnecessary for the demonstration of the phylogenetic distribution of sequences of prime interest. Phyla are numbered from 1 to 13. 1 Gammaproteobacteria; 2 Betaproteobacteria; 3 Alphaproteobacteria; 4 Epsilonproteobacteria; 5 Spirochaetes; 6 Chlorobia; 7 Bacteroidetes; 8 Cyanobacteria; 9 Actinobacteria; 10 Firmicutes; 11 Deinococcus-Thermus; 12 Thermotogae; 13 Aquificae. Numbers in brackets indicate sequenced genomes analysed. Shaded phyla are mixed, having organisms with the SahH pathway and organisms with the Pfs/LuxS pathway. Boxed strains have both LuxS and SahH. Only one or two representatives have been sequenced from other microbial phyla, so it is premature to generalize these findings. However, presently the Pfs/LuxS pathway has been found in Borrelia burgdorferi (Spirochaetes) and Deinococcus radiodurans R1 (Deinococcus-Thermus), while the organisms from other sequenced phyla (Chlorobia, Bacteroidetes, Aquificae, Thermotogae) use the SahH pathway. The second sequenced strain from the phylum Spirochates, Leptospira interrogans , uses the SahH pathway. These results are also consistent with the analysis of Winzer et al. [ 26 ]. Exploring the ERGO database [ 44 ] containing approximately 400 genomes, of which appr. 200 are microbial genomes, resulted in a consistent conclusion on the distribution pattern of SAH hydrolase or Pfs/LuxS degradation pathways (data not shown). Phylogeny of LuxS The sequences of LuxS orthologs were aligned and a phylogenetic tree was built from the alignment. There are clearly three bigger branches in the phylogenetic tree (Fig. 4 ). The first contains most Gram negatives, i.e. Gamma- and Betaproteobacteria. The second brach is comprised mainly of Lactobacillales, but contains some other groups as well. Interestingly, the LuxS ortholog from Bifidobacterium longum , which is the only species of Actinobacteria having LuxS, and which at the same time has the SahH pathway for recycling of SAH, is most closely related to that of the phylogenetically only distantly related Lactobacillus plantarum. Both bacteria share the same habitat, being commensals of the healthy human gut. There would have been ample opportunities for B. longum to acquire luxS by horizontal gene transfer from Lactobacillus . The Lactobacillus branch also contains a small subcluster with luxS from Borrelia burgdorferi , a Spirochete, which is most similar to luxS from Clostridium acetobutylicum . The third branch is dominated by Bacillales. Interestingly, it also includes two of three sequenced Epsilonproteobacteria, namely two strains of Helicobacter pylori . However, the closely related Campylobacter jejuni forms a separate, deeply branching lineage. These data confirm those of Lerat & Moran [ 28 ]. Small differences can be attributed to the treeing methods used, e.g. the position of Campylobacter jejuni luxS and the fact that γ-Proteobacterial LuxS genes were monophyletic in our analysis, but comprised two different branches in their tree. In addition, we included luxS sequences from Enterococcus faecalis and Deinococcus radiodurans . The robustness of the tree topology is caused by the high degree of conservation of luxS and strongly supports the resulting conclusions regarding gene transfer for some species. Figure 4 Phylogenetic tree of LuxS proteins from the completely sequenced genomes (138; July 2003) in the KEGG genome database. The tree was constructed using the neighbour-joining (NJ) method after alignment of orthologous genes by means of Vector NTI Advance (InforMax, United States). Transformation of homocysteine to methionine To complete the metabolic cycle of SAH, the common product homocysteine of the two degradative pathways is converted to methionine by a homocysteine methyltransferase (MetE, 5-methyltetrahydropteroyltriglutamate–homocysteine methyltransferase or MetH, 5-methyltetrahydrofolate–homocysteine methyltransferase), then to SAM by SAM synthetase (MetK). SAH is one of the products of SAM-dependent transmethylases. The distribution of MetE, MetH and MetK was analysed in a similar way to LuxS (Supplementary Table s1 [ additional file 1 ]). As a general rule, bacteria that have one of the two pathways to degrade SAH are also able to recycle homocysteine to methionine and to synthesize SAM. This conclusion again supports the importance of the degradation and recycling of SAH. As the only few exceptions, the strains Helicobacter pylori , Streptococcus pyogenes and Enterococcus faecalis lack MetE/MetH but have MetK. The overview of such enzymes in Eukarya and Archaea is more complicated probably because of their incomplete identification by searching homologues to proteins with known functions or because of the existence of other unknown pathways or enzymes in these organisms. Signal transduction pathway of AI-2 Reciprocal best hit strategy The reciprocal best-hit orthologs of the signal transduction cascade (LuxP, LuxQ, LuxU and LuxO) from Vibrio for the sequenced genomes are listed in Supplementary Table s1 and s2 [ additional file 1 and 2 ]. Unlike the broad distribution of LuxS, the orthologs of the AI-2 binding protein LuxP and the regulator LuxU were found exclusively in Vibrio strains. And only Vibrio strains have both the orthologs for the hybrid sensor kinase LuxQ and the two component response regulator LuxO. In addition, five other orthologs of LuxQ were found using the reciprocal best hit strategy, namely in Brucella melitensis , Brucella suis , Streptococcus agalactiae (two different strains), and in Methanosarcina mazei . For the two component response regulator LuxO, four additional orthologs were found in Bradyrhizobium japonicum , Listeria monocytogenes , Lactobacillus plantarum and Thermotoga maritima . Given the complexity of the proteins involved and the limited number of sequences presently available, it is not possible to draw consistent conclusions from this finding at this point. Unidirectional best hit search However, if not the reciprocal best-hit strategy but the uni-directional best-hit search was applied, 28 organisms were found to have homologues to LuxP. The LuxP homologues of 25 of these organisms were more similar to D-ribose binding proteins of Vibrio strains than to the AI-2 binding protein LuxP. We reconstructed the 3D models of LuxP proteins from different Vibrio strains by applying the method of SwissModel. All these LuxP proteins have similar 3D structures as expected from the high similarity of their sequences (data not shown). The three-dimensional structure of LuxP is very similar to that of D-ribose-binding proteins [ 6 ]. Because of the structural similarity between the ligands (AI-2 and D-ribose), and the structural similarity between the binding proteins, the question has to remain open whether the detected LuxP-homologues in the non- Vibrio organisms are actually functional AI-2 binding proteins. 104 organisms were found to have homologues both for the hybrid sensor kinase LuxQ, and the two-component response regulator LuxO. Most organisms had several homologous genes for these two-component systems, so that altogether 315 genes were found for LuxQ and 340 for LuxO. This was caused by the fact that several domains of the sensor and regulator components of signal transduction systems are highly conserved. We were not able to identify the unique binding domain from the sensor protein LuxQ specific for the detection of the AI-2 signal. Discussion Reciprocal or unidirectional best hit The reciprocal best hit strategy of sequence similarity comparisons was used here to distinguish orthologous from paralogous genes. Orthologs are genes in different species that evolved from a common ancestral gene by speciation. Paralogs are genes originating from duplication events within a genome. Orthologs tend to retain the same function in the course of evolution, whereas paralogs often evolve new functions [ 29 ]. The reciprocal best hit strategy is known to be a better method than the unidirectional best hit method to distinguish orthologs from paralogs [ 29 ]. Here, this was especially significant for the identification of components of the AI-2 signal transduction system. The number of identified orthologs for LuxO and LuxQ decreases from over 300 down to 7 and 8 by applying the reciprocal best hit strategy. The resulting orthologs for LuxO and LuxQ are close to the number of identified LuxP and LuxU, confirming that the reciprocal best hit strategy significantly filtered out the potential paralogs. However, it should be noted that in most cases the functions of these paralogs are not experimentally characterized and cannot be deduced by bioinformatics methods. Because of the high conservation among these regulatory components, it is hard to exclude the possibility that some of them may serve as alternative sensors for detecting the AI-2. On the other hand, the unidirectional best hits for the studied metabolic enzymes were basically the same as the reciprocal best hits, suggesting that the metabolic enzymes for the activated methyl cycle were seldom duplicated during the course of evolution. Distribution of LuxS The presence of either of two possible SAH degradation pathways in most living cells indicates their importance in the central cell metabolism. Both Archaea and Eukarya use exclusively a one-step detoxification pathway for SAH, indicating that this may be the ancient type of metabolism. The distribution of the two-step Pfs/LuxS pathway for detoxification of SAH within the domain Bacteria appears to be phylogenetically conserved. Since the currently sequenced genomes are biased towards pathogens, it remains to be seen if a similar phylogenetic pattern will also be found in non pathogenic bacteria from soils, sediments and marine environments. Previous studies [ 25 , 26 ] came to similar conclusions with respect to the presence of luxS. Our data strengthen the phylogenetic aspect of this distribution. Interestingly, a unique species from the genomes in the KEGG database, namely Bifidobacterium longum NCC2705, possesses both pathways. This species is a key commensal of the healthy human gastrointestinal tract and vagina. The double pathways may be helpful to recycle and use methionine more economically or to accomplish its dependence on H 2 S or methanethiol for methionine biosynthesis [ 30 ]. Another non-pathogenic species, Escherichia blattae , was also identified to have both pathways (Göttingen Genomics Laboratory, unpublished). However, their physiological roles in this species have still to be clarified. The phylogenetic tree of LuxS does not in all cases correspond to the 16S rRNA based microbial phylogeny. Thus, horizontal gene transfer might have resulted in the acquisition of LuxS genes e.g. in Bifidobacterium longum, Helicobacter pylori , Clostridium acetobutylicum and Borrelia burgdorferi , with the insect or mammalian gut serving as a melting pot of species. Production of AI-2 In most of the microbial genera other than Vibrio spp. having a LuxS enzyme the production of AI-2 has been demonstrated using the Vibrio harveyi reporter strain BB170 (reviewed by Winzer et al. [ 26 ]; otherwise citation is given); e.g. for Actinobacillus actinomycetemcomitans , Bacillus anthracis [ 31 ]; Borrelia burgdorferi , Campylobacter jejuni , Clostridium perfringens , Escherichia coli , Helicobacter pylori , Lactobacillus [ 32 ], Neisseria meningitides , Porphyromonas gingivalis , Proteus mirabilis , Salmonella typhimurium , Shigella flexneri , Staphylococcus [ 27 ], Streptococcus , Pasteurellaceae, periodontal pathogens [ 33 ] and rumen bacteria [ 34 ]. However, only in V. harveyi BB170 it was clearly shown that the active compound was a furanosyl-borate-diester. In E. coli serotype O157, Sperandino et al. [ 11 , 35 - 38 ], showed that the transcription of essential virulence factors coded on the LEE genomic island was triggered by an as yet unknown compound termed autoinducer-3 (AI-3) which depends on the presence of the luxS gene and did not elicit luminescence in V. harveyi BB170. Thus, purification of culture supernatants used for detecting AI-2 activity would be required to show that the active fraction is indeed a furanosyl-borate-diester. Conversely, the "real" universal signal might be a different, presently unknown compound produced by a different cyclization product of DPD or by an enzymatic step downstream of LuxS. Detection of AI-2 Our data clearly show that the signal transduction cascade for AI-2 is restricted to Vibrio species. This is consistent with the results of experimental studies published so far. No alternative signal transduction cascade for AI-2 has been experimentally identified in any of the strains studied, with the exception of Salmonella typhimurium [ 20 , 39 ]. Thus, there is no proof that these organisms actually respond to AI-2 in a quorum sensing related manner. However, the large diversity of two-component systems present in these organisms, for which in many cases the specific signals are not known, makes it quite possible that one of them might be devoted to the detection of AI-2 or another LuxS dependent compound. In Salmonella typhimurium , an ABC transporter with high homology to ribose transporters, which is however not homologous to LuxP, has been identified whose expression requires the presence of AI-2 and whose function is to transport it back into the cell [ 20 , 39 ]. No AI-2 induced genes other than this transporter have been found, indicating that in this organism AI-2 may not serve as a quorum sensing signal or the appropriate cultivation conditions for the expression of its activity were not met. Orthologs of the S. typhimurium lsr genes were found in most Enterobacteriales, as well as in Sinorhizobium meliloti and some Bacillus sp. (see Supplementary Table s3 [ additional file 3 ]). However, if AI-2 is indeed a universal signal molecule, it may be useful for bacteria to detect it even if they do not produce it themselves. This was shown to be actually the case in Pseudomonas aeruginosa , which does not contain the luxS gene [ 27 ]. Here, the promoters of 21 well characterized virulence associated genes were cloned into promoterless luxCDABE reporter plasmids and light induction was tested in the presence of AI-2 synthesized enzymatically from SAH or by co-culture with a luxS containing clinical isolate of Streptotoccus sp. (strain CF004). The fact that six of these virulence gene promoters were upregulated both by AI-2 and coculture with CF004 suggests a specific effect of AI-2 on the transcription of virulence associated genes in Pseudomonas aeruginosa , although the signalling cascade within the cell is presently unknown. Conclusions The presence of luxS in many phylogenetic groups within the domain Bacteria indicates that these bacteria, while recycling SAH in a two-step enzymatic process, at the same time produce a compound able to stimulate luminescence in a V. harveyi reporter strain which is most probably a furanosyl-borate-diester. The detection cascade, if any, for this compound in the producing organisms must be different from that in Vibrio strains and is presently not known. The diversity of physiological effects observed in luxS - mutants can either be interpreted as the result of a defect in a global quorum sensing regulatory mechanism, which may also be caused by a LuxS dependend compound other than AI-2, or as the result of a defect in the central methyl cycle of the cell. Thus, although there are intriguing indications for a LuxS dependent universal signal molecule in Bacteria, direct proof regarding the chemical nature of the compound and its signalling mechanism in non Vibrio organisms is presently missing. Methods Databases The protein sequences of 138 sequenced genomes were downloaded from KEGG (Status June 2003) and reformatted as local blast databases. The non-redundant protein database of NCBI (nr) [ 45 ] and the KEGG Sequence Similarity Database (SSDB) [ 46 ] were explored through their online services. Preparation of the queries To achieve a more complete finding of the proteins functionally similar to the proteins related to either the metabolic pathway (LuxS, Pfs, SahH) or the signal transduction pathway (LuxP, LuxO, LuxQ, LuxU) of autoinducter-2, the NCBI protein database was at first searched with the relevant functional terms such as "AI-2 production" or "LuxS". A phylogenic tree was constructed based on the alignment of the relevant matches by using the component AlignX of the bioinformatic software suite "Vector NTI Advance" (InforMax, United States). From each branch of the tree, one protein (mainly the protein of which the function was manually curated, for example, by SWISSPROT) was selected. All of them were put together into a file as a blast query to represent a function. It is not necessary for the members of this function to be similar to each other in sequence level. This facilitates the finding of evolutionarily far-related proteins by blast search. The protein sequences of MetK, MetE and MetH from E. coli K12 and LsrR, -A, -B, -C, -D, -E, -F and G from Salmonella typhimurium [ 39 ] were used alone as query. Blast search The queries were used to search for their respective orthologs from the local KEGG genome databases by applying the reciprocal best hit strategy [ 40 ] with a blastp cutoff E-value 1E-4. In the reciprocal best hit strategy, protein i from genome A is orthologous to protein j from genome B only under the conditions that j is the best hit when i is queried in database B and reciprocally i is also the best hit when j is queried in database A [ 40 ]. A Visual Basic script was programmed to realize this strategy automatically. All identified orthologs were submitted for further analysis. The NCBI non-redundant protein database nr was searched using the normal one-direction blastp with a cutoff E-value 1E-4. The hits were manually checked to confirm that they had the same functional annotation as the query. Phylogenetic tree construction The phylogenetic tree for the orthologs was built with the neighbour-joining (NJ) method [ 41 ] using Vector NTI Advance (InforMax, United States) after the sequences had been aligned. Authors' contributions JS conducted the data mining work and contributed to writing the manuscript. RD cooperated with the access to the ERGO database. IWD initiated the study, contributed to the concept and drafted the manuscript. APZ supervised the study and contributed to writing the manuscript. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Supplementary Table s1. Presence of AI-2 synthesis and detection genes in 138 completed genomes of the KEGG database (July 2003). Abbreviations: luxS, AI-2 synthetase/SRH cleavage enzyme; pfs, SAH-nucleosidase enzyme; sahH, SAH hydrolase; metH and metH, methionine synthetase; metK SAM synthetase; luxP, AI-2 binding protein; luxQ, membrane bound hybrid sensor kinase; luxU, histidine phosphorelay protein ; luxO response regulator. See Fig. 1 for further information on the synthesis pathway and Fig. 2 for the phosporelay detection cascade. Organisms shaded violet contain neither luxS nor sahH. Click here for file Additional File 2 Supplementary Table s2. The accession numbers of the genes in Supplementary Table s1. Abbreviations as in Supplementary Table s1. Click here for file Additional File 3 Supplementary Table s3. Presence of the Salmonella lsr gene homologs in the completed genomes of the KEGG database (July 2003). Abbreviations as in Supplementary Table s1. X denotes bi-directional hits, ? denotes unidirectional hits. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524169.xml

555960

Human macrophages limit oxidation products in low density lipoprotein

This study tested the hypothesis that human macrophages have the ability to modify oxidation products in LDL and oxidized LDL (oxLDL) via a cellular antioxidant defence system. While many studies have focused on macrophage LDL oxidation in atherosclerosis development, less attention has been given to the cellular antioxidant capacity of these cells. Compared to cell-free controls (6.2 ± 0.7 nmol/mg LDL), macrophages reduced TBARS to 4.42 ± 0.4 nmol/mg LDL after 24 h incubation with LDL (P = 0.022). After 2 h incubation with oxLDL, TBARS were 3.69 ± 0.5 nmol/mg LDL in cell-free media, and 2.48 ± 0.9 nmol/mg LDL in the presence of macrophages (P = 0.034). A reduction of lipid peroxides in LDL (33.7 ± 6.6 nmol/mg LDL) was found in the presence of cells after 24 h compared to cell-free incubation (105.0 ± 14.1 nmol/mg LDL) (P = 0.005). The levels of lipid peroxides in oxLDL were 137.9 ± 59.9 nmol/mg LDL and in cell-free media 242 ± 60.0 nmol/mg LDL (P = 0.012). Similar results were obtained for hydrogen peroxide. Reactive oxygen species were detected in LDL, acetylated LDL, and oxLDL by isoluminol-enhanced chemiluminescence (CL). Interestingly, oxLDL alone gives a high CL signal. Macrophages reduced the CL response in oxLDL by 45% (P = 0.0016). The increased levels of glutathione in oxLDL-treated macrophages were accompanied by enhanced catalase and glutathione peroxidase activities. Our results suggest that macrophages respond to oxidative stress by endogenous antioxidant activity, which is sufficient to decrease reactive oxygen species both in LDL and oxLDL. This may suggest that the antioxidant activity is insufficient during atherosclerosis development. Thus, macrophages may play a dual role in atherogenesis, i.e. both by promoting and limiting LDL-oxidation.

Introduction Oxidative modification of low density lipoprotein (LDL) plays a major role in the pathogenesis of atherosclerosis. The first stage of atherogenesis is characterized by an influx and accumulation of LDL in the intima, followed by recruitment of blood-derived monocytes and lymphocytes to the developing lesion [ 1 ]. Subsequently, LDL is oxidatively modified by free radicals that are either secreted from cells within lesions or generated extracellular in the arterial wall [ 2 ]. Oxidatively modified LDL (oxLDL) induces a multitude of cellular responses which lead to vascular dysfunction [ 3 ]. Much attention has sofar been devoted to the mechanisms by which cells oxidize LDL, since interventions targeting these mechanisms could prevent or retard the disease process. However, cells may also provide a protective effect by reducing oxidation products present in LDL and oxLDL. Murine macrophages effectively block LDL oxidation by mechanisms which include metal ion sequestration [ 4 ]. Recent studies show that macrophages decrease cholesteryl ester hydroperoxide levels in LDL, an antioxidant action that is proportional to cell number [ 5 , 6 ]. In addition, endothelial cells prevent accumulation of lipid hydroperoxides in LDL [ 7 ]. Human hepatic cells show a protective role by selective uptake and detoxification of cholesterol ester hydroperoxides present in high density lipoprotein [ 8 ]. Enzymes associated with antioxidant defense, such as manganese superoxide dismutases, catalase, and glutathione peroxidases are induced by oxidants in vitro [ 9 - 11 ]. Four selenium-dependent glutathione peroxidases (GPx) have been identified sofar: cytosolic GPx (cGPx), gastrointestinal GPx (GI-GPx), plasma GPx (pGPx), and phospholipid hydroperoxide GPx (PHGPx) [ 12 , 13 ]. The PHGPx reduces hydroperoxides present in complex lipids such as phospholipids and cholesteryl esters [ 14 , 15 ]. Interestingly, increased glutathione levels are present in macrophages derived from the human monocytic cell line THP-1, as well as in mouse peritoneal macrophages after incubation with oxLDL [ 16 , 17 ]. An increased activity of both glutathione peroxidase and superoxide dismutase occurs in the arterial wall of cholesterol-fed rabbits [ 18 ]. Furthermore, lipid-laden macrophages within atherosclerotic vessels express an extracellular form of superoxide dismutase (EC-SOD) [ 19 ]. This study tested the hypothesis that human macrophages have the ability to modify oxidation products in LDL and oxLDL. We also analyzed the activity of cellular antioxidant defenses such as catalase, glutathione peroxidase, and superoxide dismutase in these cells. We used early macrophages as cell culture model to mimic newly recruited macrophages into the intima. Materials and methods Cell isolation Mononuclear cells were isolated by the Ficoll-Hypaque procedure (Pharmacia, Uppsala, Sweden) [ 20 ] from buffy coats obtained from the blood of healthy donors from the Blood Bank at Sahlgrenska University Hospital, Göteborg. Monocytes in RPMI 1640 medium (Life Technologies, Paisley, Scotland), supplemented with non-essential amino acids, 2 mM sodium pyruvate, 100 U/mL penicillin, and 100 μg/mL streptomycin were seeded in 6 well plates at 4 × 10 6 cells per well. Non-adherent cells were removed after 1 h. RPMI 1640 containing 100 μg/mL LDL or oxLDL was incubated at 37°C in 5% CO 2 in the presence or absence of macrophages. By definition, monocytes are denoted macrophages when they are attached, thus the cells used in this study are considered early human monocyte-derived macrophages (HMDM). For chemiluminescence experiments, monocytes were allowed to adhere to cell culture flasks for 1 h. Adhered macrophages were then detached by incubation with PBS containing 5 mM EDTA and 2% fetal calf serum for 20 minutes at +4°C [ 21 ]. Cells were collected, washed, and resuspended to a density of 5 × 10 6 cells/mL in Krebs-Ringer Bicarbonate buffer supplemented with glucose (KRG) (Sigma, St. Louis, Missouri). To obtain non-viable macrophages, cells were stored at +4°C for 16 h. Trypan blue exclusion test confirmed that 100% of the cells were non-viable. Lipoproteins Fresh human EDTA-plasma was obtained from healthy male donors after overnight fasting. LDL (density 1.019–1.063 g/L) was isolated by sequential ultracentrifugation [ 22 ]. Before oxidation, native LDL was desalted on a PD-10 column equilibrated with PBS containing 100 μg/mL penicillin and 100 μg/ml streptomycin (PEST) using PBS-PEST as elution buffer. The LDL was oxidized at 37°C for 2–24 hours by 12.5 μmol CuSO 4 /mg LDL. Oxidation was terminated through the addition of 0.5 mmol/L EDTA. The oxLDL was purified on a PD-10 column with PBS as elution buffer and sterilized by filtration through a 0.22 μm filter. Native LDL was acetylated as described [ 23 ]. Oxidation of LDL was determined as the relative electrophoretic mobility (REM), i.e. the ratio between the distance oxLDL and native LDL migrate on a 0.5% agarose gel. The LDL in this study was oxidized for 2 h and had a REM ranging from 1.06 to 1.32 and TBARS values between 3 and 8 nmol MDA/mg LDL protein. Lipoprotein concentrations were determined with the BioRad protein assay using γ-globulin as standard. Chemiluminescence The chemiluminescence (CL) assay was performed at 37°C and the CL detected for at least 100 minutes with a luminescence counter (Bio Orbit Luminometer 1251, Turku, Finland) [ 21 ]. The CL response was detected in a total volume of 1.0 mL, containing 10 μg isoluminol (Sigma), 4 U horseradish peroxidase (Roche AB, Stockholm, Sweden), and 100 μg of either LDL, oxLDL, or acLDL in KRG in the presence or absence of 5 × 10 5 cells. As a specific inhibitor of hydrogen peroxide, 30 μg catalase (Roche AB) was added in some experiments to oxLDL without cells. Measurement of oxidation products Thiobarbituric acid-reactive substances (TBARS) were determined by the method of Yagi [ 24 ]. Fluorescence was measured at 553 nm with 515 nm excitation. Lipid peroxides (LPO) were determined by the Lipohydrox assay from Wak-Chemie Medical (Bad-Soden, Germany). Lipid peroxides are reduced to hydroxyl derivatives in the presence of hemoglobin, and the chromogen 10-N-methylcarbamyl-3, 7-dimethylamino-10H-phentiazine is oxidatively cleaved to form methylene blue. Lipid peroxides are quantitated by colorimetric measuring of the methylene blue at 675 nm. Levels of hydrogen peroxide equivalents (H 2 O 2eq ) were analyzed in the LDL-containing media incubated with or without macrophages. The assay is based on the oxidation of ferrous ions to ferric ions by hydrogen peroxide at acidic pH (OXIS International Inc., Portland, Oregon). The ferric ion binds to the indicator dye xylenol-orange to form a stable complex which is measured at 560 nm. Apolipoprotein B (Apo B) concentration was determined in cell culture media by immunoprecipitation enhanced by polyethylene glycol at 340 nm (Thermo Clinical Labsystems, Espoo, Finland). ApoB analyses were performed on a Konelab 20 autoanalyser (Thermo Clinical Labsystems). Analysis of antioxidant properties of macrophages The intracellular levels of glutathione and the activity of GPx were measured in crude extracts from macrophages incubated with either LDL or oxLDL. The cells were washed twice with ice-cold PBS and harvested in 0.5 mL lysis buffer. For glutathione peroxidase measurements, this buffer contained 50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 1 mM dithiothreitol. For glutathione measurements, the cells were lysed in 0.5 mL ice-cold 5% metaphosphoric acid. The lysate was spun down at 3000 × g, and the supernatants stored at -80°C. The presence of glutathione peroxidase was determined with a colorimetric assay (Bioxytech GPX-340) and glutathione (GSH) was measured with Bioxytech GSH-400, both from OXIS International Inc. The assay GPX-340 measures the functional activity of the GPx. In functional terms, all four types of GPx (cGPx, GI-GPx, pGPx, and PHGPx) appear similar in catalytic activity [ 12 , 25 ]. ROOH + 2GSH ---- GPx ----> ROH + GSSG(oxidized glutathione)+ H2O GSSG + NADPH + + H + --------> 2GSH + NADP + Catalase activity in macrophages was quantified by the method of Aebi [ 26 ]. Decomposition of H 2 O 2 was measured in cell lysates at 240 nm. One unit of catalase activity was defined as the rate constant for the reaction using purified catalase (Roche AB) as standard. In the cell lysates, the CuZn-superoxide dismutase (SOD) and Mn-SOD enzymatic activities were measured with a direct spectrophotometric method [ 27 ]. Extracellular SOD protein was determined by ELISA essentially as previously described [ 28 ]. Data analyses Data were expressed as means ± standard error. Statistical analysis was performed using Student's paired t-test or ANOVA. P values < 0.05 were considered statistically significant. Results Macrophages diminish oxidation products in LDL and oxLDL To study the effect of macrophages on LDL oxidation, we measured TBARS, LPO, and H 2 O 2 in LDL-containing culture media after 2 h and 24 h incubation with or without macrophages. Compared to cell-free controls (6.2 ± 0.7 nmol/mg LDL), there was a significant reduction of TBARS by macrophages to 4.42 ± 0.4 nmol/mg LDL after 24 h incubation with LDL (P = 0.022) (Fig. 1A ). After 2 h incubation with oxLDL, TBARS was 3.69 ± 0.5 nmol/mg LDL in cell free media, and 2.48 ± 0.9 nmol/mg LDL in the presence of macrophages (P = 0.034). Although a time-dependent increase of TBARS in the presence of cells is seen, this change was not statistically significant. Lipid peroxide levels are unaffected in LDL and oxLDL incubated in cell-free wells during 2 h to 24 h (Fig. 1B ). In the presence of cells, a reduction of lipid peroxides in LDL (33.7 ± 6.6 nmol/mg LDL) was found after 24 h compared to cell-free incubation (105.0 ± 14.1 nmol/mg LDL) (P = 0.005). In oxLDL, the cell-mediated loss of lipid peroxides was significant compared to cell-free media after 24 h. The levels of lipid peroxides in oxLDL were 137.9 ± 59.9 nmol/mg LDL and in cell-free media 242 ± 60.0 nmol/mg LDL (P = 0.012). In LDL-containing media incubated in cell-free wells, the levels of H 2 O 2eq increase with time from 116 ± 31 nmol/mg LDL after 2 h, to 270 ± 46 nmol/mg LDL after 24 h (P = 0.009) (Fig. 1C ). Levels of H 2 O 2eq increase from 303 ± 14 nmol/mg LDL to 357 ± 24 nmol/mg LDL in oxLDL-containing media (P = 0.011), suggesting both LDL and oxLDL are oxidized during culture conditions. H 2 O 2eq is significantly decreased in both LDL (P = 0.038) and oxLDL (P = 0.035) after incubation with macrophages. In macrophage-treated LDL, the H 2 O 2eq content is decreased to 14.8 ± 2.9 nmol/mg LDL after 24 h, which is similar to levels found in non-treated LDL (0 h). Taken together, these results suggest a cell-mediated loss of oxidation products in both LDL and oxLDL in the presence of macrophages. Figure 1 Effect of macrophages on oxidation products in LDL and oxLDL . RPMI 1640 containing 100 μg/mL of LDL or oxLDL (oxidized for 2 h) was incubated in cell culture wells at 37°C with and without macrophages (HMDM) for 2 h and 24 h. Values are expressed as nmol/mg LDL, TBARS (n = 5) (A), lipid peroxides (n = 6) (B), H 2 O 2 (n = 6), (C). Values of Apo B are expressed as % change compared to cell free control incubations (n = 4) (D). Apo B was analysed to study if the cell mediated decrease in oxidation products was due to a decrease of LDL or oxLDL in the cell culture media. In the presence of macrophages, the content of apo B in native LDL is reduced by 7 % after 2 h (P = 0.013) and 11 % after 24 h (P = 0.025) (Fig. 1D ). In the oxLDL medium, no significant change of apo B in the medium is found after 2 h, and the content of apo B in oxLDL decreases 18% after 24 h incubation with macrophages compared to cell-free control (P = 0.006). This result suggests that the decrease in oxidation products can not be fully explained by an increased uptake of LDL or oxLDL by macrophages. To further test the capacity of macrophages to decrease oxidation products in LDL, we used isoluminol-enhanced chemiluminescence to detect reactive oxygen species in LDL. OxLDL, acLDL, or LDL was incubated with macrophages (5 × 10 5 cells) or without cells at 37°C. OxLDL alone had a CL response of approximately 400 mV after 2 h (Fig. 2 ). AcLDL shows a low but increasing CL up to 30 mV after 2 h as does native LDL, but the response is lower than that of oxLDL. In repeated experiments, incubation of macrophages with oxLDL leads to a significant reduction in the maximum peak value to 260 ± 45 mV compared to 462 ± 127 mV for oxLDL alone (P < 0.01) (n = 6). Adding catalase to oxLDL leads to a decrease of the CL signal by 37%, suggesting peroxides in the oxLDL. Figure 2 Production of reactive oxygen species from LDL, oxLDL and macrophages as measured by isoluminol-enhanced chemiluminescence . The incubation mixture of 1.0 mL KRG contained 5 × 10 5 cells (HMDM), 10 μg isoluminol, 4 U horseradish peroxidase, and 100 μg of LDL, oxLDL (oxidized for 2 h), or acLDL. The chemiluminescence was measured every 2 minutes for a total of 100 minutes at 37°C (n = 5). LDL oxidized by Cu 2+ for 2, 8, or 20 h, shows similar maximum peak values, however a lag time of 10 min is observed with the shorter oxidation times. Incubation of oxLDL with macrophages leads to a reduction in the CL signal (Fig. 3 ). Different dilutions of oxLDL lead to a dose dependent increase in CL (Fig. 4 ). The addition of macrophages results in a 45% lower CL maximal peak value at the different concentrations of oxLDL (P = 0.0016). Figure 3 Production of reactive oxygen species from oxLDL and macrophages as measured by isoluminol-enhanced chemiluminescence . The incubation mixture of 1.0 mL KRG, containing 10 μg isoluminol, 4 U horseradish peroxidase, and 100 μg of oxLDL (oxidized for either 2 h, 8 h or 20 h), was used alone or in combination with 5 × 10 5 macrophages (HMDM) (n = 3). Figure 4 Dose-response effect of oxLDL on production of reactive oxygen species as measured by isoluminol-enhanced chemiluminescence . The incubation mixture of 1.0 ml KRG contained 10 μg isoluminol, and different concentrations of oxLDL oxidized for 2 h, alone and in combination with HMDM (5 × 10 5 cells). Data are shown as means of maximal peak values in mV ± SE for triplicate determinations within a single experiment and are representative of two independent experiments. Results of oxLDL in cell-free incubations versus oxLDL in the presence of macrophages were analyzed by ANOVA. To exclude the possibility that quenching contributes to the macrophage effect, oxLDL was incubated with non-viable macrophages. No reduction in CL signal is seen (Fig. 5 ), which suggests that no quenching occurred and that viable cells are necessary for antioxidative activity. Figure 5 Production of reactive oxygen species from oxLDL and non-viable macrophages as measured by isoluminol-enhanced chemiluminescence . The incubation mixture of 1.0 mL KRG, containing 10 μg isoluminol, 4 U horseradish peroxidase, and 100 μg of oxLDL (oxidized for 2 h), was used alone or in combination with 5 × 10 5 non-viable macrophages. Data are shown from a typical experiment with 2 different cell donors. LDL affects cellular antioxidant defences in macrophages Since the levels of peroxides are elevated in oxLDL compared to LDL, and the cellular defences against peroxides excess are catalase and GPx, we investigated the cellular activity of these enzymes in macrophages. Macrophages incubated with oxLDL for 2 h have increased intracellular activity of catalase (P = 0.006) and GPx (P = 0.0002) (Figs 6A and 6B ). In contrast, no significant increase of the intracellular activity of these enzymes occurs in macrophages incubated with LDL. In addition, the expression of glutathione, which is a cofactor for the GPx enzymes when H 2 O 2 is detoxified, is enhanced in macrophages treated with oxLDL for 2 h (P = 0.0048) (Fig. 6C ). Figure 6 Effect of LDL and oxLDL on the intracellular antioxidant defenses in macrophages . The intracellular activity of catalase (A), glutathione peroxidase (B), and the levels of glutathione (C) were measured in crude extracts from macrophages (n = 6) incubated with LDL or oxLDL (oxidized for 2 h). Control cells were incubated in the absence of LDL. Results were analyzed by ANOVA. Neither LDL nor oxLDL significantly affect the CuZn-SOD or Mn-SOD activity in macrophages (Table I ). However, the Mn-SOD activity is enhanced after 24 h compared to 2 h incubations. Although an increase in secreted EC-SOD is seen at 24 h, this change was not statistically significant. Neither LDL nor oxLDL affect this secretion. These observations suggest that macrophages respond to oxLDL by increasing their enzyme activity of catalase and glutathione peroxidase. Table 1 Effect of LDL and oxLDL on superoxide dismutase isoenzymes in macrophages. CuZnSOD U/mg cell protein MnSOD U/mg cell protein EC-SOD secreted ng/mg cell protein Control 2 h 81.1 ± 7.5 8.1 ± 0.7 0.61 ± 0.28 LDL 2 h 70.9 ± 8.1 8.2 ± 0.7 0.59 ± 0.23 oxLDL 2 h 71.8 ± 9.0 7.4 ± 0.4 0.45 ± 0.15 Control 24 h 79.3 ± 12.1 18.6 ± 1.8 (P = 0.0016) 0.86 ± 0.35 LDL 24 h 80.8 ± 9.9 15.4 ± 0.7 (P = 0.0004) 0.85 ± 0.31 oxLDL 24 h 91.2 ± 20.6 17.1 ± 1.9 (P = 0.0025) 0.80 ± 0.30 Values are means ± SE (n = 4). CuZn-SOD and Mn-SOD enzymatic activities were measured in cell lysates, and EC-SOD protein was determined in culture media. P values indicate the comparison between the effect of LDL/oxLDL on superoxide dismutase at 2 h and 24 h. Discussion Oxidation of LDL is a crucial event in the pathophysiology of atherosclerosis. Reactive oxygen species such as H 2 O 2 participate in the oxidation of LDL [ 29 ]. Although the mechanisms are not fully understood, aortic cells such as endothelial cells, smooth muscle cells, and macrophages have the capacity to oxidize LDL in vitro . We have recently shown that hypoxia enhances both macrophage-mediated LDL oxidation and the expression of the putative LDL-oxidizing enzyme 15-lipoxygenase-2 [ 30 ]. While many studies have focused on cellular oxidation of LDL, less attention has been given to the cellular antioxidant capacity of macrophages. This study shows that macrophages play an important role in limiting lipid oxidation products that accumulate in LDL and oxLDL. Since early macrophages are used in this study, this may resemble newly recruited macrophages entering into the arterial intima. Regarding oxLDL, macrophages decrease TBARS levels by about 30%, LPO decreases 43%, and the H 2 O 2eq content decreases 64% in cell culture media after 24 h compared to cell-free controls. These results are in agreement with an earlier study where human macrophages reduce the content of cholesteryl ester hydroperoxides in LDL by 43% [ 5 ]. When the apo B content was analyzed in the culture media, we found that the levels of apo B in oxLDL were 18% reduced after 24 h incubation with macrophages compared to cell free control. For LDL the corresponding figure was 11%, which implies that the decrease in oxidation products cannot be entirely explained by a higher uptake of LDL or oxLDL by macrophages. Our results suggest that macrophages respond to oxidative stress by an endogenous antioxidative activity, which is sufficient to decrease reactive oxygen species; i.e. TBARS, LPO and H 2 O 2 both in LDL and oxLDL. Our data also suggest that oxidation products accumulate in LDL and in oxLDL during regular cell culture conditions in the absence of cells. This has to be taken into consideration when cell culture data using LDL are interpreted. The chemiluminescence technique is generally used to detect cellular ROS production. The amplifying molecule isoluminol reacts with ROS to produce an excited state intermediate that emits light upon relaxation to the ground state [ 31 ]. Our data show that oxLDL contains reactive oxygen metabolites that have the capacity to induce CL. Lipid hydroperoxides, the major oxidizing species in oxLDL, are likely to cause the CL. In the presence of transition metal ions, they generate Fenton-type oxidants, which may induce the chemiluminescence. The CL response is reduced when LDL is oxidized longer than 24 h. This agrees with previous data concerning the kinetics of LDL oxidation where lipid hydroperoxides are formed during the propagation phase and are decreased during the decomposition phase [ 32 ]. In the presence of macrophages, the CL response in oxLDL is reduced by 45%, which suggests that these cells exhibit antioxidant activity. Our results further suggest that cellular viability is necessary for antioxidant activity. There are a number of cellular defences against oxidative stress, such as superoxide dismutase, catalase, and glutathione-related enzymes. In this study we sought to define the antioxidative activity of macrophages. All glutathione peroxidases reduce H 2 O 2 or soluble alkyl peroxides, by coupling its reduction of H 2 O with oxidation of glutathione [ 11 ]. We found increased glutathione peroxidase activity that coincided with enhanced glutathione levels in oxLDL-treated macrophages. Only PHGPx reduces hydroperoxy groups of lipids together with those of phospholipids and cholesteryl esters when present in lipoproteins [ 15 ]. Interestingly, overexpression of PHGPx inhibits H 2 O 2 -induced oxidation and activation of NFκB in transfected rabbit smooth muscle cells [ 33 ]. Our results may have in vivo relevance, since an increased activity of glutathione peroxidase is found in the artery wall of cholesterol-fed rabbits [ 18 ]. Catalase is a representative antioxidant enzyme and previous studies show that oxidants such as H 2 O 2 and lipid peroxides induce catalase gene expression in cultured rabbit endothelial cells, rabbit macrophages, and human smooth muscle cells [ 10 ]. This study provides further evidence that lipid hydroperoxides in oxLDL induce antioxidant defences in macrophages, since human macrophages also upregulated their catalase activity. Human macrophages induce catalase activity in response to oxidative stress [ 34 ]. Addition of catalase to oxLDL alone reduced the CL response by 37%, suggesting that peroxides are detected by the CL technique. EC-SOD occurs in high concentration in both non-diseased [ 35 ] and atherosclerotic arterial walls [ 19 ]. In non-diseased arteries, the enzyme is primarily secreted by smooth muscle cells, whereas in atherosclerotic lesions it is also expressed by macrophages [ 19 ]. Higher levels of EC-SOD are present in cell culture media after 24 h than after 2 h, but the presence of LDL or oxLDL does not effect EC-SOD expression. Similar results of EC-SOD expression have been described in human fibroblasts [ 36 ]. The expression of CuZn-SOD is neither affected by LDL, oxLDL nor time. This was not unexpected since CuZn-SOD is generally regarded as a constitutively expressed enzyme. Mn-SOD activity increases with time, but there is no additive effect of LDL or oxLDL on its activity. Cells can tolerate mild oxidative stress, which triggers the antioxidant defence system in an attempt to restore the oxidant- antioxidant balance [ 37 ]. We did not find an increase in antioxidant enzyme activity in LDL-treated macrophages, which suggests that the basal levels of catalase and GPx activity are sufficient to remove the oxidation products in LDL. In contrast, the augmented catalase and GPx activity suggests that oxLDL induces cellular adaptation when macrophages are exposed to increased oxidative stress. This study suggests that oxidative stress induced by oxLDL could be balanced by a cellular antioxidant defence by newly recruited macrophages to sites of LDL oxidation. As oxidation of LDL is implicated in the development of atherosclerosis, and oxidation products of LDL are present in advanced atheromatous lesions, this may suggest that the antioxidant activity is insufficient in vivo . Thus, it is evident that macrophages play a dual role in atherogenesis, i.e. both by promoting and limiting LDL-oxidation. It remains to be determined during which stage of lesion development these individual characteristics pertain. A strategy to intervene with the development of atherosclerosis would be to increase the endogenous intracellular antioxidant capacity of cells, which would remove and detoxify oxidized LDL. Interestingly, recent results show that Lovastatin increases hepatic catalase activity in cholesterol-fed rabbits [ 38 ]. In light of the response-to-retention hypothesis of atherosclerosis, subendothelial retention of atherogenic LDL is the initiating event of the disease [ 39 , 40 ]. It is tempting to speculate that macrophages are present in the atherosclerotic intima because of their antioxidant activity, which detoxifies and removes cytotoxic products in the retained LDL. Authors' contributions All authors have contributed to the design of the study, the data analysis, and the writing of the manuscript. The final version of the manuscript has been read and approved by all authors prior to submission. Each author's specific contribution was as follows; LMH.: study design, macrophage cell culture, CL analysis, LDL treatment. C.U.: macrophage cell culture, TBARS, LPO, GpX, and catalase analyses. A.K.: data interpretation, writing, and editing. D.v.R: CL analysis. S.L.M: SOD analyses. C.D.: setting up the CL method. O.W.: study coordination and data interpretation.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555960.xml

544845

Nitrosative stress induces DNA strand breaks but not caspase mediated apoptosis in a lung cancer cell line

Background Key steps crucial to the process of tumor progression are genomic instability and escape from apoptosis. Nitric oxide and its interrelated reactive intermediates (collectively denoted as NO X ) have been implicated in DNA damage and mutational events leading to cancer development, while also being implicated in the inhibition of apoptosis through S-nitrosation of key apoptotic enzymes. The purpose of this study was to explore the interrelationship between NO X -mediated DNA strand breaks (DSBs) and apoptosis in cultured tumor cell lines. Methods Two well-characterized cell lines were exposed to increasing concentrations of exogenous NO X via donor compounds. Production of NO X was quantified by the Greiss reaction and spectrophotometery, and confirmed by nitrotyrosine immunostaining. DSBs were measured by the alkaline single-cell gel electrophoresis assay (the COMET assay), and correlated with cell viability by the MTT assay. Apoptosis was analyzed both by TUNEL staining and Annexin V/propidium iodine FACS. Finally, caspase enzymatic activity was measured using an in-vitro fluorogenic caspase assay. Results Increases in DNA strand breaks in our tumor cells, but not in control fibroblasts, correlated with the concentration as well as rate of release of exogenously administered NO X . This increase in DSBs did not correlate with an increase in cell death or apoptosis in our tumor cell line. Finally, this lack of apoptosis was found to correlate with inhibition of caspase activity upon exposure to thiol- but not NONOate-based NO X donor compounds. Conclusions Genotoxicity appears to be highly interrelated with both the concentration and kinetic delivery of NO X . Moreover, alterations in cell apoptosis can be seen as a consequence of the explicit mechanisms of NO X delivery. These findings lend credence to the hypothesis that NO X may play an important role in tumor progression, and underscores potential pitfalls which should be considered when developing NO X -based chemotherapeutic agents.

Background Nitric oxide (NO • ) is a ubiquitous nitrogen radical species that has been found to exert protean influences on physiologic and pathophysiologic processes in a wide variety of organ systems. Vastly increasing the functional consequences of NO • production is its interrelationship between the nitroxyl anion (NO - ) and the nitrosonium cation (NO + ) depending upon the redox environment in which NO • is being produced. Each of these interrelated redox species demonstrates its own biological consequences. Collectively, nitric oxide biology attributable to the integrated actions of these three species (referred to as NO X ) demonstrates broad reaching consequences. It is generally thought that well regulated levels of NO X production is important to numerous physiologic processes, while NO X overproduction is increasingly being implicated in pathophysiologic processes via nitrosative stress. One molecular mechanism underlying these pathophysiologic processes is NO X -mediated genomic damage inducing apoptosis in susceptible cells [ 1 ]. This induction of apoptosis is thought to be dependent upon an intact p53-pathway in response to genotoxicity [ 2 ]. While the field of NO X in cancer biology is a new and developing area of research, the intricacies of NO X influence are beginning to be elucidated. We have previously demonstrated widespread increases in expression of nitric oxide synthase isoforms within human primary tumors when compared to surrounding normal tissues [ 3 - 6 ]. Both cancer promotive and cancer protective roles have been ascribed to NO X . Cancer promoting effects include: 1) the capacity to trigger mutagenesis, 2) enhance growth, invasion, angiogenesis and metastasis of tumors, 3) select for increasingly virulent tumor cell clones, and 4) suppress the host anti-tumor immune response [ 7 , 8 ]. Cancer protective actions are manifest in the ability of immune-mediated NO X production to attenuate tumor cell respiration and DNA synthesis as well as to trigger apoptosis. The mechanistic basis of these differences remains unknown, but may be explained by: 1) differences in the relevant levels of NO X -species, 2) the redox environment of NO X production, 3) the presence and proximity of downstream response elements within cells under study, 4) the susceptibility of these NO X targets, 5) as well as the rate of NO X detoxification. If DNA repair mechanisms are overwhelmed or are incapable of adequate repair, endogenous cell death is triggered via apoptosis. This "programmed cell death" thus prevents damaged cells from undergoing replication perpetuating mutational events. Chronic nitrosative stress is implicated in inhibiting apoptosis. Potential mechanisms for this inhibition of apoptosis include the alteration of protein transcription and translation, or post-translational control of protein function [ 9 ]. NO X can alter protein function through post-translational modification of thiol groups via S-nitrosation, a key motif within many enzymes and structural proteins [ 10 ]. One target of this type of post-translational modulation of apoptotic enzymatic activity are the caspases [ 11 ]. Thus, nitrosative stress can on one hand induce DNA damage while inhibiting apoptosis, on the other. This lends mechanistic support to the hypothesis that chronic nitrosative stress may be cancer promotive in specific situations. Based on these data, the purpose of this study was to investigate whether cultured tumor cells exposed to exogenously administered nitrosative stress undergo NO X -mediated DNA damage. Furthermore, we investigate the relationship between this ongoing DNA damage and cell death via apoptosis within our cell lines of interest. Methods Cell Lines The human lung adenocarcinoma cell line (A549) and the control SV40 transformed human fibroblast cell line (WI38) were obtained from the American Type Culture Collection (Manassas, VA). A549 was grown in RPMI media, while WI38 cells were grown in MEM (Gibco, Paisley, UK). All media was supplemented with 10% fetal calf serum, penicillin, streptomycin, L-glutamine, and fungizone. For various tests, cells were harvested after trypsin-EDTA treatment, washed with Dulbecco's PBS, and resuspended in serumless media. Chemicals Cells were incubated with one of four NO X -donor compounds that differed in their mode of donation of NO-equivalents and their half-lives. Two of these compounds were NONOate donor compounds [diethylenetriamine-NONOate (DETA-NONOate, Sigma Chemical Company, St. Louis, MO) and Spermine-NONOate (Oxis International, Portland, OR)] which donate NO • its pure form into the aqueous environment. The other two compounds were thiol-based nitric oxide donors [(±)-S-Nitroso-N-acetypenicillamine (SNAP) and N-(β-D-Glucopyranosyl)-N 2 -acetyl-S-nitroso-D,L-penicillamide) glyco-SNAP (Oxis International, Portland, OR)], which preferentially donate NO + -equivalents to other thiols. Additionally, these donors differed in their stability. Based upon our spectrophotometric analysis at 37°C and a pH of 7.4, Spermine-NONOate has t 1/2 = 5 hours, DETA-NONOate a t 1/2 ≈ 24 hours, SNAP a t 1/2 = 10 hours, and glyco-SNAP a t 1/2 = 28 hours [data not shown]. Furthermore, NONOate-based donors donate two molar equivalents of NO X per mole of donor compound, whereas the thiol donors deliver only one mole of NO X per mole of donor compound. These donor compounds were chosen to assess whether differences in the concentration, mode, or the rate of NO X delivery alter the genotoxicity of this radical species. Nitrite Production in Media Nitric oxide donors were added to media without cells or to cell supernatants at increasing concentrations from 75 to 600 μM. After 24 hours of incubation the amount of nitrite produced in the media was assayed by the Greiss reaction as previously described [ 12 , 13 ]. Briefly, 50 μl of media was added to 50 μl of 1% sulfanilamide in 2.5% H 3 PO 4 . Then 50 μl of 0.1% napthylethylenediamine dihydrochloride in 2.5% H 3 PO 4 was added in a 96-well microtiter plate. After incubation at room temperature for 30 minutes, the absorbance was measured on a microplate reader (Molecular Devices, Sunnyvale, CA) at 540 nm. The concentration of nitrite in the media was quantified as derived from standard curves created by adding known concentrations of NaNO 2 from 50 to 300 μM sodium nitrite. NO X delivery kinetics was confirmed by determining changes in λ maximum of each compound over time on a spectrophotometer (Beckman Coulter DU530, Fullerton, CA). Single Cell Gel Electrophoresis (COMET assay) The COMET assay was performed according to the procedure of Singh et al. with a few modifications [ 14 ]. Briefly, 120 μl of 0.5% normal melting point agarose in Ca +2 and Mg +2 -free phosphate buffer at 56°C were quickly layered onto a fully frosted slide and immediately covered with a cover-slip. The slides were kept at 4°C to allow the agarose to solidify. After gently removing the cover-slip a 50 μl aliquot of cell suspension was mixed with an equal volume of 1% low melting point agarose (Sigma, St. Louis, MO) at 37°C and quickly pipetted onto the first agarose layer in the same manner. Finally, 70 μl of 0.5% LMP agarose was added to cover the cell layer. The slide sandwiches without cover-slips were immersed in freshly prepared, cold lysing buffer [2.5 mol/l NaCl, 100 mmol/l Na 2 EDTA, 10 mmol/l Tris, 1% N -Lauroyl sarcosine sodium salt, pH 10, with 1% Triton X-100 added just before use] and kept at 4°C for 45 min to 1 hour. The slides were placed on a horizontal gel electrophoresis platform and covered with cold alkaline buffer [300 mmol/l NaOH, and 1 mmol/l Na 2 EDTA] for 8 to 20 minutes in the dark at 4°C to allow DNA unwinding and expression of the alkali-labile sites. The timing for lysis and unwinding was determined empirically for each cell line. Electrophoresis was conducted at 4°C in the dark for 20 min at 25 V and 300 mA. The slides were then rinsed gently twice with neutralizing buffer (0.4 mol/l Tris, pH 7.5). Each slide was stained with 120 μl of propidium iodine (Sigma) at a concentration of 5 μg/ml and covered with a cover-slip. COMET tail lengths were quantified as the distance from the centrum of the cell nucleus to the tip of the tail in pixel units, with the mean tail length being determined as the mean length of twelve tails. TUNEL Assay for Apoptosis A terminal deoxynucleotidyl transferase (TdT)-mediated dUTP labeling (TUNEL) method was used for the detection of apoptotic cell death. For this purpose, the Apop Tag-Plus in-Situ Apoptosis Detection Kit (peroxidase) (Oncor, Gaithersburg, MD) was used. The staining was performed according to the manufacturer's recommended procedure. In brief, cell cultures of unexposed or exposed cells were initially treated with proteinase K (20 μg/ml) for 15 minutes at room temperature prior to using TdT to label the 3'-OH ends of DNA with digoxigenin-labeled nucleotides (1 hour incubation at 37°C). The slides were then treated with anti-digoxigenin antibody-peroxidase conjugate for 30 minutes at room temperature, stained with 3'-3 diaminobenzidine tetrahydrochloride (DAB) for 5 minutes to produce the characteristic brown color of positive cells, counterstained with hematoxylin, and mounted. Sections included in the kit were stained and served as positive controls. Consecutive oil immersion (100× objective) fields were counted on an Olympus BX40 microscope. A minimum of 1000 cells were counted and the apoptotic index was calculated as the percentage of staining cells. Cells were defined as apoptotic when the whole nuclear area was labeled or when occasional labeled globular bodies (apoptotic bodies) could be observed in the cytoplasm. Apoptosis found in the untreated cell cytospins at this time point were found to be 1%. Negative controls were cytospins of our cell lines in their logarithmic phase of growth supplemented with conditioned media and FCS. No apoptotic cells were demonstrated in these cytospins. Annexin V-FITC/propidium iodine FACS Flow cytometric determination of apoptosis was performed using a commercially available (R&D Systems, Minneapolis, MN) Annexin V-FITC/propidium iodine apoptosis detection kit. Untreated and treated cells were collected after 24 hours of incubation by trypsinization and centrifugation at 500 × g for 5–10 minutes at room temperature. Cells were washed and resuspended in ice cold PBS and pelleted by centrifugation. Cells were then resuspended in Annexin V incubation reagent at a concentration of 1 × 10 6 cells/100 μl, and incubated in the dark for 15 minutes at room temperature. Binding buffer was then added to each sample. Samples were then analyzed within one hour by flow cytometry, and evaluated based on the percentage of the population of cells staining low or high for Annexin V (apoptotic cells) and propidium iodide (necrotic cells). Caspase Activity Assay Activities of caspase-3 and -9 were determined using the corresponding caspase activity detection kits (R&D Systems, Minneapolis, MN) as described previously [ 15 ]. Briefly, 100 μg of total protein was added to 50 μl of reaction buffer, and 5 μl of substrates DEVD-pNA and LEDH-pNA were used to analyze the activity of caspase-3 and -9 respectively. Samples were incubated at 37°C for 3 hours and the enzyme-catalyzed release of pNA was quantified at 405 nm using a microtiter plate reader. The values of treated samples were normalized to corresponding untreated controls allowing determination of the fold increase or decrease in caspase activity. Alterations in enzymatic activity directly attributable to NO X were determined by the change in caspase activity in the presence or absence of 1 mM DTT. Cell Viability Assay Cell viability was determined via a modified MTT assay as previously described [ 16 ]. In brief, cells were incubated with the various NO X donors for 24 hours at 37°C prior to cell viability assay. The media was aspirated and replaced with 0.2 ml of MTT (1 mg/ml in PBS), followed by incubation at 37°C for 5 hours. MTT was then aspirated and the wells were air dried for 5 minutes. The crystals were dissolved with 200 μl of DMSO, until the solution turned purple and absorbance analyzed in an enzyme-linked immunosorbent assay (ELISA) plate reader at 540 nm. Nitrotyrosine Staining Pathologic nitric oxide exposure was confirmed by immunostaining using an anti-nitrotyrosine (Upstate Biotechnology, Lake Placid, NY) monoclonal antibody by immunoperoxidase as previously described [ 17 ]. The nitrotoyrosine Mab have previously demonstrated specific activity against human nitrotyrosine. Statistical Analysis Statistical evaluation was performed using Prism-3.00 (GraphPad Software Inc., San Diego, CA). A p-value of <0.05 was considered significant. Values listed represented mean ± the standard error of the mean unless otherwise indicated. Results Increasing Concentrations of NO X donors Increases Nitrosative Stress Our first aim was to confirm that adding nitric oxide donors into aqueous media increased NO X production in a concentration dependent fashion. To investigate this, we determined the accumulation of NO 2 - in the supernatants based on the Greiss reaction. Figure 1 demonstrates the concentration of nitrite in media in the absence of and in the presence of our two cell lines. In the absence of cells, we demonstrated an increasing concentration of nitrite within the media after 24 hours of incubation of various concentrations of all four NO X donors. This increase in nitrite accumulation was the same when NO X donors were added to the media alone or to the supernatants of our two cell lines. These data proved that our NO X donors were liberating NO X in a concentration dependent fashion. Figure 1 Nitrite Production in Media and Supernatants in the Presence of Increasing Concentrations of NO X Donors: A. Nitrite production in media without cells, B. Nitrite production in WI38 supernatants, C. Nitrite production in A549 Supernatants. Measurements of nitrite production in media alone or the supernatant of WI38 and A549 as measured by the Greiss reaction demonstrate increasing concentrations of nitrite production in the presence of increasing concentrations of NO X donors. This confirms that our NO X donors are liberating NO X in a concentration-dependent fashion. Interestingly, the curves for SNAP and glycol-SNAP were shifted toward the right in the supernatants of both WI38 and A549 possibly reflecting the possibility that these thiol-based NO X donors were preferentially donating their NO X to intracellular thiols. We then confirmed that the published half-lives of our donor compounds were consistent with the kinetic release of NO X in our system. Using λ maximum specific for each donor compound, we found that the decay in absorbance to half-maximum roughly reflected the published t 1/2 for each of our donor compounds [data not shown]. To confirm that our cell lines were indeed being exposed to increasing amounts of NO X resulting from the addition of our NO X donors, we examined immunostaining for nitrotyrosine [Table 1 ]. In the absence of NO X donors, both cell lines demonstrated low levels of nitrotyrosine staining which increased with the addition of our donor compounds. The addition of exogenous NO X increased the staining intensity for nitrotyrosine in both cell lines, which appeared to saturate at a moderate level of immunostaining for nitrotyrosine. This result suggests ongoing low level of autologous nitration, and confirms that cells are being exposed to increasing nitrosative stress with the addition of NO X donors. Table 1 Cell Line Nitrotyrosine Staining (Intensity) NO • Donor Concentration (μM) Control SNAP Glyco-SNAP DETA Spermine A549 0 + 75 ++ + + ++ 150 ++ ++ ++ ++ 300 ++ ++ ++ ++ WI38 0 + 75 + ++ + ++ 150 + ++ ++ ++ 300 ++ ++ ++ ++ Mean nitrotyrosine staining intensity per 100 cells: 0+-no staining, 1+-mild staining, 2+-moderate staining, 3+-intense staining. DNA strand-breaks after exposure to NO X In order to quantify the ability of exogenous NO X to cause mutational events, we performed the in-vitro single-cell gel electrophoresis assay (the COMET assay). Increasing concentrations of these NO X donors were added to the supernatants of our cell lines and DNA strand-breaks were measured. As seen in Figure 2 , distinctly different susceptibility to NO X exposure was seen in WI38 when compared to the tumor cell lines A549. Increasing concentrations of NO X donors increased DNA strand breaks in a concentration dependent fashion in the lung adenocarcinoma cell line, A549. Exposure of WI38 cells to increasing concentrations of NO X failed to cause an increase in DNA fragmentation that was detectable by the COMET assay. Examples of COMET tail moments seen in untreated A549 cells, in cells exposed to low concentrations, and high concentrations of NO X donors are seen in Figure 3 . Figure 2 COMET Tail Moment Lengths After 24 hour Exposure to Various NO X Donors: A. COMET tail moment of WI38 after 24 hours of exposure to increasing concentrations of NO X , B. COMET tail moment of A549 after 24 hours of exposure to increasing concentrations of NO X . Measurements of the COMET tail moments for WI38 and A549 in the presence of increasing concentrations of NO X donors demonstrate that A549 was found to have increasing DNA strand breaks as nitrosative stress increased which was not seen in the WI38 control cell line. As seen in the graph, NO X donors with a long half-life [i.e. glyco-SNAP with a t 1/2 = 28 hours and DETA-NONOate with a t 1/2 of 20 hours] demonstrated significant increases in COMET tail moments only at the higher concentrations of NO donor, whereas NO donors with a shorter half-life [i.e. SNAP with a t 1/2 of 10 hours and Spermine-NONOate with a t 1/2 of 39 minutes] demonstrated significant increases in DNA strand breaks at lower concentrations. This confirms that A549 demonstrates more genomic instability in the presence of increasing nitrosative stress, which is not seen in control cells. Figure 3 Example COMET Tail Moments for A549: These are example COMET tail moments for A549 in the presence of A.) 0 μM, B.) 75 μM, and C.) 150 μM Spermine-NONOate. Note that a significant increase in tail moment can be visualized with increasing concentrations of this NO X donor. Determination of Cell Viability as a result of NO X -dependent DNA Fragmentation We then utilized the MTT assay to explore if this NO X -donor compound induced DNA fragmentation caused a decrease in cell viability. As demonstrated in Figure 4 , both cell lines were found to have a decrease in cell viability after exposure to increasing concentrations of either thiol based or NONOate-based NO X donors. Our control fibroblast cell line appeared to be more susceptible to NO X exposure, in that WI38 demonstrated a significant decrement in cell viability between 75 and 300 μM. Interestingly, the NONOate-based NO X donors were able to significantly decrease the cell viability of WI38 cells at a lower concentration than the thiol-based NO X donors. This is contrasted with our tumor cell line, A549. The decrement in cell viability for A549 was not significant until 600 μM of these NO X donors. In fact, a significant increase in cell growth was seen at the lowest concentration (75 μM) of SNAP in our A549 cells. Therefore, within the range of concentrations of NO X donor compounds studied, we have found that NO X has the ability to induce DNA fragmentation in A549 but not in the WI38 cells, while significantly decrease cell viability for WI38 relative to A549 cells. Figure 4 Cell Viability as Determined by the MTT Assay with Exposure to Increasing Concentrations of NO X Donors: A. Cell viability of WI38 as measured by the MTT assay after exposure to increasing concentrations of NO X , B. Cell viability of A549 as measured by the MTT assay after exposure to increasing concentrations of NO X . The cell viability was significantly reduced in WI38 the presence of increasing concentrations of NO X donors at all concentrations greater than 150 μM (p < 0.007 for all) for DETA- and spermine-NONOate, whereas the thiol-based glycol-SNAP and SNAP did not significantly decrease cell viability until 300 μM concentrations. These results are contrasted with A549 cells which did not demonstrate a significant decrease in cell viability (p < 0.03) until 600 μM concentrations of either thiol-based or NONOate-based NO X donors. Interestingly, the lowest concentration of SNAP (75 μM) demonstrated a significant growth advantage (p = 0.01). Determination of apoptosis after exposure to NO X donors We then sought to determine whether NO X -mediated increases in DNA strand breaks correlated with apoptosis via the TUNEL assay. Since a decrease in cell viability was not seen until A549 cells were exposed to 600 μM NO X -donor, we examined apoptosis after exposure to this concentration. Cultured cells were exposed to 600 μM of the thiol-based NO X -donor, SNAP, or the NONO-ate based NO X donor, Spermine. Either NO X donor compound exposure for 24 hours did not significantly increase the percentage of TUNEL positive cells when compared with untreated controls [data not shown]. Because TUNEL positivity represents a late event in the apoptotic cascade, we wanted to ensure that the assay time frame did not represent assessment prior to detectable apoptosis. Therefore, we confirmed our apoptosis results via Annexin-V/propidium iodide FACS sorting. After exposure of our cell lines to 600 μM of SNAP or Spermine-NONOate, we confirmed the absence of any significant increase in apoptosis (Figure 5 ). Figure 5 Annexin-V/Propidium Iodine FACS in A549 Cells Unexposed and Exposed to High Concentrations of NO X Donors: A. Untreated control, B. 600 μM Spermine NONOate treated, C. 600 μM SNAP treated. Annexin-V/Propidium Iodine FACS confirms the TUNEL results that demonstrate that no significant increase in apoptosis is noted in A549 with or without exposure to high nitrosative stress. Determination of caspase activity after NO X exposure or inhibition of NO X production In an attempt to explain these seemingly contradictory data, we examined the role of NO X in the modulation of caspase enzyme function via an in vitro caspase activity assay (figure 6 ). Using the artificial substrates DEVD-pNA and LEDH-pNA for assaying caspase-3 and -9 activity respectively, we demonstrated a basal inhibition of caspase-3 in A549 cells as demonstrated by a significant increase in caspase substrate cleavage with the addition of the NO X -specific inhibitor, N Ω -monomethyl-L-arginine (L-NMMA), when compared to control cells. This inhibition appears to be attributable to basal S-nitrosation of caspase-3, since the addition of 1 mM DTT also significantly increased caspase-3 activity. With the addition of exogenous NO X via our thiol-based donor compound, SNAP, a significant decrease in the activity of both caspases was found. In contrast, our NONOate-based NO X donor, Spermine, was unable to decrease either caspase-3 or -9 activities. The inhibition of both caspase's activity by SNAP was reversed with exposure to 1 mM DTT, suggesting that transnitrosation of these caspase enzymes may contribute to the inhibition of enzymatic activity. These data suggest that thiol-based NO X -mediated inhibition of caspase activity in A549 cells may cause a relative tolerance to ongoing DNA damage by inhibiting apoptosis. Discussion We have demonstrated that NO X genotoxicity is highly dependent upon the concentration and kinetics of delivery of NO X . In addition, tumor cells appear to have increased genomic instability but increased resistance to ongoing NO X -mediated genotoxicity when compared to more normal transformed cells which demonstrate less genomic instability but increased susceptibility to ongoing DNA damage. Moreover, the mechanism of NO X delivery appears to have profound consequences upon the function of downstream effector molecules such as caspases. Over the past 20–30 years key proteins have been increasingly implicated as targets for nitric oxide signaling. With respect to its role in cancer, both carcino-protective [ 18 ] and carcinogenic roles [ 19 ] have been attributed to NO X . The seemingly paradoxic nature of NO X in cancer has made investigations in this area quite enigmatic. Increasing evidence supports NO X 's role in the induction and promotion of cancer [ 8 , 20 ]. Chronic inflammatory nitrosative stress has been implicated in carcinogenesis in a number of other organ systems [ 21 , 22 ]. Within this lung adenocarcinoma model, it is easy to hypothesize that either endogenous chronic inflammatory nitrosative stress as seen with idiopathic pulmonary fibrosis, or exogenous NO X as a result of cigarette smoking may predispose individuals to the development of this type of cancer. One mechanism by which NO X has been hypothesized to be carcinogenic is through oxidative/nitrosative DNA damage and genotoxicity [ 23 - 25 ]. NO X can induce mutational events via DNA oxidation, deamination, point-mutations, and strand-breaks, thus contributing to the multi-step process of carcinogenesis. A number of papers have documented that NO X can cause DNA strand breaks in areas of chronic nitrosative stress [ 26 - 28 ]. Our data demonstrates that in the human lung adenocarcinoma cell line, A549, increasing concentrations of NO X donors increases DNA strand-breaks in a concentration dependent fashion. Moreover, the kinetics of NO X exposure also affects its genotoxicity. It is hypothesized that this ongoing DNA damage can lead to cancer development. In support of this notion, in a mouse model of inflammatory mediated lung carcinogenesis, studies have demonstrated that genetic ablation of the inducible isoform of nitric oxide synthase decreased the lung tumor multiplicity after exposure to the carcinogen, urethane [ 29 ]. The data presented herein of NO X -dependent DNA strand breaks in A549 cells is in stark contrast to NO X 's inability to induce DNA fragmentation in our control fibroblast cell line, WI38. Normally, DNA damage is repaired by a complex series of DNA repair mechanisms. Defects in repair mechanisms can predispose people to cancer development [ 30 ]. An increasing body of evidence supports epigenetic modulation of enzymatic function, such as S-nitrosation [ 31 ]. As a result of nitrosation of key thiol groups within DNA repair enzymes, DNA-damage repair can be inhibited potentially perpetuating mutational events. One possible explanation for these opposing sets of data is that A549 may be capable of expressing unstimulated NOS activity [ 32 ], and therefore may have saturated such intracellular thiols as glutathione. In support of this hypothesis, immunostaining for nitrotyrosine in both cell lines found low levels of nitrotyrosine staining in the absence of exogenous NO X . Research is ongoing to determine whether tumor elicited NO X production may contribute to tumor progression. If DNA damage overwhelms repair mechanisms, normal cells are triggered to undergo apoptosis. Redundant mechanisms for the induction of apoptosis exist, but the most common pathway involves the induction of p53. With the induction of p53, apoptosis is triggered through the release of cytochrome c and caspase activation. Despite possessing wild-type p53, our data demonstrated that in A549 cells we were unable to detect a decrease in cell viability as measured by the MTT assay until very high concentrations of donors, or an increase in cell apoptosis via either the TUNEL assay or Annexin V/propidium iodide FACS. There are conflicting results in the literature regarding the effects of exogenous NO X on cell viability and apoptosis. Certain studies have shown that NO X by itself does not cause A549 cells to die, but the addition of hyperoxia induces rapid cell death [ 33 ], perhaps through the production of peroxynitrite. On the other hand, other studies have demonstrated that the addition of the NO X donor SNAP decreased cell viability via apoptosis in a concentration dependent fashion [ 34 ]. This study went on to analyze apoptosis in cells other than A549, thus precluding a direct correlation with A549 and apoptosis after SNAP exposure. Further confusing the interrelationship between DNA fragmentation and apoptosis is the fact that the prevailing opinion in the literature is that there exists a direct correlation between the fragmentation seen in the COMET assay and the fragmentation seen in apoptosis. With triggering of apoptosis, DNA strands are cleaved or nicked by nucleases, exposing 3'-hydroxyl ends. Nick-end cleavage may not necessarily equate with strand breaks seen with various mutagens. A growing body of evidence suggests that the fragmentation found in the COMET assay is not necessarily related to apoptotic fragmentation [ 35 , 36 ]. Studies have demonstrated that COMET tail moments of cells undergoing apoptosis are highly fragmented to the point of loosing nuclear architecture [ 37 ]. Furthermore, nitroso-compound related DNA damage appears to be independent of more commonly accepted measures of apoptosis, such as the TUNEL assay [ 38 ]. The mechanistic basis for these disparate measures remains to be determined, but appears to represent fundamental differences in the characteristics of exposed 3' ends seen with oxidative DNA damage versus apoptosis related DNA cleavage. Therefore, the COMET assay may be a measure of genotoxicity, without necessarily detecting apoptosis. Lending credence to NO X 's possible carcinogenic role is the implication from several lines of study that NO X can influence enzymatic function of such enzymes as caspases [ 11 ]. In this way, NO X may increase the threshold by which cells that have undergone a level of genotoxicity would be triggered to undergo apoptosis. If any member of the apoptotic pathway is lost or inhibited, the threshold for apoptosis would theoretically be increased. NO X has demonstrated a capability of inhibiting caspase activity through S-nitrosation of key thiols [ 15 , 39 ]. Our data supports the NO X -mediated inhibition of both caspase-3 and -9 in A549 cells with ongoing DNA mutational events. In fact, prior data has demonstrated that basal inhibition of mitochondrial caspases by cysteine nitrosation must be removed in order to activate the caspase-mediated apoptotic pathways [ 40 ]. Our data supports this in A549 in that the addition of the antioxidant, DTT, increased caspase activity detectable by our in fluorogenic assay. One potential mechanism for this inhibition of casapase activity is the S-nitrosation of essential thiol groups [ 40 ]. The fact that our thiol-based NO X donor appeared to inhibit caspase activity more efficiently that the NONOate donor further supports this notion, since it is well known that nitroso-thiols will preferentially transnitrosate other thiols [ 41 ]. In light of the increasing interest in NO X -donor compounds as cytotoxic cancer therapy, a careful consideration of potential downstream effects of the mode of NO X delivery should be undertaken. Caspases can be inhibited not only by direct protein S-nitrosation, but also indirectly by a cGMP-mediated pathways [ 39 ]. The return of caspase activity with the addition of DTT would further suggest that caspase inactivation was mediated by S-nitrosation in our lung adenocarcinoma model. The implications of these data can be applied both to the understanding of tumor progression as well as the design of NO X -based chemotherapeutic strategies. It is well established that tumors demonstrate a high rate of cellular turnover, with the vast majority of cells undergoing cell death resulting from ongoing genomic instability. Only those clones possessing a survival advantage will be able to repopulate the tumor and thus contribute to increasing aggressiveness, the ability to invade and metastasize, and resistance to further therapeutic interventions. Chronic nitrosative stress may be contributing to this process by selecting for more virulent clones via inducing DNA mutational events and perpetuating these mutations through a relative inhibition of apoptosis. Moreover, as more investigations explore the potential use of NO X delivery as a possible cytotoxic chemotherapy, the issues of concentration, kinetics, and mechanism of delivery must be carefully considered in order to avoid untoward pro-carcinogenic side effects. Conclusions In conclusion, we demonstrate that tumor cells experience an increase in DNA strand breaks with increasing nitrosative stress relative to the concentration and kinetics of delivery. This nitrosative stress does not correlate with increased cell death or apoptosis in established tumor cell lines, which is in stark contrast to non-tumor immortalized cells. The lack of apoptosis associated with increased DNA fragmentation may in part be explained by the inhibition of caspase-3 and -9 activity by thiol-based delivery but not NONOate-based delivery of NO X . Taken together, these data support the role of NO X in nitrosative genomic instability as well as inhibiting apoptosis, implicating it in cancer promotion. The demonstration of these same findings in rodent cell lines would establish the foundation for animal models with which to fully elucidate the role of NO X in tumor development and growth. Author's contributions BB conceived of the study, participated in the design of the study, and performed the statistical analysis. GH interpreted the results of the FACS and immunoassays. NH carried out the majority of the COMET assays, as well as performing the caspase assay. JR participated in the study design and coordination. All authors read and approved the final manuscript. Figure 6 Caspase 3 (DEVD-pNa) and Caspase 9 (LEDH-pNa) Activity as Measured by the in vitro Fluorogenic Caspase Assay in the Absence and Presence of High Concentrations of NO X Donors: Both endogenous and exogenous NO X appears to influence caspase activity in A549 cells. Caspase 3 activity appears to demonstrate endogenous NO X inhibition since the addition of 3 mM L-NMMA or 1 mM DTT significantly increased caspase 3 activity (p = 0.001, p = 0.003 respectively). No similar endogenous inhibition was seen with caspase 9. The addition of the thiol-based NO X donor, SNAP (600 μM), significantly inhibited both caspase 3 and caspase 9 (p < 0.001 for both) when compared with control cell caspase activity. This caspase activity was significantly reversed with the addition of DTT (p < 0.001 for both caspase 3 and 9). Treatment groups: 1) Control, 2) 3 mM L-NMMA, 3) 1 mM DTT, 4) 600 μM Spermine-NONOate, 5) 600 μM SNAP, 6) 600 μM SNAP + 1 mM DTT.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544845.xml

526756

Improved hit criteria for DNA local alignment

Background The hit criterion is a key component of heuristic local alignment algorithms. It specifies a class of patterns assumed to witness a potential similarity, and this choice is decisive for the selectivity and sensitivity of the whole method. Results In this paper, we propose two ways to improve the hit criterion. First, we define the group criterion combining the advantages of the single-seed and double-seed approaches used in existing algorithms. Second, we introduce transition-constrained seeds that extend spaced seeds by the possibility of distinguishing transition and transversion mismatches. We provide analytical data as well as experimental results, obtained with the YASS software, supporting both improvements. Conclusions Proposed algorithmic ideas allow to obtain a significant gain in sensitivity of similarity search without increase in execution time. The method has been implemented in YASS software available at .

Background Sequence alignment is a fundamental problem in Bioinformatics. Despite of a big amount of efforts spent by researchers on designing efficient alignment methods, improving the alignment efficiency remains of primary importance. This is due to the continuously increasing amount of nucleotide sequence data, such as EST and newly sequenced genomic sequences, that need to be compared in order to detect similar regions occurring in them. Those comparisons are done routinely, and therefore need to be done very fast, preferably instantaneously on commonly used computers. On the other hand, they need to be precise, i.e. should report all, or at least a vast majority of interesting similarities that could be relevant in the underlying biological study. The latter requirement for the alignment method, called the sensitivity , counterweights the speed requirement, usually directly related to the selectivity (called also specificity ) of the method. The central problem is therefore to improve the trade-off between those opposite requirements. The Smith-Waterman algorithm [ 1 ] provides an exact algorithmic solution to the problem of computing optimal local alignments. However, its quadratic time complexity has motivated the creation of rapid heuristic local alignments tools. A basic idea behind all heuristic algorithms is to focus only on those regions which share some patterns, assumed to witness (or to hit ) a potential similarity. Those patterns are formed by seeds which are small strings (usually up to 25 nucleotides) that appear in both sequences. FASTA [ 2 ] and BLAST [ 3 , 4 ] are well-known examples of such methods. BLAST is currently the most commonly used sequence alignment tool, and is a kernel of higher-level search tools, such as PSI-BLAST [ 4 ] for instance. More recently, several new alignment methods have been proposed, such as BLAT [ 5 ], PatternHunter [ 6 ], LAGAN [ 7 ], or BLASTZ [ 8 ], to name a few. The improvement brought by all those tools results from new more efficient hit criteria that define which pattern shared by two sequences is assumed to witness a potential alignment. Two types of improvements can be distinguished. On the one hand, using two or more closely located smaller seeds instead of one larger seed has been shown to improve the sensitivity/selectivity trade-off [ 4 - 6 ], especially for detecting long similarities. On the other hand, new seed models have been proposed, such as spaced seeds [ 6 ], seeds with errors [ 5 ], or vector seeds [ 9 ]. In this paper, we propose further improvements in both those directions. In the first part (Section Group hit criterion ), we propose a new flexible and efficiently computable hit criterion, called group criterion , combining the advantages of the single-seed ([ 3 ]) and multi-seed ([ 2 , 4 - 6 ]) criteria. In the second part (Section Generalized seed models ), we propose a new more expressive seed model which extends the spaced seed model of PatternHunter [ 6 ] by the possibility of distinguishing transition and transversion mismatches. We show that this allows to obtain a further gain in sensitivity on real genomic sequences, usually rich in transition mutations. Both proposed improvements have been implemented in YASS software [ 10 ], used in the experimental part of this work. Results Group hit criterion The first preparatory step of most heuristic alignment algorithms consists of constructing a hash table of all seeds occurring in the input sequences. In this section, we assume that a seed of weight k is a word consisting of k contiguous nucleotides ( k -word), more general notions of seed will be considered in the next Section. In the simplest case, implemented in the early version of BLAST [ 3 ], an individual seed occurring in both sequences acts as a hit of a potential alignment. It triggers the X-drop algorithm trying to extend the seed to a so-called High-scoring Segment Pair (HSP), used then to obtain a larger final alignment. Gapped BLAST [ 4 ] proposes a double-seed criterion that defines a hit as two non-overlapping seeds occurring at the same dotplot diagonal within a fixed-size window. This allows to considerably increase the selectivity with respect to the single-seed approach, and at the same time to preserve, and even to improve, the sensitivity on large similarities. On the other hand, Gapped BLAST is less sensitive on short and middle-size similarities of weak score. (We will show this more formally at the end of this Section.) Most of the existing alignment programs [ 5 , 6 ] use either a single-seed or a double-seed hit criterion. Here we propose a new flexible hit criterion defining a hit as a group containing an arbitrary number of possibly overlapping seeds, with an additional constraint on the minimal number of matching nucleotides. The seeds of the same group are assumed to belong to the same similarity, and therefore should be proximate to each other. In contrast to other multi-seed hit criteria [ 4 - 6 ], we don't require seeds to occur at the same dotplot diagonal but at close diagonals, to account for possible indels. The possible placement of seeds is controlled by parameters computed according to statistical models that we describe now. Group criterion A hit criterion defines a pattern which is considered as an evidence of a potential similarity. Every time this pattern is found, its extension is triggered to compute a potential larger alignment. The extension is usually done via a dynamic programming algorithm and is a costly step. The hit criterion should be selective enough to avoid spurious extensions and, on the other hand, should be sensitive to detect as many relevant similarities as possible. The hit criterion we propose is based on a group of seeds verifying conditions (1), (2) (see Section Methods ). By the considered statistical analysis, this ensures a good sensitivity. However, many groups will contain a single seed or two strongly overlapping seeds, that either belong to a similarity with a low score, or do not belong to any similarity at all (i.e. don't belong to an alignment with a sufficiently high score). To cope with this problem, we introduce an additional criterion that selects groups that will be actually extended. The criterion, called group criterion , is based on the group size defined as the minimal number of matching individual nucleotides in all seeds of the group . The group size can be seen as a parameter specifying the maximal overlap of the seeds of a group. For example, if the group size is k + 1, then no constraint on the overlap is imposed, i.e. any group containing two distinct seeds forms a hit. If the group size is 2 k , then the group must contain at least two non-overlapping seeds, etc. Allowing overlapping seeds is an important point that brings a flexibility to our method. Note that other popular multi-seed methods [ 4 , 5 ] consider only non-overlapped seeds. Allowing overlapped seeds and controlling the overlap with the group size parameter offers a trade-off between a single-seed and a multi-seed strategies. This increases the sensitivity of the usual multi-seed approach without provoking a tangible increase in the number of useless extensions. In the next section, we will provide quantitative measures comparing the sensitivity of the YASS group criterion with BLAST and Gapped BLAST. Some comparative and experimental data In this section, we adopt the following experimental setup to estimate the sensitivity of the YASS group criterion compared to other methods. We first set a match/mismatch scoring system, here fixed to +1/-3 (default NCBI-BLAST values). The main assumption is that the sensitivity is estimated as the probability of hitting a random gapless alignments of a fixed score . Moreover, to make this model yet more close to reality, only homogeneous alignments are considered, i.e. alignments that don't contain proper sub-alignments of bigger score (see [ 11 ]). For a given alignment length, all homogeneous alignments are assumed to have an equal probability to occur. In this setting, we computed the hit probability of a single-seed criterion with seed weight 11 (default for BLAST) and compared it with multi-seed criteria of Gapped BLAST and YASS for seed weight 9 (default for Gapped BLAST). For YASS, the group size was set to 13. Figure 1 shows the probability graphs for alignment score 25. Comparing BLAST and Gapped BLAST, the former is more sensitive on short similarities (having higher identity rate), while the latter is more sensitive on longer similarities, in which two close non-overlapping runs of 9 matches are more likely to occur than one run of 11 matches. The YASS group criterion combines the advantages of both: it is more sensitive than the single-seed criterion even for short similarities, and than the non-overlapping double-seed criterion for large similarities (Figure 1 ). Note, however, that for the chosen parameters, the YASS criterion is slightly less selective than that of Gapped BLAST as it includes any two non-overlapping seeds but also includes pairs of seeds overlapped by at most 5 bp. The selectivity can be estimated by the probability of a hit at a given position in a random uniform Bernoulli sequence (see [ 5 ]). For YASS, this probability is 2.1·10 -8 , which improves that of BLAST (2.4·10 -7 ) by more than ten. For Gapped BLAST, this probability is 7.3·10 -9 . To make an accurate sensitivity comparaison of YASS and Gapped BLAST, parameters should be set so that both algorithms have the same selectivity level. To compare the sensitivity of YASS and Gapped BLAST for an equal selectivity level, we chose a parameter configuration such that both algorithms have the same estimated selectivity (10 -6 ). This is achieved with seed weight 8 for Gapped BLAST and group size 11 for YASS (while keeping seed weight 9). In this configuration, and for sequences of score 25, YASS turns out to be considerably more sensitive on sequences up to 80 bp and is practically as sensitive as Gapped BLAST on longer sequences (data not shown). At the same time, YASS is more time efficient in this case, as Gapped BLAST tends to compute more spurious individual seeds that are not followed by a second hit, which takes a considerable part of the execution time. This is because the YASS seed is larger by one nucleotide, and the number of spurious individual seeds computed at the first step is then divided by 4 on large sequences. Compared to the single-seed criterion of BLAST, the YASS group criterion is both more selective (group size 13 vs single-seed size 11) and more sensitive for all alignment lengths , as soon as the alignment score is 25 or more. If the score becomes smaller, both criteria yield an unacceptably low sensitivity, and the seed weight has then to be decreased to detect those similarities. Finally, we point out another experiment we made to bring more evidence that the group criterion captures a good sensitivity/selectivity trade-off. We monitored the partition of the execution time between the formation of groups and their extension by dynamic programming (data not shown). It appeared that YASS spends approximately equal time on each of the two stages, which gives a good indication that it provides an optimal distribution between the two complementary parts of the work. Generalized seed models So far, we defined seeds as k -words, i.e. short strings of contiguous nucleotides. Recently, it has been understood that using spaced seeds allows to considerably improve the sensitivity. A spaced seed is formed by two words, one from each input sequence, that match at positions specified by a fixed pattern – a word over symbols # and _ interpreted as a match and a don't care symbol respectively. For example, pattern ##_# specifies that the first, second and fourth positions must match and the third one may contain a mismatch. PatternHunter [ 6 ] was the first method that used carefully designed spaced seeds to improve the sensitivity of DNA local alignment. In [ 12 ], spaced seeds have been shown to improve the efficiency of lossless filtration for approximate pattern matching, namely for the problem of detecting all matches of a string of length m with q possible substitution errors (an ( m , q )-problem). The use of some specific spaced seeds for this problem was proposed earlier in [ 13 ]. Yet earlier, random spaced seeds were used in FLASH software [ 14 ] to cover sequence similarities, and the sensitivity of this approach was recently studied in [ 15 ]. The advent of spaced seeds raised new questions: How to choose a good seed for a local alignment algorithm? How to make a precise estimation of the seed goodness, or more generally, of a seed-based local alignment method? In [ 16 ], a dynamic programming algorithm was proposed to measure the hit probability of a seed on alignments modeled by a Bernoulli model. In the lossless case, an analogous algorithm that allows to test the seed completeness for an ( m , q )-problem was proposed in [ 12 ]. The algorithm of [ 16 ] has been extended in [ 17 ] for hidden Markov models on order to design spaced seeds for comparing homologous coding regions. Another method based on finite automata was proposed in [ 18 ]. A complementary approach to estimate the seed sensitivity is proposed in [ 11 ]. Papers [ 19 , 20 ] present a probabilistic analysis of spaced seeds, as well as experimental results based on the Bernoulli alignment model. Other extensions of the contiguous seed model have been proposed. BLAT [ 5 ] uses contiguous seeds but allows one error at any of its positions. This strategy is refined in BLASTZ [ 8 ] that uses spaced seeds and allows one transition substitution at any of match positions. An extension, proposed in [ 9 ], enriches the PatternHunter spaced seeds model with a scoring system to define a hit. Here we propose a new transition-constrained seed model. Its idea is based on the well-known feature of genomic sequences that transition mutations (nucleotide substitutions between purins or between pyrimidins) occur relatively more often than transversions (other substitutions). While in the uniform Bernoulli sequence transitions are twice less frequent than transversions, in real genomic sequences this ratio is often inverted. For example, matrices computed in [ 21 ] on mouse and human sequences imply that the transition/transversion rate (hereafter ti/tv ) is greater than one on average. Transitions are much more frequent than transversions in coding sequences, as most of silent mutations are transitions. ti/tv ratio is usually greater for related species, as well as for specific DNA (mitochondrial DNA for example). Transition-constrained seeds are defined on the ternary alphabet {#, @, _}, where @ stands for a match or a transition mismatch (A ↔ G, C ↔ T), and # and _ have the same meaning as for spaced seeds. The weight of a transition-constrained seed is defined as the sum of the number of #'s plus half the number of @'s. This is because a transition carries one bit of information while a match carries two bits. Note that using transition-constrained seeds is perfectly compatible with the group criterion described in Section Group criterion . The only non-trivial algorithmic issue raised by this combination is how to efficiently compute the group size during the formation of groups out of found seeds. In YASS, this is done via a special finite automaton resulting from the preprocessing of the input seed. Transition-constrained seeds for Bernoulli alignment model To estimate the detection capacity of transition-constrained seeds, we first use the Bernoulli alignment model, as done in [ 6 , 19 , 20 ]. We model a gapless alignment by a Bernoulli sequence over the ternary match/transition/transversion alphabet with the match probability 0.7 and the probabilities of transition/transversion varying according to the ti/tv ratio. The sequence length is set to 64, a typical length of a gapless fragment in DNA alignments. We are interested in seed weights between 9 and 11, as they represent a good sensitivity/selectivity compromise. Table 1 compares the sensitivity of the best spaced seeds of weight 9, 10 and 11, reported in [ 20 ], with some transition-constrained seeds, assuming that transitions and transversions occur with equal probability 0.15. The transition-constrained seeds have been obtained using a stepwise Monte-Carlo search, and the probabilities have been computed with an algorithm equivalent to that of [ 16 ]. The table shows that transition-constrained seeds are more sensitive with respect to this model. A natural question is the efficiency of transition-constrained seeds depending on the ti/tv ratio. This is shown in Figure 2 . The left and right plots correspond to the seeds from Table 1 of weight 9 and 10 respectively. The plots show that the sensitivity of transition-constrained seeds greatly increases when the ti/tv ratio is over 1, which is typically the case for real genomic sequences. Transition-constrained seeds for Markov alignment model Homologous coding sequences, when aligned, usually show a regular distribution of errors due to protein coding constraints. In particular, transitions often occur at the third codon position, as these mutations are almost always silent for the resulting protein. Markov models provide a standard modeling tool to capture such local dependencies. In the context of seed design, papers [ 16 - 18 ] proposed methods to compute the hit probability of spaced seeds with respect to gapless alignments specified by (Hidden) Markov models. To test whether using transition-constrained seeds remains beneficial for alignments specified by Markov models, we constructed a Markov model of order 5 out of a large mixed sample of about 100 000 crossed alignments of genomic sequences of distantly related species ( Neisseria Meningitidis, S. Cerevisiae, Human X chromosome, Drosophila ). The alignments were computed with different seeds of small weight, to avoid a possible bias caused by a specific alignment method. We then designed optimal spaced and transition-constrained seeds of weight 9–11 with respect to this Markov model. Table 2 shows the results of this computation providing evidence that transition-constrained seeds increase the sensitivity with respect to this Markov model too. Experiments Seed experiments In order to test the detection performance of transition-constrained seeds on real genomic data, we made experiments on full chromosomic sequences of S. Cerevisiae (chromosomes IV, V, IX, XVI) and Neisseria meningitidis (strains MC58 and Z2491). The experiments were made with our YASS software [ 10 ] that admits user-defined transition-constrained seeds and implements the group criterion described in Section Group criterion . The experiments used seeds of weight 9 and 11, obtained on Bernoulli and Markov models (reported in Tables 1 and 2 ). The search was done using group size 10 and 12 respectively for seed weight 9 and 11 (option -s of YASS). This means that at least two distinct seeds were required to trigger the extension, with no additional constraint on their overlap, which is equivalent to the double-seed criterion of PatternHunter. The scoring system used was +1/-1 for match/mismatch and -5/-1 for gap opening/extension. Both strands of input chromosomes has been processed in each experiment (-r 2 option of YASS). For each comparison, we counted the number of computed alignments with E-value smaller than 10 -3 . Table 3 reports some typical results of this experiment. They confirm that using transition-constrained seeds does increase the search sensitivity. A side (non-surprising) observation is that, in all tests, the seed designed on the Markov model turns out to be more efficient than the one designed on the Bernoulli model. Note that the similarity search can be further improved by using transition-specific scoring matrices (for example, PAM Transition/Transversion matrices or matrices designed for specific comparisons [ 21 ]) rather than uniform matches/mismatch matrices, and transition-constrained spaced seeds are more likely to detect alignments highly scored by those matrices. Another advantage of transition-constrained seeds comes from the fact that they are more robust with respect to the GC/AT composition bias of the genome. This is because purins and pyrimidins remain balanced in GC- or AT-rich genomes, and one match carries less information (is more likely to occur "by chance") than two match-or-transition pairs. Program experiments A series of comparative tests has been carried out to compare the sensitivity with traditional approaches. Several complete bacterial genomes ranging from 3 to 5 Mb have been processed against each other using both YASS and the b12seq programs (NCBI BLAST package 2.2.6.). The tests used the scoring system +1/-1 for match/mismatch and -5/-1 for gap opening/extension. The threshold E-value for the output was set to 10 (default BLAST value), and the sequence filtering was disabled. YASS was run with its default seed pattern #@#__##__#_##@# of weight 9 which provides a good compromise in detecting similarities of both coding and non-coding sequences. For each test, the number of alignments with E-value less than 10 -6 found by each program, and the number of exclusive alignments were reported. By "exclusive alignment", we mean every alignment of E-value less than 10 -6 that does not share a common part (do not overlap on both sequences) with any alignment found by the other program. To take into account a possible bias caused by splitting alignments into smaller ones (X-drop effect), we also computed the total length of exclusive alignments, found by each program. Experiments are summarized in Table 4 and show that within a generally smaller execution time, YASS detects more exclusive similarities that cover about twice the overall length of those found by b12seq. The gain in execution time increases when the sequence length gets larger. Conclusions In this paper, we introduced two independent improvements of hit criteria for DNA local alignment. The group criterion , based on statistical DNA sequence models, combines the advantages of both single-seed and double-seed criteria. Transition-constrained seeds account for specificities of real DNA sequences and allow to further increase the search sensitivity with respect to spaced seeds. Both options have been implemented in YASS software available at . Transition-constrained seeds could be further extended using the idea of vector seeds [ 9 ], i.e. by assigning weights to each seed position, but also to each type of mutation. This would give a more general mechanism to account for the information brought by different mutations. But the model is also more flexible, an thus involves a larger search space to design seeds. Another new direction for further improving the efficiency is a simultaneous use of several seed patterns [ 22 - 24 ], complementing the sensitivity of each other. However, designing such families is also hard problem, due to the involved search space. Methods Statistical analysis We first introduce some notations used in this section. Let S 1 and S 2 be the input sequences of length m and n respectively. Each of them can be considered as a succession of m - k + 1 (respectively n - k + 1) substrings of length k , called k-words . If a k -word of S 1 matches another k -word of S 2 , i.e. S 1 [ i .. i + k - 1] = S 2 [ j .. j + k - 1] for some i ≤ m and j ≤ n , then these two k- words form a seed denoted < i , j >. Two functions on seeds are considered: For a seed < i , j >, the seed diagonal d (< i , j >) is m + j - i . It can be seen as the distance between the k -words S 1 [ i .. i + k - 1] and S 2 [ j .. j + k - 1] if S 2 is concatenated to S 1 , For two seeds < i 1 , j 1 > and < i 2 , j 2 >, where i 1 < i 2 and j 1 < j 2 , the inter-seed distance D (< i 1 , j 1 >, < i 2 , j 2 >) is the maximum between | i 2 - i 1 | and | j 2 - j 1 |. The problem considered in this Section is to derive conditions under which two seeds are likely to be a part of the same alignment, and therefore should be grouped together. More precisely, we want to be able to compute parameters ρ and δ such that two seeds < i 1 , j 1 > and < i 2 , j 2 > have a probability (1 - ε ) to belong to the same similarity iff D (< i 1 , j 1 >, < i 2 , j 2 >) ≤ ρ ,     (1) | d (< i 1 , j 1 >) - d (< i 2 , j 2 >)| ≤ δ .     (2) The first inter-seed condition insures that the seeds are close enough to each other. The second seed diagonal condition requires that in both seeds, the two k -words occur at close diagonals . We now describe statistical models used to compute parameters ρ and δ . Bounding the inter-seed distance Consider two homologous DNA sequences that stem from a duplication of a common ancestor sequence, followed by independent individual substitution events. Under this assumption, the two sequences have an equal length and their alignment is a sequence of matched and mismatched pairs of nucleotides. We model this alignment by a Bernoulli sequence with the probability p for a match and (1 - p ) for a mismatch. To estimate the inter-seed shift D k , we have to estimate the distance between the starts of two successive runs of at least k matches in the Bernoulli sequence. It obeys the geometric distribution of order k called the Waiting time distribution [ 25 , 26 ]: Using this formula, we compute ρ such that the probability is (1 - ε ) for some small ε . Note that the Waiting time distribution allows us to estimate another useful parameter: the number of runs of matches of length at least k inside a Bernoulli sequence of length x . In a Bernoulli sequence of length x , the probability of the event I p , x , r of having exactly r non-overlapping runs of matches of length at least k is given by the following recursive formula: This gives the probability of having exactly r non-overlapping seeds of length at least k inside a repeat of size x . The recurrence starts with r = 0, in which case and is computed through the Waiting time distribution. The distribution allows us to infer a lower bound on the number of non-overlapping seeds expected to be found inside a similarity region. In particular, we will use this bound as a first estimate of the group criterion introduced later. Bounding the seed diagonal variation Indels (nucleotide insertions/deletions) are responsible for a diagonal shift of seeds viewed on a dotplot matrix. In other words, they introduce a possible difference between d (< i 1 , j 1 >) and d (< i 2 , j 2 >). To estimate a typical shift size, we use a method similar to the one proposed in [ 26 ] for the search of tandem repeats. Assume that an indel of an individual nucleotide occurs with an equal probability q at each of l nucleotides separating two consecutive seeds. Under this assumption, estimating the diagonal shift produced by indels is done through a discrete one-dimensional random walk model, where the probability of moving left or right is equal to q , and the probability of staying in place is 1 - 2 q . Our goal is to bound, with a given probability, the deviation from the starting point. The probability of ending the random walk at position i after l steps is given by the following sum: A direct computation of multi-monomial coefficients quickly leads to a memory overflow, and to circumvent this, we use a technique based on generating functions. Consider the function and consider the power P l ( x ) = a l . x l +…+ a - l . x - l . Then the coefficient a i computes precisely the above formula, and therefore gives the probability of ending the random walk at position i after l steps. We then have to sum up coefficients a i for i = 0,1, -1, 2, -2,..., l , - l until we reach a given threshold probability (1 - ε ). The obtained value l is then taken as the parameter δ used to bound the maximal diagonal shift between two seeds.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526756.xml

529471

Hospitalization for heart disease, stroke, and diabetes mellitus among Indian-born persons: a small area analysis

Background We set out to describe the risk of hospitalization from heart disease, stroke, and diabetes among persons born in India, all foreign-born persons, and U.S.-born persons residing in New York City. Methods We examined billing records of 1,083,817 persons hospitalized in New York City during the year 2000. The zip code of each patient's residence was linked to corresponding data from the 2000 U.S. Census to obtain covariates not present in the billing records. Using logistic models, we evaluated the risk of hospitalization for heart disease, stroke and diabetes by country of origin. Results After controlling for covariates, Indian-born persons are at similar risk of hospitalization for heart disease (RR = 1.02, 95% confidence interval 1.02, 1.03), stroke (RR = 1.00, 95% confidence interval, 0.99, 1.01), and diabetes mellitus (RR = 0.96 95% confidence interval 0.94, 0.97) as native-born persons. However, Indian-born persons are more likely to be hospitalized for these diseases than other foreign-born persons. For instance, the risk of hospitalization for heart disease among foreign-born persons is 0.70 (95% confidence interval 0.67, 0.72) and the risk of hospitalization for diabetes is 0.39 (95% confidence interval 0.37, 0.42) relative to native-born persons. Conclusions South Asians have considerably lower rates of hospitalization in New York than reported in countries with national health systems. Access may play a role. Clinicians working in immigrant settings should nonetheless maintain a higher vigilance for these conditions among Indian-born persons than among other foreign-born populations.

Background Immigrant populations in industrialized nations are on average healthier than their native-born counterparts [ 1 ]. However, there is considerable debate surrounding the risk of ischemic heart disease and other related vascular and metabolic conditions among South Asian Indians, both within India and in the countries to which they immigrate [ 2 - 7 ]. The relationship between ischemic heart disease and Indian ethnicity is complex. For instance, lower rates of smoking are juxtaposed against higher mean cholesterol levels and unfavourable cholesterol ratios, higher rates of central obesity, and a higher prevalence of diabetes mellitus among Indian immigrants relative to either native-born or other foreign-born populations[ 3 - 5 ], [ 7 - 9 ]. Studies conducted in countries with national health-care systems have noted approximately twice the risk of hospitalization for ischemic heart disease or myocardial infarction among South Asian immigrant men relative to native-born men aged 45–64 [ 10 ]. This amounts to nearly four times the risk of Caribbean-born men in the same study. However, India-born women were found to be at similar risk as native-born women. South Asian immigrants may also experience the onset of symptomatic heart disease at a younger age than native born populations [ 11 ]. Nonetheless, mortality rates of ischemic heart disease among Indian immigrants to Canada were comparable to Canadians of European ancestry [ 12 ]. There is therefore considerable confusion surrounding the actual underlying risk of ischemic heart disease among South Asian persons. We set out to describe relative differences in the rates of hospitalization for heart disease, stroke, and diabetes among Indian-born persons relative to native-born persons or other foreign-born persons living in New York City. Unfortunately, there are no national datasets that provide individual level data for the risk of hospitalisation among specific immigrant sub-groups. It is, however, possible to compare the risk of hospitalization among immigrant sub-groups on an area-level using established small area analytic techniques via local datasets with a large sample size [ 13 ]. Herein, we quantify the risk of hospitalization for heart disease among Indian-born persons relative to other groups using a large hospitalization dataset covering New York City. Methods We obtained all hospital admission records for residents of New York City from the Statewide Planning and Research Cooperative System (SPARCS) and population data for New York City from Census 2000 [ 14 ]. The SPARCS dataset contains information on 2.45 million hospitalizations for the civilian population of New York State, including basic demographic variables and diagnosis codes [ 15 ]. A logistic model was used to calculate rate ratios and risk ratios (RR) rather than odds ratios. All analyses were conducted on SAS version 8 (Cary, NC). We used this model to calculate hospitalizations for manifestations of atherosclerotic heart disease (myocardial infarction, angina, and congestive heart failure, International Classification for Disease, 9 th Revision codes 410, 413, and 428 respectively), stroke (430–438), and diabetes (250) as the dependent variables in each regression model. The following categorical variables were entered as independent predictors: age (0–6, 7–17, 18–44 [reference], 45–64, 65+), sex (reference = female), education level (completed high school versus no high school [reference]), and country of birth (born in India, other foreign-born persons, and native-born [reference]). While the SPARCS dataset includes information on patients' age, sex, race, and ethnicity, it does not contain information on country of origin, income, or education level. To obtain country of origin and income data, we matched each patient's zip code from SPARCS hospitalization records to information obtained from the Census 2000 Long Form [ 14 ]. For instance, to calculate age-specific hospitalization rates, all hospitalizations among persons aged 65 and older residing within a particular zip code were divided by the number of persons aged 65 and older living within that same zip code. We then entered the proportion of foreign-born persons and Indian-born persons living within each zip code in New York City, as well as the proportion graduating from high school as independent variables in our regression analysis. The model took the general form: Logistic models were constructed for each category of hospitalization and the variables representing the proportion of persons from each region of the world. We tested the models using Hosmer and Lemeshow goodness of fit test, and index plots of the residuals. Results In 2000, there were 70,598 persons in New York City who were born in India (see Table 1 ). Of these, 45% were female, 63% were married, and 14.4% were living below the poverty line. Table 2 presents the relationship between age, sex, and education level on the risk of hospitalization for heart disease alone. As expected, there is an increase in heart disease risk with increasing age, male sex, and lower educational attainment. Table 1 Socio-demographic characteristics of Indian-born persons in New York City. Number Percent Total Population 70,598 - Age 0–17 years 7,489 10.6% 18–24 years 7,681 10.9% 25–34 years 17,950 25.4% 35–44 years 15,938 22.6% 45–54 years 11,789 16.7% 55–64 years 6,499 9.2% 65–74 years 2,320 3.3% 75 + years 932 1.3% Sex Female 31,749 45.0% Male 38,849 55.0% Marital Status Now married 44,665 63.3% Widowed 2,093 3.0% Divorced 1,352 1.9% Separated 561 0.8% Never married* 21,927 31.1% Poverty status Group quarters 583 0.8% Below poverty line 10,150 14.4% 100–200% poverty line 11,797 16.7% >200% poverty line 48,068 68.1% *Includes persons < age 15. Table 2 Risk ratios of hospitalization for heart disease by age, sex, and education (reference levels and 95% confidence interval are indicated in brackets) Variable Female Male Total Age 0–6 0.218 (0.153, 0.310) 0.534 (0.361, 0.788) 0.299 (0.230, 0.388) 7–17 0.192 (0.130, 0.285) 0.336 (0.126, 0.892) 0.219 (0.152, 0.316) 18–44 (ref) 1.00 1.00 1.00 45–64 13.58 (12.99, 14.19) 15.42 (14.42, 16.48) 13.98 (13.47, 14.51) 65+ 44.75 (42.84, 46.75) 77.37 (72.55, 82.50) 55.67 (53.71, 57.70) Male Sex (Female) N/A 1.62 (1.60, 1.65) N/A High School Grad. (Non-Graduate)* 0.943 (0.935, 0.951) 0.857 (0.850, 0.865) 0.901 (0.896, 0.906) *Because some values were small, risk ratios for this category were rounded to 3 decimal places. Table 3 presents the risk of hospitalization for heart disease, stroke, and diabetes mellitus by country of birth controlling for age, sex, education, and percentage of foreign-born persons in zip code. After controlling for covariates, foreign-born persons were less likely than native-born persons to be hospitalized for heart disease (RR = 0.70, 95% confidence interval 0.67, 0.72). As with native-born persons, foreign-born women (RR = 0.67, 95% confidence interval 0.63, 0.71) were at lower risk of hospitalization for heart disease than foreign-born men (RR = 0.74, 95% confidence interval 0.70, 0.78); however, the difference was much smaller than it is with native-born persons (RR = 1.62, 95% confidence interval 1.60, 1.65, see Table 1 ). Persons born in India were at similar risk of hospitalization as native-born persons for heart disease (RR = 1.02, 95% confidence interval 1.02, 1.03), but 47% more likely to be hospitalized for heart disease than other foreign-born persons (data not shown). Differences in the risk of hospitalization between India-born men and women were minimal. Table 3 Risk ratios of hospitalization for heart disease, stroke, and diabetes among Indian persons after controlling for age, sex, education, and percentage of foreign-born persons in zip code (reference levels and 95% confidence interval are indicated in brackets) Variable Male Female Total Heart Disease Foreign-born (Native-born) 0.74 (0.70, 0.78) 0.67 (0.63, 0.71) 0.70 (0.67, 0.72) Born in India (Native-Born) 1.03 (1.02, 1.04) 1.01 (1.00, 1.02) 1.02 (1.02, 1.03) Stroke Foreign-born (Native-born) 0.75 (0.68, 0.83) 0.80 (0.73, 0.87) 0.78 (0.73, 0.83) Born in India (Native-Born) 1.00 (0.98, 1.02) 0.99 (0.98, 1.01) 1.00 (0.99, 1.01) Diabetes Mellitus Foreign-born (Native-born) 0.37 (0.33, 0.40) 0.42 (0.38, 0.47) 0.39 (0.37, 0.42) Born in India (Native-Born) 0.97 (0.95, 0.99) 0.95 (0.92, 0.97) 0.96 (0.94, 0.97) The risk of hospitalization for stroke is similar among India-born persons and native-born persons, with a risk ratio of 1 (95% confidence interval 0.99, 1.01). Indian-born persons, however, are more likely than other foreign-born persons to be hospitalized for stroke; the average foreign-born person has a risk ratio of 0.75 (95% confidence interval 0.73, 0.83). While the risk of hospitalization for diabetes similar among Indian-born persons and native-born persons (RR = 0.96, 95% confidence interval 0.94, 0.97), the likelihood of hospitalization for this condition is considerably higher on average than that of other foreign-born persons regardless of sex. Foreign-born persons overall are at 39% the risk of hospitalization for diabetes mellitus (95% confidence interval 0.37, 0.42). Discussion While Indian immigrants to industrialized nations are at similar risk of hospitalization for heart disease, diabetes, and stroke as their native-born counterparts, they are at much greater risk of hospitalization for heart disease, stroke, and diabetes than are immigrants from other countries. Although earlier studies have found differences in the risk of heart disease among Indian-born persons to England and Canada by sex, we found that risks were similar among males and females[ 10 , 11 ]. In an earlier paper, a higher age-adjusted mortality rate due to ischemic heart disease and stroke was noted among foreign-born women relative to native-born women [ 16 ]. In that study, it was postulated that the higher observed mortality rates might be due to changing demographic trends among recent immigrants. Specifically, it was hypothesized that mortality differences may be attributable to a higher proportion of female immigrants from higher risk countries. This hypothesis does not seem correct in light of current findings, and may be attributable to random error in the census sample. Our study was subject to a number of important limitations. While the hospitalization data contained insurance status, the census data did not. We were therefore unable to control for insurance status in our study. Given that foreign-born persons in the U.S. approach three times the rate of lacking health insurance as native-born persons (33.4% versus 12.2% respectively)[ 17 ], it is possible that many persons are at relatively lower risk of hospitalization and relatively higher risk of death. Still, since heart disease is a serious and potentially fatal condition, it seems unlikely that many immigrants would forgo hospital-based care for fear of incurring expenses. Moreover, the abundance of public hospitals in New York City reduces the probability that seriously ill persons would avoid seeking hospital care for fear of incurring medical costs. Nonetheless, some Indian-born persons may avoid seeking care due to a lack of health insurance. To the extent that avoidance of care occurs, the risk of hospitalization among Indian-born persons is underestimated in our analysis. A second limitation was our use of zip-code-only data for the proportion of foreign-born in a neighbourhood, which may be confounded by ecological factors. This was minimized somewhat by examining individual-level hospitalizations for all variables but education and country of birth. It is notable that our data on the foreign-born overall are constant with a growing body of literature showing lower morbidity and mortality among immigrants to the United States than native-born persons[ 1 ]. The risk differences we found between native-born persons and persons born in India were not large – for every 100 hospitalizations for heart disease among native-born persons, we would expect 2–3 additional hospitalizations among persons from India. However, the risk of heart disease, cancer and diabetes is significantly higher than those of other immigrant groups with a similarly low prevalence of smoking[ 1 ]. Previous studies of behavioural risk factors suggest that diet is the overwhelming risk factor in the India-born population – a finding consistent with our subjects' higher risk of hospitalization for diabetes mellitus relative to native-born persons and other foreign-born persons [ 5 , 10 , 11 ]. The presence of a single overriding risk factor simplifies public health efforts; it is easier to target diet and exercise regimens, for instance, than a broader range of health behaviours. Conclusions Indian-born New Yorkers are at substantially higher risk of heart disease, stroke, and diabetes than other foreign-born persons. However, these risks are comparable to that of heart disease, stroke, and diabetes among native-born persons. It is also conceivable that access to health services plays a role in the risk of hospitalisation for heart disease or related conditions given that higher hospitalization rates have been observed in countries with universal health coverage. Prospective data are needed to refine our understanding of the risk factors associated with heart disease in persons from India. As with interpreting data with any epidemiological study, results should not be generalized to clinical practice without first considering the individual patient's socio-demographic risk profile. Nevertheless, clinicians working in facilities serving foreign-born populations may wish to maintain a higher degree of vigilance for heart disease and its risk factors among persons from India. Competing interests The author(s) declare that they have no competing interests. Authors' contributions PM planned the analysis, outlined the study, reviewed the data and regression analyses, and contributed to the development of the manuscript. HJ compiled the data, designed the small area analyses, and conducted the regression analyses. KK contributed to the planning of the analysis and the preparation of the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529471.xml

545597

Complete genome sequence of the industrial bacterium Bacillus licheniformis and comparisons with closely related Bacillus species

The complete sequence of the Bacillus licheniformis ATCC 14580 genome was determined, revealing 4,208 predicted protein-coding genes, 7 rRNA operons and 72 tRNA genes.

Background Bacillus licheniformis is a Gram-positive, spore-forming bacterium widely distributed as a saprophytic organism in the environment. This species is a close relative of Bacillus subtilis , an organism that is second only to Escherichia coli in the level of detail at which it has been studied. Unlike most other bacilli, which are predominantly aerobic, B. licheniformis is a facultative anaerobe, which may allow it to grow in additional ecological niches. Certain B. licheniformis isolates are capable of denitrification; the relevance of this characteristic to environmental denitrification may be small, however, as the species generally persists in soil as endospores [ 1 ]. There are numerous commercial and agricultural uses for B. licheniformis and its extracellular products. The species has been used for decades in the manufacture of industrial enzymes including several proteases, α-amylase, penicillinase, pentosanase, cycloglucosyltransferase, β-mannanase and several pectinolytic enzymes. The proteases from B. licheniformis are used in the detergent industry as well as for dehairing and bating of leather [ 2 , 3 ]. Amylases from B. licheniformis are deployed for the hydrolysis of starch, desizing of textiles and sizing of paper [ 3 ]. Specific B. licheniformis strains are also used to produce peptide antibiotics such as bacitracin and proticin in addition to a number of specialty chemicals such as citric acid, inosine, inosinic acid and poly-γ-glutamic acid [ 4 ]. Some B. licheniformis isolates can mitigate the affects of fungal pathogens on maize, grasses and vegetable crops [ 5 ]. As an endospore-forming bacterium, the ability of the organism to survive under unfavorable environmental conditions may enhance its potential as a natural biocontrol agent. B. licheniformis can be differentiated from other bacilli on the basis of metabolic and physiological tests [ 6 , 7 ]; however, biochemical and phenotypic characteristics may be ambiguous among closely related species. Recent taxonomic studies indicate that B. licheniformis is closely related to B. subtilis and Bacillus amyloliquefaciens on the basis of comparisons of 16S rDNA and 16S-23S internal transcribed spacer (ITS) nucleotide sequences [ 8 ]. Lapidus et al. [ 9 ] recently constructed a physical map of the B. licheniformis chromosome using a PCR approach, and established a number of regions of colinearity where gene content and organization were conserved with the B. subtilis genome. Given that B. licheniformis is an industrial organism used for the manufacture of enzymes, antibiotics, and chemicals, important in nutrient cycling in the environment, and a species that is taxonomically related to B. subtilis , perhaps the best studied of all Gram-positive bacteria, we derived the complete nucleotide sequence of the B. licheniformis type strain (ATCC 14580) genome. With this data in hand, functional and comparative genomics studies can be initiated that may ultimately lead to new strategies for improving industrial strains as well as better understanding of genome evolution among the species within the subtilis-licheniformis group. Results and discussion General features of the B. licheniformis genome The genome of B. licheniformis ATCC 14580 consists of a circular chromosome of 4,222,336 base-pairs (bp) with an average G+C content of 46.2% (Table 1 ). No plasmids were found during the genome analysis, and none were found by agarose gel electrophoresis (data not shown). Using a combination of several gene-finding programs and manual inspection, 4,208 protein-coding sequences (CDSs) were predicted. These CDSs constitute 87% of the genome and have an average length of 873 bp (ranging from 78 to 10,767 bp). They are oriented on the chromosome primarily in the direction of replication (Figure 1 ) with 74.4% of the genes on the leading strand and 25.6% on the lagging strand. Among the 4,208 protein coding genes, 3,948 (94%) had significant similarity to proteins in PIR, 3,187 (76%) of these gene models contain Interpro motifs, and 2,895 (69%) contain protein motifs found in PFAM. The number of hypothetical and conserved hypothetical proteins in the B. licheniformis genome with hits in the PIR database was 1,318 (212 conserved hypothetical proteins). Among the list of hypothetical and conserved hypothetical gene products, 683 (52%) have protein motifs contained in PFAM (148 conserved hypothetical proteins). There are 72 tRNA genes representing all 20 amino acids and seven rRNA operons. The likely origin of replication (Figure 1 ) was identified by similarities to several features of the corresponding regions in B. subtilis and other bacteria. These included co-localization of four genes ( rpmH , dnaA , dnaN , and recF ) near the origin, GC nucleotide skew ((G-C)/(G+C)) analysis, and the presence of multiple dnaA -boxes and AT-rich sequences immediately upstream of the dnaA gene [ 10 - 12 ]. On the basis of these observations we assigned a cytosine residue of the Bst BI restriction site between the rpmH and dnaA genes to be the first nucleotide of the B. licheniformis genome. The replication termination site was localized near 2.02 megabases (Mb) by GC skew analysis. This region lies roughly opposite the origin of replication (Figure 1 ). Unlike B. subtilis , there was no apparent gene encoding a replication terminator protein ( rtp ) in B. licheniformis . The Bacillus halodurans genome also lacks an obvious rtp function [ 13 ]; therefore, it seems likely that B. subtilis acquired the rtp gene following its divergence from B. halodurans and B. licheniformis . Transposable elements, prophages and atypical regions The genome of B. licheniformis ATCC 14580 contains nine identical copies of a 1,285 bp insertion sequence element termed IS3Bli1 [ 9 ]. This sequence shares a number of features with other IS3 family elements [ 9 ] including direct repeats of 3-5 bp, a 10-bp left inverted repeat, and a 9 bp right inverted repeat (Figure 2 ). IS3Bli1 encodes two predicted overlapping CDSs, designated orfA and orfB in relative translational reading frames of 0 and -1. The presence of a 'slippery heptamer' motif, AAAAAAG, before the stop codon in orfA may indicate that programmed translational frameshifting occurs between these two coding sequences, resulting in a single gene product [ 14 ]. The orfB gene product harbors the DD(35)E(7)K motif, a highly conserved pattern among insertion sequences. Eight of these IS3Bli1 elements lie in intergenic regions, and one interrupts the comP gene as noted previously [ 9 ]. In addition to these insertion sequences, the genome encodes a putative transposase that is most closely related (E = 1.8 × 10 -11 ) to one identified in the Thermoanaerobacter tengcongensis genome [ 15 ]; however, similar transposase genes are also found in the chromosomes of B. halodurans [ 13 ], Oceanobacillus iheyensis [ 16 ], Streptococcus agalactiae [ 17 ] and Streptococcus pyogenes [ 18 ]. The presence of several bacteriophage lysogens or prophage-like elements was revealed by Smith-Waterman comparisons to other bacterial genomes and by their AT-rich signatures (Figure 3 , Table 2 ). Prophage sequences, designated NZP1 and NZP3 (similar to B. subtilis prophages PBSX and φ-105), were discovered by noting the presence of nearby genes that code for the large subunit of terminase, a signature protein that is highly conserved among prophages [ 19 ]. Interestingly, a terminase gene was not observed in the third putative prophage, termed NZP2 (similarity to B. subtilis phage SPP1); however, its absence may be the result of genome deterioration during evolution. Interestingly, we observed that regions in which the G+C content is less than 39% usually encoded proteins that have no B. subtilis ortholog and share identity only with hypothetical and conserved hypothetical genes. Two of these AT-rich segments correspond to the NZP2 and NZP3 prophages. An isochore plot (Figure 3 ) also revealed the presence of a region with an atypically high (62%) G+C content. This segment contains two hypothetical genes whose sizes (3,831 and 2,865 bp) greatly exceed the size of an average CDS in B. licheniformis . The first gene encodes a protein of 1,277 amino acids for which Interpro predicts 16 collagen triple-helix repeats, and the amino acid pattern TGATGPT is repeated 75 times within the polypeptide. The second CDS is smaller, and encodes a protein with 11 collagen triple-helix repeats; the same TGATGPT motif recurs 56 times. The primary translation products from these genes do not contain canonical signal peptides for secretion, and they do not contain motifs for the twin-arginine or sortase-mediated translocation pathways. Therefore, it is not likely that they are exported to the cell surface or the extracellular medium. Interestingly, the chromosomal region (19 kb) adjacent to these genes is clearly non-colinear with the B. subtilis genome, and virtually all of the predicted genes encode hypothetical or conserved hypothetical proteins. There are a number of bacterial proteins listed in PIR that also contain collagen triple-helix repeat regions, including two from Mesorhizobium loti (accession numbers NF00607049 and NF00607035) and three from B. cereus (accession numbers NF01692528, NF01269899 and NF01694666). These putative orthologs share 53-76% amino-acid sequence identity with their counterparts in B. licheniformis , and their functions are unknown. Extracellular enzymes and metabolic activities In the Bacillus licheniformis genome, 689 of the 4,208 gene models have signal peptides forecast by SignalP [ 20 ]. Of these, 309 have no transmembrane domain predicted by TMHMM [ 21 ] and 134 are hypothetical or conserved hypothetical genes. Based on a manual examination of the remaining 175 genes, at least 82 are likely to encode secreted proteins and enzymes. Moreover, there are 27 predicted extracellular proteins encoded by the B. licheniformis ATCC 14580 genome that are not found in B. subtilis 168. In accordance with its saprophytic lifestyle, the secretome of B. licheniformis encodes numerous secreted enzymes that hydrolyze polysaccharides, proteins, lipids and other nutrients. Cellulose is the most abundant polysaccharide on Earth, and microorganisms that hydrolyze cellulose contribute to the global carbon cycle. Interestingly, two gene clusters involved in cellulose degradation and utilization were discovered in B. licheniformis , and there are no counterparts in B. subtilis 168. The enzymes encoded by the first gene cluster include two putative endoglucanases belonging to glycoside hydrolase families GH9 and GH5, a probable cellulose-1,4-β-cellobiosidase of family GH48, and a putative β-mannanase of family GH5. The β-mannanase (GH5) and endoglucanase (GH9) both harbor carbohydrate-binding motifs. With the exception of the cellulose-1,4-β-cellobiosidase (GH48), all of the gene products encoded in this cluster have secretory signal peptides, and all have homologs in Bacillus species other than B. subtilis . The overall G+C content of this cluster (48%) does not appear to differ appreciably from that of the genome average (46%). The second gene cluster encodes a putative β-glucosidase (GH1) and three components of a cellobiose-specific PTS transport complex. A second β-glucosidase (GH3) gene is present at an unlinked locus in the genome. Collectively, the genes in these two clusters should enable B. licheniformis to utilize cellulose as a carbon and energy source, converting it to cellobiose and ultimately glucose. In this regard, we have confirmed that B. licheniformis ATCC 14580 is capable of growth on carboxymethyl cellulose as a sole carbon source (not shown). The chromosome of B. licheniformis ATCC 14580 encodes a number of additional carbohydrase activities that may allow the organism to grow on a broad range of polysaccharides. These include xylanase, endo-arabinase and pectate lyase that may be involved in degradation of hemicellulose, α-amylase and α-glucosidase for starch hydrolysis, chitinases for the breakdown of chitooligosaccharides from fungi and insects, and levanase for utilization of β-D-fructans (levans). Several of these activities are marketed as industrial enzymes. Saprophytic organisms must utilize a variety of nitrogenous compounds as nutrients for growth and metabolism. On the basis of the information encoded in its genome, B. licheniformis ATCC 14580 possesses the ability to acquire nitrogen from exogenous proteins, peptides, amino acids, ammonia, nitrate and nitrite. Like B. subtilis , the repertoire of extracellular proteases produced by B. licheniformis includes serine proteases ( aprE , epr , vpr ), metalloprotease ( mpr ), and an assortment of endo- and exopeptidases ( yjbG , ydiC , gcp , ykvY , ampS , bpr (two copies), yfxM , yuiE , yusX , ywaD , pepT ). However, B. licheniformis also has the capacity to produce a number of additional proteases and peptidases that are not encoded in the B. subtilis genome. These include a clostripain-like protease, a zinc-metallopeptidase, a probable glutamyl endopeptidase, an aminopeptidase C homolog, two putative dipeptidases and a zinc-carboxypeptidase. B. licheniformis also has the ability to utilize amino and imino nitrogen from arginine, asparagine and glutamine via arginine deiminase, arginase, asparaginase and glutaminase activities. Interestingly, there appear to be two genes each for arginase, asparaginase and glutaminase. Presumably, the arginine deiminase activity is expressed during anaerobic growth on arginine, whereas the arginase activities are predominant during aerobic growth. The occurrence of putative arginase genes is somewhat of an enigma in B. licheniformis , because there are no genes encoding urease activity for the hydrolysis of urea that is generated by the arginase reaction. In addition to the absence of urease gene homologs ( ureABC ) in B. licheniformis , the glutamine ABC transporters ( glnH , glnM , glnP , glnQ gene products) are also lacking. It appears that nitrogen assimilation and transport pathways may be coordinated similarly in B. licheniformis and B. subtilis owing to the presence of key genes such as glnA , glnR , tnrA and nrgA in both species. Likewise, the pathways for nitrate/nitrite transport and metabolism in B. licheniformis appear to be analogous to the corresponding pathways in B. subtilis as suggested by the presence of nasABC (nitrate transport), narGHIJ (respiratory nitrate reductase), and nasDEF (NADH-dependent nitrite reductase) genes. Unlike B. subtilis , B. licheniformis evidently possesses the capability for anaerobic respiration using nitric oxide reductase. Moreover, the gene encoding this activity lies in a cluster that includes CDSs for narK (nitrite extrusion protein), two putative fnr proteins (transcriptional regulators of anaerobic growth), and a dnrN -like gene product (nitric oxide-dependent regulator). These observations are consistent with previous findings that certain B. licheniformis isolates are capable of denitrification [ 22 ]. While denitrification is a process of major ecological importance, the contribution of B. licheniformis may be small as the species exists predominantly as endospores in soil [ 1 ]. Microbial D-hydantoinase enzymes have been applied to the industrial production of optically pure D-amino acids for synthesis of antibiotics, pesticides, sweeteners and therapeutic amino acids [ 23 ]. This enzyme catalyzes the hydrolysis of cyclic ureides such as dihydropyrimidines and 5-monosubstituted hydantoins to N -carbamoyl amino acids. Hydantoinase activities have been detected in a variety of bacterial genera, and a cluster of six genes in B. licheniformis appears to confer a similar capability. This gene cluster encodes N -methylhydantoinase (ATP-hydrolyzing), hydantoin utilization proteins A and B ( hyuAB homologs), a possible transcriptional regulator (TetR/AcrR family), a putative pyrimidine permease, and a hypothetical protein that contains an Interpro domain (IPR004399) for phosphomethylpyrimidine kinase. Protein export, sporulation and competence pathways Kunst et al . [ 10 ] listed 18 genes that have a major role in the secretion of extracellular enzymes by the classical (Sec) pathway in B. subtilis 168. This list includes several chaperonins, signal peptidases, components of the signal recognition particle and protein translocase complexes. All members of this list have B. licheniformis counterparts. In addition to the Sec pathway, some B. subtilis proteins are directed into the twin-arginine (Tat) export pathway, possibly in a Sec-independent manner. Curiously, the B. licheniformis genome encodes three tat gene orthologs ( tatAY , tatCD , and tatCY ), but two others ( tatAC and tatAD ) are conspicuously absent. Furthermore, specific proteins may be exported to the cell surface via lipoprotein signal peptides or sortase factors. Lipoprotein signal peptides are cleaved with a specific signal peptidase (Lsp) encoded by the lspA gene in B. subtilis . An lspA homolog can be found in B. licheniformis as well, suggesting that this species may possess the ability to export lipoproteins via a similar mechanism. Lastly, surface proteins in Gram-positive bacteria are frequently attached to the cell wall by sortase enzymes, and genome analyses have revealed that more than one sortase is often produced by a given species. In this regard, three possible sortase gene homologs were detected in the genome of B. licheniformis ATCC 14580. Together these observations suggest that the central features of the protein export machinery are principally conserved in B. subtilis and B. licheniformis . From the list of 139 sporulation genes tabulated by Kunst et al . [ 10 ], all but six have obvious counterparts in B. licheniformis . These six exceptions ( spsABCEFG ) comprise an operon involved in synthesis of a spore coat polysaccharide in B. subtilis . In addition, the response regulator gene family ( phrACEFGI ) appears to have a low level of sequence conservation between B. subtilis and B. licheniformis . Natural competence (the ability to take up and process exogenous DNA in specific growth conditions) is a feature of few B. licheniformis strains [ 24 ]. The reasons for variability in competence phenotype have not been explored at the genetic level, but the genome data offer several possible explanations. Although the B. licheniformis genome encodes all of the late competence functions ascribed in B. subtilis (for example, comC , comEFG operons, comK , mecA ), it lacks an obvious comS gene, and the comP gene is punctuated by an insertion sequence element ( IS3Bli1 ), suggesting that the early stages of competence development have been pre-empted in B. licheniformis ATCC 14580. Whether these early functions can be restored by introducing the corresponding genes from B. subtilis is unknown. In addition to an apparent deficiency in DNA uptake, two type I restriction-modification systems were discovered that may also contribute to diminished transformation efficiencies. These are distinct from the ydiOPS genes of B. subtilis , and could participate in degradation of improperly modified DNA from heterologous hosts used during construction of recombinant expression vectors. Each of these loci in B. licheniformis (designated as BliI and BliII ) encode putative HsdS, HsdM and HsdR subunits that share significant amino-acid sequence identity to type I restriction-modification proteins in other bacteria. Curiously, the G+C-content for these loci (37%) is substantially lower than the overall genome average (46%) which may hint that they are the result of gene acquisitions. Lastly, the synthesis of a glutamyl polypeptide capsule has also been implicated as a potential barrier to transformation of B. licheniformis strains [ 25 ]. While laboratory strains of B. subtilis usually do not produce significant capsular material, the genome sequence of B. subtilis 168 indicates that they may harbor the genes required for synthesis of polyglutamic acid. In contrast, many B. licheniformis isolates produce copious amounts of capsular material, giving rise to colonies with a wet or slimy appearance. Six genes were predicted ( ywtABDEF and ywsC orthologs) that may be involved in the synthesis of polyglutamic acid capsular material in B. licheniformis . Antibiotics, secondary metabolites and siderophores Bacitracin is a cyclic peptide antibiotic that is synthesized non-ribosomally by some B. licheniformis isolates [ 26 ]. While there is variation in the prevalence of bacitracin synthase genes among laboratory strains of this species, one study suggested that up to 50% may harbor the bac operon [ 27 ]. Interestingly, the bac operon is not present in the type strain (ATCC 14580) genome. Seemingly, the only non-ribosomal peptide synthase operon encoded by the B. licheniformis type strain genome is that responsible for lichenysin biosynthesis. Lichenysin structurally resembles surfactin from B. subtilis [ 28 ], and their respective biosynthetic operons are highly similar. Surprisingly, we found no B. licheniformis counterparts for the pps (plipastatin synthase) and polyketide synthase ( pks ) operons of B. subtilis . Collectively, these two regions represent sizeable portions (80 kb and 38 kb, respectively) of the chromosome in B. subtilis , although they are reportedly dispensable [ 29 ]. Unexpectedly, a cluster of 11 genes was found encoding a lantibiotic, with its associated modification and transport functions. We designated this peptide of 75 amino acids as lichenicidin, and its closest homolog is mersacidin from Bacillus sp. strain HIL-Y85/54728 [ 30 ]. Lantibiotics are ribosomally synthesized peptides that are modified post-translationally so that the final molecules contain rare thioether amino acids such as lanthionine and/or methyl-lanthionine [ 31 ]. Like mersacidin, lichenicidin appears to be a type B lantibiotic, comprising a rigid globular peptide with no net charge (7 acidic residues, 7 basic residues) and a leader peptide with a conserved double glycine cleavage site (GG-type leader peptide). These antimicrobial compounds have attracted much attention in recent years as models for the design and genetic engineering of improved antimicrobial agents [ 32 ]. However, since several post-translational modifications and product-specific export functions are required, a dedicated expression system is a prerequisite to provide all the factors necessary to synthesize, modify and transport the lantibiotic peptide. With its history of use in industrial microbiology, B. licheniformis may be an attractive candidate for the development of such an expression system. Like B. subtilis 168, the B. licheniformis ATCC 14580 chromosome harbors a siderophore biosynthesis gene cluster ( dhbABCEF ), and the organization of the cluster is similar to the corresponding chromosomal segment in B. subtilis . In addition, the B. licheniformis genome contains a second gene cluster of four genes ( iucABCD ) that show significant similarity to proteins involved in aerobactin biosynthesis in E. coli . Surprisingly, a gene encoding the receptor protein ( iutA homolog) was not found in B. licheniformis . The B. halodurans genome also contains genes that are homologous to iucABCD , but like B. licheniformis , no iutA homolog could be found using BLAST or Smith-Waterman searches. Comparison of the B. licheniformis genome with those of other bacilli The B. licheniformis ATCC 14580 gene models were compared to the list of essential genes in B. subtilis [ 33 ]. Predictably, all of the essential genes in B. subtilis have orthologs in B. licheniformis , and most are present in a wide range of bacterial taxa. In pairwise BLAST comparisons, 66% of the predicted B. licheniformis genes have orthologs in B. subtilis , and 55% of the gene models are represented by orthologous sequences in B. halodurans (E-value threshold of 1 × 10 -5 ; Figure 4 ). Using a reciprocal BLASTP analysis we found 1,719 orthologs that are common to all three species (E-value threshold of 1 × 10 -5 ). As noted by Lapidus et al. [ 9 ], there are broad regions of colinearity between the genomes of B. licheniformis and B. subtilis (Figure 5 ). Less conservation of genome organization exists between B. licheniformis and B. halodurans , and substantial genomic segments have been inverted in B. halodurans with respect to B. licheniformis and B. subtilis . These observations clearly support previous hypotheses [ 8 ] that B. subtilis and B. licheniformis are phylogenetically and evolutionarily closer to each other than to B. halodurans . Conclusions In comparisons of shared regions, the genomes of B. licheniformis ATCC 14580 and B. subtilis 168 are approximately 84.6% identical at the nucleotide level and show extensive organizational similarity. Accordingly, their genome sequences represent potentially useful instruments for comparative and evolutionary studies among species within the subtilis-licheniformis group, and they may offer new information regarding the evolution and ecology of these closely related species. Despite the broad colinearity of B. licheniformis and B. subtilis genomes, there are local regions that are individually unique. These include chromosome segments that comprise prophage and insertion sequence elements, DNA restriction-modification systems, antibiotic synthases, and a number of extracellular enzymes and metabolic activities that are not present in B. subtilis . It is tempting to speculate that the presence of these genes forecasts the ability of B. licheniformis to grow on an expanded array of substrates and/or in additional ecological niches compared to B. subtilis . Together, the similarities and differences may hint at overlapping but non-identical environmental niches for these taxa. The subtilis-licheniformis group of bacilli includes many strains that are used to manufacture industrial enzymes, antibiotics and biochemicals. The availability of a complete genome from B. licheniformis should permit a thorough comparison of the biochemical pathways and regulatory networks in B. subtilis and B. licheniformis , thereby offering new opportunities and strategies for improvement of industrial strains. When considering the safety of B. licheniformis as an industrial organism it should be noted that the species is considered neither a human pathogen nor a toxigenic microorganism [ 34 ]; however, there are reports in the literature implicating it as a causal agent of food poisoning. In these isolated cases, specific strains were shown to produce a toxin similar to cereulide, the emetic toxin of B. cereus [ 35 ]. Cereulide is a cyclic depsipeptide synthesized non-ribosomally [ 36 ]. Importantly, the only non-ribosomal peptide synthase genes found in the B. licheniformis ATCC 14580 genome are those that involved in synthesis of lichenysin. Similarly, we detected no homologs of the B. cereus hemolytic and non-hemolytic enterotoxins (Swiss-Prot accession numbers P80567, P80568, P80172, and P81242). In a comparison of the genotypic and phenotypic characteristics among 182 soil isolates of B. licheniformis , Manachini et al. [ 37 ] observed that while this bacterial species appears to be phenotypically homogeneous, clear genotypic differences are evident between isolates. They postulated the existence of three genomovars for B. licheniformis . Similarly, De Clerck and De Vos [ 38 ] proposed that this species consists of two lineages that can be distinguished using several molecular genotyping methods. The genome sequence data presented in this work should provide a solid foundation on which to conduct future studies to elucidate the genotypic variation among B. licheniformis isolates. Materials and methods Shotgun DNA sequencing and genome assembly The genome of B. licheniformis ATCC 14580 was sequenced by a combination of the whole-genome shotgun method [ 39 ] and fosmid end sequencing [ 40 ]. Plasmid libraries were constructed using randomly sheared and Mbo I-digested genomic DNA that was enriched for fragments of 2-3 kb by preparative agarose gel electrophoresis. Approximately 49,000 random clones were sequenced using dye-terminator chemistry (Applied Biosystems) with ABI 377 and ABI 3700 automated sequencers yielding approximately 6× coverage of the genome. A combination of methods was used for gap closure, including sequencing on fosmids [ 40 ] and primer-walking on selected clones and PCR-amplified DNA fragments. We also incorporated data from both ends of approximately 1,975 fosmid clones with an average insert size of 40 kb to aid in validating the final assembly. In total, the number of input reads was 62,685, with 78.6% of these incorporated into the assembled genome sequence. Individual nucleotides were called using TraceTuner 2.0 (Paracel), and sequence reads were assembled into contigs using the Paracel Genome Assembler using optimized parameters and the quality score set to >20. Phrap, Crossmatch and Consed were used for sequence finishing [ 41 ]. Prediction and annotation of CDSs Protein-coding regions in the assembled genome sequence were identified using a combination of previously described software tools including EasyGene [ 42 ], Glimmer [ 43 ] and FrameD [ 44 ]. EasyGene was used as the primary gene finder in these studies. It searches for protein matches in the raw genome data to derive a good training set, and an HMM with states for coding regions as well as ribosome-binding sites (RBSs) is estimated from the dataset. This HMM is used to score all the predicted CDSs in the genome, and the score is converted to a measure of significance (R-value) which is the expected number of CDSs that would be predicted in 1 Mb of random DNA. Gene models with R-values lower than 10 and a log-odds score of greater than -10 were included/considered significant. The principal advantage of this significance measure is that it properly takes into account the length distribution of random CDSs. EasyGene has been shown to match or exceed other gene finders currently available [ 42 ]. Glimmer was used as a secondary gene finder to aid in identification of small genes (< 100 bp) that were sometimes missed by EasyGene. Glimmer models were post-processed with RBSFINDER [ 45 ] to pinpoint the positions of start codons by searching for consensus Shine-Dalgarno sequences. According to the RBS states in the EasyGene HMM model, the bases with the highest probability were AA AAGGAG (the bases in bold type had distinctly higher probabilities compared to the initial AA). This motif concurs with the consensus Shine-Dalgarno sequence for B. subtilis (AAAGGAGG) [ 46 ]. RBSFINDER identified the core AAGGAG motif in around 80% of the cases for Glimmer gene predictions and adjusted the start codon accordingly. Manual inspection and alignments to B. subtilis homologs were also used to determine the incidence of specific genes. During the gene-finding process, possible errors and frameshifts were detected by both visual inspection of the CDSs to look for interrupted or truncated genes and by deploying FrameD software [ 44 ]. Frameshifts were resolved by re-sequencing of PCR-amplified segments or subclones. After re-sequencing and manual editing a total of 27 frameshifts remain in the genome assembly (excluding those contained in the IS3Bli1 elements). It is not known at present whether these represent pseudogenes or instances of programmed translational frameshifting. The positions of rRNA operons in the genome assembly were confirmed by long-range PCR amplification using primers that annealed to genes flanking the rRNA genes. These PCR fragments were sequenced to high redundancy and the consensus sequences were manually inserted into the assembly. Among the seven rRNA operons, the nucleotide sequences of 16S and 23S genes are at least 99% identical, differing by only one to three nucleotides in pairwise comparisons. Protein-coding sequences were annotated in an automated fashion with the following software applications. Predicted proteins were compared to the nonredundant database PIR-NREF [ 47 ] and the B. subtilis genome [ 48 ] using BLASTP with a E-value threshold of 1 × 10 -5 . InterProScan was used to predict putative function [ 49 ]. The InterPro analysis included comparison to PFAM [ 50 ], TIGRFAM [ 51 ], Interpro [ 52 ] signal peptide prediction using SignalP [ 20 ] and transmembrane domain prediction using TMHMM [ 21 ]. These CDSs were assigned to functional categories based on the Cluster of Orthologous Groups (COG) database [ 53 ] with manual verification as described [ 54 , 55 ]. Phage gene boundaries were predicted using gene finding algorithms and by homology to known bacteriophage genes. Transfer RNA genes were identified using tRNAscan-SE [ 56 ]. B. licheniformis genes that shared significant homology with B. subtilis counterparts were named using the nomenclature in the SubtiList database [ 48 ] with updated gene names from the BSORF [ 57 ] and UniProt [ 58 ] databases. Comparative analyses VisualGenome software (Rational Genomics) was used for comparisons of ortholog distribution among B. licheniformis , B. subtilis and B. halodurans genomes with precomputed BLAST results stored in a local database. Accession of genome sequence information The GenBank accession number for the B. licheniformis ATCC 14580 genome is CP000002. An interactive web portal for viewing and searching the assembled genome based on the generic genome browser developed by Stein et al. [ 59 ] is available at [ 60 ].

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545597.xml

516440

Flagellin acting via TLR5 is the major activator of key signaling pathways leading to NF-κB and proinflammatory gene program activation in intestinal epithelial cells

Background Infection of intestinal epithelial cells by pathogenic Salmonella leads to activation of signaling cascades that ultimately initiate the proinflammatory gene program. The transcription factor NF-κB is a key regulator/activator of this gene program and is potently activated. We explored the mechanism by which Salmonella activates NF-κB during infection of cultured intestinal epithelial cells and found that flagellin produced by the bacteria and contained on them leads to NF-κB activation in all the cells; invasion of cells by the bacteria is not required to activate NF-κB. Results Purified flagellin activated the mitogen activated protein kinase (MAPK), stress-activated protein kinase (SAPK) and Ikappa B kinase (IKK) signaling pathways that lead to expression of the proinflammatory gene program in a temporal fashion nearly identical to that of infection of intestinal epithelial cells by Salmonella . Flagellin expression was required for Salmonella invasion of host cells and it activated NF-κB via toll-like receptor 5 (TLR5). Surprisingly, a number of cell lines found to be unresponsive to flagellin express TLR5 and expression of exogenous TLR5 in these cells induces NF-κB activity in response to flagellin challenge although not robustly. Conversely, overexpression of dominant-negative TLR5 alleles only partially blocks NF-κB activation by flagellin. These observations are consistent with the possibility of either a very stable TLR5 signaling complex, the existence of a low abundance flagellin co-receptor or required adapter, or both. Conclusion These collective results provide the evidence that flagellin acts as the main determinant of Salmonella mediated NF-κB and proinflammatory signaling and gene activation by this flagellated pathogen. In addition, expression of the fli C gene appears to play an important role in the proper functioning of the TTSS since mutants that fail to express fli C are defective in expressing a subset of Sip proteins and fail to invade host cells. Flagellin added in trans cannot restore the ability of the fli C mutant bacteria to invade intestinal epithelial cells. Lastly, TLR5 expression in weak and non-responding cells indicates that additional factors may be required for efficient signal propagation in response to flagellin recognition.

Background Intestinal epithelial cells serve as a barrier between the luminal microflora and the body and as such are perfectly positioned to monitor the approach/invasion of pathogens. These intestinal epithelial cells (IECs) serve as innate immune sentinels and monitor their environment and constantly give out innate host defense instruction to local immune effector cells [ 1 , 2 ]. Pathogens such as Salmonella and other enteroinvasive pathogenic bacteria such as enteroinvasive E. Coli , Shigella and Yersinia upon infection of IECs leads to the up-regulation of the expression of host genes, the products of which activate mucosal inflammatory and immune responses and alter epithelial cell functions [ 3 - 6 ]. Previously we and others demonstrated that IKK via NF-κB and the SAPK signaling pathways via Jun N-terminal kinase (JNK) and p38 kinase were key regulators of the up-regulation of the proinflammatory gene program [ 3 , 7 - 9 ], with NF-κB appearing to be the most critical [ 3 ]. Typically Salmonella infects thirty-forty percent of IECs in culture models of infection [ 10 ], however, we and others have found that Salmonella infection activates NF-κB DNA binding activity to levels equivalent to that of TNFα which activates NF-κB in all of the cells [ 3 ]. Previous studies examining NF-κB activation by Salmonella in HT29 colonic intestinal epithelial cells, which serve as model colonic epithelial cells in culture, indicated that delivery of Salmonella proteins into the host cell via its type III secretion system (TTSS), such as SopE and SopE2, the bacterially encoded exchange factors for the Rho-family members Rac1 and CdC42, result in exchange factor activation, cytoskeletal rearrangements and activation of the MAPK, SAPK and NF-κB signaling pathways [ 7 , 8 , 11 - 15 ]. Recent observations that utilized Salmonella strains that were defective in invasion and delivery of invasion proteins by the TTSS but not attachment indicated that additional factors other than those delivered by the TTSS could lead to NF-κB activation [ 16 ]. Presently it is not clear what protein(s) dictate the activation of key signaling pathways that lead to the temporal expression of the proinflammatory gene program, although the SopE proteins have been given extreme attention recently [ 7 , 8 , 15 ]. In searching for additional Salmonella proteins that could activate the proinflammatory gene expression program, bacterial flagellin was recently found to be such a protein [ 16 - 19 ] and had been shown previously to activate IL-8 expression in monocytes [ 19 - 21 ]. Flagellin was found to activate NF-κB in polarized epithelial cells only when flagellin was present on their basolateral surface [ 22 ] consistent with the idea that a cell surface receptor was present there and could recognize it. The toll-like receptors (TLRs) have been found to recognize pathogen associated molecular patterns (PAMPs) reviewed in [ 23 - 26 ]. TLR2 interacts with TLR1 and TLR6 to recognize bacterial lipopeptides and zymosan respectively [ 27 , 28 ]. TLR4 recognizes LPS only when associated with its co-receptor MD2 and CD14 [ 29 - 32 ]. Recently, flagellin was demonstrated to be recognized by TLR5 and activate an innate host response [ 22 , 33 ]. However, little was known or demonstrated about the endogenous levels of TLR5 in cells used in those studies and why those cells failed to respond to flagellin. Here we have identified flagellin as the primary initiator and temporal regulator of not only the major signaling pathways activated during Salmonella infection but also of key target genes of the proinflammatory gene program too. We have also found flagellin expression to be required for Salmonella bacterial invasion. Independently we found TLR5 recognizes flagellin but its signaling activity toward this PAMP is consistent with either the aid of another flagellin-recognizing co-receptor (as TLR4 utilizes for LPS) or the use of another adapter protein, perhaps similar to MyD88, that is absent or present at low levels in flagellin non-or low-responding cells. Results Salmonella infection leads to a minority of cells invaded but activates NF-κB in nearly all cells Previously, we have noted that pathogenic Salmonella sp . infection leads to potent IKK and NF-κB activation and activation of the proinflammatory gene program [ 3 ]. Previous studies suggest that about thirty-forty percent of the intestinal epithelial cells are infected during a typical Salmonella infection in cultured intestinal epithelial cells [ 10 ]. We wished to address the question of how bacterial infection of about thirty percent of the host cells could give rise to NF-κB DNA binding activity equivalent to activation of NF-κB in nearly all of the host cells as TNFα treatment of the cells does. To examine this phenomenon in detail HT29 cells either mock-infected or infected at a MOI of fifty for one-hour with wild-type S. typhimurium that had been transformed with the plasmid pFM10.1, that encodes green fluorescent protein (GFP) under the control of the Salmonella ssaH promoter and only functions once the bacteria has invaded the host cell [ 34 ]. As can be seen in Fig. 1A , GFP expression occurs in about thirty to forty percent of the cells. We next examined the localization of the NF-κB subunit p65 (RelA) in non-treated (mock-infected), Salmonella infected or TNFα (10 ng/ml) stimulated cells and found that p65 (RelA) was localized to the cytoplasm in non-treated cells whereas, in Salmonella infected cells or in TNFα treated cells p65 (RelA) had localized to the nucleus (Fig. 1B ). These results demonstrate that Salmonella infection activates NF-κB in virtually all of the cells even though only a minority of them become infected and is consistent with and aids in explanation of our previous results examining Salmonella infection and NF-κB activation [ 3 ]. Figure 1 Salmonella infection leads to NF-κB nuclear localization even in non-infected cells. HT29 cells were grown on glass coverslips and either mock-infected, left untreated, infected with Salmonella typhimurium , or treated with TNFα (10 ng/ml). Cells fixed after 30 min (TNF) and 1 h ( Salmonella ) as described in Experimental Procedures and Salmonella that had invaded HT29 cells were detected by direct fluorescence microscopy of GFP expression, p65(RelA) localization was monitored by indirect immunoflourescence of rabbit anti-p65 antibody detected with FITC-conjugated donkey anti-rabbit antibody. DAPI was used to stain nuclei. A, HT29 cells were mock-infected or infected at an MOI of 50 with Salmonella typhimurium strain SJW1103G which expresses GFP from the ssaH promoter that is only active inside infected host cells [10,34]. Cells were photographed using bright field microscopy (BF), and immunoflourescence to detect GFP or DAPI staining as indicated. Images were merged (overlay) to reveal cells that were infected. B, HT29 cells were left untreated, infected with Salmonella typhimurium strain 1103 or treated with TNFα. NF-κB p65(RelA) localization under various conditions as indicated was monitored by indirect immunofluorescence. Cells were visualized by bright field microscopy (BF), cell nuclei were stained with DAPI and p65(RelA) was visualized with FITC. DAPI staining was falsely colored red to make visualization of the merge (overlay) easier to distinguish. Soluble bacterial product identified as flagellin can activate NF-κB in intestinal epithelial cells Since Salmonella sp . infection of intestinal epithelial cells in culture led to only roughly thirty percent infection but activation of NF-κB in nearly all of the cells, we anticipated that NF-κB activation was in response to host cell recognition of bacteria structural components or products produced by the bacteria and not by the invasion process. Invasion itself has been demonstrated not to be required for activation of the proinflamatory gene program as had previously been thought [ 16 ]. To investigate this possibility sterile-filtered S. dublin culture broth left either untreated or boiled for twenty minutes was used to challenge HT29 intestinal epithelial cells and NF-κB DNA binding activity was monitored by electromobility shift assays (EMSAs) of whole cell extracts (WCE) prepared forty-five minutes after exposure [ 3 , 35 ]. Potent activation of NF-κB in response to the broth under both conditions was observed indicating the activating factor was heat-stable (AD, TT and JD, personal observations) and is not LPS since HT29 cells are not responsive to LPS [ 3 , 35 ]. The native sterile-filtered concentrated broth was subsequently treated with DNase, RNase, proteinase K or crudely size fractionated on 100 kDa centricon filters. The variously treated broths were then used to challenge HT29 intestinal epithelial cells and WCEs were prepared after forty-five minutes and NF-κB DNA binding activity was analyzed by EMSA (Fig 2A ). Direct infection of HT29 cells by S. typhimurium 1103 or exposure to the culture broths (supt), as indicated, induced NF-κB DNA binding activity, while the activity-inducing factor was found to be sensitive to protease digestion and was retained by a 100 kDa filter (Fig. 2A ). To further determine the identity of the NF-κB inducing activity, sterile-filtered concentrated broth culture was fractionated by Superose 12 gel permeation chromatography (Fig. 2B ) and by anion exchange chromatography (Fig. 2C ). Aliquots of chromatography fractions were assayed for their ability to activate NF-κB in HT29 cells and analyzed by EMSA. As can be seen from the Coomassie blue stained gel (Fig. 2B , top panel) increased NF-κB DNA binding activity (Fig. 2B , lower panel lanes 4–6) corresponded to the increased abundance of an approximately 55 kDa protein. Anion exchange chromatography on POROS HQ matrix and elution of bound proteins with an increasing salt gradient as indicated (Fig. 2C ) demonstrated that NF-κB DNA binding-inducing activity corresponded to chromatographic fractions containing an increased abundance of the 55 kDa protein (Fig. 2C top panel, and data not shown). Eluted fractions observed in Fig. 2C were concentrated and fractionated on preparative 12% SDS-PAGE gels and bands corresponding to B1-B6 were cut from the gels and the proteins eluted, precipitated and renatured as described in Experimental Procedures and used to stimulate HT29 cells. Whole cell extracts from these cells were assayed for NF-κB DNA binding-inducing activity by EMSA and only band 2 (B2) corresponding to the 55 kDa protein (Fig. 2C lower panel) was able to elicit NF-κB DNA binding activity while buffer from the beginning or end of the salt gradient failed to activate NF-κB DNA binding activity. Figure 2 Protein factor in Salmonella culture broth leads to NF-κB activation. A, Salmonella dublin culture broth concentrated 100-fold was treated as indicated or infectious bacteria, as indicated was used to challenge HT29 cells. NF-κB DNA binding activity was assayed by EMSA from whole cell extracts prepared 45 min after treatment. Authenticity of the NF-κB DNA:protein complex was determined using p65(RelA)-specific and p50-specific antibody supershifts. B, Concentrated Salmonella dublin culture broth (IN) was chromatographed by gel permeation on a Superose 12 column. Eluted protein fractions were analyzed by fractionation on 10% SDS-PAGE and visualized by Coomassie blue (CB) staining. Molecular weight markers for chromatography and on the gels are indicated. Aliquots of each fraction as indicated was used to stimulate HT29 cells and resultant WCEs were analyzed by EMSA for NF-κB DNA binding activity. C, Concentrated Salmonella dublin culture broth (IN) was chromatographed by anion exchange chromatography on POROS HQ matrix. Proteins were eluted with an increasing NaCl gradient as indicated and analyzed on 10% SDS-PAGE and visualized by Coomassie blue (CB) staining. Input and aliquots of each fraction as indicated was used to stimulate HT29 cells and resultant WCEs were analyzed by EMSA for NF-κB DNA binding activity. Eluted material corresponding to protein bands B1-B6, a blank portion of the gel was isolated from a duplicate 10% SDS-PAGE gel as described in Experimental Procedures along with buffer samples from the beginning and end NaCl buffer gradient and used to stimulate HT29 cells and resultant WCEs were analyzed by EMSA for NF-κB DNA binding activity. Proteins corresponding to protein bands B1-B6 and blank areas of the gel were further processed for peptide sequencing as described in Experimental Procedures. Trypsin digestion of the protein corresponding to B2 and analysis by electrospray ion trap LC/MS identified the amino acid sequence of twenty-one peptides. Flagellin (seventy-five percent coverage by the twenty-one peptides) was unambiguously identified as the protein consistent with inducing NF-κB DNA binding activity (Fig 3 ). Figure 3 Identifcation by mass spectrometry of flagellin as the NF-κB activating factor in Salmonella culture broth. Microcapillary HPLC tandem mass spectrometry of Band 2 digested by trypsin. Peaks corresponding to Salmonella peptides are numbered and identified with the corresponding numbered peptide sequence to the right. Flagellin is required to activate NF-κB in intestinal epithelial cells To determine if flagellin was indeed the factor that was responsible for triggering activation of NF-κB after exposure of intestinal epithelial cells to direct bacterial infection or to filtered culture broths of pathogenic Salmonella sp . we prepared infectious bacteria and boiled and filtered culture broths from the non-flagellated E. Coli DH5α, pathogenic S. dublin strain 2229, an isogenic S. dublin 2229 SopE - mutant, isogenic S. dublin 2229 SopB - mutant, isogenic S. dublin 2229 double SopE - /SopB - mutant (strain SE1SB2), S. typhimurium strain 1103, and isogenic S. typhimurium fliC :: Tn 10 insertion mutant (strain 86) and a S. typhimurium 1103 isogenic double mutant fliC - /fljB - and were used to challenge HT29 cells. Bacteria and culture broths were used to challenge HT29 intestinal epithelial cells and WCE extracts were prepared after forty-five minutes and analyzed for NF-κB DNA binding activity by EMSA. Salmonella strains could activate NF-κB, while Salmonella strains failing to produce flagellin (fliC and fliC - /fljB - mutants as indicated) also failed to activate NF-κB (Fig. 4A & 4B ). E. Coli DH5α is non-flagellated and does not produce flagellin failed to activate NF-κB. We also noticed through numerous experiments that S. dublin direct infections always activated NF-κB to a greater extent than S. typhimurium as observed in Fig. 4A while culture broths from both species activated NF-κB almost equally well (Fig. 4B ). We believe this difference is due perhaps to S. dublin releasing more flagellin into the cell culture media than S. typhimurium during infection since purification of flagellin from both S. dublin and S. typhimurium and addition of equivalent amounts of chromatographically purified flagellin gave similar NF-κB activation profiles (TT & JD, unpublished observations). Figure 4 Flagellin mutants fail to activate NF-κB. EMSAs assaying for NF-κB DNA binding activity in WCEs prepared 45 min from non-infected cells (UN) and after direct infection of HT29 cells with wild-type E. coli DH5α, wild-type Salmonella dublin or SopE - mutant, SopB - mutant, the SopE - /SopB - double mutant, wild-type Salmonella typhimurium strain 1103, the fliC - mutant (fliC::Tn 10 ), the fliC - /fljB - double mutant as indicated at an MOI of 50. B, EMSAs assaying for NF-κB DNA binding activity in WCEs prepared 45 min after challenge of HT29 cells from non-infected cells (UN) or with sterile-filtered concentrated culture broths from wild-type and mutant bacteria as indicated. Of note is the total failure of the double flagellin gene mutants to activate NF-κB as compared to the very minor activation observed in the single Phase I flagellin fliC::Tn 10 insertion mutant (next to last lanes in Fig. 4A & 4B ) which likely is due to the extremely limited expression of the phase II flagellin (from fljB), although the strains of Salmonella used here genetically are unable or rarely shift phases of flagellin production. These results are consistent with previous reports identifying flagellin as a potent inducer of the proinflammatory response and IL-8 production [ 16 - 19 ]. Since flagellin appears required for activation of the NF-κB pathway upon direct infection of intestinal epithelial cells it appeared possible that flagellin may also be the major determinant of other major mitogenic and stress activated signaling pathways activated upon pathogenic Salmonella infection of intestinal epithelial cells. Previously others and we have demonstrated that direct Salmonella infection of intestinal epithelial cells results in JNK activation [ 8 ] and also the activation of NF-κB via IKK [ 3 ]. The identification of flagellin as a potent NF-κB activator is significant since SopE had previously been shown to be a pathogenic Salmonella bacteriophage encoded protein that is injected into the host cell and acts as an exchange factor for the small Rho GTPases Rac1 and CdC42 initiating cytoskeleton rearrangements and eventual activation of the MAPK, SAPK and NF-κB pathways [ 7 , 15 ], while SopB is a Salmonella protein that functions as an inositol phosphate phosphatase and participates in cytoskeletal rearrangements and stimulates host cell chloride secretion [ 36 ]. Flagellin triggers activation of the mitogen activated protein kinase, stress activated protein kinase and IKK signaling pathways Intestinal epithelial cells act as sentinels for invasion of luminal surfaces and orchestrate the attraction of effector immune cells to the area by production of chemokine genes like IL-8 and macrophage chemoattractant protein 1 (MCP1) proinflammatory cytokine genes such as TNFα, IL-1 and IL-6 [ 1 , 4 - 6 ]. Expression of these genes primarily depends upon the action of transcription factors that are activated in response to the transmission of signals via the MAPK, SAPK and IKK signaling pathways. Since NF-κB is considered a central regulator/activator of the proinflammatory gene program we decided to examine the effect that non-flagellin producing mutant strains of Salmonella had on activation of the MAPK, SAPK and IKK signaling pathways compared to infection of intestinal epithelial cells with wild-type Salmonella or by exposure of the intestinal epithelial cells to purified flagellin. Infection of HT29 cells with wild-type S. typhimurium resulted in activation of MAPKs ERK1&2, the SAPKs p38, JNK and IKK (Fig. 5 ) as determined by use of activation-indicating phospho-specific antibodies in immunoblot (IB) analysis or antibody-specific immuno-kinase assays (KA) for JNK and IKK using their respective substrates GST-cJun 1–79 and GST-IκBα1–54 [ 37 - 39 ]. Interestingly, MAPK stimulation is transient in nature as activation declines beginning at forty-five minutes while p38, JNK and IKK activity increases with time through one hour. As seen in Fig. 4 , the fliC - /fljB - double mutant Salmonella also failed to induce IKK and NF-κB activity (Fig. 5 as indicated). Surprisingly, the fliC - /fljB - double mutant Salmonella failed to induce the SAPKs p38 and JNK and only briefly (fifteen minutes) activated MAPK. This result is puzzling since other Salmonella proteins such as SopE and SopE2 can activate the small GTPases Rac and CdC42, and these Rho family GTPases have been linked to JNK and p38 activation [ 7 , 8 , 14 , 15 ] yet appear not to function in the flagellin minus strain. Figure 5 Flagellin is required for activating multiple signaling pathways during Salmonella infection and leads to nuclear localization of NF-κB. HT29 cells were left untreated, stimulated with TNFα (10 ng/ml) or a cocktail of anisomycin [An] (20 μg/ml)/PMA (12.5 ng/ml) for 15 min, or infected with either wild-type (WT) Salmonella typhimurium strain 1103 or the Salmonella typhimurium double fliC - /fljB - mutant strain 134 as indicated. WCE were prepared at the indicated times or at 10 min for TNF-treated cells or 15 min for anisomycin/PMA treated cells and used in EMSAs to analyze NF-κB DNA binding activity, or in immuno-kinase assays (KA) using anti-IKK or anti-JNK antibodies to measure IKK and JNK kinase activity on their respective substrates GST-IκBα 1–54 and GST-cJun 1–79 (as indicated). Immunoblot (IB) analysis of equivalent amounts (40 μg) of protein from each extract was fractionated on SDS-PAGE gels and transferred to PVDF membranes and probed with the indicated antibodies to detect bulk IKK, JNK, ERK and p38 as indicated. Immunoblot analysis using phospho-specific antibodies for ERK and p38 to detect activated ERK and p38 are indicated. B, Immunofluorescence demonstrating that flagellin mutant Salmonella fail to infect HT29 cells and that purified flagellin stimulation of HT29 cells leads to NF-κB nuclear p65 (RelA) localization as determined by indirect immunofluorescence. Imaging of the treatment indicated HT29 cells grown on coverslips was essentially the same as in Fig. 1A & 1B. False coloring of the DAPI stain was used to enhance the visualization of both DAPI stained nuclei and p65 nuclear localization. The fliC - /fljB - double mutant Salmonella failed to invade HT29 cells compared to the wild-type Salmonella strain as determined by gentamycin protection/invasion assay (see Experimental Procedures). The flagellin fliC - /fljB - double mutant displayed a four orders of magnitude difference in its ability to invade HT29 cells (TT & JD, unpublished observations). To demonstrate this point further, we infected HT29 cells with either wild-type Salmonella or the fliC - /fljB - double mutant Salmonella (strain 134), both strains were transformed with the plasmid pFM10.1 that encodes GFP under the control of the Salmonella ssaH promoter and only functions once the bacteria has invaded the host cell [ 10 , 34 ]. The wild-type Salmonella clearly was able to infect HT29 cells (GFP, Fig. 5B ) while the flagellin mutant bacteria failed to invade HT29 cells as evidenced by the lack of GFP expression (Fig. 5B ). To determine if flagellin is sufficient or that other bacterially produced proteins are required for invasion, we added either purified flagellin or sterile-filtered culture broths or a combination of both to HT29 cells that were challenged with the Salmonella fliC - /fljB - double mutant and assayed for invasion. Intestinal epithelial cells failed to be invaded using all tested combinations of purified flagellin and/or culture broths with the fliC - /fljB - double mutant strain (TT & JD, unpublished observations). To our knowledge there is no known direct connection between expression of flagellin genes and the effectiveness of the type III secretion system to deliver bacterially produced proteins such as SopE, SopE2 and SipA or other Sip or Sop proteins [ 7 , 14 , 15 , 40 , 41 ] that play important roles in initiating bacterial internalization. Furthermore, to evaluate the effectiveness of flagellin to stimulate p65 (RelA) nuclear localization in intestinal epithelial cells we challenged HT29 cells with purified flagellin and examined p65 (RelA) localization using indirect immunofluorescence and found p65 (RelA) nuclear localization in nearly every cell (Fig. 5B as indicated). Purified flagellin (0.5 μg/ml) was added to the culture media of HT29 cells and WCE were prepared at various times as indicated after exposure and assayed for NF-κB DNA binding activity in EMSAs (Fig. 6A ). Flagellin potently activated NF-κB in a time dependent manner similar to that observed for TNF (10 ng/ml) treatment of HT29 cells (Fig. 6A ). Analysis of the MAPK, SAPK and IKK signaling pathways (Fig. 6B ) at various times after flagellin treatment of HT29 cells using activation-specific phospho-antibodies to monitor MAPK and p38 kinase activation or antibody-specific immunoprecipitation kinase assays for JNK and IKK activities demonstrated that JNK and IKK activity increased through time to one-hour while p38 and MAPK (ERK1&2) activity peaked at thirty minutes and began to decline to noticeably lower levels by one-hour (Fig 6B as indicated). The activation profile of the MAPK, SAPK and IKK signaling molecules ERK1&2, p38, JNK and IKK in intestinal epithelial cells in response to purified flagellin exposure remarkably resembled that of intestinal epithelial cells infected with wild-type Salmonella (Fig. 5A ). From these observations we conclude that the temporal activation of the signaling pathways examined here (MAPK, SAPK and IKK), which reflect early events in Salmonella infection, are determined almost exclusively by recognition and response of intestinal epithelial cells to flagellin. Figure 6 Purified flagellin activates signaling pathways and proinflammatory gene expression in intestinal epithelial cells mimicking that of wildtype a wild-type Salmonella infection. HT29 cells were left untreated or treated with TNFα (10 ng/ml) or a cocktail of anisomycin [An] (20 μg/ml)/PMA (12.5 ng/ml) for 10 min, or with flagellin (1 μg/ml) for the indicated times. WCE were prepared and analyzed by EMSA for NF-κB DNA binding activity, immuno-kinase assays (KA) or immunoblot analysis using phospho-specific antibodies for ERK or p38 to detect activation and with kinase-specific antibodies as described in Fig. 5A to detect bulk kinase abundance as indicated. A, EMSA to detect NF-κB DNA binding activity. Authenticity of the NF-κB bandshift was tested with supershift of the complex with p65(RelA)-specific antibody (α p65), normal rabbit serum (NRS) served as an irrelevant antibody control. B, immunoblot and kinase assays to detect IKK, JNK, ERK and p38 kinase activities and protein abundance as in Fig. 5A. C, semi-quantitative RT-PCR of proinflammatory gene expression of non-treated, wild-type and flagellin double mutant Salmonella typhimurium infected, TNFα (10 ng/ml) or flagellin (1 μg/ml) stimulated cells. HT29 cells were harvested at the indicated times after the indicated treatments and isolated RNA was used to make first strand cDNA that subsequently used in RT-PCR reactions (as described in Experimental Procedures) using gene-specific primers for IL1α, IL1β, IL-8, TNFα, MCP1 and β-actin. β-actin was used as a standard for normalizing expression patterns. Resulting PCR products were fractionated on 2% agarose gels and visualized by eithidium bromide staining. We wished to further examine the effect of purified flagellin and flagellin present on Salmonella on the temporal pattern of proinflammatory cytokine gene expression in intestinal epithelial cells in order to differentiate the effects of flagellin alone vs. flagellated Salmonella or non-flagellated Salmonella infection. HT29 cells were left untreated, stimulated with TNFα (10 ng/ml), or stimulated with flagellin (0.5 ug/ml), or infected with wild-type Salmonella typhimurium or the Salmonella fliC/fljB double mutant (at MOI of 50). After the indicated times after treatment or infection, HT29 cells were harvested in ice-cold PBS and the cell pellets lysed in Trizol and RNA was purified and used to prepare first-strand cDNA (see Experimental Procedures). Aliquots of the cDNA were used in semi-quantitative RT-PCR reactions using IL1α, IL-1β, IL-8, TNFα, MCP1 and β-actin gene specific primers (sequences available upon request) and the products were fractionated on ethidium bromide containing 1.2% agarose gels. Expression of the known NF-κB target genes IL-1β, IL-8, TNFα and MCP1 was increased in response to TNFα or purified flagellin exposure (Fig. 6C ). Wild-type Salmonella infection also led to activation of these same genes although the expression of TNFα and MCP1 was transient in comparison and occurred immediately after infection. The Salmonella fliC - /fljB - double mutant failed to induce IL-1β, IL-8 and TNFα expression, however MCP1 expression was induced, although at lower levels than that induced by wild-type Salmonella , and also, the expression of MCP1 was not transient in nature and continued throughout the time course (9 h) (Fig. 6C ). The expression level of β-actin served as an internal standard for comparison. Interestingly, IL-1α, which is not an NF-κB target gene was stimulated in response to HT29 cell challenge by all of the treatments. Obviously, the Salmonella fliC - /fljB - double mutant can activate other signaling pathways leading to IL-1α expression. We presently do not know what these signaling pathways are. Flagellin activates NF-κB DNA binding in a MyD88-dependent manner Flagellin was capable of activating the requisite signaling pathways consistent with proinflammatory gene activation similar to that of a cytokine like TNFα that activates all cells on which a functional cell surface receptor for it is present (see p65 [RelA] nuclear localization in Fig. 1 and Fig. 5C ) we decided to examine the potential of the Toll-like receptors, to activate the NF-κB pathway in response to flagellin exposure. To quickly test this hypothesis we examined the effect that a dominant-negative MyD88 (aa 152–296) [ 42 ] expressing adenovirus had on flagellin-mediated NF-κB activation in HT29 cells. MyD88 is an adapter protein utilized by the IL-1 receptor and all of the known TLRs, which share homology to IL-1 through their cytoplasmic signaling domain and is required for immediate activation of the NF-κB pathway [ 43 , 44 ]. We found that expression of DN-MyD88 in HT29 cells blocked the activation of NF-κB DNA binding activity assayed by EMSA analysis in response to IL-1 or flagellin exposure, consistent with the action of a TLR-mediated activation of NF-κB (TT & JD, unpublished observations). To examine this possibility further we initially used wild-type, MyD88 -/- and TLR2 -/- /TLR4 -/- MEFs (a gift of S. Akira, Univ. of Osaka, JA) to verify the role of MyD88 and to examine the potential role of two of the TLRs to respond to flagellin or to direct wild-type Salmonella infection and lead to NF-κB activation (Fig. 7 ). Wild-type Salmonella infection activates NF-κB potently in both the wild-type and TLR deficient MEFs (lanes 2 & 15) but this activation is somewhat defective in the MyD88 deficient MEFs (lane 10). Challenge of all three types of cells with concentrated sterile-filtered wild-type S. dublin or the double SopE - /SopB - isogenic mutant S. dublin strain SE1SB2 culture broths activated NF-κB in wild-type MEFs and TLR2/4 double deficient cells but failed to activate NF-κB in MyD88 deficient cells (compare lanes 11 and 12 with lanes 3, 4, 6, 7, 16 and 17). NF-κB was potently activated in wild-type MEFs by exposure to purified flagellin (0.5 μg/ml) and therefore eliminated the possibility that LPS played a role in NF-κB activation in these experiments. The exclusion of LPS as a major contributor to NF-κB activation is also provided by the potent activation of the TLR2/4 double deficient MEFs (lanes 16 & 17). TLRs 2 and 4 respond to bacterial lipopeptides, peptidoglycans, certain LPSs and gram negative LPS respectively [ 45 - 47 ]. IL-1 stimulation verified the functional requirement of MyD88 in transmission of IL-1 and flagellin-mediated signals. Figure 7 Flagellin-mediated activation of NF-κB is MyD88 dependent. Infectious wild-type Salmonella Dublin (MOI of 100), IL-1 (20 ng/ml), purified flagellin (1 μg/ml) (as indicated), sterile-filtered and concentrated 100 kDa filter retentate supernatant (spt) from wild-type Salmonella dublin and SopE - /SopB - double mutant Salmonella dublin strain SE1SB2 (S2, as indicated) was used to challenge wild-type, MyD88 -/- knockout or TLR2 -/- /TLR4 -/- double knockout MEFs as indicated. WCEs were prepared 45 min after treatments and examined by EMSA to analyze NF-κB DNA binding activity. IL-1 (20 ng/ml) was used as a positive control to monitor MyD88 function. To further define a possible role for the TLRs in flagellin recognition we assayed for the ability of overexpressed TLRs to activate NF-κB in cells that normally respond poorly to flagellin exposure. Choosing cells that responded slightly to purified flagellin ensured that the signaling components and adapters that flagellin uses were present and functional and that the limiting factor was likely only to be the receptor that responds to flagellin. We found that HeLa cells and HEK293T cells activated NF-κB DNA binding activity in response to IL-1 stimulation but poorly to flagellin exposure (TT & JD, unpublished observations) (but see Fig. 9B ) and we chose HEK293T cells to use further because of their greater transfection efficiency. Amino-terminus FLAG epitope-tagged TLRs 1–9 (kind gifts of R. Medzhitov, Yale Univ. and R. Ulevitch, TSRI) [ 48 , 49 ] were overexpressed in HEK 293T cells in transient transfections along with the 2×-NF-κB-dependent promoter driven luciferase reporter gene [ 50 ] and the expression of luciferase in response to no treatment, flagellin (0.5 μg/ml) or TNFα (10 ng/ml) was determined. TLR5 was the only TLR whose expression resulted in a noticeable response to flagellin challenge of the cells (Table 1 ). Figure 9 TLR5 is expressed in numerous cell types and has variable responses to flagellin. A, whole cell extracts were prepared from non-stimulated T84, HT29, A549, HeLa, 293T and T98G cells and fractionated on a 8% SDS-PAGE gel, proteins were transferred to PVDF membrane and probed with anti-TLR5 antibody for immunoblot analysis (IB). Protein loading was examined by probing with anti-actin antibody. B, HT29, A549, HeLa, 293T and T98G cells were left untreated (--), treated with flagellin (F) or TNFα (T) and WCEs were prepared after 45 min and used in EMSA to monitor NF-κB DNA binding activity. Authenticity of the NF-κB bandshift was tested with supershift of the complex with p65(RelA)-specific antibody (αp65), normal rabbit serum (NRS) served as an irrelevant antibody control. C, HT29, A549, HeLa, 293T and T98G cells WCEs (50 μg) were fractionated on a 8% SDS-PAGE gel, proteins transferred to Immobilon P and immunoprobed with anti-muc1 (1:450, Santa Cruz). Size markers are listed and muc1 position is indicated with an arrow. Table 1 TLR 5 reponds to flagellin and activates NF-κB No Stim TNF Flic Vector 1 13.5 4.9 TLR1 1.7 ND 5.1 TLR2 1.6 ND 5.3 TLR3 1.5 ND 5.0 TLR4 1.8 ND 5.4 TLR5 1.6 ND 9.2* TLR7 1.5 ND 5.2 TLR8 1.4 ND 5.0 TLR9 1.5 ND 5.1 293T cells were transfected with empty vector (pCDNA3.1) or the individual listed wild-type TLR alleles in triplicate in 6-well dishes. Cells were left untreated (No Stim), TNFα (10 ng/ml) or flagellin (1 μg/ml). NF-κB reporter activity was adjusted by normalizing expression to control Renilla luciferase activity and fold induction was calculated as reporter gene activity in treated cells/reporter gene activity in non-stimulated cells. ND is not determined. To further determine the likelihood of TLR5 being the TLR through which flagellin activated NF-κB, we constructed dominant-negative signaling mutations by deletion of the carboxyl portion of each TLR to a conserved tryptophan in the TIR domain (see Materials and Methods). A similar mutation in the IL-1 receptor abrogates its ability to lead to NF-κB activation [ 51 , 52 ]. Each DN-TLR along with a reverse cloned TLR5 (AS-TLR5) were cloned into the mammalian expression vector pCDNA3.1 (Invitrogen, Carlsbad, CA). All mutant proteins were expressed well (TT & JD unpublished observations). Each DN-TLR mammalian expression vector and empty expression vector along with 2× NF-κB Luc was transfected as previously described [ 3 ] into HT29 cells which respond very well to flagellin. The transfected cells were left untreated, stimulated with TNFα (10 ng/ml) or with purified flagellin (0.5 μg/ml). Reporter gene expression was observed not to be affected by DN-TLR expression in response to TNFα stimulation of transfected cells (Fig. 8A ) however, only expression of either the DN-TLR5 or an antisense TLR5 construct resulted in a nearly fifty percent and twenty-five percent inhibition of flagellin-mediated reporter gene activation respectively (Fig. 8B ), while DN-TLR2 also was found to mildly inhibit flagellin-mediated reporter expression. These results imply that TLR5 takes part in cell surface recognition of flagellin and initiates the signaling pathway leading to NF-κB activation. The effect of DN-TLR2 on NF-κB-dependent reporter gene activation may be non-specific since its expression also inhibited TNFα-mediated reporter activation as compared to the other DN-TLRs. DN-TLR2 may also compete for an unknown adapter protein that both TLR2 and TLR5 might share. In any event, TLR2 and TLR4 were shown by the results presented in Fig. 7 not to be required for flagellin-mediated activation of NF-κB. Figure 8 TLR5 inhibits flagellin-mediated NF-κB reporter gene activity. HT29 cells were transfected in triplicate in 6-well dishes using the indicated DN-TLR mammalian expression vectors or antisense TLR5 (AS TLR5) (2 μg/well), 2× NF-κB Luc reporter gene (100 ng/well), pRL-TK Renilla luciferase for normalization (50 ng/well) adjusted to 4 μg total DNA/well with empty vector pCDNA3.1 DNA. A, Fold-induction of 2× NF-κB Luc reporter gene in non-stimulated cells (light shading) and in TNFα (10 ng/ml) treated cells (dark shading). Lysates were prepared 12 h after stimulation. Results of a representative experiment are shown. B, HT29 cells transfected as in A were treated with flagellin (1 μg/ml) and cell lysates were prepared and analyzed as in Fig. 8A. Results of a representative experiment are shown. Flagellin-mediated activation of NF-κB in intestinal epithelial cells leads to increased and decreased expression of a subset of TLRs Stimulation of intestinal epithelial cells with S. typhimurium or with purified flagellin led to activation of the proinflammatory gene program (Fig. 6C ). We wished to examine whether or not expression of TLR genes would also be altered in flagellin stimulated cells. HT29 cells were treated or not with purified flagellin (0.5 μg/ml) or with TNFα (10 μg/ml) and total RNA was isolated from non-treated and treated cells three hours after stimulation and used to make first-strand cDNA. Real-time RT-PCR using gene-specific primers for each of the TLRs (Superarray, Frederick, MD) and first-strand cDNA prepared from non-stimulated or flagellin stimulated cells was used to generate SYBR-green (Perkin-Elmer) labeled DNA products that were detected in an iCycler™ (Bio-Rad). Interestingly, flagellin only mildly activated the expression of TLR2, while expression levels of TLRs 5, 6, 9 and 10 were decreased by 2-fold (Table 2 ). Contrastingly, TNF stimulation led to increased expression of TLRs 3 and 4 (1.6- and 3.5-fold respectively), while TLRs 2, 9 and 10 were decreased by approximately 2-, 5- and 3-fold respectively. GAPDH expression served as comparative standard. Table 2 Change in TLR mRNA levels following TNFα or FliC stimulation. Fold change Fold change Gene TNFα stimulated FliC stimulated TLR1 ND ND TLR2 0.6 1.3 TLR3 1.6 0.6 TLR4 3.5 1.1 TLR5 0.9 0.5 TLR6 0.9 0.6 TLR7 M M TLR8 ND ND TLR9 0.2 0.5 TLR10 0.3 0.5 GAPDH 1.0 1.0 ND None detected by RT 2 PCR M None detected above level of minus RT control. Reverse Transcription and Real Time PCR (RT 2 PCR)-RNA was prepared from cells left untreated, stimulated with TNFα (10 ng/ml) or flagellin (1 μg/ml) for 3 h. RT 2 PCR was performed with an iCycler (Bio-Rad) to quantify SYBR-green labeled products generated from PCR products of 1 st strand cDNA prepared from TLR1 through TLR10 mRNA, 18S rRNA, and GAPDH mRNA. RT 2 PCR (25 ul reaction volume) was performed with the appropriate primers (SuperArray) in triplicate with HotStart Taq DNA polymerase (SuperArray) at 95°C for 5 min to activate Taq and amplified for 40 cycles (95°C, 30 sec, 55°C, 30 sec, 72°C, 30 sec). RT 2 PCR was performed on the minus RT controls with TLR5 primers to detect DNA contamination. Fold change in mRNA expression was expressed as 2 ΔΔCt . ΔCt is the difference in threshold cycles for the TLR mRNAs and 18S rRNA. ΔΔCt is the difference between ΔCt non-simulated control and ΔCt stimulated sample. TLR5 is expressed in cells that don't respond well to flagellin This study and others [ 22 , 33 ] have identified TLR5 as the likely TLR through which flagellin activates NF-κB. Previous reports made no determination on the presence or abundance of TLR5 in the cells that they used to ascertain its function [ 22 , 33 ]. We wished to determine if TLR5 protein abundance was absent or greatly decreased in cells that failed to respond or responded poorly to challenge by flagellin. TLR5 abundance in a number of cell lines was examined by immunoblot analysis using a TLR5-specific antibody and compared with the ability of purified flagellin to induce NF-κB DNA binding activity of WCEs prepared from them. Intestinal epithelial cell lines T84 and HT29 were used as was the lung adenocarcinoma cell line A549, the human cervical adenocarcinoma cell line HeLa, the human embryonic kidney cell line expressing large T antigen HEK293T, and the glioblastoma cell line T98G. TLR5 protein was detected in all cell lines examined by immunoblot with TLR5-specific antibody (Fig. 9A ). T84 cells exhibited the highest abundance while expression levels of the other cell lines were similar and appeared not to differ by more than two-fold (Fig. 9A ). NF-κB DNA binding activity in non-stimulated, TNFα and flagellin stimulated cells was analyzed by EMSA assays of WCEs prepared from each cell type (Fig 9B ). HT29 and A549 cells responded strongly to flagellin and to TNFα stimulation while HeLa, 293T and T98G cells responded poorly (HeLa, 293T) or not at all (T98G) to flagellin stimulation. The authenticity of the NF-κB DNA binding complex was determined using p65-specific antibody to supershift the NF-κB DNA:protein complex. It is of interest that some cells that express TLR5 either do not respond at all or do so very poorly. This may be due to either lack of receptor presence at the plasma membrane and intracellular localization, inactivating or detrimental mutations in the TLR5 gene in these cell lines or lack of or low abundance of a required co-receptor or adapter protein (as is the case in some cells for TLR4 and its co-receptor/adapter MD2 [ 30 , 53 , 54 ]). IL-1 can activate NF-κB DNA binding activity in all of the examined cell lines so it appears that the signaling apparatus downstream of MyD88 to NF-κB is intact. Recently Muc1 a secreted and membrane bound mucin protein was shown to serve as a receptor that bound Pseudomonas aeruginosa and its flagellin, leading to activation of the MAPK pathway [ 55 , 56 ] although NF-κB activity was not examined. We examined the muc1 abundance levels in HT29 (strong flagellin responder), A549 (strong flagellin responder), HeLa, 293T (both weak flagellin responders) and T98G (no flagellin response) to determine if its expression correlated with the activation profile of NF-κB and MAPK in these cells in response to flagellin [ 55 ]. Should this be the case then muc1 might serve as a viable co-receptor for TLR5 in propagating activation signals leading to NF-κB activation. We observed that only HT29 cellular proteins gave a strong signal by immunoblot analysis using an muc1-specific antibody while muc1 was barely detectable in the other cell lines (Fig. 9C ). These results suggest that muc1 does not serve the role of a TLR5 co-receptor that leads to NF-κB activation and likely plays little to no role activating MAPK pathways in A549 cells where we have observed similar temporal MAPK activation in response to flagellin exposure as we do in HT29 cells (TT and JD, unpublished results). Further examination of muc1's role in HT29 cells in regards to NF-κB and MAPK signaling using siRNA is warranted. Discussion Intestinal epithelial cells at mucosal surfaces serve as innate immune sentinels controlling the innate host defense instruction to the immune effector cells inside the body in response to the external environment [ 1 , 2 ]. Previous studies examining pathogenic Salmonella invasion of intestinal epithelial cells demonstrated activation of the proinflammatory gene program and invasion of only a minor portion of the cells [ 10 ]. We previously demonstrated that NF-κB is as potently induced in pathogenic Salmonella .sp infected cells similar to those treated with the proinflammatory cytokines that are potent NF-κB activators such as TNFα and IL-1β and that this activity was IKK-mediated [ 3 ]. Here we examined how bacterial invasion of only a third of the cells could give rise to NF-κB activity profiles consistent with activation of NF-κB in every cell such as the profile TNFα stimulation provides. We found that bacterial infection activates nuclear translocation of p65 (RelA) in nearly all of the intestinal epithelial cells consistent with the hypothesis that a cell surface receptor was recognizing either a soluble product that bacteria were producing, or a bacterial product on the bacteria, or both. We examined bacterial culture broths and found a bacterial product that was protein in composition and when used to challenge intestinal epithelial cells it potently activated NF-κB DNA binding activity (Fig. 2A ). We further purified this protein by gel permeation and anion exchange chromatography and found the protein to be flagellin by electrospray ion trap mass spectroscopy (Fig. 2B & 2C and Fig. 3 ). While our studies were in progress, flagellin was identified as being a potent proinflammatory mediator leading to IL-8 production and secretion [ 16 - 18 ]. We demonstrate in this study that flagellin appears to be exclusively responsible for activating NF-κB in intestinal epithelial cells since flagellin mutant strains do not activate NF-κB (Fig. 4 ) nor lead to their internalization (Fig. 5B ). Furthermore, flagellin challenge of intestinal epithelial cells leads to p65 (RelA) nuclear localization in nearly all of the treated cells (Fig. 5B ). Transcription factors like activator protein 1 (AP-1) and NF-κB, which are key regulators/activators of the proinflammatory gene program [ 57 , 58 ] are activated by engagement of the MAPK, SAPK and IKK signaling pathways. We demonstrate that the MAPK, SAPK and IKK signaling pathways activation fails to occur in host cells by infection/exposure to Salmonella strains devoid of flagellin or products in the culture broths derived from those mutant Salmonella strains (Figs. 5A and 6B ). We also demonstrate here that combined mutants of both fliC and fljB exhibit a severe lack of invasion (10 -4 less than wild-type) and failure to activate stress response signaling, which has not been revealed previously. It is likely that the lack of flagellin production interferes with the functioning of the type III secretion system (TTSS) although flagellin is not known to effect expression of TTSS-required gene products. This hypothesis seems credible since supply of flagellin or bacterial culture components from wild-type Salmonella cultures in trans to the double fliC - /fljB - mutant bacteria fails to compliment their lack of infectivity in gentamycin invsion assays (TT & JD, unpublished observations and see Fig. 5B ). In fact, we found the abundance of a subset of Sip and Sop proteins (SipA and SopD) released into the bacterial culture media to be drastically reduced in the flagellin mutant strains used here (TT & JD, unpublished observations). These two proteins have not previously been identified as activators of NF-κB nor are they considered as such here. The TTSS translocates the Salmonella invasion proteins (Sips) and the SopE proteins into the host cell initiating cytoskeletal rearrangements that ultimately lead to bacterial internalization, [ 11 , 12 , 41 ]. In any event, it is clear that purified flagellin activates a similar cadre of proinflammatory genes as does infection of intestinal epithelial cells with wild-type flagellated Salmonella . The temporal expression pattern of these genes was found to be remarkably similar (Fig. 6C ) indicating that flagellin-mediated temporal activation of the MAPK, SAPK and IKK signaling pathways can suffice for signaling pathways activated by Sips or SopE and SopE2 and largely recapitulates the temporal activation of key proinflammatory genes as does infection of intestinal epithelial cells with wild-type flagellated Salmonella . The rapid, and potent activation of the MAPK, SAPK and IKK signaling pathways by flagellin was consistent with and indicative of the activation of a cell surface receptor. In this study and in other studies TLR5 has been demonstrated to play an integral role in the recognition of flagellin leading to activation of NF-κB and expression of the IL-8 gene (Fig. 6C ) [ 22 , 33 ]. Identification of TLR5 utilized transfection of TLRs 1 – 9 into cell lines which responded poorly to flagellin (this study) or not at all [ 22 , 33 ] and challenging the transfectants with flagellin to identify which TLR responded to this PAMP. Previous studies that identified TLR5 as the receptor for flagellin did not examine the abundance of TLR5 in these cells or account for the lack of TLR5-mediated signaling in response to flagellin [ 22 , 33 ]. We demonstrate here that cells which respond poorly (HeLa and HEK293T) or not at all (T98G) contain TLR5 in at least equivalent abundance as HT29 cells which are highly responsive to flagellin. We propose at least three possibilities to account for this discrepancy, first, this may be due to either lack of TLR5 receptor presence at the plasma membrane and intracellular localization; second, inactivating or detrimental mutations in the TLR5 gene in these cell lines; and lastly, lack of or low abundance of a required co-receptor or adapter protein required for either efficient ligand recognition and/or signaling. These possibilities are currently being investigated. We favor the last possibility since surface biotinylation experiments indicate that TLR5 is present on the cell surface in both flagellin responding cells and in non-responders mentioned above (data not shown). Invocation of the second hypothesis would require inactivating mutations be present in three different cell lines, a highly improbable outcome. How do the findings presented here correlate with events during a "normal" Salmonella infection? We have indicated in this study that defective type III secretion system functioning leads to loss of host cell infectivity and underscores the importance of this system in the normal course of infection. In the in vivo setting, polarized epithelial cells express TLR5 on the basolateral surface [ 48 ] and flagellin can only reach the receptor either after either breaching the tight junction barrier by physical damage or by loosening of the junctions in response to Sips and Sops delivered into the intestinal epithelial cells by the TTSS or by delivery of flagellin across the intestinal epithelial cell by the bacteria itself [ 17 , 59 - 61 ]. This scenario would imply the main function of the type III secretion system would be to trigger stress response signaling facilitating invasion and lead to loosening the tight junctions and result in flagellin/ flagellated bacteria to passing through the junctions and infected cells allowing access the basolateral surface and then systemic dispersion. TLR5 on the basolateral surface of the intestinal epithelial cells, in response to flagellin, could then lead to activation of NF-κB and the proinflammatory gene program and host protection. This model is consistent with activation of the proinflammatory gene program observed in response to flagellated Salmonella sp . infection in many reports too numerous to cite here and would allow the innate host defense system a fail-safe way to recognize pathogen exposure. In instances where infection of intestinal epithelial cells by naturally occurring non-flagellated Salmonella occurs, a strong proinflammatory response would not initially be presented but the Salmonella would instead lead to systemic infection as is the case in chickens with S. galinarum and S. pollorum and result in typhoid-like disease [ 62 ]. Infection of chicken epithelial cells does not lead to proinflammatory gene expression by these non-flagellated pathogens but does when infected with S. typhimurium or S. dublin [ 62 ]. Argument for the existence of an additional TLR5 co-receptor/adapter being in limited abundance or absent might be in evidence from the transfection results presented in Table 1 which demonstrated that overexpression of cell surface localized FLAG-tagged TLR5 only resulted in slightly over a two-fold increase in NF-κB reporter gene expression in response to flagellin. If only TLR5 was required for activation of the signaling pathway should not a much more robust response been observed? We have also used DN-TLR5 transfections and NF-κB-dependent reporter gene assays or overexpresssion of DN-TLR5 using recombinant adenoviruses and analysis of resulting NF-κB DNA binding activity in response to flagellin to examine its effectiveness to completely inhibit TLR5-mediated flagellin activation of NF-κB. We have found it difficult to gain more than a fifty-percent reduction in either reporter gene activation or NF-κB DNA binding activity in HT29 cells (TT, AD & JD, unpublished observations). These results suggest that the resting TLR5 signaling complex may be quite stable as has recently been suggested [ 63 ]. Should the endogenous TLR5 signaling complex be extremely stable it would therefore be expected that titration of a required pre-stimulus bound adapter or co-receptor away would be inefficient and this is what we have observed. Expression of a DN-MAL (TIRAP), a MyD88-related TLR adapter [ 64 , 65 ] had little to no effect on flagellin-mediated NF-κB activity in transient transfection NF-κB reporter gene assays (TT & JD, unpublished observations). Recently, TLR5 has been shown to bind flagellin [ 66 - 68 ] and that this is likely a direct interaction due to failure of the human TLR5 to respond to a purified flagellin derived from a mouse-specific Salmonella strain [ 68 ]. These observations still do not preclude the existence of a co-receptor or adapter that is critical for signal transmission. Detailed biochemical characterization of the TLR5 signaling complex will resolve this issue. Muc1, a recently described flagellin interacting membrane protein by virtue of its ability to trigger activation of the MAPK pathway in response to flagellin exposure [ 55 ] was considered a viable candidate for such a co-receptor but our observations suggest that it can not serve as the putative TLR5 co-receptor as it is expressed at similar levels in flagellin non-responding cell lines examined here as it is in A549 cells which respond strongly to flagellin and both cell line types express similar levels of TLR5 (Fig. 9 ). Conclusion In conclusion, our data clearly demonstrates that flagellin can act as the major determinant in activating key stress response signaling pathways and proinflammatory gene program expression in a temporal and qualitative fashion as observed during infection of intestinal epithelial cells by wild-type Salmonella sp . that express flagellin, a point that was not well established until this study. In addition, expression of the fli C gene appears to play an important role in the proper functioning of the TTSS since mutants that fail to express fli C are defective in expressing a subset of Sip proteins and fail to invade host cells. Flagellin added in trans cannot restore the ability of the fli C mutant bacteria to invade intestinal epithelial cells. Flagellin is "sensed" by TLR5 and in response propagates signaling pathways culminating in potent proinflammatory gene expression. Interestingly we found that TLR5 is expressed in weakly responding and also in some flagellin non-responding cells, 293T, HeLa and T98G cells respectively at levels similar to cells such as HT29 and A549 cells that respond strongly to flagellin and can be found on the cell surface, raising a strong possibility that productive TLR5 signaling may require an additional factor/adaptor other than those already known to be key in the IL-1 signaling pathway, which shares extensive similarities to the TLRs signaling pathways. Methods Materials Human tumor necrosis factor alpha (TNFα) and human IL-1 were purchased from R&D Systems (Minneapolis, MN). Tris [hydroxymethyl]aminomethane (Tris) was purchased from Fisher Scientific (Fairlawn, NJ). Fetal calf serum was purchased from US Biotechnologies Inc. (Parkerford, PA). Para-nitro-phenylphosphate (PNPP) was purchased from Aldrich Chemical (Milwaukee, WI). The Polyacrylamide gel electrophoresis (PAGE) supplies: acrylamide, bis-acrylamide, sodium dodecyl sulfate (SDS), TEMED, and ammonium persulfate were purchased from Bio-Rad Laboratories (Hercules, CA). Dulbecco's modified essential medium (DMEM), DMEM:F12, phosphate buffered saline (PBS), glutamine, penicillin G, streptomycin, amphotericin B, and Grace's Insect medium were purchased from Invitrogen (Carlsbad, CA). Luria Broth (LB) was purchased from Becton Dickson and Co (Sparks, MD). The protease inhibitors: aprotinin, bestatin, leupeptin, pepstatin A, and phenylmethylsulfonyl fluoride (PMSF) were purchased from Cal Biochem (La Jolla, CA). Protease inhibitor cocktail contained 10 μg/ml aprotinin, 2.5 μg/ml leupeptin, 8.3 μg/ml bestatin, and 1.7 μg/ml pepstatin A. Phorbol 12-myristate 13 acetate (PMA), N-[2-hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid] (Hepes), anisomycin, and 2-[N-morpholino]ethanesulfonic acid (MES) were purchased from Sigma Chemical (St. Louis, MO). All other reagents were purchased from Sigma Chemical or Fisher Scientific unless stated otherwise. Cell culture HT29 human intestinal (colorectal adenocarcinoma) epithelial cells (ATCC HTB-38), HeLa cervical epithelial adenocarcoma cells (ATCC CCL-2), 293T kidney cells (CRL-11268), A549 lung carcinoma cells (ATCC-185), and T98G glioblastoma cells (ATCC CRL-1690) were cultured in DMEM with 2 mM glutamine, 10% Fetal Calf Serum, 100 Units/ml Penicillin G, and 100 μg/ml Streptomycin at 37°C in a humidified 5% CO 2 atmosphere. T84 colorectal carcinoma cells (ATCC CCL-248) were cultured in DMEM:F12 with 2 mM glutamine, 5% Fetal Calf Serum, 100 Units/ml Penicillin G, and 100 μg/ml Streptomycin at 37°C in a humidified 5% CO 2 atmosphere. H5 insect cells (Invitrogen) were cultured in Grace's medium with 2 mM glutamine, 10% Fetal Calf Serum, 100 Units/ml Penicillin G, 100 μg/ml Streptomycin, and 0.25 μg/ml amphotericin B at 28°C. MyD88 -/- & TLR2 -/- /TLR4 -/- double knockout cells were obtained from Shizuo Akira and Osamu Takeuchi (Univ. of Osaka, Japan) and grown in DMEM with 2 mM glutamine, 10% Fetal Calf Serum, 100 Units/ml Penicillin G, and 100 μg/ml Streptomycin at 37°C in a humidified 5% CO 2 atmosphere. Bacterial strains Salmonella typhimurium strain SJW1103 (FliC, phase 1 flagellin, stabilized) [ 69 ] is a wild-type Salmonella typhimurium and can only express the Phase I fliC flagellin, SJW86 (SJW1103 FliC::TN10), and SJW134 (SJW1103 FliC and FljB deletions) were obtained from Robert Macnab (Yale Univ., Conn) and have been described [ 70 ]. Salmonella serovar dublin strain 2229, strain SE1 (2229 SopE mutant), strain SB2 (2229 SopB mutant), and SE1SB2 (2229 SopE and SopB mutant) were obtained from Edward Galyov (Compton Laboratory, Berkshire, UK) and have been described [ 14 , 15 ]. Salmonella strains for stimulation were grown in LB at 37°C without agitation for 16 hours, centrifuged at 6,000 × g for 1 minute, gently washed with PBS, and gently suspended in DMEM to maintain cells with attached flagella. Plasmid pFM10.1 (ampicillin resistance), encodes a green fluorescent protein (GFP) expressed after the Salmonella host is internalized by mammalian cells, obtained from Stanley Falkow (Stanford Univ., Stanford, CA) [ 10 , 34 ] and was transformed into strains SJW1103 and SJW134 by electroporation. Strains containing pFM10.1 were designated SJW1103G and SJW134G. Preparation and analysis of Salmonella cell free culture supernatant Native flagellin was harvested from S. dublin 2229 or S. typhimurium SJW1103. Starter cultures were grown in Luria broth (LB) for 18 hours at 37°C with aeration, diluted 1:5000 in fresh LB, and grown for 12 hours under the same conditions. All subsequent procedures were performed at 4°C. Cells were removed from the medium by centrifugation at 10,000 × g for 5 min and discarded. The supernatant containing flagellin was filtered through a 0.8 micron filter (Millipore, Bedford, MA) to remove residual cells. Supernatant was concentrated 100 fold using an Amicon 100 kiloDalton (kDa) cutoff membrane (Millipore). Initial studies used concentrated culture supernatant from S. dublin strain 2229 that was treated with DNase, RNase, Protease K, boiled for 20 min or 100 mM DTT at 37°C for 2 hours and used for stimulation of cultured cells. Concentrated S. typhimurium 1103 bacterial culture supernatant was washed 4 times by 1:10 dilution with 50 mM MES, pH 6.0, 50 mM NaCl and re-concentrated. Material not retained by the 100 kDa membrane was discarded. Washed culture supernatant was fractionated by gel permeation or anion exchange chromatography for analysis. For long-term storage, washed culture supernatant was supplemented with protease cocktail and stored at -20°C. Fractionation by gel permeation chromatography was performed with a Superose 12HR column (Pharmacia) on a Bio-Logic system (Bio-Rad). One-half mililiter of 100× washed supernatant (equivalent of 50 ml original culture supernatant) was separated on the column at 0.4 ml/minute in 50 mM Hepes, pH 7.4, 200 mM NaCl. Fractions (0.5 ml) were collected, and 50 μl was fractionated by SDS-PAGE and stained with Bio-Safe Coomassie (Bio-Rad). Thirty microliters of each fraction was used for stimulation of HT29 cells (60 mm dishes) for 45 min and NF-κB DNA binding activity in the resulting whole cell extracts extracts were assayed by EMSA. The column was standardized with catalase (232 kDa), aldolase (158 kDa), abumin (67 kDa), ovalbumin (43 kDa), and Chymotripsinogen A (25 kDa), all obtained from Amersham-Pharmacia. Fractionation by anion exchange chromatography was performed with Poros HQ matrix (2 ml column, PerSeptive Biosystems, Farmingham, MA) on a Bio-Logic system. Five mililiters of 100× washed supernatant (equivalent of 500 ml original culture supernatant) was separated at 1 ml/minute in 50 mM Hepes, pH 7.4, and a NaCl gradient from 50–500 mM. Fractions were collected and 5 μl of each fraction was examined by 10% SDS-PAGE. Proteins were fractionated on duplicate 10% SDS-PAGE precast gels (BioRad). One gel was stained with Bio-Safe Coomassie (Bio-Rad) and the protein bands were isolated for Mass Spectroscopy analysis (CCF Mass spectroscopy core facility) from the other identical non-stained gel, by electro-elution with a whole gel eluter (Bio-Rad) and SDS was removed with SDS-Out (Pierce, Rockland, IL) per the manufacturers directions. Proteins isolated from bands B1 to B6 were acetone precipitated by addition of 20 μg Aprotinin and 5 μg of BSA to each eluted fraction, ice-cold acetone (-20°C) was added to 80%, mixed well and precipitated overnight at -20°C. Proteins were pelleted by centrifugation at 14,000 × g in the cold for 30 min, acetone/liquid was removed and the pellets washed 2× with 1 ml acetone (-20°C). After removal of the acetone, protein pellets were air dried and then resuspended and denatured in 5 μl of 6 M guanidinium hydrochloride (Gu-HCl) at room temperature for 30 min. Resuspended proteins were two-fold serially diluted in DMEM to a final Gu-HCl concentration of 55 mM to renature the proteins. Two hundred fifty microliters of individual renatured proteins/DMEM were added per ml to HT29 cells (60 mm dishes) and whole cell extracts were prepared 45 min after stimulation and were assayed for NF-κB DNA binding activity by EMSA. Purification of flagellin (purified flagellin) The washed and concentrated culture supernatant from S. typhimurium 1103 containing flagellin was boiled for 20 minutes and precipitants removed by centrifugation at 15,000 × g. The supernatant containing flagellin was diluted 1:2 with 50 mM MES, pH 6.0, 50 mM NaCl and mixed with 2 ml Poros SP cation exchange matrix (PerSeptive Biosystems) per 1 liter of original culture. The Poros SP matrix was prepared as a 50% slurry and equilibrated with 50 mM MES, pH 6.0. The flagellin preparation and matrix were mixed on a roller at 12 to 14 RPM for 2 hours. The matrix along with bound contaminants was removed by filtration through a 0.85 micron filter and discarded, flagellin failed to bind to the cation exchange matrix at pH 6.0 and eluted in the flowthrough and was collected. The pH of the flowthrough was adjusted by five-fold dilution of the sample with 50 mM Hepes, pH 7.8, 50 mM NaCl, and loaded onto a Poros HQ anion exchange column (2 ml column, PerSeptive Biosystems) equilibrated with 50 mM Hepes, pH 7.4, 50 mM NaCl. The column was washed with 2 volumes 50 mM Hepes, pH, 7.4, 50 mM NaCl, and eluted with a 10 column volume linear gradient of 50–500 mM NaCl in 50 mM Hepes, pH 7.4. Flagellin eluted from the column between 200–275 mM NaCl. Fractions containing flagellin were pooled and concentrated. The preparation was determined to be pure by electrophoresis of 5 μg protein by SDS-PAGE and stained with Bio-Safe Coomassie (Bio-Rad). Samples were stored at -80°C in 50 mM Hepes, pH 7.4, approx 225 mM NaCl, 10% glycerol and protease cocktail. A 4 liter preparation of culture supernatant yielded 2 mg purified flagellin. In-gel tryptic digestion and protein identification by LC-MS Gels were fixed and stained (Bio-Safe Blue, BioRad). All of the following procedures were performed by the CCF Mass spectroscopy core facility. Excised gel bands were reduced (100 mM DTT), and alkylated (100 mM iodoacetamide). Proteins in the gel bands were digested with modified trypsin (Promega, 20 μg/mL) with an overnight incubation at 37°C. Tryptic peptides were extracted from the gel with 50% acetonitrile, 0.1% acetic acid, concentrated in a SpeedVac (Thermo Savant) to remove acetonitrile, and reconstituted to 20 uL with 0.1% acetic acid. Extracted peptides were subjected to reversed phase (50 uM ID packed with Phenomenex Jupiter C18, 6 cm capillary column) liquid chromatography (2%–70% solvent B; Solvent A, 50 mM acetic acid, aqueous, Solvent B acetonitrile), coupled to a Finnigan LCQ DECA ion trap mass spectrometer for peptide sequencing, as described [ 38 ]. Preparation of GST-IκBa1-54 and GST-cJUN1-79 kinase substrates IκBα amino acids 1 to 54 fused to GST or cJUN amino acids 1–79 fused to GST were prepared as previously described [ 37 - 39 ] and stored in kinase buffer (20 mM Hepes, pH 7.6, 10 MM MgCl2, 10 mM NaCl, 2 mM beta-glycerophosphate, 10 mM PNPP). Preparation of cells for microscopy HT29 cells for microscopic examination were grown in 6 well plates on sterile cover slips to a density of 50–75%. Cells were stimulated as described above. After stimulation, cover slips with HT29 cells were washed 2 times with ice cold PBS and fixed with 4% w/v formalin at room temperature for 20 minutes. Cells were washed 4 times with PBS prior to mounting for visualization of Salmonella invasion. Cover slips were mounted with Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA), and cover slips sealed to slides. Cells for antibody staining were treated with absolute methanol for 20 minutes following formalin fixation, then washed 3 times with PBS supplemented with 0.1% BSA (PBSB) and used directly or stored in the cold after azide was added to 0.02%. For p65(RelA) localization, cells on coverslips were blocked for 1 h at 37°C with PBS supplemented with 1% BSA. The PBSB was removed, washed once with PBSB and coverslips were placed cell-side down onto 150 μl of p65 antibody (Zymed, South San Francisco, CA) diluted 1:1500 in PBSB on a square of parafilm and placed in a humidified chamber at 37°C for 1.5 h. Coverslips were removed and placed cell-side up in 6-well dishes and washed 3 × 5 min with PBSB. Coverslips were then removed and placed cell-side down onto 150 μl of FITC-labeled donkey anti-rabbit secondary antibody (Jackson Immunoresearch Laboratories, West Grove, PA) (1:300 in PBSB) on a square of parafilm and placed in a humidified chamber at 37°C for 1.5 h. Coverslips were removed and placed cell-side up in 6-well dishes and washed 5 × 5 min with PBSB, removed and placed cell-side down onto slides mounted with Vectashield (Vector Laboratories, Burlingame, CA) with DAPI and then sealed. NF-κB localization was determined by indirect immunofluorescence. Samples were observed on a Leica DMR upright microscope (Leica Microsystems Inc., Heidelberg, Germany) at 400× with oil immersion and equipped with FITC and UV filters. Images were collected with a MicroMax RS camera (Princeton Instruments Inc., Princeton, NJ), and Image Pro plus, version 4.5, software (Media Cybermetics Inc., Carlsbad, CA). Color enhancements were performed with Image Pro plus software. Visible light plus color overlays for Fig. 1 and Fig. 5B were performed with MetaMorph Software (Universal Imaging Corp., Downington, PA). Bacterial infection and cell stimulation Mouse embryo fibroblasts (MEFs) or HT29 cells were grown in DMEM as above to a density of 90% prior to stimulation. All cells were washed with warm PBS and supplemented with DMEM without serum or antibiotics in preparation for stimulation. Cells were stimulated with; 10 ng/ml TNFα, 1 μg/ml flagellin unless specified otherwise, 20 μg/ml Anisomycin, 12.5 ng/ml PMA, or 10 8 Salmonella /ml at 37°C for desired times and extracts prepared as below. Cells harvested beyond one hour were washed with warm PBS and supplemented with warm DMEM, 2 mM glutamine, and 200 ug/ml gentamycin after 1 hour and returned to 37°C until extract preparation desired. Whole cell extract preparation Cells were washed with ice-cold PBS and all subsequent steps carried out at 4°C or on ice. Cells were scraped from the dish in ice-cold PBS, and collected by centrifugation at 1000 × g for 1 minute. Cells were lysed by suspension in 50 mM Tris-HCl, pH 7.6, 400 mM NaCl, 25 mM beta-glycerol phosphate, 25 mM NaF, 10 mM PNPP, 10 % glycerol, 0.5 mM sodium orthovanadate, 0.5% nonidet-40 (NP-40), 5 mM benzamidine, 2.5 mM metabisulfite, 1 mM PMSF, 1 mM DTT and protease inhibitor cocktail as described [ 3 ]. Electromobility shift assays (EMSA) NF-κB DNA binding assays were carried out as previously described [ 3 , 35 , 38 ]. Anti-p65 antibody (Zymed, South San Francisco), anti-p50 antibody (Santa Cruz Biotechnologies, Santa Cruz, CA), and anti-STAT3 antibody (Santa Cruz) were used for EMSA supershifts. Invasion assay HT29 cells, 90–95% confluent in 35 mm round dishes, were prepared for stimulation as above and treated with a 1 ml suspension of Salmonella SJW1103 or SJW134 or left untreated in triplicate as above. After one hour, HT29 cells were washed 4× with warm PBS, supplemented with warm DMEM, 2 mM glutamine, and 200 μg/ml gentamycin, and incubated at 37°C for 4 hours. Cells were then harvested as above and lysed by suspension in 1 ml sterile distilled water. Ten-fold serial dilutions were prepared in PBS and 100 μl of each dilution was plated on LB agar plates and grown at 37°C for 20 hours. Colonies were counted and averaged. Kinase assays Whole cell extracts (250 μg) were supplemented with 150 μl of Buffer A (20 mM Hepes, pH 7.9, 20 mM beta-glycerophosphate, 10 mM NaF, 0.1 mM orthovanadate, 5 mM PNPP, 10 mM 2-mercaptoethanol, 0.5 mM PMSF, and protease inhibitor cocktail), and immuno precipitation kinase assays carried out as described [ 3 ] using either IKKα monoclonal antibody (PharMingen – Becton Dickson), anti-JNK1 (Santa Cruz Biotechnologies, Santa Cruz, CA), or anti-hemagglutinin (HA) epitope antibody (Covence Antibodies, Princeton, NJ) as indicated. Protein G immunopellets were collected by centrifugation at 500 × g for 30 sec, washed 3 times with Buffer B (Buffer A plus 250 mM NaCl), and one time with Buffer C (Buffer A plus 50 mM NaCl and 10 mM MgCl 2 ). Immunopellets were resuspended in 30 μl Kinase buffer with 0.1 mM orthovanadate, 50 μM "cold" ATP, 5 μCi γ- 32 P-ATP, 2 mM DTT, and 2 μg of soluble GST-IκBα1–54 or GST-cJUN1-79, and incubated at 30°C for 30 minutes. Reactions were stopped by the addition of 15 μl 4× SDS-PAGE loading buffer, heated at 95°C for 5 minutes, and resolved on 10% SDS-PAGE gels by standard procedures. Gels were rinsed, stained with Bio-Safe Coomassie (Bio-Rad) to visualize protein bands, rinsed, photographed then dried and exposed to Kodak X-OMAT AR film (Eastman Kodak Co., Rochester, NY) to detect substrate phosphorylation. Immunoblotting Protein samples (40 μg) were resolved by SDS-PAGE on a 10% acrylamide gels by standard procedures, and proteins transferred to PVDF membrane (Millipore) and probed with antibodies as described [ 3 ]. Membranes were washed 3× briefly with TBST, incubated with a 1:1000 dilution (1:800 for anti-TLR5) of the primary antibody in TBST, 1% non-fat milk for 1 hour, washed 3 × 5 min with TBST, and then incubated with a 1:2000 dilution of the appropriate HRP-conjugated secondary antibody in TBST, 0.5% non-fat milk for 1 hour. Primary antibodies used were: anti-IKKα/β (H-470, Santa Cruz), anti-JNK1, anti-ERK2 (K-23, Santa Cruz), anti-phospho-ERK (E-4, Santa Cruz), anti-p38MAPK (Cell Signaling Technologies, Beverly, MA), anti-phosopho-p38MAPK (Cell Signaling), anti-TLR5 (H-127, Santa Cruz), anti-muc1 (H-295, Santa Cruz) and anti-actin (C-11, Santa Cruz). Secondary antibodies used were: anti-mouse IgG HRP conjugate (Amersham-Pharmacia), anti-rabbit IgG HRP conjugate (Amersham-Pharmacia), anti-goat IgG-HRP conjugate (Santa Cruz). HRP activity was detected by ECL (Amersham-Pharmacia) as per manufacturers instructions, on Kodak X-OMAT AR film. Construction of dominant-negative TLRs All DN-TLRs were constructed using PCR. The universal 5' primer consisted of a 5'KPN I restriction site followed by sequences encoding the kozak sequence, translational start site, and preprotrypsin leader sequence of pCMV-1 (Sigma) that all the wild-type TLRs were initially cloned into. The 3' anti-sense (AS) primers were human TLRgene-specific primers (sequences available upon request) that created a stop codon immediately after a conserved tryptophan in Box 9 of the TLR TIR homology domain according to Bazan [ 71 ], thus creating carboxy terminus deletions. The 5' end of the AS primer contained a number of convenient restriction sites to allow directional cloning. PCR was performed with turbo-Pfu polymerase (Stratagene, La Jolla, CA) using standard procedures on individual wild-type TLR pCMV-1 plasmid DNAs (5 ng each, kind gifts of R. Medzhitov, Yale Univ. and R. Ulevitch, TSRI) [ 48 , 49 ] with the 150 ng each of the universal 5' sense primer and individual gene-specific TLR 3' primers. PCR products were cleaned-up with PCR cleanup kit (Qiagen, Germany) digested with appropriate restriction enzymes, gel purified and then ligated into the mammalian expression vector pCDNA3.1 (Invitrogen). Positive clones were sequenced to verify the mutations and tested for expression in transient expression assays and detected on immunoblots by probing with anti-FLAG M2 monoclonal antibody (Sigma). All wild-type and DN-TLR alleles are amino terminus FLAG epitope-tagged. Transfections HT29 cells were transfected with Lipofectamine Plus (Invitrogen) as previously described [ 3 ]. In transfections monitoring reporter gene expression, transfections were performed at least three times in 6 well dishes in triplicate with the total DNA mass kept constant at 4 μg (2 μg effector plasmid DNA, 100 ng 2× NF-κB Luc reporter gene, 50 ng pRL-TK, a thymidine kinase promoter driven Renilla luciferase normalization reporter and 1.85 μg pCDNA3.1 plasmid DNA as bulk filler DNA) and fire-fly luciferase expression was normalized to Renilla luciferase expression using the dual-luciferase assay (Promega, Madison, WI). Fold inductions were calculated and values between experiments did not vary more than 15%, a representative experiment is presented. Transfection of 293T cells was performed with lipofectamine 2000 (Invitrogen) in 6-well dishes in triplicate as per the manufacture's protocol. TLR expression plasmids were added at 2 μg/well, and NF-κB and normalization control plasmids were as above with HT29 cells and pCDNA3.1 plasmid DNA as bulk filler DNA to a final DNA mass of 4 μg/well. Fold inductions were calculated and values between experiments (N of 3) did not vary more than 10%, a representative experiment is presented. Real Reverse Transcription and Real Time PCR (RT 2 PCR) Cells (N = 3) were stimulated 3 hours at 37°C with TNFα or FliC or left untreated and harvested for total RNA isolation. Total cellular RNA was extracted from cells with Trizol reagent (Invitrogen) [ 3 ] and reverse transcribed with ReactionReady first strand cDNA synthesis kit (SuperArray Bioscience Corp., Fredrick, MD). RNA (2.5 ug per 20 ul reaction) was reverse transcribed using random primers and Moloney murine leukemia virus reverse transcriptase per manufacturer specified conditions. Controls without reverse transcriptase (minus RT) was also generated for each RNA sample. RT 2 PCR was performed with an iCycler (Bio-Rad) to quantify TLR1 through TLR10 mRNA, 18S rRNA, and GAPDH mRNA. RT 2 PCR (25 ul reaction volume) was performed with the appropriate primers (SuperArray) per manufacturers instructions in triplicate with HotStart Taq DNA polymerase (SuperArray) at 95°C for 15 min to activate Taq and amplified for 40 cycles (95°C, 30 sec, 55°C, 30 sec, 72°C, 30 sec). RT 2 PCR was performed on the minus RT controls with TLR5 primers to detect DNA contamination. Real-time PCR analysis was performed using SYBR-green (Perkin-Elmer) according to manufacture's instructions with the specific primer pairs indicated above and primer pairs for 18S ribosomal RNA as reference RNA (Classic 18S primer pairs – Ambion Inc). Cycle time (Ct) was measured using the iCycler™ and its associated software (Bio-Rad). Relative transcript quantities were calculated by the ΔΔCt method using 18S ribosomal RNA as a reference amplified from samples using the Classic 18S primer pairs from Ambion, Inc (Austin, TX). Normalized samples were then expressed relative to the average ΔCt value for untreated controls to obtain relative fold-change in expression levels. Fold change in mRNA expression was expressed as 2 ΔΔCt . ΔCt is the difference in threshold cycles for the TLR mRNAs and 18S rRNA. ΔΔCt is the difference between ΔCt non-simulated control and ΔCt stimulated sample. Values for fold-induction varied less than 5% among replicates. Abbreviations The abbreviations used are: FBS, fetal bovine serum; IL-1, interlukin-1, SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; EMSA, electromobility shift assay; IB, immunoblot; KA, kinase assay; GST, glutathione S-transferase; PBS, phosphate-buffered saline; TNFα, tumor necrosis factor α; NF-κB, nuclear factor kappa B; IKK, Ikappa B kinase; IκB, Ikappa B; PCR, polymerase chain assay; RT-PCR, reverse transcription polymerase chain assay; Gu-HCl, guanidinium hydrochloride ; MAPK, mitogen activated protein kinase; SAPK, stress-activated protein kinase; ERK, extracellular regulated kinase; TLR, toll-like receptor; DN, dominant-negative; JNK, Jun N-terminal kinase; AP-1, activator protein-1; MEF, mouse embryo fibroblast; WCE, whole cell extract; IEC, intestinal epithelial cell; MCP1, macrophage chemoattractant protein 1; TTSS, type III secretion system; Sip, Salmonella invasion protein; PMA, phorbol 12-myristate 13 acetate; PNPP, para nitrophenyl phosphate; TK, thymidine kinase; BF, bright field; NP-40, nonidet-40; NRS, normal rabbit serum; IN, input; Ct, cycle time. Authors' contributions TT and AD initiated the study and performed the majority of the experiments and contributed equally and were assisted by NK and JL. MD constructed a number of DN-TLRs and JD developed the study, provided funding support, oversaw the project and also constructed a number of mutant TLRs.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC516440.xml

544886

Is DRE essential for the follow up of prostate cancer patients? A prospective audit of 194 patients

Background Prostate cancer follow up forms a substantial part of the urology outpatient workload. Nurse led prostate cancer follow up clinics are becoming more common. Routine follow-up may involve performing DRE, which may require training. Objectives The aim of this audit was to assess the factors that influenced the change in the management of prostate cancer patients during follow up. This would allow us to pave the way towards a protocol driven follow up clinic led by nurse specialists without formal training in DRE. Results 194 prostate cancer patients were seen over a period of two months and all the patients had DRE performed on at least one occasion. The management was changed in 47 patients. The most common factor influencing this change was PSA trend. A change in DRE findings influenced advancement of the clinic visit in 2 patients. Conclusions PSA is the most common factor influencing change in the management of these patients. Nurse specialists can run prostate cancer follow-up clinics in parallel to existing consultant clinics and reserve DRE only for those patients who have a PSA change or have onset of new symptoms. However larger studies are required involving all the subgroups of patients to identify the subgroups of patients who will require DRE routinely.

Introduction Prostate cancer ranks first amongst all male urological cancers [ 1 ]. In the UK, 26027 new patients were diagnosed with prostate cancer during 2001 [ 1 ]. The evidence suggests an increasing trend in the incidence in the recent years, being 18201 in 1997 [ 2 ]. Nonetheless, better treatment modalities and earlier detection has resulted in a decrease in cancer related mortality [ 3 ]. This is shown in the age-standardized death rate per million population for prostate cancer, being 302 and 274 in 1991 and 2001 respectively. Widespread PSA testing and increased awareness has led to the detection of early prostate cancer in many patients [ 4 ]. This has probably resulted in more patients requiring long periods of follow up. Nurse Specialists in UK health care system have evolved to share the increasing demand on the clinicians to meet the targets and waiting times in all the specialties. In urology, Nurse Specialists have assumed various roles including prostate assessment clinics, urodynamics and flexible cystoscopy [ 5 ]. In some health care trusts, Nurse Specialists are involved in the follow up of treated prostate cancer patients. Faithfull et al studied the use of telephone follow up of prostate cancer patients by nurse specialists. They found that this method of follow-up at 3, 6 and 12 weeks post radiotherapy was effective and economical [ 6 ]. In addition a study on the follow-up of prostate cancer patients by on-demand contact with a nurse specialist was found to be as effective as traditional outpatient follow up by urologists [ 7 ]. The EAU guidelines [ 8 ] suggest that prostate cancer patients should be followed at regular intervals with a disease specific history and PSA estimation supplemented by digital rectal examination. This would suggest that all Nurse Specialists undertaking the role of follow-up of such patients should be trained in DRE. Data on the role of DRE in the follow up of prostate cancer patients is available only for the subgroup of patients who have had treatment with curative intent (radical prostatectomy or radical radiotherapy) and these studies show that PSA trend plays a more important role than DRE. However there is limited data available on the role of DRE and other factors (e.g. LUTS, Bone pain etc) in the follow up of diagnosed prostate cancer patients in the general setting involving all treatment varieties which is likely to be encountered in a nurse led follow up clinic. The aim of this audit was to prospectively assess the various factors that influence a change in the management of the prostate cancer patients on follow up and to highlight the feasibility of nurse led clinics for the follow up of prostate cancer patients. Methods Over a two-month period (Dec 2002–Jan 2003) all the prostate cancer patients being followed up in the Urology outpatient clinics at our institution were audited prospectively. The patients were seen by a Consultant, Specialist Registrar or Senior House Officer. The period of follow-up, initial stage of the disease, management modality, consecutive PSA values and consecutive DRE findings (if available) were recorded on specifically designed data collection forms. All the patients had DRE done on at least one occasion. The change in the management was defined as any alteration in the follow-up pattern; either as an advancement or postponement of a future appointment, the need for further investigation or treatment, the admission of a patient and the referral to a different specialist, for example an Oncologist or Palliative Care specialist The attending physicians were requested to record whether there was any change in the management and which factors influenced the change. They were specifically requested to record whether DRE influenced a change. Results During the period studied 194 patients being followed up for treated prostate cancer were included. The mean age was 74.8 years and the stages at initial diagnosis were: T1 (n = 73), T2 (n = 63), T3 (n = 44), T4 (n = 14). Ten patients had metastatic disease. The management modalities that these patients had undergone included: hormonal manipulation (68), orchidectomy (8), radical radiotherapy with hormonal manipulation (15), radical radiotherapy (48), radical prostatectomy (21), brachytherapy (1) and active surveillance (33) (Table 1 ). The management changed in 47 of 194 (24%) patients. The factors that influenced the changes included PSA trend (n = 27), LUTS (n = 10), bone pain (n = 4), change in DRE findings (n = 2) and other factors namely abnormal renal functions (n = 1), hematochezia (n = 1), pruritis (n = 1) and erectile dysfunction (n = 1) (Table 2 ). Table 1 Management categories of the follow up prostate cancer patients Management Number of patients Percentage of the total number (n = 194) Active surveillance 33 17 Radical prostatectomy 21 10.8 Radical radiotherapy 48 24.8 Radical radiotherapy With hormones 15 7.8 Brachytherapy 1 0.5 Hormone therapy 68 35 Orchiectomy 8 4.1 Table 2 Factors that influenced a change in management Factors Number of patients Percentage of the total number of changes (n = 47) PSA trend 27 57.5 Lower urinary tract symptoms 10 21.3 Bone pains 4 8.5 DRE findings 2 4.3 Pruritis 1 2.1 Altered renal functions 1 2.1 Erectile dysfunction 1 2.1 Bleeding per rectum 1 2.1 In this audit PSA trend was the most common factor that resulted in a management change. In the two patients there was a change in DRE findings (progression from T 2 b disease to T 3 disease as observed by the assessor). This only resulted in the subsequent visit being sooner than planned. Discussion The follow up of patients with prostate cancer has traditionally included a disease specific history, serial PSA estimations and a DRE. The roles of PSA and DRE have been extensively evaluated in the diagnosis of prostate cancer patients [ 9 , 10 ]. There have only been a few studies questioning the importance of DRE in the follow up of patients treated with a curative intent [[ 11 - 13 ] and [ 14 ]]. These have been based on groups of patients undergoing specific treatments. These studies concluded that DRE is unnecessary in the follow up of patients if PSA is undetectable. However there have been rare case reports describing local or systemic recurrence in the absence of detectable PSA [ 15 , 16 ]. There are no reported studies in the English language assessing the role of routine DRE in the follow up of all treated prostate cancer patients in a general urology outpatient setting. In addition, studies assessing the various factors (e.g LUTS, bone pains etc) that influence a change in the management of these patients have not been reported. The present audit shows that PSA trend is the most common factor influencing a change in management whilst DRE plays a very limited role. Further, there are other factors that influence a change in the management of these patients' e.g. Bone pain and LUTS. Although the numbers of patients involved in this audit are moderate it would suggest that Nurse Specialists could deliver the optimum care in following up treated prostate cancer patients. Such Nurse led clinics could be carried out in parallel to the existing Consultant clinics thereby allowing the availability of medical personnel to perform DRE where deemed necessary. A protocol to perform DRE when there is an increase in PSA, onset of new symptoms or worsening of existing symptoms would be suitable for such a clinic. This audit suggests that Nurse Specialists need not be trained to perform DRE before the establishment of such clinics. However larger studies are required to identify subgroups of treated prostate cancer patients who may require a DRE on a regular basis. Alternatively nurses could be taught to undertake DRE thereby further reducing clinician workload. This would require a standardised and validated teaching method, which currently does not exist. In our hospital this audit has influenced the initiation of Nurse led prostate cancer follow up clinics conducted in parallel to the consultant clinics. Competing interests The author(s) declare that they have no competing interests. Authors' contributions NR – Along with VKS conceived the study, collected the data and jointly prepared the text with VKS – Along with NR conceived the study, assessed the data and prepared the text, SG – Participated in collecting patients details and in the preparation of the text, JH – helped in approaching the patients and data collection, SSM – Advised regarding the design of the study and contributed to the text, MEW – Advised regarding the design of the study and contributed to the text, RAB – Overall supervision of the project with periodic assessment on progress and preparation of text All 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/PMC544886.xml

544892

Strategies to prevent HIV transmission among heterosexual African-American men

Background As part of qualitative research for developing a culturally sensitive and developmentally appropriate videotape-based HIV prevention intervention for heterosexual African- American men, six focus groups were conducted with thirty African-American men to determine their perceptions of AIDS as a threat to the African-American community, characteristics of past situations that have placed African Americans at risk for HIV infection, their personal high risk behaviors, and suggestions on how HIV intervention videotapes could be produced to achieve maximum levels of interest among African-American men in HIV training programs. Methods The groups took place at a low-income housing project in Houston, Texas, a major epicenter for HIV/AIDS. Each group was audiotaped, transcribed, and analyzed using theme and domain analysis. Results The results revealed that low-income African-American men perceive HIV/AIDS as a threat to their community and they have placed themselves at risk of HIV infection based on unsafe sex practices, substance abuse, and lack of knowledge. They also cite lack of income to purchase condoms as a barrier to safe sex practice. They believe that HIV training programs should address these risk factors and that videotapes developed for prevention should offer a sensationalized look at the effects of HIV/AIDS on affected persons. They further believe that programs should be held in African-American communities and should include condoms to facilitate reduction of risk behaviors. Conclusions The results indicate that the respondents taking part in this study believe that HIV and AIDS are continued threats to the African-American community because of sexual risk taking behavior, that is, failure to use condoms. Further, African-American men are having sex without condoms when having sex with women often when they are under the influence of alcohol or other mind-altering substances and they are having sex with men while incarcerated and become infected and once released resume unprotected sexual relations with women. According to the men, substance abuse is an important part of the problem of HIV in the African-American community. This is in keeping with research that shows that drug use, especially crack cocaine, is linked to sexual risk taking among African Americans and to increased likelihood of becoming infected with other sexually transmitted diseases (STDs) including HIV. Thus, interventions for men should address condom use, condom availability, skills for using condoms, eroticizing condoms and substance abuse prevention. Men in the present study also strongly recommended that HIV/AIDS videotaped messages should include footage of the sensational effects of the disease.

Background The HIV/AIDS epidemic continues to be a major public health challenge in the African-American community. Although African Americans constitute only 13% of the United States population, they have accounted for 39% of all the 886,000 estimated AIDS cases that have been diagnosed since the epidemic began in 1981 [ 1 ]. In 2002, African Americans comprised about 50% of the 42,000 AIDS cases diagnosed among adults in the United States, and also accounted for more than half of the HIV diagnoses that were reported to the Centers for Disease Control and Prevention (CDC) [ 1 ]. Amongst these, the rate of diagnoses among African-American men was almost nine times greater than the rate for white men [ 1 ]. The CDC has identified sexual contact between African-American men who have sex with men, injection drug use, and heterosexual contact respectively, as the leading causes of HIV infection among this group [ 1 ]. While several successful intervention programs have been developed, evaluated, and implemented for African-American men who have sex with men, African-American women, and African-American adolescents [ 2 ], only a limited number of prevention programs have been implemented for heterosexual African-American men [ 2 ]. This is unusual given that heterosexual transmission has been identified as the leading cause of HIV infection among African-American women [ 1 ]. Several factors other than those reported by the CDC have placed African-American men at an increased risk of HIV infection. Essien et al. [ 3 ] explored the role of misperceptions about HIV transmission among African-American men and women by conducting eleven focus groups with sixty-nine men and women. They reported myths and misperceptions about HIV transmission such as denial of personal risk, perceptions that HIV was a disease that happened to outsiders and "others" and the perceived role of the government in the development of HIV as contributing factors to the efficiency of HIV transmission in this population. In a follow-up study to assess the range of discrepancy in self-reported sexual identity and sexual behavior in men and women of four racial/ethnic groups, Ross et al. [ 4 ] found that 43% of the African-American men in their sample who self-identified as heterosexuals were also having sex with other men. This finding is consistent with a recent report by Montgomery et al. [ 5 ] which showed that, among African-American men who had sex with men, 34% had also been engaged in a bisexual relationship. Research has shown that bisexual African-American men may not relate to HIV prevention messages that have been developed for gay men [ 1 ]. Thus, intervention programs for African-American men must be innovative and firmly grounded in the socio-cultural contexts of sexual risk taking in this population. An array of socioeconomic and cultural factors exacerbates high-risk behaviors that place African-American men at risk for HIV. First, nearly 25% of African Americans live in poverty [ 6 ], and several studies have shown a direct relationship between poverty and increased prevalence of HIV/AIDS [ 7 - 10 ]. This may be due to high rates of unemployment, associated drug related economic activities, money-for-sex and drugs-for-sex exchanges, and inadequate access to health care and HIV prevention programs [ 1 , 11 , 12 ]. Secondly, injection drug use has been identified as the next leading cause of HIV infection among African-American men [ 1 ]. Aside from sharing injection paraphernalia, African Americans who are substance abusers often engage in high risk behaviors such as unprotected sex when they are under the influence of drugs and alcohol [ 13 ]. Thirdly, cultural and social norms in African-American communities are not supportive of homosexual behavior. It has been argued that negative attitudes associated with homophobia may lead to psychological distress and sexual risk taking among African-American men leading them to having relationships with both women and men [ 14 ]. A major challenge facing HIV prevention interventionists is the issue of developing HIV prevention programs that are firmly grounded in the cultural nuances of African Americans and the translation of these programs into effective HIV prevention practice. To date, several HIV intervention programs have been developed for African Americans [ 15 - 18 ], and the most promising interventions have been programs that are based on social cognitive principles. However, Kalichman et al. [ 19 ] have noted that HIV prevention programs that are based on social cognitive principles and proven to be effective in the scientific literature have not been widely utilized because of their reliance on expert interventionists for implementation in face-to-face formats, making them difficult to transfer to community-based organizations [ 19 ]. In contrast, social cognitive theory principles applied to HIV prevention can be delivered effectively by videotape and community-based organization personnel with minimal training in skills building techniques [ 19 ]. The rationale for using videotapes as part of an HIV intervention delivery system is provided by the emerging literature which demonstrates the feasibility of this medium in changing high risk sexual behaviors [ 20 - 23 ]. The research presented here results from qualitative studies conducted in Houston, Texas, to examine the sociocultural contexts of sexual risk taking among African-American men and to determine how a videotape-based HIV prevention intervention could be tailored so that it is effective in preventing HIV transmission among African-American men. Methods Design This study used a qualitative exploratory research design to elicit information about the strategies for preventing HIV transmission among African-American men, ages 18–29. Thirty low-income African-American men participated in five focus groups that were conducted at a housing project targeted for convenience sampling and because of its location in close proximity to the research institution in Houston, Texas, a leading HIV/AIDS epicenter in the United States [ 1 ]. Approval for the study was obtained from the relevant university Committee for the Protection of Human Subjects. Procedure The investigative team recruited focus group participants by displaying posters and flyers at strategic locations in the housing project identified by the manager. The flyer listed the study inclusion criteria: African-American heterosexual male, aged 18 – 29, self-reported unprotected vaginal sex in the last six months or having been diagnosed or treated for a sexually transmitted disease in the past year. The flyer also listed a university telephone number that prospective participants could call to obtain additional information and/or schedule their participation in a group. When calls were received, a trained research assistant confirmed eligibility. Those selected for participation were then encouraged to invite their friends to participate. Of the 97 prospective participants who contacted the university, 54 agreed to participate. Of that number, 30 appeared on the scheduled day and formed the study sample. The focus groups were conducted at the housing project's clubhouse by a facilitator with approximately 15 years experience conducting qualitative research studies with African-American men. Assistance was provided by a trained research assistant who took notes to ensure that pertinent information not captured by the audiotape was obtained and to record major points discussed. Participants were informed of the benefits of their involvement, that involvement was voluntary, that they could refuse to answer any question, and that they could cease participation at any time without penalty. Prior to the start of each group, written informed consent was obtained and agreement obtained to record the session. The participants were also told about the confidentiality of the information discussed at the meeting, especially important given that some of the men were acquaintances as a result of residency in or near the housing complex. They were advised that they would receive a $25 mall gift certificate and condoms as incentives for their participation. Respondents were also advised that the tapes, which were anonymous, would be destroyed following transcription and checking. All questions as well as the informed consent were provided in English. Data collection The investigative team utilized semi-structured and open-ended questions to elicit information from participants based on our research interest in determining perceptions of AIDS as a threat to African-American communities, characteristics of past situations which placed African Americans at risk for HIV infection, suggestions on ways to recruit African Americans into HIV/AIDS prevention programs and on how HIV intervention videotapes could be produced to achieve maximum levels of interest. A three-step staging process was used to generate the questions used in the focus groups: (1) generation of working hypotheses on facilitators and barriers to HIV prevention and program messages and methods by conducting interviews with six key informants experienced in HIV/AIDS prevention among African Americans; (2) reshaping of hypotheses until no new information was received by testing them in interviews with men similar to the target population; and (3) using the resulting hypotheses to generate questions and field testing them with members of the target population after which final changes were made. A guide consisting of a written list of questions and probes was used to conduct the interviews. According to McCracken [ 24 ], the advantages of using such a guide include the assurance that all areas of interest are covered and allowing the researcher to focus his attention on listening to the informants, thereby enabling a better understanding of their lines of thought and possibly unanticipated explanations of the concepts. The duration of each group was about two hours. Data analysis The standard grounded theory approach of Glaser and Strauss was performed to conduct data analysis [ 25 ]. Grounded theory procedure and techniques were utilized for data analysis by doing a line-by-line analysis of the focus group transcripts and identifying the emerging themes [ 26 ]. The thematic concepts representing ideas expressed by a majority of the members of three or more focus groups were characterized as a domain and are reported below. Results Study participants The study sample was comprised of 30 men between the ages of 18 and 29. Educational attainment ranged from tenth grade to college graduate with the greatest number of men stating that they had at least one year of post-secondary education. Most men identified themselves as day laborers attempting to earn income by doing various odd jobs as they become available. The remainder was unemployed. No men identified themselves as HIV positive, and none of the respondents had health insurance. Perception regarding AIDS and African Americans The situational determinants of HIV risk taking and their impact on HIV/AIIDS prevention behaviors and education programs were examined. The discussion began by asking the men for their overall view regarding HIV and its impact on the African-American community. Specifically, they were asked if they believe that AIDS is a threat to the African-American community and to what they attribute its rapid spread. About one-half of men stated that it is due to lack of unsafe sex practices including having sex with multiple partners and failure to use condoms. A large number of men believed that African Americans have higher rates of HIV and AIDS due to their use of alcohol and drugs while a smaller number stated that AIDS is not a threat to the African-American community, but rather it is a universal problem or that African Americans have feelings of invincibility because they think that HIV happens to others and that God will protect them. A lesser number attribute the rapid spread of HIV and AIDS among African Americans to incarcerated men having sex with men while behind bars and returning to their female partners when they are released. Clarence, a 26-year old forklift operator, believes that AIDS is a threat to the African-American community due to failure to use protection from lack of understanding about the disease: " A lot don't understand the disease or what the effect of it may be so they have unprotected sex because they don't understand the disease really well as they supposed to. " Ben, a 22 year old promotional worker, expressed his concerns about men having sex with men and also with women. He states: " It is not a problem on the outside, but a lot of men have sex without condoms in the jails and they are bringing it right back home, and they are not telling anybody that they had sex in jail, and they are bringing it and sharing it among the girls. " Kenny, a 27 year old temporary worker, adds: " I was recently locked up a couple of years ago. I met some guys you'd never think would be gay, undercover homosexuals, they pump weights with you, run track, and play basketball with you, but as soon as they are behind closed doors, their cellmate might be gay and they are banging in there. Even black gangs, sometimes when they initiate them, they do that, and they bring the disease back home. " Thus, incarceration and gang membership may facilitate HIV transmission and may be partially responsible for the high rates of bisexual activities in the African-American community. Jason, a 20 year old laborer explained that AIDS is a universal threat but that it is an epidemic among African Americans due to associated substance abuse and failure to use protection: " It's a threat to everyone but we have a tendency in the black race, especially now with another epidemic that we have in our neighborhood called crack. People have a tendency now to be even careless with this drug. So they do things that they would not naturally do. Just having sex. The trust thing – we feel that we won't use condoms because we feel like they're uncomfortable, yeah, they kill the sensation and so then we're gonna jump in and do the do and so it has a stronger effect on us now than a lot of races. " Terry, a 28-year old home improver, asserts that African Americans do not believe that God will inflict added suffering (e.g., AIDS) upon them because they have already experienced hardship, God would not cause something else to make life difficult. He said: " It's because of the color of our skin. People believe that, seems to believe that, oh, God is going to let them know or it's not going to happen to me or something like this. We done had enough bad happen to us. " Karl, a 29 year old college graduate did not believe that AIDS is a threat to the African-American community, but rather that statistics are inflated for this group due to the places where they obtain medical treatment. He believes the government keeps data that private physicians do not which results in inflated numbers for African Americans. Mistrust of healthcare system reporting about African Americans leads him to believe the situation is not as bleak as reports indicate. He stated: " Most African Americans go to clinics which are state or federally funded and they keep that data. And most people with private health insurance go to their doctors and those private doctors don't submit that same data. So when you look at the quote-unquote per capita, it always looks like it is higher in the African-American community. " Risk situations for HIV transmission When asked to describe the means by which they had personally placed themselves at risk for HIV infection, almost half of participants cited unprotected sex or not taking the time to use protection "in the heat of the moment." These men did not have condoms in their possession, did not take the time to use a condom, or simply disregarded condom use because of dislike for them. Similarly, about one-third of men described specific situations with risks known to them at the time they engaged in unprotected intercourse. They included having sex with an HIV infected person, having sex with a stranger, sex with a drug user, and sex in exchange for money or drugs. The remainder of men stated that their drinking and drug use had created HIV risk situations. Austin, a 25 year old pipe fitter, explained that when he went to clubs and took women home, the result was being irresponsible and having unprotected sex: " I've been drinking, just being in the room... the heat of the moment ... the heat of passion, just the heat of the moment taking away all the common sense. " Clarence is living with the consequences of having sex with strangers in a mood altered state. Although he did not contract HIV, he expressed that alcohol and unprotected sex resulted in him contracting hepatitis A.: " Alcohol did that to me. I had sex with many women that I didn't know had the disease or nothing and I caught hepatitis A behind it once. " Like Austin and Clarence, Jason attributed his involvement in HIV risk situations as resulting from substance abuse. He stated: " You don't mean to do it, but any mind-altering drug will lead you into a situation where you will have unprotected sex. " Joe, a 24 year old truck driver explained that he has unprotected sex in the heat of the moment and once he becomes involved in a sexual act, contact will occur without condom use. " Once you get in a certain mood and you get stuck, I'm in that freak mode; there ain't no protection. " Harry, a 23 year old landscaper, described how he was told that his partner has AIDS but he continues to have sex with her but gets himself tested for HIV every 6 months. He, in effect, continues to take a chance and uses the test as a way to keep himself informed. " In fact, I had two girls and they telling me that you know your partner tell you where they got AIDS from her, but yet and still you get yourself checked and you don't come up with AIDS. You know there's a doubt there somewhere. Every six months I have myself checked for almost a year because I supposed to get it from her. How did I miss it? I don't know. " Barriers to practicing safe sex Like their perceptions about the reasons for the rapid spread of HIV and the situations that have placed them at risk for infection, one-third of men stated that substance abuse is a barrier to safe sex practices. Other reasons stated as barriers were lack of money for condoms and judging a partner on appearance alone. About one-fifth of participants said they had no barriers to practicing safe sex. Karl believes that alcohol and drug use, specifically crack cocaine, creates a feeling of power over women which results in the women failing to negotiate safe sex practices. He said: " I think alcohol and drugs, mood altering, mind altering, especially with that crack cocaine. When you get that crack cocaine, the fist thing you get, and you take you a hit, you automatically assume power. You know with that crack, you have power over women. You know you can make them do what you want them to do. " Freddy, a 28-year old laborer, explained that when a choice between drugs and condom has to be made, he will choose drugs. He stated that he did not have money to buy both condoms and drugs: " I don't have enough money to go out and buy condoms that everybody wants. I need to spend money on condoms or I'm a spend it on getting the next hit. " Although infrequently stated, Jason said that he had, in the past, judged a woman's HIV/AIDS status based on her appearance: " What prevented me from practicing safe sex in the past was, like I said, the way she looked. You feel that you know her, but now I see that you can't go on that. " Barry, a 26-year old construction worker, stated that he always uses protection. He said: " Like I said earlier, if I can't find a plastic baggie or the plastic that you break on in, I wear a sandwich bag, hey, if I can't use that, I won't mess around. I won't do it! " Facilitators to safe sex practices Close to one-half of men were motivated to practice safe sex because they were personally acquainted with someone who had been affected by AIDS and had seen its effects firsthand. A quarter of men was adamant about remaining disease free or had previously acquired sexually transmitted infections. About 15% simply stated that life was their motivation. Barry has seen the effects of AIDS firsthand and the images have caused him to use condoms. He said: " I was about to do it with no condom and that seriously woke me up, put a stop and then God stepped in and then about a year later, she passed. The loss of friends and hearing people over the news. Their body just shrivel up cause the body can't hold their bladder, their bowels I mean it's the ugliest sight to see. You know and that's enough to wake you up. " Austin, who in the past had taken strangers home and engaged in unprotected sex, explains that his concern for his future facilitates his use of safe sex practices: " My motivation is AIDS, not contracting AIDS. The fact that I don't have any kids or I'm not married ... something that will affect my future. I mean I want kids. I wanna be married. I mean if I have AIDS or what have you, don't none of that happen. " Jason reiterates Austin's comments quite simply when he states: " I want to live. That's the bottom line. " Motivation for safe sex practices was not always associated with action even in the face of prior sexually transmitted infection. Joe adds: " One thing that motivates me is, I don't necessarily take heed to it though, but the only venereal disease that I have had in the past is them crabs or whatever you want to call it. One thing that motivates me is that I had got crabs and I felt dirty, I mean I felt real bad. " Strategies for preventing HIV infection In light of their experiences, participants were asked to describe what they would recommend to curtail the spread of AIDS in the African-American community including what specific governmental, media and community interventions they would consider effective. A variety of recommendations were offered, but the majority believed that more education is needed. Other men suggested that condoms are distributed without charge and consistent condom use is urged. To recruit African Americans into HIV training programs, financial and social incentives were highly recommended as were the use of a community-based, community-friendly approach to program recruitment and implementation. Trevor, a 27-year old who is unemployed said: ' The way that AIDS can be prevented is just through education, you know, education and that's basically the only tool that we have is education. Uh, I think the parents play a good role in it [education] too. " Wally, a 20-year old bricklayer adds: " There's only one way and it's through education and I mean education will sum it up. We have to educate on drugs, we have to educate on protective sex. I mean, education is it. " Freddy echoes the sentiments of Trevor and Wally regarding education but also discussed the need to provide condoms: " We need to get more condoms out on the streets. All STDs have to be prevented so we have to get more education on the streets, not just about AIDS because they [AIDS and other STDs] work hand in hand. " Incentives such as gifts and social events were recommended as a means to get African-American participation in HIV prevention and risk reduction interventions. Wally stated the need to provide incentives to increase program participation: " You can give and receive at the same time. What I'm saying is you may have to start having a little gift or something to get them to come in the first few times, or a meal. " Barry adds: " You gotta give something to get something. Another thing, have you a little barbecue, have it sitting out or something and you can get a whole lot of people. And believe it or not, a crowd of people will open up to a conversation, too. " Karl stressed incentives and programs for adolescents: " Sometimes especially with the younger generation, you have to give them something that they want. Give them caps, give them T-shirts, but at the same time, push the condoms. " Outreach workers were advised to eliminate formality and go into the community to provide their programs. They should make an effort to ensure the target community members are comfortable with their presence. Jason said: " And go into the neighborhoods and walk around and talk to folks. People have a tendency to stay away from people in a suit and tie in a neighborhood. " Gene, a 20-year old laborer added: " Instead of making it formal, just come in off the streets. Go to the people and give them that incentive and sit down and talk to them. " It is believed among two-thirds of men participating in these discussion groups that the government can assist AIDS prevention by firstly, providing funding for community-based education programs, secondly, providing convenient and free testing, and thirdly, free condom distribution. Other suggestions included quarantining or having a salient means of identifying those affected by HIV and AIDS, providing more medication and treatment for those affected, and having stiffer penalties for those who knowingly transmit the virus. Jason believes in the efficacy of outreach programs that provide free testing and incentives: " Just have three of four vans, just go to different neighborhood, have free testing, give T-shirts. " Terry believes in the community approach to governmental intervention: " I think the greatest weapon we have against AIDS is knowledge and we must bring some type of community-based program. " Roger asserted his belief that the entire US population should be tested and those with HIV and AIDS identified and placed in isolation: " Either quarantine, mandatory testing for HIV for the whole US population. You go to get tested. You're going to a quarantine island for the rest of your life. " Terry believes that the government should impose stringent punishment against persons who knowingly infect others with HIV: " I think we can make stiffer penalties for people with AIDS that are willingly passing the disease on. Probably the government can open some more clinics or help the doctors find a cure. " To create a video that addresses HIV prevention, close to one-half of participants recommended using a sensational approach that includes those morbidly affected by HIV and AIDS. They believe that the video should include footage of people in pain, being shunned by friends and family, and suffering tremendous physical pain. They also recommended testimonials from infected persons. A quarter of respondents believed that the video should be educational providing information about transmission and prevention. The remainder suggested the use of rappers to promote the message of prevention and acknowledgement that HIV happens to African Americans. Karl who has a college degree stated that a video developed for AIDS prevention should include the disease's effects on the person: " They need to see those people suffering to let them know that this is how you're going to end up if you don't begin to practice safe sex. But also how people are being discriminatory to you, treating you like crap like you're this or that or you're contagious or something. " Roger has similar views but adds that education about high risk behaviors should be included: " Stay away from the dope and don't go tricking and also go to the hospitals. A lot of African Americans are not educated on the full blown AIDS and how it destroys the major organs in your body. And let them see first hand on what you're dealing with here. " Jason would like a video that includes the various modes of transmission as well as prevention information: " Just about everything that you can get on it from drug use to safe sex to having sex with the same partner. You know it's different ways you can catch it and a lot of people don't know that. " When asked what can be done to recruit African Americans to HIV training programs, most men suggested the use of participation incentives, especially those that are financial and preventive (condoms). The men also recommended recruiting participants from and having the programs take place within the communities in which the target population resides. Wally said: " You may have to start with a few passes; they can be little small things. Then everyone would gradually grab a hold and then their minds would be off on what you're giving personally and they could see what they can receive. You can give and receive at the same time. What I'm saying is you may have to start having a little gift or something to get them to come in the first few times, or a meal. " Jason suggests community-based recruitment: " Stand at every corner. AIDS is gonna get too bad in a couple of years. We don't want it to be too late before we say, "Hey, let's put education groups here and people just stop here and get rubbers. " Conclusions The results indicate that the respondents taking part in this study believe that HIV and AIDS are continued threats to the African-American community because of sexual risk taking behavior, that is, failure to use condoms. Further, African-American men are having sex without condoms when having sex with women often when they are under the influence of alcohol or other mind-altering substances and they are having sex with men while incarcerated and become infected and once released resume unprotected sexual relations with women. According to the men, substance abuse is an important part of the problem of HIV in the African-American community. This is in keeping with research that shows that drug use, especially crack cocaine, is linked to sexual risk taking among African Americans and to increased likelihood of becoming infected with other sexually transmitted diseases (STDs) including HIV [ 27 , 28 ]. African Americans are disproportionately affected by STDs [ 1 ] and African-American crack users are more likely to have multiple partners and to participate in drug for sex exchanges [ 29 , 30 ]. According to the men interviewed, they were less likely to use a condom when under the influence. Research shows that crack cocaine users may not perceive sexual risk taking as an important self-threat compared to other social and health issues they confront on a daily basis [ 27 ]. Although most respondents indicated they have knowledge of behaviors that place one at risk for HIV and they are motivated by reasons such as life and not wanting to contract an STD, they still fail to consistently use condoms and may even have sex with someone they know is HIV positive and they continue to suggest that more educational programs are needed. This shows that knowledge and motivation are not enough for behavior change to occur. These men need education but education that is provided in a way that has been different from what has been offered in the past because it seems not to change their behaviors. Interventions for men should address condom use, condom availability, skills for using condoms, and eroticizing condoms. Substance use programs should incorporate sexual risks. Men in the present study strongly recommended that HIV/AIDS videotaped messages should include footage of the sensational effects of the disease. There is a trend toward reality broadcasting in the US and development and field testing of such a videotape for AIDS prevention might prove worthwhile [ 31 ]. Low-income women taking part in a similar focused group discussion also recommended the use of sensationalism. Contrary to research indicating fear or sensationalism does not work [ 32 - 34 ], based on these results it may be worthwhile to field test such a videotape with groups of African Americans to evaluate the utility of such a teaching tool. If the tide of HIV and AIDS infection among African Americans is to be reduced, programs must incorporate culturally relevant contextual information presented to the target audience in a setting and in a manner that addresses their norms and beliefs and provides them the knowledge and skills needed to make correct decisions. The message could be presented using behavioral journalism, an approach espoused by McAlister [ 35 ] that offers a balance between the message source and the audience constructing messages within a theoretic framework and tailoring them to specific audiences. It may well be that this approach works with low-income African Americans who are engaged in high risk behaviors. Such a program could also include information about transmission routes, a subject identified as important to discuss when developing training programs. Prior research has identified the leading causes of HIV infection among African-American men that includes sexual contact between African-American men who have sex with men, injection drug use and heterosexual contact. These high-risk behaviors are exacerbated by the high rate of poverty among African Americans that may be due to high rates of unemployment and associated drug related economic activities, injection drug, and substance abusers' engagement in unprotected sex when they are under the influence of drugs and alcohol [ 1 , 13 ]. The present study sought to identify the sociocultural contexts of sexual risk taking among African-American men and to gather information that could be used to develop and disseminate a videotape-based HIV prevention intervention for African-American men. Although the CDC identifies the leading cause of HIV transmission among African-American men as homosexual contact, none of the participants in the present study acknowledged this cause [ 1 ]. Men of color who have sex with other men often do so without disclosure and may consider themselves heterosexual [ 36 ]. Further research is needed in this area to determine how best to address this issue within the cultural context of African Americans and non-disclosing men. Lack of money for condoms was identified as a barrier to condom use. Many of the men believed that condoms should be a part of any HIV training program and that they could also be given as participation incentives for coming to a training program. Prior research has indicated that a combined approach to HIV prevention that includes HIV testing and counseling, educational and behavioral interventions delivered through community outreach, condom distribution and substance abuse treatment are effective means for reducing transmission [ 37 ]. However, such interventions have not yet been widely implemented in a sustained and integrated fashion. Although most of the men had completed high school or beyond, they were mainly homeless or living in substandard housing. Their conditions of poverty may exacerbate their risk for infection because they may engage in illegal or high-risk behaviors. The barriers these men face related to poverty should also be addressed when developing training programs. This may call for the involvement of social service workers. Like programs developed for African-American women, they must be presented in settings that are familiar, comfortable and easily accessible to the target audience and incorporate culturally relevant contextual information [ 38 ]. Before developing programs, health educators should become familiar with the facilitators and barriers to HIV risk reduction behaviors among this population. They should also recognize that African Americans are a heterogeneous group and that not all messages will work with all audiences. The messages must be tailored and celebrate the diversity that exists within this population. Interventions such as videotapes should have mass appeal yet contain contextual, cultural, and gender specific messages. There are several limitations to this study. Qualitative data collection methodology was employed and it is therefore not possible to make assumptions or draw inferences. Next, the data cannot be generalized to other African American men. The participants were homeless and low-income and selected using a convenience sampling approach. The goal of this study, however, was to recruit low-income African-American men because by virtue of their income status they may be at increased risk for HIV infection. Although the findings from this research are limited, they may provide a foundation for conducting future research among and developing a videotape-based HIV intervention for low-income African-American men. Competing Interests The author(s) declare that they have no competing interests. Authors' contributions EJE, AFM and GOO conceived and designed the study. EJE, AFM, RJP jointly planned and executed the data analyses. EJE and AFM wrote the paper with assistance from RJP, GOO and NIO. Table 1 Focus group guide questions 1. Do you perceive AIDS as a threat to the African American community and why? 2. What are the perceived roles of men in heterosexual relationships in the African American community? 3. What are the expectations for personal and sexual responsibilities for contraception and sexually transmitted diseases prevention among African American men? 4. What situations have placed you at risk for HIV infection in the past? 5 How have alcohol and drug use placed you at risk of HIV infection? 6. What are the things that motivate you to practice safe sex? 7. What are the things or barriers that prevent you from practicing safe sex 8. Why do you think that AIDS is spreading so rapidly in the African American community? 9. What information do you think we need to include in a videotape developed to train African Americans about HIV prevention that will encourage them to watch the videotapes? 10. Do you have any other suggestions on how AIDS can be prevented in the African American community? 11. What can we do to get people to sign up for focus groups such as this one and also get them to participate in HIV/AIDS training programs? 12. What can we do to make these training programs most useful to you? Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544892.xml

523835

How Should the Health Community Respond to Violent Political Conflict?

Violent political conflict is on the front pages, in Iraq, Afghanistan, and Sudan. This provocative piece discusses lessons we can learn from past conflicts in dealing with future ones

Violent political conflict, and its impact, is again on the front pages—in Iraq, Afghanistan, and Sudan. While the situation in Darfur is now particularly urgent (see sidebar) [ 1 , 2 , 3 , 4 , 5 ], there are many other settings in which complex political emergencies are undermining health service provision and threatening human rights. Such emergencies have a direct impact on health (see Table 1 ). They also impair the functioning of health systems through, for example, destruction of infrastructure (such as clinics and vehicles), reduced access to medicines, death of health workers, and weakened national capacity for health policy-making [ 6 ]. Table 1 Examples of the Direct Impact of Conflict on Health Adapted from [ 7 ], with permission from the World Health Organization Such violent political conflicts stir us—the global health community—to discover our own humanity amidst the bloodshed. How best should we respond? Despite unique features in each setting, we must learn lessons from previous conflicts to help guide our response to current and future ones. There are six key lessons that emerge from studying health in conflict settings. Lessons from Conflict Settings Violent conflict is driven by politics and economics [ 7 ]. Complex political emergencies (1) occur within and across state boundaries, (2) have political antecedents typically relating to competition for power and resources, (3) are protracted in duration, (4) express existing social, political, economic, and cultural structures and cleavages, and (5) are often characterised by one sector preying on other parts of the community [ 8 ]. Damage to health is not just a side effect but may be the objective for violent groups. In complex political emergencies, we can typically identify three groups: the winners, the ‘conflict entrepreneurs’ (who seek the perpetuation of conflict because they profit economically or politically), and the losers, whose lives and livelihoods are imperilled. Humanitarian and relief agencies increasingly recognise that belligerents may seek to control or manipulate the inflow of humanitarian and relief resources [ 9 ]. A political economy perspective helps identify those interests, which may impede the transition to peace [ 7 ]. Sudan—Conflict and Health The current crisis in Darfur reflects a devastatingly acute episode in the chronic internal conflict that has plagued Sudan since 1983. The cost of this conflict has been enormous: over 2 million lives lost, over 628,000 refugees from Sudan in neighbouring countries, and over 4 million people internally displaced [ 1 ]. In southern Sudan, the conflict has led to widespread ill health and has severely compromised the well-being of women and children. Indicators of immunisation, nutrition, primary school completion, and antenatal care are among the worst in the world. About 95,000 children under five years old died last year, most from preventable disease [ 2 ]. Statistics from UNICEF are chilling: ‘A girl born in southern Sudan has a better chance of dying in pregnancy or childbirth than of completing primary school….One in nine women dies in pregnancy or childbirth but only one in a hundred girls completes primary school’ [ 2 ]. Communities in Darfur face ongoing violence from militia supported by the government of Sudan. The fighting has resulted in large-scale destruction of villages, rape, and kidnapping. About 15,000–30,000 lives are estimated to have been lost from January 2003 to June 2004 [ 3 ]. Surveys by Médecins Sans Frontières found death rates of three to five per 10,000 people/day in Mornay and Zalinge villages (the emergency threshold level is set at one death per 10,000/day) [ 4 ]. Over 300,000 people are at risk if humanitarian access remains restricted. Of displaced Darfurians, 90% need shelter and latrines, and over half lack access to primary health care [ 3 ]. Food insecurity is widespread and is being used as a ‘weapon of war’ [ 5 ] resulting in widespread nutritional problems. Despite widespread concern, information gaps abound, and humanitarian agencies report having access to only a fraction of those most affected. Yet, genocide is taking place in real time. Appreciating context is crucial. The nature of the conflict—its background, history, and the different forms of violence involved—will greatly influence health outcomes. Most conflicts are today intra-national rather than international [ 10 ]. Internal conflicts affect populations through forced migration, violence, and human rights abuses including torture, disappearances, and rape. The forms of violence and types of health damage relate to the phase of the conflict, the sophistication of weapons used, the degree of involvement of regular military forces, the extent of terrorism employed, and the extent to which genocide is intended. Ongoing insecurity and instability may be present even after the ostensible end to the conflict, as in latter-day Afghanistan and Iraq. Challenges to governance, to service delivery, and to the reestablishment of livelihoods may persist for years. A 2003 survey in Iraq found that despite the brief duration of the war and the intent to spare hospitals and clinics from direct attack, many people suffered in the post-war period, primarily as a result of disruption to civil order [ 11 ]. Recent reports highlight the difficulties of re-establishing the health system in Iraq—partly because of a failure to appreciate the cultural and health services context [ 12 ]. Better care can save lives. Emergency relief efforts are increasingly based upon empirical evidence, and priority health issues are much more effectively addressed than previously. Emphasis is typically placed upon disease surveillance, immunisation, control of infectious diseases, reproductive health, water and sanitation, shelter, and nutrition [ 13 ]. Mental health, sexually transmitted infections, and HIV have recently attracted additional attention. Standards have improved, can be further improved, and warrant widespread dissemination and application. The more-established humanitarian agencies have accepted that their relief efforts must be as evidence-based as possible. This principle should also apply to the post-conflict period, during which the health of affected communities continues to suffer [ 14 ]. About 70% of structures were destroyed in Dili, East Timor, in the violence wrought by Indonesian militia after the referendum in 1999 (Photo: Anthony Zwi.) We need enhanced accountability for humanitarian action. Despite a developing evidence base for health-related humanitarian action, evaluations of humanitarian activities have found ongoing problems. These include poor standards of delivery, duplication of efforts by different agencies, lack of coordination, and failing to learn from prior experience. The Sphere Project has advocated minimum standards for the delivery of humanitarian assistance, and has established a “Humanitarian Charter” ( http://www.sphereproject.org ). The project's objectives and achievements have been to improve the quality of humanitarian action and promote a movement concerned with the rights and dignity of those caught up in war and disaster [ 15 , 16 ]. The Active Learning Network for Accountability and Performance in Humanitarian Action ( http://www.alnap.org ) seeks to ensure that lessons are learned, distilled, and disseminated. At a meeting in Stockholm in June 2003, key international donors committed themselves to ‘good humanitarian donorship’, which recognises the importance of promoting standards in humanitarian action [ 17 ]. However, recent sober reflection suggests that donors and humanitarian agencies could do better: ‘An ailing humanitarian enterprise is labouring under pressures from the external environment over which it has little control, while struggling with issues internal to its own function for which it should take greater responsibility’ [ 18 ]. Militarization of humanitarian efforts is problematic. Multinational military forces have played a major part in recent conflicts in Kosovo, East Timor, Sierra Leone, Iraq, and Afghanistan. The military has become increasingly involved not only in waging war but also in seeking to win the peace; it is increasingly active in delivering emergency relief. It not only provides services—sometimes necessary to deliver needed relief—but also seeks to ‘win hearts and minds’ while operating within structures responsive to military and foreign policy directives. The result has seen a blurring of the separation between military and humanitarian efforts [ 19 ]. This can make humanitarian agencies a target—recent examples include the bombing of United Nations headquarters and the International Committee of the Red Cross in Iraq and the recent, reluctant withdrawal of Médecins Sans Frontières from Afghanistan following the murder of five aid workers [ 20 ]. Emerging evidence and good practice in civil-military cooperation highlights the importance of (1) promoting needs-based assistance free of discrimination, (2) civilian-military distinction in humanitarian action, (3) independence of humanitarian organisations from political pressures and interference, and (4) the security of humanitarian personnel [ 19 ]. The transition from emergency relief to development is poorly managed. The objectives of humanitarian relief activity (saving lives and livelihoods) differ from those of development (building sustainable systems, promoting equity, building systems of governance, and eradicating poverty). In each phase there are different actors, strategies, and approaches. The increasing politicisation of humanitarian intervention [ 21 , 22 ] brings threats and dangers, undermining key humanitarian principles. The balance between relief and development will vary over time and place; getting the balance right and adequately resourcing the transition warrants careful research, documentation, reflection, and the commitment of appropriate longer-term funding. What Gaps Remain in Our Knowledge? Despite the knowledge we have gained on responding to violent political conflict, many important gaps remain. In southern Sudan, the conflict has affected the well-being of women and girls (Illustration: Margaret Shear, Public Library of Science.) We still do not hear the voices of those most affected or of the service providers seeking to assist. The reality of people's experiences is inadequately appreciated [ 23 ]; whatever we learn of their fears, challenges, and suffering is typically represented and reported through sanitised language and media. The language used dehumanises ‘the enemy’ and blunts our senses to the reality of atrocity and to the negative effects of our own countries’ interventions. Within the health sector, ensuring that we hear the voices of service providers and carers will help bring home the reality of system disruption, destruction, and damage and will simultaneously document the mechanisms and potential for effective responses. The new communication technologies provide immense opportunity to ensure that experience is placed in the public domain from where lessons can be drawn and better practice promoted. We know little about how communities and systems survive adversity. In most settings, the inherent ability and ingenuity of people and systems allows them to withstand instability and insecurity. Health personnel and health systems could play a valuable role in these fragile settings—assisting individuals, communities, and systems to further develop their coping strategies, adaptations, and responses. But whether health systems do so and how is unclear. Failing to support and maintain these systems may result in much greater challenges when we seek at a later stage to resuscitate them. We also know relatively little about whether the health sector can indeed make a special contribution to building the peace. While it has been forcefully argued that the health sector is uniquely placed to play a role in peace building [ 24 ], the evidence for this remains limited [ 25 ]. We know little about how health workers see and respond to these challenging roles. The health sector could play a role in demonstrating the values and priorities of government, reflecting the relationship between those with and without resources, and the relationship between those who do and do not have protection. In the aftermath of major periods of violence, the health sector could also help to ensure that the structural inequities that preceded the violence and may have contributed to it, are not reinforced and the same injustices not recreated. But, engagement around health is not always positive: the health system is open to abuse and has been abused by repressive systems. From Learning Lessons to Sound Policy Perhaps the most important gap of all is between observing lessons and putting them into practice. We urgently need to transform evidence and experience into sound policy. We need more sophisticated policy analyses, more sensitive policy-making, and more relevant research. Policy in these difficult areas will never be entirely evidence-based—often it will at best be ‘evidence-informed’. Our objective must be to promote organisations and systems that are able to reflect on experience, work with partners to critically analyse and learn, and thereby formulate better responses. Violent political conflict will continue to challenge the global health community. International policy-makers and funders must support more extensive documentation and reflection: the building blocks of better practice.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC523835.xml

555747

Differences between men with screening-detected versus clinically diagnosed prostate cancers in the USA

Background The advent of prostate specific antigen (PSA) testing in the United States of America (USA) has led to a dramatic increase in the incidence of prostate cancer in the United States as well as the number of men undergoing aggressive treatment with radical prostatectomy and radiation therapy. We compared patient characteristics and treatment selection between American men with screening-detected versus clinically diagnosed prostate cancers. Methods We evaluated 3,173 men with prostate cancer in the USA. Surveys and medical records provided information on demographics, socioeconomic status, comorbidities, symptoms, tumor characteristics, and treatment. We classified men presenting with symptoms of advanced cancer – bone pain, weight loss, or hematuria – as "clinically diagnosed"; asymptomatic men and those with only lower urinary tract symptoms were considered "screening-detected." We used multivariate analyses to determine whether screening predicted receiving aggressive treatment for a clinically localized cancer. Results We classified 11% of cancers as being clinically diagnosed. Men with screening-detected cancers were more often non-Hispanic white (77% vs. 65%, P < 0.01), younger (36% < 65 years vs. 25%, P ≤ 0.01), better educated (80% ≥ high school vs. 67%, P < 0.01), healthier (18% excellent health vs. 10%, P < 0.01), and diagnosed with localized disease (90% vs. 75%, P < 0.01). Men with screening-detected localized cancers more often underwent aggressive treatment, 76% vs. 70%, P = 0.05. Conclusion Most cancers were detected by screening in this American cohort. Appropriately, younger, healthier men were more likely to be diagnosed by screening. Minority status and lower socio-economic status appeared to be screening barriers. Screening detected earlier-stage cancers and was associated with receiving aggressive treatment.

Background Prostate-specific antigen (PSA) testing was introduced in the United States of America (USA) in the late 1980s with Federal Drug Administration (FDA) approval for prostate cancer surveillance [ 1 ]. However, the test indications were soon expanded to include prostate cancer screening. By the early 1990s, the American Urologic Association and the American Cancer Society were recommending PSA testing, along with digital rectal examination (DRE), as part of annual prostate cancer screening [ 2 , 3 ]. The advent of PSA testing led to a dramatic increase in the incidence of prostate cancer in the USA, with the number of new cases rising from 152,811 in 1990 to over 230,000 in 1992 [ 4 , 5 ]. During the past decade, the number of American men undergoing aggressive treatment with radical prostatectomy and radiation therapy also increased substantially [ 6 , 7 ]. Urologic screening studies provide the most comprehensive information about the men undergoing PSA screening [ 8 - 10 ]. Several trials have taken place in both Europe and the USA. In general, study subjects usually were recruited through advertisements and they were screened with combinations of PSA, DRE, and transrectal ultrasound. The average age of these study participants was in the mid-60s, and minority subjects were not well represented. Minimal data were provided on symptoms, comorbidity, or socioeconomic status. Among American men diagnosed with clinically localized prostate cancers, approximately 90% underwent treatment with radical prostatectomy or radiation therapy. Population-based data on PSA screening are largely unavailable, including information on the proportion of prostate cancers diagnosed by screening, the demographic, socioeconomic, and clinical characteristics of men with screening-detected cancers, and the association of screening detection with treatment decisions. We used data from the United States-based Prostate Cancer Outcomes Study (PCOS) to 1) determine the proportion of screening-detected prostate cancers in a population-based cohort, 2) compare baseline demographic, socioeconomic, and clinical characteristics between men with screening-detected versus clinically diagnosed cancers, and 3) determine whether men with screening-detected clinically-localized prostate cancers were more likely to undergo aggressive treatment (radical prostatectomy or radiation therapy). Methods Study population The American National Cancer Institute instituted the PCOS in 1994 to measure practice patterns and health-related quality of life among men diagnosed with prostate cancer in the United States. Methods for this multi-site, longitudinal project are described elsewhere [ 11 ]. Briefly, PCOS subjects were men histologically diagnosed with prostate cancer between October 1, 1994 and October 31, 1995. Subjects were identified using a rapid case ascertainment system by the six participating National Cancer Institute Surveillance, Epidemiology and End Results (SEER) cancer registries (Atlanta, Georgia metropolitan area; Los Angeles County California; King County, Washington; Connecticut; Utah; and New Mexico). Eligible subjects were residents of the areas covered by these registries at the time of diagnosis and were between the ages of 39 and 89 years, except in King County, where only men over 60 years were eligible. The institutional review board of each participating institution approved the study. Eligible patients were sampled within strata of age, race/ethnicity, and tumor registry to approximate a sample representative of the United States population of prostate cancer patients. The PCOS oversampled younger men and minorities and excluded patients with race/ethnicity other than non-Hispanic white, African American, or Hispanic, because their sample sizes were small. A total of 11,137 men with prostate cancer comprised the eligible patient population for the study and the PCOS randomly selected 5,672 of these men. Among these selected patients, 3173 (55.9%) completed a health-related quality-of-life survey questionnaire 6 months after initial diagnosis. We used survey and medical record data collected from these subjects to evaluate differences in patient characteristics and treatments between men with screening-detected cancers and those who were diagnosed clinically. Responders to the PCOS survey were younger than non-responders and more likely to be non-Hispanic white and have a higher socioeconomic status. A substantial proportion of the responders had regional stage and moderately differentiated cancers, while non-responders had a greater proportion of distant stage and poorly differentiated cancers. Responders also were more likely to receive radical prostatectomy [ 11 ]. Data collection Investigators contacted eligible subjects by mail and/or telephone requesting them to sign a release form allowing review of all medical records from any physicians and facilities diagnosing and/or providing care for prostate cancer. Records were obtained from private and public hospitals, freestanding radiological or surgical centers, Veterans Administration hospitals, Health Maintenance Organizations, and private physician offices. Certified Tumor Registry abstractors collected baseline information on demographics, clinical symptoms before diagnosis (systemic and urinary), comorbidity, diagnostic procedures and results (including PSA levels and digital rectal examination findings), clinical staging, tumor characteristics, and treatment details. The PCOS re-abstracted a random sample of 5% of records to assess and correct any systematic coding errors. The PCOS also collected data on general and disease-specific measures of health-related quality of life, symptoms, comorbidity, and specific treatments received for prostate cancer using a mailed self-administered questionnaire. Most respondents completed the self-administered questionnaire (91%); those who did not return the questionnaire were contacted by telephone and asked to complete the survey by telephone or in person. Subjects were asked to recall their health-related quality of life and symptoms, including the domains of urinary, bowel, and sexual function, just before their prostate cancer was diagnosed. Demographic and socioeconomic questions from this survey were used to determine race/ethnicity, employment status, educational level, household income, insurance status, and marital status. A question assessing comorbidity asked about 12 medical conditions that were likely to affect prostate cancer treatment decisions and long-term quality of life. The conditions were derived from the Charlson index as well as the expert opinion of the PCOS investigators [ 12 ]. If the patient reported being told by a doctor that he had cerebrovascular disease, inflammatory bowel disease, liver disease, or ulcers, he received one point on his comorbidity score for each condition. If the patient reported that any of eight conditions – arthritis, diabetes, depression, hypertension, chest pain, heart attack, heart failure, or chronic lung disease – limited his activity or required prescription medications, he received 1 additional point for each of these conditions. In the analyses, comorbidity scores were divided into the categories of 0, 1, 2, and greater than or equal to 3 points. We assigned screening status using information from the medical record abstract and the patient questionnaire. We considered men presenting with symptoms consistent with advanced prostate cancer, including bone pain, weight loss or hematuria, to be "clinically diagnosed." We initially created separate categories for men with only irritative or obstructive symptoms consistent with benign prostatic hyperplasia and an asymptomatic group who had neither prostate cancer nor lower urinary tract symptoms. Clinical cancer stage was based on an algorithm using information abstracted from medical records. The algorithm was necessary because the community-based medical records were not detailed enough to classify cases by TNM (tumor, node, metastases) staging [ 13 ]. The algorithm defined T1 tumors as confined to the prostate with a normal digital rectal examination and no positive scans (magnetic resonance imaging, computed tomography, bone scan) or evidence of metastases. T2 tumors were defined as confined to the prostate, with abnormal or suspicious digital rectal examinations, but no positive scans or evidence of metastases. We defined clinically localized cancers as either T1 or T2 tumors. Initial treatment, based on medical record abstractions, was defined as treatment received within the first six months after diagnosis. We defined aggressive treatment as either radical prostatectomy or radiation therapy. We defined conservative management as androgen deprivation, either surgical or chemical, or watchful waiting. Statistical analyses Descriptive statistics were calculated for ethnicity/race, age, stage at diagnosis, education, marital status, employment, income, digital rectal exam and PSA results, Gleason score from biopsy or transurethral resection of the prostate, comorbid conditions and self-reported general health status. We used contingency tables to compare men presenting without any symptoms, those with lower urinary tract symptoms alone, and those with prostate cancer symptoms. Although screening is defined as applying a diagnostic test to asymptomatic people [ 14 ], the prevalence of benign prostatic hyperplasia is very high among men at risk for prostate cancer [ 15 ]. We found that the men with only lower urinary tract symptoms were much more similar to asymptomatic men than to men we classified as having clinically diagnosed cancers. Therefore, we also considered cancers diagnosed in men who reported only lower urinary tract symptoms at the time of PSA testing to be "screening-detected." We used this combined screening-detected group to compare baseline characteristics against clinically diagnosed cases and in modeling treatment selection for clinically localized cancers. Logistic regression analyses were used to determine whether screening history was independently associated with selecting aggressive treatment versus conservative management among men with clinically localized prostate cancer. Covariates for this multivariate model, based on previous literature, included age, race/ethnicity, marital status, study site, education, insurance status, annual income, comorbidity, health status, and tumor characteristics [ 16 , 17 ]. We also examined interactions between screening status with age, comorbidity, PSA level, and Gleason score. The results of the logistic regression models are shown as percentages receiving the treatment of interest, adjusting for the independent variables included in the model. These percentages were directly adjusted to the distribution of the variables among the weighted sample used in each model [ 18 ]. The probability of receiving the treatment of interest can then be directly compared across levels of the variables included in the model. All analyses were performed with the Survey Data Analysis statistical package (Research Trial Institute, Research Triangle Park, North Carolina, 1997) to account for the complex survey design. We obtained unbiased estimates of parameters for all eligible prostate cancer patients in the PCOS areas by using the Horvitz-Thompson weight, which is the inverse of the sampling proportion for each sampling stratum (defined by age, race/ethnicity, and study area). A two-tailed P-value of < .05 was considered statistically significant. Results The baseline demographic, socioeconomic, and clinical characteristics of the PCOS subjects are shown in Tables 1 and 2 . The majority of subjects were non-Hispanic white men, older than sixty-five, and married at the time of diagnosis. Socioeconomic status was relatively high; a majority had more than a high school education, and a substantial proportion of subjects had private insurance. Among the study subjects, 10.7% presented with symptoms consistent with prostate cancer and were considered to be clinically diagnosed cases. Nearly two-thirds of subjects had lower urinary tract symptoms, while 30.9% were completely asymptomatic. Overall, 83.1% of men rated their general health at "good" or "excellent" before their cancer diagnosis. We compared baseline characteristics of asymptomatic men, those with lower urinary tract symptoms alone, and men with clinically diagnosed cancers in Table 3 . We found that the characteristics of men with lower urinary tract symptoms alone were closer to the asymptomatic men than to the clinically diagnosed cancer cases for race/ethnicity, socioeconomic status, health status, and cancer grade and stage. When we combined these two groups into a single category of screening-detected cases, we found significant differences between the screening-detected and clinically diagnosed cases. Men with screening-detected cancers were more likely to be non-Hispanic white, were younger age, and had a higher socioeconomic status. They also reported being healthier and were more likely to have early stage disease. We then evaluated whether screening status independently predicted receiving aggressive treatment among the 2796 men who were diagnosed with clinically localized cancer. The primary treatment for these men was radical prostatectomy for 1535 (53.4%), while 518 (20.6%) underwent radiation therapy, 671 (26.0%) were treated conservatively; we had no treatment information for 72 subjects (2.5%). The results of the multivariate analysis are shown in Table 4 . After adjusting for age, race/ethnicity, marital status, area of the country, education, insurance coverage, annual income, comorbidity, self-reported health status, and tumor characteristics, we found that men with screening-detected cancers were more likely to receive aggressive treatment. The adjusted percentage of men with screening-detected cancers undergoing aggressive treatment was 76% (95% CI 0.74, 0.78) vs. 70% (95% CI 0.64, 0.76), in men with clinically diagnosed cancers, OR = 1.5 (95% CI 1.1, 2.3), P = 0.05. Other factors that were significantly associated with aggressive treatment included geographic area, ethnicity, age, marital status, comorbidity, health status, and tumor characteristics. We found no significant interactions for treatment selection between screening status with age, comorbidity, PSA level, or Gleason score. Discussion We found that the majority of cancers (89.3%) in a population-based PCOS cohort were detected by screening. Compared to men with clinically diagnosed prostate cancer, men with screening-detected cancers were younger, more likely to be married, less likely to be a member of a minority group, and in better health. The cancers detected by screening were more likely to be clinically localized and less likely to be poorly differentiated. Among men with clinically localized prostate cancers, those with screening-detected cancers were significantly more likely to undergo aggressive treatment, even after adjusting for demographics, comorbidity, and tumor characteristics Our finding that a high proportion of prostate cancers diagnosed in 1994 and 1995 were detected by screening is consistent with the temporal correlation between the increased use of PSA testing and the increased incidence of prostate cancer in the USA beginning during the early 1990s [ 4 , 5 ]. Although prostate cancer incidence rates decreased for several years in the mid 1990s, more recent data show that incidence rates are again increasing [ 5 , 19 , 20 ] and survey results from the Centers for Disease Control's Behavioural Risk Factor Surveillance System (BRFSS) show that a high proportion of American men continue to undergo PSA testing [ 20 ]. These data suggest that our findings are still relevant for prostate cancers being diagnosed in the USA. We also found that men with screening detected cancers were more likely to have early stage cancers, again mirroring the epidemiologic data showing an increased incidence of early stage cancers and a decreased incidence of advanced stage cancers [ 4 , 5 ]. The majority of screening-detected tumors were moderately to poorly differentiated; however, a significantly higher proportion of clinically diagnosed cancers were poorly differentiated. Previous data, including an analysis of the PCOS cohort, have shown African Americans to be twice as likely as non-Hispanic whites to present with advanced stage cancers [ 4 , 5 , 21 ]. In the current analysis, we found a greater prevalence of ethnic/racial minorities in the clinically diagnosed versus screening-detected cancers. This disparity may reflect ethnic/racial differences in accessing preventive health care services, particularly arising from socioeconomic barriers. This in turn could contribute to disparities in cancer stage at diagnosis [ 22 - 24 ]. However, African Americans also have been reported to demonstrate more skeptical attitudes towards screening [ 25 ] and the stage disparity could be due to racial differences in tumor aggressiveness [ 26 ]. Men with screening-detected clinically localized cancers were more likely to undergo aggressive treatment with radical prostatectomy or radiation therapy than men with clinically diagnosed cancers. The odds ratio for receiving aggressive treatment was statistically significant at 1.5, but the adjusted absolute difference between screening-detected and clinically diagnosed cases was only 6 percentage points. This modest association between screening status and treatment selection suggests that clinical practice may be only partly consistent with the American College of Physicians' view that "aggressive treatment is necessary to realize any benefit from the discovery of a tumor [ 27 ]." Our findings may reflect the scientific uncertainty about whether and how to treat screening-detected prostate cancers [ 28 ]. Our study has some potential limitations. We classified men presenting with symptoms of advanced cancer as being clinically diagnosed. We do not know that these symptoms actually prompted diagnostic PSA testing. However, the tumor registry medical record abstractors are trained to identify the events leading to a cancer diagnosis; they would attempt to record only symptoms consistent with cancer. Classifying PSA as a screening test is also difficult given the high prevalence of lower urinary tract symptoms in older men [ 15 ]. Few members of our study cohort were truly asymptomatic because nearly two-thirds reported lower urinary tract symptoms. However, our classifications for clinical diagnosis and screening detection were internally valid because men diagnosed with symptoms of advanced cancer were significantly more likely to present with advanced stage and more aggressive cancers than the combined group of men who were either asymptomatic or had only lower urinary tract symptoms. Additionally, when we compared demographic and socioeconomic characteristics across groups, we generally found that the men with lower urinary tract symptoms alone most closely resembled the asymptomatic men. Selection bias may have occurred because 44% of the sampled patients did not complete the 6-month survey. Responders were younger than non-responders, more likely to be non-Hispanic white, had higher socioeconomic status, had earlier stage disease, and were more likely to receive radical prostatectomy. Results may be less generalizable to older men, those with lower socioeconomic status, or members of racial/ethnic groups other than non-Hispanic white. However, these were also the groups who were less likely to have screening-detected cancers. We do not believe that including these non-responders would have altered our findings on the differences between screening-detected and clinically diagnosed cancers. However, based on their demographics, socioeconomic status, and advanced disease stage, the non-responders were not likely to have a high proportion of screening-detected cancers and thus we may have overestimated the proportion of screening-detected cancers. Another potential limitation arose from asking subjects to recall their baseline symptoms 6 months after diagnosis. Recall errors could lead us to misclassify screening status. However, Legler and colleagues prospectively studied a subset of PCOS subjects and found high concordance for symptom recall at 6-months after diagnosis compared with reports at the time of diagnosis [ 29 ]. Finally, we may have had incomplete symptom data, particularly for questions appearing only in the medical record abstract. The abstracts would report a symptom if it appeared in the medical records; the absence of a symptom could be due to either the patient being asymptomatic or the physician's failure to ask about or record the symptom. We performed a sensitivity analysis by using only subject reported symptoms from the survey, then only symptoms reported on the medical record abstract, and then ultimately using a combination of both sources. The results for all analyses were essentially the same. Conclusion The great majority of prostate cancers diagnosed in our study cohort were detected by screening. Appropriately, younger and healthier men were more likely to be diagnosed by screening. Minority status and lower socioeconomic status appeared to be screening barriers. Screening detected earlier stage and less histologically aggressive prostate cancers. After adjusting for baseline demographic, socioeconomic, clinical, and tumor factors, men with a screening-detected clinically localized cancer were slightly more likely to receive aggressive treatment, either radical prostatectomy or radiation therapy, than men with clinically diagnosed cancers. Competing interests The author(s) declare that they have no competing interests. Authors' contributions RMH and ALP initiated the project. SNS co-ordinated the data collection and was responsible for the data analyses. RMH, SNS, DE and ALP prepared 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/PMC555747.xml

528784

Germinating fission yeast spores delay in G1 in response to UV irradiation

Background Checkpoint mechanisms prevent cell cycle transitions until previous events have been completed or damaged DNA has been repaired. In fission yeast, checkpoint mechanisms are known to regulate entry into mitosis, but so far no checkpoint inhibiting S phase entry has been identified. Results We have studied the response of germinating Schizosaccharomyces pombe spores to UV irradiation in G1. When germinating spores are irradiated in early G1 phase, entry into S phase is delayed. We argue that the observed delay is caused by two separate mechanisms. The first takes place before entry into S phase, does not depend on the checkpoint proteins Rad3, Cds1 and Chk1 and is independent of Cdc2 phosphorylation. Furthermore, it is not dependent upon inhibiting the Cdc10-dependent transcription required for S phase entry, unlike a G1/S checkpoint described in budding yeast. We show that expression of Cdt1, a protein essential for initiation of DNA replication, is delayed upon UV irradiation. The second part of the delay occurs after entry into S phase and depends on Rad3 and Cds1 and is probably due to the intra-S checkpoint. If the germinating spores are irradiated in late G1, they enter S phase without delay and arrest in S phase, suggesting that the delay we observe upon UV irradiation in early G1 is not caused by nonspecific effects of UV irradiation. Conclusions We have studied the response of germinating S. pombe spores to UV irradiation in G1 and shown that S phase entry is delayed by a mechanism that is different from classical checkpoint responses. Our results point to a mechanism delaying expression of proteins required for S phase entry.

Background Checkpoint mechanisms are important for cell survival and genetic stability. They prevent cell cycle transitions until previous events have been completed or damaged DNA has been repaired [ 1 ]. Checkpoint pathways and proteins are evolutionarily conserved from yeast to man, underlining their importance in maintaining genomic integrity. In fission yeast several checkpoint pathways monitor the status of the DNA and arrest the cell cycle in response to DNA damage or inhibition of DNA replication [ 2 , 3 ] They include mechanisms to inhibit mitosis when the DNA is damaged (the G2/M checkpoint) or when S phase has not been completed (the S/M checkpoint) as well as a mechanism to inhibit ongoing DNA replication when the DNA is damaged (the intra-S checkpoint). Screens designed to reveal elements of the checkpoint pathways have led to the identification of the so-called checkpoint rad genes as well as crb2 / rhp9 , mrc1 , chk1 and cds1 [ 4 - 13 ] The checkpoint rad genes consist of rad1 , rad3 , rad9 , rad17 , rad26 and hus1 (reviewed in ([ 14 ]). Rad3 is a member of the phosphatidylinositol 3-kinase family of proteins and the closest mammalian homologue is the ATR ( AT M and R ad3 related) protein [ 15 , 16 ]. Rad3 forms a complex with Rad26 and this association is required for activation of Rad3 kinase activity in response to DNA damage or replication arrest [ 17 , 18 ] The Rad1, Rad9 and Hus1 proteins have similarities to PCNA, the sliding clamp of the replicative DNA polymerase, and the three proteins may form a similar ring-shaped structure [ 19 - 21 ]. Rad17 has similarities to all five subunits of replication factor C [ 22 ], a complex which loads PCNA onto chromatin. There are two known effector kinases downstream of the checkpoint Rad proteins, Chk1 and Cds1. Chk1 is phosphorylated in response to DNA damage induced in late S or G 2 in a Rad3 dependent manner [ 12 , 23 , 24 ]. Phosphorylation of Chk1 leads to an increase of Chk1 kinase activity [ 25 ] and is often used as a convenient molecular marker for Chk1 dependent checkpoint activation. Cds1 is activated only in S phase as part of the intra-S and the S/M checkpoints [ 8 , 26 ] Activation of either kinase leads to inhibition of Cdc2 activity by maintaining the inhibitory phosphorylation on Tyr15 [ 27 - 29 ]. Crb2 and Mrc1 act upstream of Chk1 and Cds1, respectively. Crb2 shares homology with the budding yeast RAD9 protein [ 10 ], which is involved in delaying entry into S phase upon DNA damage in G1 [ 30 - 32 ] In fission yeast, Crb2 is required both for activation of Chk1 and for subsequent inactivation of Chk1 for reentry into the cell cycle [ 10 , 33 ] Mrc1 plays a parallel role by binding to and activating Cds1. Expression of Mrc1 is regulated in the cell cycle, thus linking Cds1 activation to S phase [ 6 , 11 ]. In addition to the G2/M, S/M and intra-S checkpoints, three papers have reported the existence of G1 checkpoints that inhibit mitosis when the cells are arrested in G1 using cell cycle mutants. Arrest at the cdc10 arrest point was shown to depend on Chk1 [ 34 ] and Rum1 [ 35 ]. Arrest of orp1 mutant cells depends on the checkpoint Rad proteins and on Chk1 [ 36 ]. It should be noted that neither of these cell cycle mutants is able to replicate their DNA at the restrictive temperature, and failure of the checkpoints responsible for cell cycle arrest results in aberrant entry into mitosis and not into S phase. A G1/S checkpoint has so far not been detected in S. pombe . The drop of CDK activity at the M/G1 transition allows the assembly of the pre-Replication Complex, preRC, which is the first step leading to initiation of S phase. The preRC consists of the ORC (Origin Recognition Complex), Cdc18, Cdt1 and the MCM proteins. Expression of Cdc18 and Cdt1 is cell cycle regulated, thus providing one of the means to regulate initiation of S phase [ 37 , 38 ] Once the preRC is assembled, the chromatin is competent to replicate, but replication is not initiated until other replication proteins are loaded, and two kinases, Cdc2 and Hsk1, are activated. It has been shown both in fission yeast and in Xenopus that the intra-S phase checkpoint cannot be engaged until polα-primase is loaded and replication begins [ 39 , 40 ] This observation poses the question whether the cells have any means to respond to DNA damage sustained in G1. G1 in fission yeast is very short under standard laboratory growth conditions, rendering the investigation of a G1/S checkpoint(s) difficult. However, G1 might be much extended in the natural habitat of S. pombe due to poor nutrient availability. We decided to use several approaches to synchronise the cells and/or to extend G1. Recently we reported the existence of a mechanism that delays entry into S phase when cycling cells are UV-irradiated in G1, using cdc10 and cdc25 mutants to synchronise the cells or growing the cells in medium where G1 is extended [ 41 ]. Here we show that germinating S. pombe spores delay entry into S phase upon UV irradiation in early but not late G1. We demonstrate that there is a G1/S delay that is not dependent on any of the known checkpoint proteins and does not target Cdc2 phosphorylation. We argue that the delay is due to a novel mechanism that leads to delayed expression of Cdt1 and possibly other replication proteins. Results Entry into S phase is delayed by UV irradiation Spores made from diploid cells were allowed to germinate for 3.5 h at 30°C before UV irradiation. At this time point, 1 – 2 hours before S phase entry, the spores showed visible signs of germination by phase contrast microscopy. The dose of UV light was 1200 J/m 2 , which gave a cell survival of about 30% in wild type cells (data not shown). At the time of irradiation, the majority of germinating spores had a 1C DNA content. The timing of S phase was measured, by flow cytometry, as an increase in cellular DNA content from 1C to 2C. The decrease of the 1C population was plotted against time, and the graphs for unirradiated control and UV irradiated cells were compared at the point where 50% of the cells had 1C DNA content. An inherent problem in the present experiments is that the time of germination varies both within each population of spores (low degree of synchrony) and between the different preparations (experiment-to-experiment variation). Thus, the time from resuspension in medium until the cells enter S phase is variable, and it is difficult to ensure that irradiation occurs at exactly the same time point relative to S phase entry. Therefore the experiments were repeated at least twice and the averages were determined (Table 1 ). The experiments revealed that UV irradiation made wild type cells delay their S phase entry by 76 minutes relative to unirradiated control cells (Fig. 1A , Table 1 ). When the irradiated cells started to increase their DNA content, they did not delay appreciably within S phase compared to unirradiated control cells, suggesting that when they started to synthesise DNA, the DNA damage had been removed. However, when the germinating spores were irradiated shortly before S phase entry, they entered S phase without a delay and were arrested with a DNA content between 1C and 2C (Fig. 1B , Table 1 ), presumably due to the intra-S checkpoint. Indeed, the later the irradiation was performed the more pronounced the intra-S phase delay was (data not shown). We conclude that cells irradiated in early G1 arrest temporarily with 1C DNA content, then replicate their DNA with normal timing. Cells irradiated in late G1 exit from 1C with the same kinetics as unirradiated control cells do, but they are unable to complete S phase in normal time. Table 1 Length of the delay in the investigated mutants Mutant Length of the 1C delay (min) (*) Average length of the 1C delay (min) wt early irradiation 70, 80, 90, 60, 80 76 wt late irradiation <10, <10 <10 caffeine <10, <10 <10 rad3 40, 20, 45, 15 30 rad26 45, 40 43 rad1 50, 40 45 rad9 40, 30 35 hus1 55, 40 50 rad17 40, 30 35 cds1 50, 40 45 chk1 90, 55, 70 72 chk1 cds1 35, 40 38 rum1 55, 45, 25 42 res2 <10, <10, <10 <10 (*) The length of the delay was measured at the point where 50% of the cells had a 1C DNA content on the quantitations. The second column shows the results of individual experiments, the third column shows the average lengths of the delay. Irradiations were carried out 3.5 hours after inoculation in medium, except for the entry "wt late irradiation", which was performed 4.5 hours after inoculation. Figure 1 Irradiation of germinating wild type spores delays entry into S phase. Germinating spores were irradiated with UV light 3.5 h ( A ) and 4.5 h ( B ) (time 0) after inoculation into EMM medium, as described in Materials and Methods. Samples were taken for flow cytometry at the indicated times after treatment. The uppper panels show DNA histograms for the unirradiated control (shaded) and the irradiated cells (bold outline without shading). The lower panels show the quantification of cells with a 1C DNA content. Filled symbols represent the control cells, open symbols represent the irradiated cells. The cells delay with low levels of Cdt1 Flow cytometry cannot distinguish between G1 and early S phase cells, therefore we sought to confirm that the cells arrest prior to S phase. PreRC formation is a prerequisite for initiation of S phase. The first step towards preRC formation is de novo synthesis of Cdc18 and Cdt1, which in turn are required for MCM loading. We investigated Cdt1 levels in germinated spores treated as above to establish the timing of the 1C delay relative to Cdt1 expression. Wild type spores carrying myc-tagged Cdt1 were UV irradiated as described above and samples of irradiated and control cells were removed for analysis by flow cytometry and by immunoblotting. Cdt1 expression was induced already at 30 minutes in the control cells, but not until 55 minutes later in the UV irradiated cells (Fig. 2 , Table 2 ). These observations suggest that cells irradiated in early G1 may delay entry into S phase at least in part by delaying preRC formation. Figure 2 The cells delay with low levels of Cdt1. Wild type spores carrying myc tagged Cdt1 were germinated and irradiated as described in the legend to Figure 1A. Samples were taken for protein extracts and flow cytometry at the times indicated. Total protein extracts were prepared and the amount of Cdt1-myc was investigated by SDS-PAGE and immunoblot analyses against total Cdc2, which served as loading control, and Cdt1-myc (top panel). Quantification of the fraction of cells with a 1C DNA content is also shown in the bottom panel (filled symbols: control; open symbols: UV). Table 2 Cdt1 expression and Cdc2 phosphorylation are delayed upon UV irradiation Event Length of the delay (min) (*) Average length of the delay (min) Cdt1 expression 60, 50 55 Cdc2 phosphorylation 40, 50 45 (*) The levels of Cdt1 expression and Cdc2 phosphorylation were quantified and the length of the delay was measured at the point where Cdt1 expression and Cdc2 phosphorylation, respectively, reached 50% of its maximal value. Irradiations were carried out 3.5 hours after inoculation in medium. Rum1 is required for part of the delay Rum1 inhibits the mitotic CDK, Cdc2-Cdc13, and is required for efficient proteolysis of Cdc13 [ 42 - 44 ]. Furthermore, Rum1 is required for all G1 arrests and delays investigated so far. We irradiated germinating rum1Δ spores as above and progression into S phase was followed by flow cytometry. Irradiated spores delayed with a 1C DNA content for 40 minutes (Fig. 3A , Table 1 ). At the 60–90 minute timepoints the irradiated cells display a distinct delay in S phase, consistent with activation of the intra-S-phase checkpoint. The absence of Rum1 shortens G1, therefore some of the germinating spores were in fact in late G1 or early S at the time of irradiation, giving rise to significant activation of the intra-S-phase checkpoint. Figure 3 Rum1 is required for part of the delay. A. rum1Δ spores were irradiated and analysed as described in the legend to Figure 1A. B . Wild type spores were germinated and irradiated as described in the legend to Figure 1A. Samples were taken for protein extracts and flow cytometry at the times indicated. Total protein extracts were prepared and the amount of Rum1 was investigated by SDS-PAGE and immunoblot analyses against total Cdc2, which served as loading control, and Rum1. Quantification of the fraction of cells with a 1C DNA content is also shown in the bottom panel (filled symbols: control; open symbols: UV). Rum1 expression is cell cycle regulated such that it is only expressed in G1 [ 45 ]. We investigated Rum1 levels in germinated spores UV irradiated as described above and samples of irradiated and control cells were removed for analysis by flow cytometry and by immunoblotting. Rum1 expression was induced at 30 minutes in both cultures, but was maintained to a higher extent and longer in the irradiated cells (Fig. 3B ). Both increased expression of Rum1 and the requirement for Rum1 for part of the delay demonstrate that part of the delay takes place in G1. Is the 1C delay checkpoint dependent? The definition of a checkpoint calls for the existence of mutations or chemicals that eliminate the delay. We addressed this issue by treating the germinating spores with caffeine. Caffeine is known to abolish checkpoint function in both higher eukaryotes and fission yeast, possibly through the inhibition of Rad3 [ 46 ]. Caffeine was added to the culture 15 minutes before UV irradiation. Flow cytometric analyses showed that the caffeine-treated spores entered S phase with the same kinetics as unirradiated cells (Fig. 4A , Table 1 ). This observation indicates, (but does not prove, see Discussion), that the delay might be caused by a checkpoint mechanism. Figure 4 Checkpoint Rad proteins in the G1 checkpoint. A. Wild type spores germinating in the presence of caffeine were irradiated and analysed as described in the legend to Figure 1A. B. Germinating rad3 spores were irradiated and analysed as described in the legend to Figure 1A. C . The indicated mutants were sporulated and the germinating spores were irradiated and analysed as described in the legend to Figure 1A. Given that caffeine can inhibit Rad3 related kinases, we investigated whether the G1/S delay is also abolished in rad3 mutant cells. Irradiated rad3 germinating spores delayed with a 1C DNA content for 30 minutes, in contrast to the 76 minute delay of wild type cells (Fig. 4B , Table 1 ). We have investigated whether the other checkpoint Rad proteins are involved in the G1/S delay. rad26 , rad1 , rad9 , hus1 , and rad17 spores were germinated and UV irradiated as described above. Figure 4C (and Table 1 ) shows that rad26 , rad1 , rad9 , hus1 and rad17 cells delay much less than wild type cells do (35–50 versus 76 min). We conclude that Rad26, Rad1, Rad9, Hus1 and Rad17 are required for at least a part of the 1C delay. Mrc1 and Cds1, but not Crb2 and Chk1, are required for part of the delay The products of the checkpoint genes cds1 and chk1 are both known downstream targets of the Rad3 protein kinase and they are required for Cdc2 phosphorylation in the DNA damage and replication checkpoints. We irradiated germinating spores carrying mutations of cds1 , chk1 or both. In cds1 spores the delay was reduced to 45 minutes (Fig. 5A , Table 1 ). In chk1 spores (Fig. 5B , Table 1 ) the length of the delay was not reduced compared to that found in wild type cells. In cds1 chk1 double mutant spores the delay was somewhat shorter than in either single mutant, 35 minutes versus 45 and 73 minutes (Fig. 5C , Table 1 ). The shorter delay in the cds1 chk1 double mutant compared to that in cds1 indicates that chk1 might have a synthetic effect with the cds1 mutation. Figure 5 Cds1, but not Chk1, is required for part of the delay. cds1Δ ( A ), chk1Δ ( B ) and chk1Δ cds1Δ ( C ) spores were irradiated and analysed as described in the legend to Figure 1A. Crb2 and Mrc1 are required for activation of Chk1 and Cds1, respectively. We irradiated germinating crb2 and mrc1 spores [ 11 ]. Consistent with the above findings, in mrc1 the delay was reduced to 50 minutes, while in crb2 the delay was not reduced compared to that in wild type cells (data not shown). We conclude that Mrc1 and Cds1 are required for part of the delay, while Crb2 and Chk1 are not required. The arrested cells maintain Cdc2 in the unphosphorylated form Cdc2 kinase activity is required for the initiation of S phase and is inhibited by phosphorylation on Tyr15 as DNA replication commences [ 35 , 47 ]. We investigated whether the Cdc2 protein is phosphorylated when the cells are delayed with a 1C DNA content. Germinating wild type spores were treated as above and samples of irradiated and control cells were removed for analysis by flow cytometry and by immunoblotting. The results show an increase in the phosphorylation signal as the unirradiated cells enter S phase (Fig. 6 ), in agreement with previous findings [ 35 , 47 ] The irradiated cells increased phosphorylation of Cdc2 45 minutes later (Table 2 ). We conclude that the irradiated cells arrest with a 1C DNA content for a significant length of time with unphosphorylated Cdc2. Figure 6 Cdc2 is not phosphorylated in the arrested cells. Wild type spores were germinated and irradiated as described in the legend to Figure 1A. Samples were taken for protein extracts and flow cytometry at the times indicated. Total protein extracts were prepared and the amount of phosphorylated Cdc2 was investigated by SDS-PAGE and immunoblot analyses against total Cdc2, which served as loading control, and phosphorylated Cdc2 (top panel). Quantification of the fraction of cells with a 1C DNA content is also shown in the bottom panel (filled symbols: control; open symbols: UV). res2 mutant cells do not delay S phase entry after irradiation A number of genes required for DNA replication are transcribed as the cells prepare for S phase. This activation depends on the cell cycle regulated transcription factor Cdc10/Res1/Res2 [ 48 , 49 ]. In the absence of Res2, transcription is constitutively active [ 50 ]. If the G1/S delay in fission yeast cells is brought about by inhibiting this transcription factor, constitutive activation of transcription in a res2 mutant should override the UV-induced G1 delay. We have irradiated germinating res2Δ spores as described above and found that S phase entry was not delayed compared to unirradiated control cells (Fig. 7 , Table 1 ). Figure 7 The res2Δ mutant cells do not display the delay. res2Δ mutant spores were germinated and irradiated as described in the legend to Figure 1A. Cdc10 dependent transcription is not inhibited during the delay The above result indicates that either constitutive expression of Cdc10 dependent genes required for S phase entry can override the delay or inhibiting Cdc10 dependent transcription might be the mechanism of the delay. A prediction of the latter alternative is that cells arrested with 1C DNA content upon UV irradiation should not have performed the Cdc10-dependent transcriptional events, including induction of the cdc18 , cdt1 , and cig2 genes. We isolated total RNA from irradiated and unirradiated germinating wild type spores and followed the transcription of cig2 , cdt1 and cdc18 . There was no delay in the appearance of the Cdc10 dependent transcripts upon UV irradiation (Fig. 8 , only data for cdt1 are shown). We conclude that Cdc10 dependent transcription is not the mechanism of the delay. Figure 8 Cdc10 dependent transcription is not inhibited during the delay. Wild type spores were germinated and irradiated as described in the legend to Figure 1A. Samples were taken for RNA extracts and flow cytometry at the indicated times. Total RNA extracts were prepared and the amount of cdt1 mRNA was investigated by Northern analysis. Discussion We have provided evidence for the existence of a mechanism in germinating fission yeast spores that delays entry into S phase upon UV irradiaton in early G1. Germinating wild type spores displayed a pronounced delay in entering S phase after UV irradiation. The delay was observed only when irradiation was carried out in early but not in late G1. We have investigated the dependence of the delay on classical checkpoint proteins and showed that they are required for some but not all of the delay with 1C DNA content. We argue that the observed delay is caused by two separate mechanisms, the first taking place before entry into S phase, and the second in early S phase (see below). The delay in exit from the 1C population was demonstrated by means of flow cytometry, which does not allow us to distinguish between a G1/S and an early S delay. The following data represent strong evidence that part of the delay takes place before entry into S phase. First, the irradiated cells delay expression of Cdt1. In the absence of Cdt1 the cells cannot form preRCs and thus cannot initiate S phase. Second, the irradiated cells express Rum1 longer than unirradiated control cells. Since Rum1 expression is cell cycle regulated such that it is only expressed in G1, [ 45 ], this observation implies that the irradiated cells do delay in G1. Furthermore, the delay is shorter in a rum1 mutant, which presumably loses the G1 part of the delay. Third, mutants lacking Mrc1 or Cds1, which are essential for S-phase checkpoints reported so far in fission yeast [ 6 ], still delay for a significant length of time, pointing to the existence of a non-S mechanism [ 11 ]. Fourth, cells delay with the Cdc2 kinase in an unphosphorylated state. Cdc2 is normally inhibited by phosphorylation on Tyr15 as DNA replication commences [ 35 , 47 ], arguing that the cells arrest before S phase. Fifth, the delay is not observed in a res2 mutant, which can not turn off the Cdc10 dependent transcription signal. The finding that a mutation affecting expression of proteins crucial for preparation for S phase abolishes the delay argues that the wild type cells first stop in G1 and only later stop in S phase. On the basis of these results we conclude that there is a UV-induced G1 delay, which is not checkpoint Rad dependent and is brought about by an as yet undescribed mechanism. This part of the total delay with 1C DNA content is ca 40 minutes, since rad , cds1 and mrc1 mutants delay 30–50 minutes and rum1Δ cells lose ca 40 minutes of the delay, compared to wild type cells. The remaining ca 40 minutes of the total delay requires the checkpoint Rads, Mrc1, Cds1 and Cdc2 is phosphorylated. We argue that this part of the delay is brought about by the intra-S checkpoint. However, the resolution of our experiments is not high enough to exclude the possibility that some of the checkpoint Rad- and Cds1-dependent part of the delay occurs in late G1. We consider this possibility unlikely for two reasons; first, previous work has shown that the role of Cds1 is specific for S phase [ 26 ] and second, we have shown in the current paper that if irradiation occurs later, the cells enter S phase without delay and delay in S phase. Since the level of synchrony is low in germinating spores, we have not emphasised minor differences in the timing of S phase entry. In spite of poor synchrony, we deem germinating spores a good model system, since spore germination is a natural phenomenon, it involves an extended G1 period and we observed clear-cut effects. Furthermore, this model system allowed us to investigate the effects of a number of mutations that would have not been possible using synchronisation by other methods. We have explored whether the G1/S delay is caused by a checkpoint mechanism. We have shown that caffeine abolishes the delay, but this is not entirely due to inhibition of Rad3 activity, since a rad3 mutation does not abolish all of the G1/S delay. Since we have not identified a checkpoint mutation which abolishes the delay, we attribute the effect of caffeine to another effect than the inhibition of checkpoint proteins. Interestingly, recent data suggest that caffeine inhibits checkpoint responses without inhibiting the ATR and ATM kinases in human cells [ 51 , 52 ]. Previously, Rhind and Russell [ 53 ] showed that UV-irradiation during G1 delays passage through S-phase. However, this checkpoint arrests cells in S phase, requires Cds1 function and probably represents the intra-S checkpoint. We have recently discovered a mechanism that delays entry into S phase in cells irradiated in early G1 in synchronised or in cycling cells [ 41 ] This inhibitory mechanism has several features in common with that described here. Both pathways are activated in early but not in late G1; both inhibit entry into S phase; both pathways are independent of classical checkpoint genes and of Cdc2 phosphorylation. These similarities argue that the G1/S mechanism demonstrated in germinating spores and in cycling cells is one and the same. In budding yeast there is a G1 DNA damage checkpoint response that depends upon Mec1 [ 31 , 32 , 54 ], a homologue of the mammalian ATM/ATR and the fission yeast Rad3 protein. However, the budding yeast G1 checkpoint response depends on Rad53 [ 55 , 56 ], whereas its homologue in S. pombe , Cds1, is not involved in the present pathway. The budding yeast G1/S checkpoint delays entry into S phase by phosphorylating and thereby downregulating Swi6, the homologue of Cdc10 [ 57 ]. In contrast, in fission yeast Cdc10 dependent transcription is not delayed during the G1/S delay (Fig. 8 ). Other possible mechanisms for the G1/S delay include inhibition of Cdc2 by Rum1 or an as yet unidentified mechanism such as preventing the formation of Cdc2-cyclin complexes or by restricting the availability of cyclins. We have shown that Rum1 is expressed during the delay and is required for the G1 delay. This observation does not imply that Rum1 is a direct target of the G1/S delay, but this remains an attractive possibility. Another possible mechanism for the delay is delaying expression of proteins required for the initiation of DNA replication. In particular, the findings that (1) irradiation in late G1 does not cause a delayed entry into S phase, (2) increased transcription of Cdc10 dependent genes in res2Δ overrides the delay, (3) transcription of Cdc10 dependent genes is not downregulated during the delay and (4) expression of Cdt1 is delayed, suggest that the G1/S delay is caused by delayed expression of Cdt1 and probably also of Cdc18 and Cig2. Since we have shown that transcription of Cdc10 regulated genes is not downregulated, the most likely mechanism of the delay is reduced translation rate of Cdt1 and possibly other proteins required for initiation of DNA replication. Conclusions We studied the response of Schizosaccharomyces pombe cells to UV irradiation in G1. We used germinating spores to exploit a natural phenomenon where the cells have a long G1. In this paper we provide evidence for the existence of a mechanism in fission yeast that delays entry into S phase upon UV irradiaton in early G1. The G1 delay is independent of classical checkpoint proteins and Cdc2 phosphorylation. Our results point to a mechanism that delays translation of proteins required for S phase entry. Methods Fission yeast strains and methods All our strains are derivatives of the Schizosaccharomyces pombe L972h - strain. All basic growth and media conditions were as described [ 58 ]. Sporulation and spore germination Diploids were made by interrupted mating of h + and h - strains carrying the met3 or ade1 complementing markers. The rum1:::ura4 + / rum1 + diploid was made by protoplast fusion since rum1Δ is sterile [ 44 ]. All diploids, with the exceptions of res2Δ (which is deficient in meiosis [ 59 ]) and rum1Δ , were homozygous for the respective mutations. In case of these two mutants Δ/wild type ura4-D18 / ura4-D18 diploids were sporulated and the spores were germinated in the absence of uracil. The diploids were sporulated in liquid malt extract medium at 30°C, then incubated with 3 μl/ml β-glucuronidase ( Helix pomatia juice, Biosepra) at 30°C overnight. The spores were washed twice in water and resuspended in EMM2 supplemented with adenine and methionine for germination and UV irradiation. UV irradiation Cells were irradiated with 254 nm UV light while rapidly stirred in a thin layer (3 mm) of liquid medium. The dose administered was measured with a radiometer (UVP instruments) and an exposure time of 4 minutes gave an incident dose of about 1100 J/m 2 . Cell survival was monitored by conventional plating on YE plates. The incident dose does not reflect the dose absorbed by the cells because UV light of this wavelength penetrates poorly into water. However, since irradiation conditions were constant, the incident dose was proportional to the absorbed dose. Protein extracts and western blots Protein extracts for western blotting were made by TCA extraction, as described previously [ 19 ]. For western blot analysis the following antibodies were used: anti-phosphotyrosine Cdc2 (Sigma C0228) at a dilution of 1:400, anti-PSTAIRE against Cdc2 (Santa Cruz sc-53) at a dilution of 1:2000, anti-myc (PharMingen) at a dilution of 1:1000. The secondary antibodies were either HRP or AP conjugates, used at a dilution of 1/5000. Detection was performed using the enhanced chemiluminescence procedure (NEN ECL kit). Cdc2 and phosphorylated Cdc2 was measured using ECF detection (Amersham) and quantified with the Image Quant software. RNA preparation and blotting Total RNA was isolated as described [ 58 ], resolved on formaldehyde agarose gels and blotted onto a nitrocellulose membrane (NitroPure, Osmonics). All blots were hybridized with 32 P-labelled RNA probes, generated with T7 RNA polymerase (Riboprobe System T7 Kit, Promega). For cig2 and cdc18 , the ORFs were inserted into pGEM-3 MCS to serve as template for producing the RNA probes. For cdt1 , a PCR fragment of the ORF with T7 promoter sequence attached to the lower primer was used as template. Hybridisation was carried out using standard procedures and visualised by a STORM 860 Phosphoimager (Molecular Dynamics). Flow cytometry About 10 7 cells were spun down for each sample and fixed in 70% ethanol before storing at 4°C. Samples were processed for flow cytometry as described [ 60 ] and stained with Sytox Green (Molecular Probes S-7020) [ 61 ], and analysed with a Becton-Dickinson FACSCalibur. The fraction of 1C cells was quantified using the CellQuest software (BD Biosciences). Authors' contributions EAN showed the existence of the delay and investigated the roles of Rad3, Chk1 and Cds1, Cdc2 phosphorylation and Cdc10-dependent transcription as a potential mechanism. MS, TT and HV investigated the roles of further checkpoint proteins and that of Res2. EB participated in the design and coordination of the study and in writing the manuscript. BG devised the study, analysed Rum1 and Cdt1 expression and drafted the manuscript. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC528784.xml

551535

Polarized monocyte response to cytokine stimulation

A comprehensive study of the transcriptional response of mononuclear phagocytes to cytokines reveals distinct classes of cytokines that elicit either the classical or alternative pathway of monocyte activation.

Background Resident and recruited mononuclear phagocytes (MPs) display a versatile phenotype that reflects the plasticity of these cells in response to microenvironmental signals. This heterogeneity spans a continuous spectrum that can be polarized into two extremes recently described by Mantovani et al . [ 1 ]. Pathogen stimulation exemplified, for instance, by lipopolysaccharide (LPS) stimulation in the presence of interferon (IFN)-γ induces M1 MPs through engagement of Toll-like receptors (TLRs). M1 MPs are true antigen-presenting cells capable not only of killing invading organisms but of concomitantly recruiting and activating immune effector cells [ 2 , 3 ]. Treatment of MPs with type II cytokines such as interleukin (IL)-4, IL-13 and, partly, IL-10 [ 4 ], polarizes their function towards tissue repair, angiogenesis and containment of collateral damage through reduction of inflammation (the M2 macrophage phenotype). This alternative mode of macrophage activation accounts for a distinct phenotype with a key role in humoral immunity and tissue repair [ 5 ]. It has been suggested that the extreme dichotomy between a classical M1 and an alternative M2 polarization of macrophage function may not take into account intermediate regulation by cytokines such as IL-10, transforming growth factors (TGF-α and TGF-β), macrophage colony-stimulating factor (M-CSF), IFN-α and IFN-β and tumor necrosis factor (TNF) [ 5 ]. Most important, a rigid dichotomy in MP function may not directly apply to physiological and/or pathological conditions in which these cells are exposed to an array of cytokines produced by innate or adaptive immune mechanisms during infection, tissue damage or in other conditions. Indeed, a comprehensive overview of the modulatory properties of cytokines on the MP reaction to a pathogen is missing. In addition, little is known about the transcriptional changes occurring in MPs on exposure to pathogen components such as LPS. A recent study analyzed the transcriptional profile induced by the exposure of circulating MP conditioned in vitro for 7 days with IL-4 and GM-CSF (immature dendritic cell, (DC)) to pathogen components [ 6 ]. Bacterial, viral and fungal components elicited distinctive pathways that were, however, largely overlapping. The predominant response of these DCs to most pathogen components encompassed a rapid upregulation of genes associated with the innate arm of the immune response followed by induction of adaptive immune response genes. The response of circulating MP-derived DCs is short-lived as these cells can exhaust their production of effector molecules (cytokines and chemokines) within a few hours of LPS stimulation [ 7 ]. The transience of mRNA and protein expression can cause DCs to redirect the immune response in different ways at different time points. For instance, soon after stimulation, DCs elicit T-cell responses of the Th-1 type, whereas at later stage of activation they prime T-cell responses of the Th-2 type, suggesting that their function is strongly dependent on the timing and duration of exposure to individual and/or combined stimulatory conditions in the surrounding microenvironment. At the transcriptional level, the dual function of DCs shifted from an early pre-inflammatory phase occurring within 3 hours to a later regulatory phase occurring approximately 8 hours following LPS exposure. During these evolving stages of DC activation, cytokines play a dominant role in shaping the function of DCs and other immune cells, providing a malleable link between the innate and adaptive immune responses [ 8 ]. In natural conditions, circulating or resident MPs may encounter a pathogen before the surrounding microenvironment has a chance to influence their maturation. Therefore, it is unknown whether non-conditioned circulating CD14 + MPs would react similarly to DCs on engagement with infectious agents. Thus, a preliminary aim of this study was to evaluate the kinetics of the response of non-conditioned circulating CD14 + MP to LPS. The results suggested that these cells respond to LPS similarly to immature CD14 - DCs, with a surge in transcriptional activity that peaks around 3 hours after stimulation and in which the activation of genes associated with a classical activation of innate immune mechanisms predominates [ 6 ]. The stringent dichotomy describing a classical activation of MPs into mature antigen-presenting cells caused by IFN-γ and an alternative induction into macrophages induced by IL-4 and IL-13 may not apply to physiological conditions in which the microenvironment responds to pathogen exposure with a broad array of cytokine secretion. We therefore investigated whether polarized responses of MPs to pathogens are extreme behaviors that can be observed in vitro by studying a few illustrative cytokines or whether they represent a mandatory molecular switch through which most cytokines operate. Thus, we stimulated non-conditioned CD14 + MPs with LPS for 1 hour. The MPs were then exposed to an array of different cytokines that may be expressed in distinct pathologic conditions by different immune-cell subsets. The 1-hour interval was empirically selected to induce a biphasic model in which the presumed modulation by cytokines occurred during an ongoing reaction to LPS. This allowed mapping of cytokines into conditional subclasses based on their effects on the global transcriptional changes responsible for MP activation and differentiation. Two main classes of cytokines were identified that induced a classical or alternative pathway of MP activation, respectively. An intermediate class (including IL-10) was also identified, while TNF-α, TNF-β and GM-CSF displayed a quite distinct behavior from the other cytokines. Expression of genes affected by NFκB activation was most predictive of the two main classes, suggesting that in most cases the NFκB pathway is a central target of cytokine regulation that modulates the cascade of events following LPS stimulation. Overall, it seems that MP maturation/differentiation goes through a molecular switch that is partly independent of the fine differences in the stimulatory properties of the various cytokines and, with few exceptions, is pre-programmed towards a classical or an alternative route. As LPS itself induces a classical type of activation, the most dramatic modulation in this model was observed toward the alternative pathway, suggesting that a broad array of cytokines may counteract the pro-inflammatory effects of bacterial components. Results Effect of LPS stimulation on circulating CD14 + MPs Enriched MP preparations (about 90% CD14 + ) were exposed to LPS. Total RNA extracts were obtained 4 and 9 hours after. These time points were selected to catch salient stages of the biphasic response of MP to LPS described by others [ 6 , 7 ]. Amplified antisense RNA (aRNA) [ 9 ] was hybridized to a custom-made 17,000 (17K)-clone cDNA microarray chip enriched with genes relevant to immune function. The transcriptional profile of LPS-induced MPs was similar to that described by others in DCs [ 6 ]. In particular, genes associated with the innate response of CD14 - , immature DCs to pathogen components [ 6 , 10 ] were similarly upregulated in CD14 + MPs (data not shown). This finding suggests that the differential expression of the LPS co-receptor CD14 between the two cell populations has a relatively minor impact on the transcriptional regulation of the innate immune response [ 11 ]. Kinetics of the response of CD14 + MPs to LPS and its modulation by cytokines Aliquots of CD14 + MPs were stimulated in parallel with LPS and exposed 1 hour later to individual cytokines selected from a library of recombinant proteins possibly relevant to MP regulation. MPs were kept in culture for 4 and 9 hours, at which times aRNA was prepared for transcriptional analysis. Unsupervised Eisen's clustering [ 12 ] was applied to the complete dataset (Figure 1 ). The kinetics of the response to LPS had the greatest influence on the global transcriptional profile of MP induction; samples preferentially clustered according to time of stimulation rather than type of treatment. This was underlined by the observation that MPs stimulated with LPS alone clustered with the cytokine-stimulated MPs according to the time elapsed after stimulation. In addition, a cluster containing most of the samples obtained 9 hours after stimulation (9') included three control samples consisting of CD14 + MPs not exposed to LPS or cytokines (no stimulation). These three samples were prepared at times 0, 4 and 9 hours to parallel the culture conditions used for stimulation. This finding suggested that the transcriptional profile of CD14 + MPs 9 hours after LPS stimulation and 8 hours after treatment with most cytokines converges toward a less reactive metabolic state closer to that of unstimulated MPs. Several cytokines, however, maintained a more active metabolic profile and after 9 hours retained a transcriptional footprint relatively close to that of samples treated for 4 hours (9"). This group included the genes for most IFN-α isoforms, IFN-β, vascular endothelial growth factor (VEGF), FLT-3 ligand, TGF-α, the chemokine RANTES (CCL5), IL-2, IL-4, IL-15 and the chemokines MIP1α (CCL3) and MIP1β (CCL4), suggesting that these cytokines may have relatively prolonged kinetics of MP activation. The average number of genes whose expression was increased compared to unstimulated MPs was higher (420 genes) after 4 hours than after 9 hours (265 genes). About half of the genes upregulated after 4 hours remained upregulated at 9 hours (223 genes). This is in accordance with Huang's observation [ 6 ] that transcriptional changes in DCs occur predominantly in the early phase of the response to pathogen components, with about half the genes displaying transitory expression and the other half sustained expression. For this reason, subsequent analyses were limited to the 4-hour time point. Transcriptional modulation by cytokines 4 hours after LPS stimulation of CD14 + MPs Although LPS alone strongly affected the transcriptional program of MPs, it was possible to discern the contribution of individual cytokines. This analysis was limited to samples treated for 4 hours, when the most dramatic effects on gene modulation were noted. Genes differentially expressed in samples treated with cytokines compared to those treated with LPS alone were identified. Stringent criteria were applied to select genes expressed in at least 80% of samples with a threefold or greater increase or reduction in expression over LPS in at least one of the cytokine-treated samples. This gave 2,057 genes that were deemed most relevant to the analysis. Using these genes, all cytokine-treated samples were subjected to unsupervised clustering to evaluate their relatedness (Figure 2a ). Two main classes of cytokines were identified. One class (Figure 2a , blue horizontal line) included IL-4, IL-13 TGF-α, TGF-β and VEGF, which are unquestionably associated with the alternative pathway of MP activation [ 1 , 5 ]. The second class (Figure 2a , red line) included IFN-α2, IFN-β, IFN-γ, CD40 ligand (CD40L), and FLT-3 ligand, which are generally associated with the classical pathway of MP activation [ 2 , 3 ]. Therefore, we considered the first cluster representative of the alternative and the second of the classical pathway of MP activation in response to LPS. As predicted by Gordon [ 5 ], few cytokines (Figure 2a , green line) did not properly belong to either class. These included IL-10, IL-1β, IL-15 and two IFN-α isoforms. Because several of these cytokines have previously been associated with the alternative pathway, we referred to this class as alternative II. Not surprisingly [ 5 ], MPs stimulated with TNF-α (Figure 2a , purple line) and, most dramatically, TNF-β and GM-CSF (Figure 2a , black line) had a totally independent effect on the transcriptional regulation of LPS-induced CD14 + MPs [ 5 ]. Cytokines classified according to the previous groups were tested for class prediction by applying unsupervised principal component analysis (PCA) to the global, unfiltered 17K gene dataset (Figure 2b ). This analysis independently classified cytokines in two groups corresponding to the alternative (Figure 2b , blue circles) and classical (Figure 2b , red circles) cytokine classes. Most of the cytokines belonging to the alternative II class (Figure 2b , green circles) grouped with the alternative group, whereas TNF-α (Figure 2b , purple circle and arrow), TNF-β and GM-CSF (Figure 2b , gray circles) remained separate. The sample treated with LPS alone (Figure 2b , yellow circle and arrow) grouped with the classical cytokines, confirming the predominant pro-inflammatory effects of this bacterial product and its alignment with the classical pathway of MP activation [ 1 , 11 ]. This finding based on the complete dataset indicates the intrinsic bias of this study aimed at exploring the alternative modulation of the MP response to LPS. Particular mention should be made of the erratic behavior of various IFN-α subtypes, which clustered indiscriminately between the two main cytokine classes. Interestingly, however, alignment of the IFN-α protein sequences through the EMBL-EBI Clustal W database identified, with the exception of IFN-αG, a close relationship among the IFN-α subtypes that clustered with the alternative type of cytokines (data not shown). This subclassification was also supported by the phylogenetic relationship among interferons described by Henco [ 13 ]. This information suggests that specific domains of the IFN-α molecules may have dramatically different effects in the modulation of the MP response to pathogen [ 14 , 15 ]. Interestingly, IFN-α 2 , which is the one most commonly used in clinical trials as a pro-inflammatory cytokine, clustered with the classical cytokines adjacent to IFN-γ. Cytokine-mediated modulation of LPS-stimulated CD14 + MPs predominantly affects pathways downstream of NFκB Signatures associated with several pathways of immune-cell activation were constructed by selecting genes from the global pool of 17K clones according to literature information without pre-existing information about the association of their expression to either class of cytokines. Signature genes were then subjected to supervised clustering according to the cytokine classification shown in Figure 2a . This independent process identified virtual signatures, in some cases portraying opposite transcriptional regulation by the two classes. The signature that most strongly discriminated the two classes comprised 121 genes whose expression is closely dependent on NFκB modulation [ 11 ] (Figure 3a ). This is not surprising as LPS acts through engagement of Toll receptor 4 (TLR4) and CD14, with resulting activation of NFκB [ 3 ]. It would, therefore, seem intuitive that the strongest modulation in the present experimental conditions would target this pathway. In particular, several TNF- and IL-1-related genes classically modulated by NFκB during the acute phases of the innate immune response [ 11 ] were strongly and inversely modulated by the two cytokine classes. The same 121 genes were used for unsupervised class prediction by reclustering cytokine-treated samples (Figure 3b ). This independent analysis segregated cytokines into two classes that with the exception of one (IFN-α2b) matched the respective original classical and alternative classification (Figure 3a , red and blue horizontal bars, respectively). Interestingly, the cytokines that belonged to the alternative II class clustered with the alternative cytokines (Figure 3a , green horizontal bars) while TNF-α (Figure 3a , purple horizontal bar), TNF-β, and GM-CSF (Figure 3a , dark gray horizontal bar) clustered separately but in proximity of the classical group. Analysis of early signaling events occurring 1 hour and 30 minutes after LPS stimulation and, therefore, 30 minutes after the additional cytokine exposure, demonstrated significantly increased levels of the free p50 subunit of NFκB in MP whole-cell extracts treated with alternative class cytokines (IL-4 and IL-13). In addition, IL-1α and TNF-α significantly upregulated p50, whereas no significant changes were caused by classical cytokines (IFN-γ, IL-6 and IL-3). Extracts from MPs stimulated only with LPS also failed to demonstrate changes in NFκB subunit release; this is probably related to the 90-minute period from stimulation that allowed a return of signaling molecules to baseline conditions (Figure 3c ). The similarity of the IL-1α and TNF-α effects on p50 to those of alternative cytokines contrasts with the dramatic differences observed on the respective transcriptional profiles, suggesting that other pathways induced by these cytokines may prevail in the conditions tested here. Among the additional pathways tested, those mediated through STATs, Janus kinases (JAKs), and interferon regulatory factor (IRF) did not appear consequential to the experimental conditions tested in this study (data not shown). Cytokine-mediated modulation of metalloproteinase expression in LPS-stimulated CD14 + MPs Matrix metalloproteinases (MMP) are tightly connected to MP activation. MMP released by MPs contribute to normal and pathological tissue remodeling and MP migration. In addition MMP function as regulatory proteins by promoting the activation or degradation of cytokines. Finally, MMP are susceptible to cytokine stimulation. Thirty-six MMP and MMP-related genes (filtered from a larger group of 184 MMPs, disintegrins, α-defensins, TGF-β, TNF-α, insulin-like growth factor 1 (IGF-1), epidermal growth factor (EGF), fibroblast growth factor (FGF), IL-1 and monocyte chemotactic protein-3 (CCL7/MCP-3) genes) [ 16 , 17 ] were clustered according to the alternative and classical group denomination and the 11 most representative are shown in Figure 4 . The alternative group of cytokines induced the transcription of MMPs (MMP 7, 9, 10, 19), enzymes related to MMP function (disintegrin ADAM 9, pro-collagen proline dioxygenase, MEK1 kinase, serine protease inhibitor, cathepsin L) and structural proteins (gap junction connexin 26, laminin A/C). This confirms the role of alternative MP activation in promoting tissue remodeling, cell-cell interactions and local control of the inflammatory process through activation (via MMP-7, 9) [ 18 - 21 ] or degradation (via MMP-19) of cytokine activity [ 17 , 22 ]. In addition, these observations suggest a role for the cytokines in the alternative group (other than the well documented IL-4 or IL-13) in polarizing MPs toward an M2 regulatory phenotype. Characterization of the alternative and classical groups of cytokines Comparison of gene-expression patterns induced by the two cytokine classes (classical and alternative) identified 2,007 genes that were differentially expressed at a less than 0.001 significance level ( t -test, p 2 -value). Genes associated with immune function were proportionally over-represented. A selection of genes relevant to MP function is shown in Figure 5 . In most cases, the pattern of expression echoed the class allocation suggested by the literature [ 23 ]. Classical cytokines induced genes responsible for the cytotoxic and migratory properties of MPs such as those for CD95, TRAIL, granzymes, perforin, CD16 (stimulatory Fcγ receptor) and CD62L. In addition, antigen presentation was enhanced as suggested by the coordinate expression of several HLA class II genes. IFN-γ and the other classical cytokines inhibited the expression of the macrophage-derived chemokine CCL22/MDC, while upregulating the expression of CXCR3, as previously reported [ 23 ]. Alternative cytokines induced the expression of several cytokines and their respective receptors involved in the chemotaxis and activation of neutrophils, MPs, natural killer (NK) cells, DCs, helper T lymphocytes (Th2) and B lymphocytes. Several genes known to be associated with alternative MP activation [ 5 , 23 ] were consistently upregulated. These included the mannose receptor, the inhibitory Fc-IIb receptor CD32, and the cell-surface molecule CD44 [ 24 ], which is associated with the disposal of inflammatory cell corpses without expansion of the inflammatory process. Several inducible chemokines were expressed in response to alternative stimulation such as CCL22/MDC, supporting the emerging role of this cytokine as an enhancer of polarized Th2 responses [ 25 ]. Other chemokines known to be induced by master type II cytokines and associated with the induction of Th2 responses were also induced by alternative activation; these included CCL11 (eotaxin), CCL1 (I-309), CCL2 (MCP-1) and CCL7 (MCP-3) [ 23 ]. Of interest was the relatively higher expression of IL-24, a cytokine belonging to the IL-10 family constitutively expressed by MPs [ 26 ]. Although the true role of this cytokine in inflammatory processes is not known, it is likely that its pro-apoptotic and angioregulatory properties have important roles in tissue repair and remodeling during inflammation. Finally, various chemokine receptors responsible for MP trafficking and localization were differentially regulated by the two cytokine groups, including in particular CCR1 and CCR5, which were induced by alternative stimulation, and CCR2, induced by classical stimulation. This early transcriptional profile underlines the primary role of MPs during the acute phases of the response to pathogen as effector cells that can kill pathogens, take up antigen and migrate to local regional lymph nodes to recruit adaptive immune responses (classical activation). In contrast, alternative cytokines may be produced in the microenvironment to maintain a resident MP phenotype rich in chemokine production, which can attract Th2-type immune responses while continuing pathogen clearance through retention of phagocyte properties and promoting tissue remodeling. Surprisingly, genes of the IL-1 family and its receptors, TNF and IL-6 were consistently upregulated by the alternative class of cytokines, suggesting that LPS alters MP polarization with regard to these cytokines [ 5 , 10 ]. Particularly interesting is the alternative induction of IL-6 and its receptor, which may play a central role in mediating the transition from neutrophils to MP recruitment during progression from acute to chronic inflammation [ 27 ]. Validation of microarray analysis by TaqMan real-time PCR To define the validity and accuracy of our global microarray analysis, quantitative TaqMan real-time PCR was performed on amplified RNA material isolated after stimulation of monocytes obtained from five additional normal donors with representative cytokines/soluble factors selected from the alternative and from the classical groups. Comparison of monocyte stimulation with a candidate cytokine from the alternative/M2 (IL-4) and the classical/M1 (IFN-γ) group, respectively, is shown in Figure 6 . Ten genes whose expression was upregulated by alternative cytokines were tested, as the thrust of the analysis was the evaluation of the effect of alternative cytokines on LPS-stimulated MPs. The relative expression of five genes out of 10 (IL-1α, IL-1 receptor, mannose receptor, NFKb-p105/50 and TRAF) was significantly higher after treatment with IL-4, as suggested by the array data. Also, the expression profiles of the other genes tested reproduced the pattern observed in the array experiments even if they did not reach statistical significance for each individual gene. This is not surprising because array experiments summarize in signature fine differences in gene expression, that often describe patterns rather than absolute significance for individual genes. Reproducibility of the estimates of mannose receptor, NFKb-p105/50 and TRAF gene expression in MPs obtained from the same five donors was evaluated after stimulation with four cytokines/soluble factors selected from the alternative (IL-13, IL-1α, IL-4, TGF-β) and four from the classical groups (FLT-3 ligand, IFN-γ, IL-3, CD40L) (Figure 7 ). In all cases gene expression significantly reflected the pattern of gene expression detected by microarray analysis (Figures 3 , 5 ). Discussion It has been suggested that MP activation and maturation progresses through a polarized mechanism whereby two extreme products result that promote inflammation on one side and tissue repair on the other [ 1 , 5 ]. The first mechanism has been called the classical pathway of MP activation. It induces M1 monocytes specialized for pathogen killing and activation of innate and adaptive immune effector cells. A second, alternative, pathway induces M2-type monocytes committed to clearing pathogen through internal metabolism while reducing inflammation. This process limits the collateral damage induced by an excessive immune response, and, upon cessation of the pathogenic stimulus, promotes tissue repair. This dichotomy is based on the study of a few cytokines deemed representative of the classical mode (LPS, IFN-γ) or the alternative mode (IL-4, IL-13) of MP activation. Other cytokines such as IL-10, TGF, M-CSF, IFN-α/β and TNF, although partially overlapping both pathways, display functional effects that diverge significantly enough that they would be inaccurately grouped in either class [ 5 ]. In this study we evaluated whether MP commitment is dependent on an early bipolar switch through which most cytokines operate. This was done by testing in parallel a library of 42 stimulatory molecules possibly present in the tissue microenvironment following a pathogenic insult. This modulation was tested on MPs triggered by a pathogenic stimulus, exemplified in this case by LPS. This was done on the assumption that in most circumstances resident or migratory MPs reaching an infected area are exposed concomitantly to a pathogen and to the cytokine milieu resulting from the infection. The transcriptional profile identified by this study cannot distinguish between the direct effect of each soluble factor analyzed and the downstream activation of transcriptional pathways by secondary paracrine or autocrine secretion of biological modifiers by MPs. However, the main goal of this study was to identify the overall effect on MPs of the exposure to individual cytokines. Future analyses, focused on specific cytokine patterns described here, should possibly include the addition of blocking antibodies to segregate secondary from primary MP responses. The results suggest that MP activation is in most cases a bipolar process regulated by an internal switch through which cytokines modulate the yin and yang of the MP transcriptional program. In fact, most cytokines preferentially induced one or other pattern of transcriptional activation. We, therefore, mapped most of the cytokines within a classical or an alternative classification according to their effects on CD14 + , LPS-induced MPs. In particular, it appeared that transcriptional programs down-stream of NFκB activation [ 11 ] were mostly associated with either class, suggesting that the NFκB system is at the center of the switch regulating MP activation/differentiation in the conditions tested in the present study. This is not surprising as LPS signaling is mediated through the TLR-4 and CD14, which in turn directly regulate the IκB kinase (IKK)-NFκB pathway [ 3 , 11 , 28 , 29 ]. It appears that cytokine regulation modulates the release of the p50 subunit of NFκB, whch is in turn responsible for the downstream effects on the transcriptional program. Interestingly, reclustering of cytokines based on NFκB-dependent genes consolidated the two classes of cytokines adding TNF-α and TNF-β, IFN-α 2b and GM-CSF to the classical group and the alternative type II cytokines with the alternative group, suggesting that NFκB may be central to MP polarization. This information cannot of course be generalized to other conditions as the model tested is strongly biased by NFκB induction by LPS. This is underlined by the unexpected alternative upregulation of genes associated with IL-1, TNF and IL-6 [ 11 ]. This observation suggests that in different conditions, cytokines may differently modulate MP activation, possibly through various modulatory feedback mechanisms [ 5 ]. In addition, genes associated with the interrelated arginine and tryptophan pathways that modulate nitric oxide induction and are indirectly associated with NFκB function were, at least in part, differentially regulated by the two classes of cytokines [ 30 - 32 ]. It has been suggested that inducible nitric oxide is produced rapidly after LPS stimulation of MPs and inhibits NFκB through the stabilization of IκB [ 16 ]. It is possible that cytokines may counteract this effect through modulation of this metabolic junction. Cytokine and chemokine effects on MP function are tightly intertwined with the enzymatic activities of MMP. Membrane-bound cytokine receptors and adhesion molecules can be released from the cell surface by MMPs acting as 'sheddases' or 'convertases'. This, in turn, can downregulate cell-surface signaling by removal of receptors, or induce paracrine activity by release of soluble proteins. Not surprisingly, IL-13 overexpression results in production of several MMPs [ 33 ]. For instance, MMP-9 activates latent TNF-α on the surface of MPs or soluble VEGF [ 19 , 34 ]. Furthermore, Yu and Stamenkovic [ 21 ] observed that gelatinase B/MMP9 bound to CD44 activates latent TGF-β stored in the pericellular matrix. MMP-7 can enhance tissue repair by facilitating migration of epithelial cells [ 16 ]. In agreement with these reports we observed that the transcription of MMP-7, MMP-9 and CD44 are coordinately induced (Figures 4 , 6 ) by the alternative activation of MPs. Conversely, the downregulation of MMP-7 and MMP-9 by the classical cytokines confirms their inhibitory effects on MMP expression. Inhibition of MMP-9 production by IFN-β and IFN-γ has been recently reported by Sanceau et al . [ 35 ], who noted that interferons regulate MMP expression through IRF/NFκB interaction. Binding of NFκB p50/p65 to the MMP-9 promoter is competitively inhibited by IFN-β and IFN-γ-induced IRF-1. Possibly, NFκB regulation of MMP promoters through release of p50 (Figure 3 ) was responsible for the transcriptional activation of MMP expression by alternative cytokines observed in this study (Figure 4 ). These observations confirm the tight specificity of the relationship between cytokine and MMP regulation, which is finely toned at several check points and strongly polarized, in these experimental conditions, toward an M2 phenotype. Enhanced transcription of gap junction/connexin 26 by the alternative cytokines is also of particular interest in view of the hypothesized junctional communications among MPs or between MPs and endothelial cells [ 36 , 37 ]. The finding that MPs activated by alternative cytokines induce the transcription of genes for gap junction components is of physiological importance and may be the missing link in the identification of factors that regulate the expression of gap junction connexins in MPs. In addition, it opens up the possibility that in a milieu dominated by alternative cytokines, where the ultimate goal is to return to homeostasis, the induction of gap junctions increases the ability to transmit or receive regulatory signals [ 38 ] that could facilitate the return to normal housekeeping functions. Conclusions The early-phase transcriptional profile presented in this study may not comprehensively parallel the plethora of biological effects that a given cytokine can induce under in vitro or, most importantly, in vivo conditions. Secondary, autocrine and paracrine modulation through the cytokine network following a primary stimulation may introduce novel on and off switches that could override the original signal. Nevertheless, this analysis is directly informative on the primary effect of individual cytokines on the early stages of LPS stimulation and, therefore, may be most informative on the way MP maturation may be polarized at the early stages of the immune response. The clustering of most cytokines into two main groups suggests that their control of central switches (NFκB), or regulatory molecules (cytokines, MMPs, gap junctions, cytotoxic molecules, migratory markers) is essentially bimodal. This polarization program turns MPs to a 'cytotoxic' or a 'symbiotic' phenotype [ 18 ]. In physiologic conditions, this dualism is probably modulated by a multiplicity of factors: the extent and duration of the environmental insult and the conditions of the resulting microenvironment. Possibly, predominant and persistent stimulation by pathogen components (such as LPS) may polarize MP towards the cytotoxic phenotype. A predominantly regulatory response is then mounted by the host, mediated by alternative cytokines that would take over at a later stage to induce a symbiotic phenotype aimed at resuming homeostasis upon pathogen clearance. Materials and methods MP separation and FACS staining Peripheral blood mononuclear cells (PBMCs) from an HLA-A*0201-positive healthy caucasian male donor age 35 were collected at the Department of Transfusion Medicine, NIH. PBMCs were isolated by Ficoll gradient separation and frozen until analysis. After thawing, PBMCs were kept overnight in 175-cm 2 tissue-culture flasks (Costar) in complete medium (CM) consisting of Iscove's medium (Biofluids) supplemented with 10% heat-inactivated human AB-serum (Gemini Bioproducts), 10 mM HEPES buffer (Cellgro; Mediatech), 0.03% l-glutamine (Biofluids), 100 U/ml penicillin/streptomycin (Biofluids), 10 μg/ml ciprofloxacin (Bayer), and 0.5 mg/ml amphotericin B (Biofluids). Adherent and non-adherent cells were gently removed from the flask and centrifugated. MPs were separated by negative selection using the MP isolation kit and an autoMACS system (Miltenyi). Before and after separation cells were stained with anti-CD14-FITC (Becton Dickinson), and analyzed using a FACScalibur flow cytometer and CellQuest software (Becton Dickinson). Stimulation of MPs and RNA isolation Negatively selected CD14 + cells were washed twice with serum-free OPTI-MEM (OM) medium (Gibco-BRL) prepared similarly to CM. CD14 + cells were then seeded at a concentration of 1 × 10 6 /ml in 10 ml OM in 25 cm 2 flasks (Falcon) and stimulated with 5 μg/ml LPS (Sigma) for 1 h. LPS was used at 5 μg/ml to simulate maximal pathogen exposure as in [ 6 ]. No LPS was added to the non-stimulation control flask. After 1 h, 42 cytokines, chemokines and soluble factors were added individually to the MP suspensions (Table 1 ). Then, 4 and 9 h after LPS stimulation, MPs were harvested, washed twice in PBS and lysed for RNA isolation using 700 μl RNeasy lysis buffer (Qiagen) per 25 cm 2 flask, according to the manufacturer's protocol. Probe preparation, amplification and hybridization to microarrays Total RNA was isolated using RNeasy minikits (Qiagen). Amplified antisense RNA (aRNA) was prepared from total RNA (0.5-3 μg) according the protocol previously described by us [ 9 , 39 ]. Test samples were labeled with Cy5-dUTP (Amersham) while the reference sample (pooled normal donor PBMCs) was labeled with Cy3-dUTP. Test-reference sample pairs were mixed and co-hybridized to 17K cDNA microarrays. Microarrays and statistical analyses Hybridized arrays were scanned at 10-μm resolution on a GenePix 4000 scanner (Axon Instruments) at variable PMT voltage to obtain maximal signal intensities with less than 1% probe saturation. Resulting jpeg and data files were analyzed via mAdb Gateway Analysis tool [ 40 ]. Data were further analyzed using Cluster and TreeView software [ 12 ] and Partek Pro software (Partek). The global gene-expression profiling of 4- and 9-h treated and untreated MP consisted of 98 experimental samples. Subsequent low-stringency filtering (80% gene presence across all experiments and removal of genes that did not have a log 2 ≥ 1.2: 2.3 ratio in at least one of the samples) selected 10,370 genes for further analysis. Clustering of experimental samples according to Eisen et al . [ 12 ] was based on these genes. Gene ratios were average corrected across experimental samples and displayed according to the central method for display using a normalization factor as recommended by Ross [ 41 ]. NFκB protein activation analysis MPs separated from peripheral blood by adherence were stimulated for 1 h with LPS and for an additional 30 min with cytokines selected from the alternative group (IL-4, IL-13, IL-1α) or the classical group (IFN-γ, IL-6, IL-3). In addition, TNF-α was tested. After 90 min stimulation, cytoplasmic cell extracts were isolated using a cytoplasmic and nuclear extract kit (Active Motif), and the TransAM NFκB transcription factor kit (Active Motif) was used to detect activation of NFκB subunits p50, p52, p65, c-Rel and RelB, according to the manufacturer's protocol. Real-time quantitative RT-PCR MPs obtained from PBMC of five normal caucasian donors (three males, two females, age range: 35-55 years old) were stimulated with four cytokines/soluble factors selected from the alternative group (IL-13, IL-1α, IL-4, TGF-β) and four from the classical groups (FLT-3L, IFN-γ, IL-3, CD40L). TaqMan real-time PCR was performed on amplified RNA material isolated after stimulation for 4 h in conditions identical to those applied for the cDNA array study to validate the expression of the following 10 genes: TRAF binding protein, NFκB-p105/50, MMP9, MMP19, MCP-1, mannose receptor, IL-24, IL-1R, IL-1A and FADD-MORT. An ABI Prism 7900 HT sequence detection system with 384-well capability (Applied Biosystems) was used for detection. Primers and TaqMan probes (Biosource) were designed to span exon-intron junctions and to generate amplicons of less than 150 bp. TaqMan probes were labeled at the 5' end with the reporter dye molecule FAM (6-carboxyfluorescein; emission λ max = 518 nm) and at the 3' end with the quencher dye molecule TAMRA (6-carboxytetramethylrhodamine; emission λ max = 582 nm). The following are the sequences for forward (f) and reverse (r) primer and probe (p) pairs: IL-1α f: TGTATGTGACTGCCCAAGATGAA IL-1α r: ACTACCTGTGATGGTTTTGGGTATC IL-1α p: FAM-AGTGCTGCTGAAGGAGATGCCTG-TAMRA IL-1 rec. f: TGTCACCGGCCAGTTGAGT IL-1 rec. r: GCACTGGGTCATCTTCATCAATT IL1 rec p: FAM-ACATTGCTTACTGGAAGTGGAATGGGTCAG-TAMRA TRAF bp f: TTGCTTACAG AGGTGTCTCAACAAG TRAF bp r: CTCCGGATTTGTTCTGTCAGTTC TRAF bp p: FAM-AGCAAAGTGTATTCCAGCAATGGTGTGTCC-TAMRA MMP9 f: TGGATCCAAAACTACTCGGAAGA MMP9 r: GAAGGCGCGGGCAAA MMP9 p: FAM-CGCGGGCGGTGATTGACGAC-TAMRA MMP19 f: GACGAGCTAGCCCGAACTGA MMP19 r: TTTGGCACTCCCGTAAACAAA MMP19 p: FAM-TCAGCAGCTACCCCAAACCAATCAAGG-TAMRA Mannose receptor f: CTAAACCTACTCATGAATTACTTACAACAAAAG Mannose receptor. r: CTCCGGCCACGTTGGA Mannose receptor p: FAM-ACACAAGGAAGATGGACCCTTCTAAACCGTC-TAMRA FADD-MORT f: GGTGGCTGACCTGGTACAAGA FADD-MORT r: ACATGGCCCCACTCCTGTT FADD-MORT p: FAM-TTCAGCAGGCCCGTGACCTCCA-TAMRA NFκB p105/50 f: CTACACCGAAGCAATTGAAGTGA NFκB p105/50 r: CAGCGAGTGGGCCTGAGA NFκB p105/50 p: FAM-CAGGCAGCCTCCAGCCCAGTGA-TAMRA IL-24 f: AAGAAAATGAGATGTTTTCCATCAGA IL-24 r: CTGTTTGAATGCTCTCCGGAAT IL-24 p: FAM-ACAGTGCACACAGGCGGTTTCTGC-TAMRA MCP-1 f: CATGGTACTAGTGTTTTTTAGATACAGAGACTT MCP-1 r: TAATGATTCTTGCAAAGACCCTCAA MCP-1 p: FAM-AACCACAGTTCTACCCCTGGGATG-TAMRA Standards for the selected genes were amplified by reverse transcriptase primer-specific amplification of 6 μg antisense RNA obtained from PBMCs stimulated in vitro with IL-2 (300 IU/ml and Flu M1 peptide) and reverse transcribed using random dN 6 primers (Boehringer Mannheim). Amplified cDNA standards were quantified by spectrometry and the number of copies was calculated using the Oligo Calculator software [ 42 ]. Six micrograms of test antisense RNA samples were converted to cDNA using random primers and were immediately used for quantitative real-time PCR (RT-PCR). RT-PCR reactions of cDNA samples were conducted in a total volume of 20 μl, including 1 μl cDNA, 1x TaqMan Master Mix (Applied Biosystems), 2 μl of 20 μM primers and 1 μl of 12.5 μM probe. Thermal cycler parameters included 2 min at 50°C, 10 min 95°C and 40 cycles involving denaturation at 95°C for 15 sec, annealing-extension at 60°C for 1 min. Linear regression analyses of all standard curves were 0.98 or greater. Standard curve extrapolation of copy number and quantity means were performed using the ABI Prism SDS 2.1 software (Applied Biosystems). Normalization of samples was performed by dividing the quantity mean of the gene of interest run in duplicate by the quantity mean of reference actin filament associated protein (AFAP) gene × 10 5 [ 43 ]. Additional data files The following additional data are available with the online version of this article. Additional data file 1 is a spread sheet containing microarray raw data (intensity ratios of test versus reference samples) subsequently subjected to cluster analysis. Supplementary Material Additional data file 1 A spread sheet containing microarray raw data (intensity ratios of test versus reference samples) subsequently subjected to cluster analysis Click here for additional data file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC551535.xml

535931

Social Phobia in an Italian region: do Italian studies show lower frequencies than community surveys conducted in other European countries?

Background The lifetime prevalence of Social Phobia (SP) in European countries other than Italy has been estimated to range from 3.5% to 16.0%. The aim of this study was to assess the frequency of SP in Sardinia (Italy) in order to verify the evidence of a lower frequency of SP in Italy observed in previous studies (from 1.0% to 3.1%). Methods A randomised cross sample of 1040 subjects, living in Cagliari, in rural areas, and in a mining district in Sardinia were interviewed using a Simplified version of the Composite International Diagnostic Interview (CIDIS). Diagnoses were made according to the 10 th International Classification of Diseases (ICD-10). Results Lifetime prevalence of SP was 2.2% (males: 1.5%, females: 2.8%) whereas 6-month prevalence resulted in 1.5% (males: 0.9%, females: 2.1%). Mean age at onset was 16.2 ± 9.3 years. A statistically significant association was found with Depressive Episode, Dysthymia and Generalized Anxiety Disorder. Conclusions The study is consistent with findings reported in several previous studies of a lower prevalence of SP in Italy. Furthermore, the results confirm the fact that SP, due to its early onset, might constitute an ideal target for early treatment aimed at preventing both the accumulation of social disabilities and impairments caused by anxiety and avoidance behaviour, as well as the onset of more serious, associated complications in later stages of the illness.

Background Several epidemiological studies have attempted to describe the prevalence, socio-demographic characteristics, comorbidity, and severity of clinical manifestations of Social Phobia (SP). Quality of life and functional status of affected individuals have also been investigated [ 1 ]. Lifetime prevalence of SP in European societies other than Italy ranges from 3.5% to 16.0% [ 2 ]. As discussed in several excellent review papers, these rate differences were partly attributed to probable genetic or cultural factors [ 3 ]. Furthermore, major methodological differences (type of diagnostic criteria used, assessment tools, age of the sample) affecting the estimates have been demonstrated [ 1 ]. This study, part of an extended epidemiological investigation "Health in Sardinia," aimed to assess the prevalence rates of SP in Sardinia (Italy) in order to confirm the evidence of low SP prevalence rates ranging from 1.0% to 3.1% observed in previous research projects in Italy [ 4 - 6 ]. The study also intended to evaluate the treatments and to verify the comorbid psychiatric disorders in the identified people with Social Phobia. Methods The sample, which had already been examined and described in greater detail in a previously published study [ 7 ], consisted of 1040 subjects recruited throughout the island of Sardinia, Italy. 393 subjects came from the city of Cagliari, 344 from rural areas, and 303 from an industrial mining district, thus representing fairly well all socio-economic strata present on the island. The age distribution of the sample is shown in table 1 , with age range from 18 to 89 years. 79.2% out of a total of 1313 subjects approached, agreed to take part in the study, 12.5% refused to participate, and 8.3% could not be traced; the final sample did not differ respect to the population of origin with reference to the variables applied in stratification (Table 1 ). All subjects were interviewed "face to face" by trained physicians using a Simplified version of Composite International Diagnostic Interview (CIDI) [ 8 ], hence the acronym CIDI "Simplified" (CIDIS) [ 9 , 10 ]. The version used in this study had been translated into Italian and back-translated into the original French language under blind conditions respect to the first translation, by a bilingual researcher; approval of the final version was obtained from the original authors [ 11 ]. The CIDI structured interview in its various versions currently represents the most widely used diagnostic tool in psychiatric epidemiological studies conducted on the general population [ 12 ]. The CIDIS is a highly structured tool made up of 5 sections which investigate respectively: Somatoform Disorders and General Medical Conditions, Anxiety Disorders, Depressive Disorders, Substance-Related Disorders (Alcohol-Related Disorders and Substance-Related Disorders) and Eating Disorders. The computer elaboration of data obtained through application of the CIDIS enables calculation of both "lifetime prevalence" and, for the preceding six months, a series of psychiatric disorders (those more frequently observed in the general population) according to the ICD-10 diagnostic system [ 13 ]. This interview moreover enhances identification of both the type of therapist referred to and treatment already used by each patient for his or her specific problem and definition of degree of impact of the problem on the subject's daily routine. An ad hoc computer algorithm ascertained the presence of the disorders according to ICD-10 criteria [ 13 ], both in the past six months and in the lifetime. The items of the CIDIS concerning Social Phobia and the related algorithm is reported in Figure 1 . Table 1 Percentage of subdivision according to age, sex, and marital status of the sample. N (1040) MALES 461 (44.3%) FEMALES 579 (55.7%) AGE 18–24 146 (14.0%) AGE 25–44 353 (33.9%) AGE 45–64 310 (29.8%) AGE >64 231 (22.2%) UNMARRIED 370 (35.6%) MARRIED 571 (54.9%) WIDOWED/SEPARATED/DIVORCED 99 (9.5%) Figure 1 The Composite International Diagnostic Interview Simplified (CIDIS) [9, 10] algorithm for the diagnosis of Social Phobia. Results Estimates of lifetime SP prevalence of 2.2% were found (males: 1.5%, females: 2.8%), with no statistically significant difference between the sexes (χ 2 = 1.2, 1DF, P = 0.25). 6-month prevalence rates were lower (total: 1.5%, males: 0.9%; females: 1%; no significant difference between the sexes, χ 2 = 1.6, 1DF, P = 0.13). Table 2 compares results obtained in this study with those of the major research projects carried out in Europe and the USA [ 4 , 5 , 14 - 23 ]. Table 3 illustrates the lifetime prevalence according to age and sex, and Table 4 shows the 6-month prevalence according to age and sex. An increased frequency of SP among the younger age groups was observed, although distribution in males resulted as being less homogeneous. In no case statistical significance was reached. Lifetime prevalence rates of SP respect to marital status of subjects studied were as follows: 3.0% among the unmarried; 1.2% among married subjects; and 3.5% among the divorced, separated and widowed (χ 2 = 5.8, 1DF, P = 0.06). Mean age at onset was 16.2 ± 9.3 years. Table 2 Lifetime prevalence of Social Phobia in the general population of Europe and USA. Country Reference Diagnostic criteria N Male Female Total Italy Faravelli et al., 1989 [4] DSM-III-R 1110 1.0 Switzerland Wacker et al., 1992 [16] DSM-III-R 470 16.0 ICD-10 9.6 ECA (USA) Schneier et al., 1992 [15] DIS 18571 2.0 3.1 2.4 Iceland Lindal and Stefansson, 1993 [32] DSM-III 862 2.5 4.5 3.5 Switzerland Degonda and Angst, 1993 [21] DSM-III 591 3.1 5.7 4.4 NCS (USA) Kessler et al., 1994 [17] CIDI 8098 11.1 15.5 13.3 France Lepine and Lellouch, 1995 [19] 2.1 5.4 4.1 Germany Wittchen et al., 1998 [23] DSM-IV 3021 2.2 4.8 3.5 Italy Carta and Rudas, 1998 [6] CIDI 783 1.7 4.6 3.1 Spain Arillo et al., 1998 [cited in 33] DIS 237 8.9 Netherlands Bijl et al., 1998 [20] DSM-III-R 7076 5.9 9.7 7.8 France Lépine and Pélissolo, 1999 [22] DSM-IV 7.3 Italy Faravelli et al., 2000 [5] FPI/CIDI 2355 1.9 4.0 3.1 Italy Carta et al., 2002 [7] CIDI 1040 1.5 2.8 2.2 Table 3 Lifetime prevalence N (%) of Social Phobia according to age and sex. Age Male OR Female OR Total OR <25 1 (1.3) 0.9 4 (5.0) 2.3 5 (3.4) 1.7 25–44 4 (2.5) 2.5 4 (2.1) 0.7 8 (2.2) 1.1 45–64 1 (0.7) 0.4 4 (2.4) 0.8 5 (1.6) 0.6 >65 1 (1.1) 0.7 4 (2.8) 1.1 5 (2.2) 0.9 Male according to age χ 2 with Yathes correction = 1.6, 3 df, p = 0.89; female according to age χ 2 with Yathes correction = 2.2, 3 df, p < 0.71; total sample according to age χ 2 with Yathes correction = 1.8, 3 df, p = 0.82 Table 4 Six month prevalence N (%) of Social Phobia according to age and sex. Age Male OR Female OR Total OR <25 1 (1.5) 1.8 3 (3.7) 2.0 4 (2.7) 2.1 25–44 1 (0.6) 0.6 3 (0.7) 0.5 4 (1.1) 0.6 45–64 1 (0.7) 0.7 2 (1.1) 0.5 3 (0.9) 0.5 >65 1 (1.1) 1.4 4 (2.8) 1.4 5 (2.1) 1.6 Male according to age χ 2 = 0.6, 3 df, p = 0.98; female according to age χ 2 with Yathes correction = 3.4, 3 df, p = 0.44; total sample according to age χ 2 with Yathes correction = 3.4, 3 df, p = 0.44 During the week prior to the study, 8 (50.0%) out of the 16 subjects who had been diagnosed with SP over the past six months had been taking low doses of anxiolytics (less than the equivalent of 2 mg of Lorazepam). 3 (18.7%) were on antidepressants, one of whom (6.2%) at non-therapeutic doses, and 1 subject (6.2%) was undergoing cognitive behavioural psychotherapy. The remaining 6 subjects (37.5%) were not on any treatment. Six out of the 10 treated subjects (60%) were being supervised by their general practitioners (GP), 1 (10%) by a neurologist and 2 (20%) by psychiatrists. All subjects presented some degree of comorbidity with Depressive Episodes (DE), Panic Attack Disorder (PAD), and agoraphobia (AP). Only 1 subject (10%) was undergoing cognitive behavioural therapy with a psychologist/psychotherapist. Table 5 illustrates the rate of comorbidity with major psychiatric disorders observed in the general population, as well as the degree (OR) of associated disorders observed with regard to frequency reported for the latter in populations not affected by SP. A statistically significant difference was revealed for association with DE, Dysthymia (DD), and Generalized Anxiety Disorder (GAD). In spite of their increased frequency among patients affected by SP, disorders such as PAD and Specific Phobia do not represent a statistically significant association. The mean age at onset of comorbid DE was 6.5 ± 6.6 years subsequent to onset of SP, whereas GAD was manifested at a mean of 4.3 ± 7.8 years later. Table 5 Lifetime comorbidity of Social Phobia. N (%) OR χ 2 Depressive Episode (DE) 9 (39.1) 4.3 11.1* Dysthymia (DD) 5 (21.7) 7.1 14.1* Generalized Anxiety Disorder (GAD) 10 (43.4) 6.5 20.9* Panic Attack Disorder (PAD) 2 (8.7) 3.3 1.1 Specific Phobia 1 (4.3) 8.6 1.6 *p < 0.001 Discussion Several epidemiological studies carried out in Europe (Switzerland [ 16 , 21 ], France [ 19 , 22 ], Germany [ 23 ]) and in the USA [ 15 , 17 ], recently reviewed by Furmark [ 2 ], suggest that SP is one of the more frequently observed anxiety disorders in the general population in Western countries. However, frequency rates reported in the various studies differ from country to country and according to time and evaluation methods used. Indeed, the two American studies [ 15 , 17 ] carried out at an interval of approximately 15 years, illustrate distinctly contrasting results, and it is hard to establish what factors really determine variance in findings. The present study is consistent with the tendency towards rather low lifetime prevalence rates of SP observed in other Italian research projects. If we take into account only those European researches that adopted ICD-10 diagnosis, our results seem to indicate a lower frequency than a study carried out in Formentera, Spain (lifetime prevalence of 2.8% against 8.9% in females [cited in [ 33 ]]) and Basel, Switzerland (lifetime prevalence in the total sample 2.2 against 9.6% [ 16 ]). However, lower frequencies emerged also in Italian surveys conducted using different methods [ 5 , 7 ]. The Italian studies were carried out over a considerable period of time and in two distinct areas: the Florence area [ 4 , 5 ] and Sardinia [[ 6 ], the present study]. It is therefore quite likely that the lower frequencies observed may be the result of an effectively reduced vulnerability of Italians to SP. The results of this study can indeed be assumed as being influenced by several cultural variables – the genetic diversity of the two populations examined, as well as the considerable genetic variance of the Sardinian population respect to other European populations [ 24 , 25 ]. Findings for six-month prevalence show lower frequencies than those evidenced in the recent studies: 1.5% against 4.0% in Iceland [ 26 ] and against 4.0% in Munich [ 14 ]. Our rates are lower than E.C.A. findings: 2.7% in Duke [ 27 ] and 2.2% in Baltimore, New Haven, and Saint Louis [ 28 ]. The 6-month prevalence rates obtained in Sardinia are similar only to data reported for Edmonton (1.2%) from a survey carried out more than 15 years ago [ 29 ] and published in 1994 [ 30 ]. However, the prevalence data emerging from this study further justify the considerable interest shown in this disorder from both a medical and a social point of view. This research confirms the fact that onset of the illness occurs primarily during childhood and adolescence [ 31 ], thus underlining the suitability of the condition as a candidate for early treatment aimed at preventing both the slow accumulation of social disabilities and impairment caused by anxiety and by avoidance behaviour, as well as the onset of more serious complications (e.g., DE or GAD), which may be manifested many years after onset of SP. Indeed, subjects affected by social phobia presented a high risk of comorbidity with both the latter disorders and DD [ 1 ] . It should be underlined that 60% of subjects undergoing treatment (not all affected subjects) chose their general practitioner (GP). This view is corroborated by the fact that those patients treated by a psychiatrist invariably presented comorbidity with DE, PAD, and AP, thereby representing the more severely affected from a psychopathological point of view. Overall however, the low rate of patients with SP treated with first line-treatments is alarmingly low. In the future, serious attempts should be made to improve the GPs' abilities to recognise SP in order to prevent the use of inappropriate treatments, such as insufficient doses of benzodiazepines, which may be linked to the physician's incorrect diagnosing of the disorder. Due to the fact that subjects affected by SP most frequently refer to their GP, the importance of preventing SP as opposed to other types of disorders, as well as the markedly incapacitating nature of SP reinforce this necessity for a better training of GPs. It is however mandatory to briefly acknowledge some potential limitations of this study. First, the number of cases and the sample size is too small to allow firm conclusions to be drawn concerning the true rate and the degree to which they might actually differ from previous studies with higher estimates. Secondly, differences in the assessment strategy might have resulted in a diminished comparability. Conclusions The study is consistent with findings reported in several previous studies of a lower prevalence of Social Phobia in Italy and confirms the fact that onset of the illness occurs primarily during childhood and adolescence. Furthermore, the results confirm the fact that SP, due to its early onset, might constitute an ideal target for early treatment aimed at preventing both the accumulation of social disabilities and impairments caused by anxiety and avoidance behaviour, as well as the onset of more serious, associated complications in later stages of the illness, or many years after onset of SP. Competing interests The authors declare that they have no competing interests. Authors' contributions MGC participated in the design of the study, performed the statistical analysis and drafted the manuscript. UHW participated in the statistical analysis and drafted the manuscript. MCH, MC, BC, LDO, MAR conceived of the study, and participated in its design and coordination. 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/PMC535931.xml

544931

Why Bad Boys Always Get the Girl and Other Tales of Evolutionary Madness

Biologists studying the evolution of mate choice have their work cut out for them. Not only is there little agreement on how best to determine the interplay between mate choice and fitness, there's not even consensus on how to estimate fitness. (Fitness being an individual's success in passing their genes on to future generations.) Why females fall for captivating males that give them nothing but trouble is especially puzzling. Such males tend to contribute their genes and little else, leaving the female to spend precious resources bearing and raising her young, efforts that often cut her life short. Of course, females are not all innocence and light in the mating game: some black widow species, for example, famously make dad a post-coitus snack. Still, females typically incur more costs than males in rearing offspring, especially when they choose flashy mates. So why do they do it? One model holds that females put up with deadbeat dads because the benefits, though indirect, outweigh the costs: that is, attractive males are more likely to grace their offspring with good genes that increase survival (“I'm so fit I can afford to waste energy on this excessively exuberant tail”) or with sexy traits that confer mating success (“my magnificent plumage may decrease my survival, but I get lots of dates”). Another model argues that selection for such indirect benefits is much weaker than direct selection on genes that affect mate preference and thus is likely to exert little influence on mating preference. In a new study, Megan Head and her colleagues navigate this intellectual minefield by studying the mating behavior of crickets. The authors paired females with either “attractive” or “unattractive” males (see below) and measured a variety of fitness components to estimate the overall fitness consequences of the various unions. Female crickets, they found, pay a high price for mating with attractive males. But when the fitness consequences for their sons and daughters are taken into account, mate costs are balanced by, and may even be outweighed by, the indirect benefits of spawning offspring with elevated fitness. This benefit stems in large part, the authors argue, from siring sexy sons. How does one distinguish lothario from loser in the cricket world? By running a cricket tournament, of course. For crickets to mate successfully, the female must mount the male so their genitalia align. Noting that females produce more eggs for males they mount quickly, the authors use time to mount as a measure of male attractiveness. An attractive cricket? In the first round of the tournament, Head and colleagues paired males with a randomly assigned female; after mounting, but before copulation, the couples were separated. This continued until half of all females had mounted a male. (Under tournament rules, crickets had to be in the first half of a given category to qualify for the next round.) In round two, a new female was randomly assigned to each male. Males that had been mounted in the first round and remounted in the second were deemed “attractive.” Males rebuffed in the first round that remained unmounted longest in round two were “unattractive.” Females were randomly assigned to males that were either attractive or unattractive. Equivalent males were swapped out every seven days to control for any individual quirks that might bias the results. To estimate the total fitness of the participants, the authors measured both direct and indirect fitness components, such as female hatching success and reproductive effort (egg number and size), as well as sons' attractiveness and the number of eggs laid by daughters. Females that mated with attractive males produced daughters that laid more eggs within a given time and sons that were more attractive, though they had lower survival. Thus, by evaluating both the direct effects of female lifetime fecundity and the indirect effects of offspring fitness, the authors determined the net consequences of a mating strategy. And once again, it's mom's sacrifices that keep things on track. With this approach, Head and colleagues bridge the gap between empirical studies of mating choice evolution, which rely largely on rate-insensitive measures (such as counting grandchildren), and theoretical studies, which typically use rate-sensitive measures. Their results suggest that there may be selection for choosing costly mates and that generating a reliable analysis of the fitness consequences requires a long view: look at the reproductive success of mom's sons and daughters before judging her bad taste in mates.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544931.xml

524182

Utilization of a ts-sacB selection system for the generation of a Mycobacterium avium serovar-8 specific glycopeptidolipid allelic exchange mutant

Background Mycobacterium avium are ubiquitous environmental organisms and a cause of disseminated infection in patients with end-stage AIDS. The glycopeptidolipids (GPL) of M. avium are proposed to participate in the pathogenesis of this organism, however, establishment of a clear role for GPL in disease production has been limited by the inability to genetically manipulate M. avium . Methods To be able to study the role of the GPL in M. avium pathogenesis, a ts-sacB selection system, not previously used in M. avium , was employed as a means to achieve homologous recombination for the rhamnosyltransferase ( rtfA ) gene of a pathogenic serovar 8 strain of M. avium to prevent addition of serovar-specific sugars to rhamnose of the fatty acyl-peptide backbone of GPL. The genotype of the resultant rtfA mutant was confirmed by polymerase chain reaction and southern hybridization. Disruption in the proximal sugar of the haptenic oligosaccharide resulted in the loss of serovar specific GPL with no change in the pattern of non-serovar specific GPL moieties as shown by thin layer chromatography and gas chromatography/mass spectrometry. Complementation of wild type (wt) rtfA in trans through an integrative plasmid restored serovar-8 specific GPL expression identical to wt serovar 8 parent strain. Results In this study, we affirm our results that rtfA encodes an enzyme responsible for the transfer of Rha to 6d-Tal and provide evidence of a second allelic exchange mutagenesis system suitable for M. avium . Conclusion We report the second allelic exchange system for M. avium utilizing ts-sacB as double-negative and xylE as positive counter-selection markers, respectively. This system of allelic exchange would be especially useful for M. avium strains that demonstrate significant isoniazid (INH) resistance despite transformation with katG . Through the construction of mutants in GPL or other mycobacterial components, their roles in M. avium pathogenesis, biosynthesis, or drug resistance can be studied in a consistent manner.

Background Mycobacterium avium is a frequent cause of disseminated infection among patients with end-stage AIDS [ 9 , 11 , 19 ]. M. avium can also present with a similar spectrum of pulmonary and extra pulmonary syndromes as Mycobacterium tuberculosis [ 27 ] including the establishment of latent infection that can reactivate concomitant with immune suppression [ 13 ]. While significant advances have been made in deciphering the host responses against M. avium infection, there is only a rudimentary understanding of the bacterial factors involved in the pathogenesis of infection [ 14 , 22 ]. Numerous studies have implicated the cell wall lipids in mycobacterial pathogenesis. For M. avium , there is evidence that the glycopeptidolipids (GPL), as the dominant lipid for this species, may negatively affect host immunity [ 4 , 25 ]. Study of GPL in M. avium pathogenesis has been limited by a lack of suitable genetic techniques to be able to create site directed knockouts. Further, as reviewed below, there is controversy as to which portion of GPL predominates in disease production. The GPLs are comprised of a lipopeptide (LP) core of D-phenylalanine-D- allo threonine-D-alanine-alaninol with a fatty acyl group N-linked to the phenylalanine residue and a methylated rhamnose modifying the terminal alaninol. The LP core is glycosylated at D- allo threonine with 6-deoxytalose (6dTal) to form non-specific GPL (nsGPL) and is further glycosylated at 6dTal with a haptenic oligosaccharide to yield serovar-specific GPL (ssGPL). All serovars maintain a common α-L-rhamnopyranosyl-(1→2)-6dTal [ 6 ]. Historically, the predominance of serovars 1, 4, and 8, among patients with disseminated infection [ 10 , 26 ] has been suggested as evidence to support a role for the oligosaccharide moiety of GPL in pathogenesis, but may conversely represent the fact that a restricted set of clones are disease producing. More direct evidence of a role of the GPL oligosaccharide in pathogenesis is provided by the study of Minami that demonstrated that heat-killed Staphylococcus aureus coated with M. avium GPL promote phagocytosis and inhibit phagolysosomal fusion in relation to serovar [ 17 ]. Other studies have, however, suggested a dominant role for the lipopeptide core in pathogenesis [ 5 ]. Significantly limiting the development of a consistent framework of the role of GPL in mycobacterial pathogenesis has been the inability to construct isogenic strains differing in GPL structure, necessitating the comparison of genetically distinct strains of differing serotypes. To study the role of the serovar-specific oligosaccharide moiety of GPL in the pathogenesis of M. avium , an allelic exchange mutant in rtfA was created for a pathogenic serovar 8 strain to yield a strain deficient in ssGPL. Homologous recombination was performed using a novel allelic exchange vector that incorporated a temperature-sensitive mycobacterial origin of replication ( ts-oriM ) and sacB as counter selective markers [ 21 ] and xylE [ 7 ] as a positive selection marker. Complementation of rtfA in trans through an integrative plasmid restored serovar-8 specific GPL expression identical to wild type (wt) serovar 8 smooth opaque (SmO) parent strain. In addition to reaffirming our results for serovar 2 [ 15 ] that rtfA encodes an enzyme responsible only for the transfer of Rha to 6d-Tal to form the serovar-8 specific oligosaccharide, this study delineates a second system of allelic exchange mutagenesis for M. avium . Methods Bacterial strains and plasmids Escherichia coli strain DH5α was used as the host strain for plasmid construction and propagation. Wild type and recombinant M. avium and Mycobacterium smegmatis strains were grown in Middlebrook 7H9 broth or 7H11 agar supplemented with 10% OADC (Difco Laboratories, Detroit, MI) at 37°C, except where indicated. M. smegmatis mc 2 155 [ 24 ] was employed as a test strain for mycobacterial shuttle vectors. M. avium 920A6 is a serovar 8 bloodstream isolate cultured from a patient with AIDS [ 1 ]. Transformation of E. coli and M. smegmatis was performed as described [ 23 , 24 ]. Transformation of M. avium was performed according to the protocol of Lee et al . [ 12 ]. For E. coli , selection was carried out using ampicillin at 50 μg ml -1 and kanamycin at 25 μg ml -1 . For M. smegmatis and M. avium , selection was accomplished using hygromycin at 100 μg ml -1 , gentamicin at 100 μg ml -1 , and kanamycin at 50 μg ml -1 . Construction of allelic exchange vector pVAP39 The allelic exchange vector pVAP39 was created in a manner similar to allelic exchange vector pVAP41 [ 15 ] to include counter selection markers ts-oriM , sacB , and the hygromycin resistance gene ( hyg ); and positive selection marker xylE . Construction of allelic exchange vector pVAP39 is shown in Fig. 1 . The 1.1 kb BamHI-XbaI fragment of pXYL4 containing the xylE gene [ 21 ] was ligated into the BamHI site of pPR27, containing a temperature-sensitive origin of replication of M. fortuitum plasmid pAL5000 and sacB [ 21 ] to create pVAP38 (10.8 kb). The 3.2 kb XbaI-XhoI fragment containing rtfA::hyg , isolated from pVAP37 [ 15 ], was blunt-ligated into pVAP38 to create pVAP39 (14.1 kb). The presence of rtfA::hyg in pVAP38 was confirmed by PCR and Southern blot analysis as described [ 15 ]. Expression of XylE was detected by applying one drop of filter-sterilized 1.1% catechol solution (1.1% catechol in 50 mM potassium phosphate buffer, pH 7.5) to individual colonies to detect a yellow color [ 20 ]. Plasmid pVAP42 was constructed by ligating the amplified wt rtfA gene with HindIII overhangs into the HindIII site of plasmid pMVGFP (kanamycin-resistant, GFP-positive, [ 12 ]). Plasmid pVAP52 was constructed by ligating the amplified wt rtfA gene with HindIII overhangs into the HindIII site of plasmid pIGFP2 (kanamycin-resistant, GFP-positive, [ 12 ]). Figure 1 Construction of allelic exchange vector pVAP39. See Methods and Reference 15 for details. Isolation and analysis of GPL Colonies of wt, mutant, and complemented strains were collected from 7H10 plates. Procedures for purification of alkaline stable GPLs, and alditol acetate analyses of sugar moieties by gas chromatography/mass spectrometry (GC/MS) were performed as described by Eckstein et al [ 8 ]. Results Selection of rtfA allelic exchange mutants of M. avium 920A6 strain The expression of xylE was first examined for M. avium since there is minimal published data on the use of this marker in mycobacteria. After construction, pVAP38 was first electroporated into M. smegmatis strain mc 2 155 to assess expression in a test system. All (100%) of gentamicin-resistant colonies expressed XylE as determined by a yellow color change after application of catechol. M. avium 920A6 SmO was then transformed with pVAP39 and selected at 32°C on 7H11 medium containing 100 μg ml -1 hygromycin. Yellow colonies were easily detected, indicating that xylE represents a suitable marker for M. avium . However, only 25–40% of hygromycin-resistant colonies yielded a yellow color, indicating a high rate of spontaneous hygromycin resistance when M. avium is transformed at 32°C, with a final efficiency of transformation of 1–8 × 10 2 transformants per μg of DNA. These results contrasted with our earlier observations that transformation of M. avium with non-temperature sensitive plasmids yielded less than 5% of spontaneously resistant colonies and an efficiency of transformation 1.6-log higher (8 × 10 3 transformants per μg of DNA, [ 12 ]). To derive an allelic exchange mutant, a representative hygromycin-resistant, XylE-positive colony of M. avium 920A6 SmO/pVAP39 was inoculated into 7H9 medium for 3 weeks at 32°C to late log phase. Growth in hygromycin-free medium allowed for spontaneous loss of plasmid DNA. Moreover, expansion in hygromycin-free medium limited the appearance of spontaneous hygromycin resistance (unpublished data). Selection for allelic exchange mutants was performed at 39°C on 7H11 medium containing 100 μg ml -1 hygromycin and 2% (w/v) sucrose to isolate single colonies. Incubation at 39°C precludes the replication of the temperature-sensitive origin of replication of pVAP39. Hygromycin-resistant, XylE-positive colonies that arose at this non-permissive temperature represented single crossover mutants or illegitimate recombinants, whereas XylE-negative colonies represented either double crossover mutants or colonies that had lost plasmid DNA and had developed spontaneous hygromycin resistance. Growth on sucrose was used as a means to eliminate strains that retained plasmid DNA. Of >10 4 colonies screened, three XylE-negative colonies were identified of which only one (213R.4) colony with an SmO morphotype yielded a single 3.2 kb PCR product corresponding to the 1.9 kb rtfA gene interrupted with the 1.3 kb hyg cassette. Southern blot analysis confirmed that strain 213R.4 possessed only a chromosomal copy of rtfA::hyg (Fig 2a,2b ). The remaining 2 XylE-negative colonies yielded a single 1.9 kb band corresponding to the native rtfA gene, indicating spontaneous hygromycin-resistant strains. Figure 2 PCR and Southern hybridization of wild type 920A6 and ΔrtfA mutant, 213R.4. (A) PCR of wild type M. avium 920A-6 (lane 3) yielded a single 1.9 kb band corresponding rtfA whereas the rtfA mutant 213R.4 yielded a 3.2 kb band corresponding to rtfA with an inserted 1.3 kb hygromycin resistance gene cassette ( rtfA::hyg , lane 2). Vector pVAP39 served as a positive control (lane 4). Lane 1 represents a molecular weight marker. (B) Southern blot analysis of genomic DNA from wt M. avium 920A6 and clone 213R.4 was digested with HindIII and probed for rtfA . M. avium 920A6 yielded a 11.13 kb band (lane 2) whereas clone 213R.4 (lane 1) yielded a 12.49 kb band, corresponding to the incorporation of the 1.3 kb hyg gene. GPL analysis of serovar 8 M. avium strains by TLC and GC-MS Total lipids were isolated from wild type and mutant strains. Alkaline stable lipids analyzed by TLC demonstrated that wt strains 920A6 SmO and 920A6 SmT expressed serovar 8 specific GPL (Fig. 3 , lanes 1, 2). Strain 213R.4 was devoid of ssGPL but produced an identical pattern of nsGPL as the wt strains (Fig. 3 , lane 3). To confirm that the loss of serovar 8 specific GPL resulted from the disruption of rtfA , clone 213R.4 was transformed with integrative (pVAP52) plasmid to complement rtfA in trans . Strain 233R.1 created by transformation of 213R.4 with pVAP52 and thus containing only a single copy of rtfA , demonstrated a pattern of ssGPL and nsGPL similar to wild-type, serovar 8 M. avium (Fig. 3 , lane 5). Strain 277R.1 created by transformation of 213R.4 with rtfA on an episomal plasmid (pVAP42) expressed ssGPL but not nsGPL (Fig. 3 . lane 4). Figure 3 Thin layer chromatography (TLC) of alkaline-stable lipids from GPL mutants of 920A6. GPL were isolated from each strain and 100 μg of lipid was applied to each lane on a silica gel TLC plate, developed in CHCl 3 :CH 3 OH:H 2 O (65:35:4), and sprayed with H 2 SO 4 in ethanol. Lane 1, 920A6 SmO; Lane 2, 920A6 SmT; Lane 3, ΔrtfA mutant 213R.4; Lane 4, 227R.1; Lane 5, 233R.1. Wild-type strains 920A6 SmT and 920A6 SmO both expressed ssGPL and nsGPL whereas 213R.4 did not express serovar-8 specific GPL (arrow). Complementation of rtfA with a single copy integrant restored ssGPL expression for 233R.1 to a pattern similar to wild-type M. avium . Strain 227R.1 complemented with rtfA on an episomal plasmid expressed ssGPL but did not express nsGPL. Analysis of the glycosyl residues was performed by GC-MS of alditol acetate derivatives of GPL (Fig. 4 ). Relative to wt 920A6 SmO (Fig. 4 , panel B), clone 213R.4 (Fig. 4 , panel A) demonstrated loss of both the non-methylated rhamnose (Rha) of the haptenic oligosaccharide (peak 4) and the terminal glucose residue (peak 6). Clone 213R.4 however retained 3,4-O-diMe-Rha (peak 1), 3-O-Me-6dTal (peak 2), 3-O-Me-Rha (peak 3), and 6dTal (peak 5) associated with nsGPL. Individual nsGPL and ssGPL band(s) were isolated from a preparative TLC gel and each band resolved separately by TLC (Fig. 5 ) and analyzed by GC. Band α, absent from 213R.4, contained serovar 8 specific GPL. Bands β, γ2, γ3, and δ represented nsGPL bands demonstrating the sequential addition of methyl groups to Rha attached to the alaninol of the nsGPL, and 6dTal to generate serovar 8 specific GPL (band α). These data further confirm our previous results for serovar 2 [ 15 ] that rtfA encodes for the transfer of Rha to 6dTal as the proximal sugar in the oligosaccharide moiety of GPL and does not encode for the transfer of Rha to the alaninol of the GPL lipopeptide core. Figure 4 Gas chromatography of alditol derivatives of GPL of 920A6 SmO and ΔrtfA mutant 213R.4. Panel A, 213R.4; Panel B, 920A6 SmO, Panel C rhamnose (Rha) standard. Peaks 1, 2, 3, 4, 5 and 6 represent 3,4-O-dimethylrhamnose (diMe-Rha), 3-O-methyl-6dtalose (3-O-Me-6dTal), 3-O-methylrhamnose (Me-Rha), rhamnose (Rha), 6dTal, and glucose, respectively. The peak at 18.5 min. in all the panels represents the Rha standard. Peaks with asterisks (*) do not represent pattern associated with alditol acetates of known sugars. Figure 5 TLC and GC analyses of individual GPL bands (α, β, γ, γ2, γ3, δ) of wt M. avium 920A6, confirms the role of rtfA in ssGPL biosynthesis. GPL from 920A6 SmO was resolved by preparative TLC and individually resolved by analytical TLC. Each band was collected from the plate, and the lipids analyzed by GC/MS. Bands α represented serovar-8 specific GPL, bands β, γ (mix of γ2, γ3), and δ represented nsGPLs. GPL from the ΔrtfA mutant, 213R.4 was analyzed similarly and yielded identical nsGPL (data not shown). Discussion Here we report on the generation of allelic exchange mutants using a double negative-selection system utilizing a temperature sensitive origin of replication of plasmid pAL5000 and the Bacillus subtilis sacB gene. This vector (pPR27) has been used successfully to generate homologous recombinants of M. tuberculosis [ 21 ]. The inclusion of the reporter gene xylE [ 7 ], that encodes for catechol 2,3-dioxygenase and converts catechol into 2-hydroxymuconic semialdehyde [ 20 ], provided identification of true transformants and allowed for differentiation of putative double crossover mutants (XylE-negative) from single crossover mutants or illegitimate recombinants (XylE-positive). We observed a high background of hygromycin-resistant, sucrose-resistant, XylE-positive colonies after selection at 39°C on sucrose-containing media suggesting a high frequency of a single crossover or illegitimate recombination. The high number of XylE-positive clones either represented a high degree of spontaneous mutation in sacB or the inability of this gene to provide efficient counter selection as a single copy. The latter was consistent with our observation that katG expressed as a single copy-integrant did not confer INH-susceptibility to M. avium sufficient to serve as a counter selection marker [ 15 ]. Additionally, we observed a high degree of spontaneous hygromycin resistance at 32°C. Although poorly efficient, this system of allelic exchange would be useful for strains that exhibit significant isoniazid resistance despite transformation with katG , as we have observed with the smooth transparent (SmT) morphotype of 920A6 (unpublished data). In this system, for INH-resistant strains, it may be prudent to perform selections as a two-step process, i.e., growth in broth at 39°C to eliminate plasmid replication followed by selection in solid medium containing hygromycin to increase the yield of double crossover mutants in relation to spontaneous hygromycin-resistant strains. Although sacB is a useful marker for M. tuberculosis , it appears to be minimally useful as a counter-selection marker for allelic exchange in M. avium . Also, this is the first reported instance of using xylE as a marker for allelic exchange in M. avium . The glycopeptidolipids represent the most abundant cell wall component of M. avium . Studies have suggested a role for serovar-specific GPL in the pathogenesis of M. avium infection as highly antigenic molecules [ 16 ] affecting host immune function. However, these data have relied on comparisons of strains representing different serovars [ 18 ] or have used purified and/or chemically modified GPL and GPL components [ 2 , 3 , 5 , 25 ]. In this study, we disrupted the rtfA gene via homologous recombination to block the addition of rhamnose as the proximal sugar common to ssGPLs resulting in construction of isogenic mutants expressing only non-specific GPL. Complementation of the rtfA gene as a single copy integrant in trans restored ssGPL synthesis and maintained nsGPL synthesis. Complementation of the ssGPL-null mutant with rtfA on an episomal plasmid, however yielded only serovar-8 specific GPL. In the latter case, all nsGPL components (bands β, γ2, γ3, δ) were utilized as substrates for generation of ssGPL and thus were lost due to over-expression of rtfA . Also, since we do not observe any serovar-1 ssGPL (6dTal-Rha) on TLC or GC analyses, this would suggest that the serovar-8 specific GPL disaccharide Rha-Gluc was generated prior to its addition to 6dTal for the generation of ssGPL. Conclusion Insertion mutagenesis via the ts-sacB double negative and xylE counter-selection system was reported for M. avium and we were able to construct isogenic mutants devoid of serovar-8 GPL. Due to limitations of various genetic manipulation techniques, this is the second only reported allelic exchange system for M. avium . With a few experimental modifications, this system of allelic exchange would be especially useful for M. avium strains that demonstrate high levels of INH drug resistance. Finally, through the construction of mutants in GPL (or any other cellular component) synthesis, the role of M. avium GPLs (and other components) in host-pathogen interaction, immunogenesis, and other qualities such as drug resistance can be determined. List of abbreviations used GPL: glycopeptidolipid 6-d Tal: 6-deoxytalose nsGPL: non-specific glyopeptidolipid ssGPL: serovar-specific glycopeptidolipid rtfA : rhamnosyltransferase wt: wild type LP: lipopeptide SmO: smooth opaque Rha: rhamnose Hyg: hygromycin GC/MS: gas chromatography/mass spectrometry PCR: polymerase chain reaction TLC: thin-layer chromatography 3,4-O-diMe-Rha: 3,4-O-dimethyl-Rhamnose 3-O-Me-6dTal: 3-O-Methyl-6-deoxytalose 3-O-Me-Rha: 3-O-Methyl-Rhamnose SmT: smooth transparent INH: isoniazid hydrazide 6dTal-Rha: 6-deoxytalose-rhamnose Rha-Gluc: rhamnose-glucose μg ml -1 : microgram per milliliter Authors' contributions VRI: Writing and submission of this manuscript, molecular genetic analysis of the wt, rtfA mutant, and complemented strains, generation of pVAP52 and complemented rtfA mutant as well as selection techniques of the rtfA mutant. SHL: Generation and initial characterization of plasmids and the serovar 8 rtfA mutant. TME, JMI, and JTB: Isolation and analysis of GPL, critical reading of the manuscript, and assistance in experimental techniques and study design. JNM: Principal Investigator in whose lab this research was conducted.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524182.xml

524035

Journal of Experimental & Clinical Assisted Reproduction: shaping the future of research and practice in reproductive endocrinology/infertility

Journal of Experimental & Clinical Assisted Reproduction is an open access, online, peer-review journal publishing papers on all aspects of research into reproductive endocrinology, infertility, bioethics and the advanced reproductive technologies. The journal reports on important developments impacting the field of human reproductive medicine and surgery. The field exists as a sub-specialty of obstetrics & gynecology, focusing on the diagnosis and treatment of complex human reproductive problems. The continued growth of this relatively new field depends on quality research by proven scientists as well as junior investigators who, together, make contributions to this area of medical and surgical practice. The publishing revolution made possible by internet technology presages a bright future for continued interdisciplinary collaboration among researchers. Against this background, Journal of Experimental & Clinical Assisted Reproduction exists for the scientific community to facilitate this scholarly dialogue.

Introduction In early 2002, planning for Journal of Experimental & Clinical Assisted Reproduction began in Atlanta after discussions with London-based publisher BioMed Central to develop a new electronic peer-reviewed journal reporting on the rapidly growing field of reproductive endocrinology and infertility. Working with colleagues who shared this vision, an exploratory organizational group was established to design a journal that would not only meet the current demands of the specialty, but also anticipate its challenges and needs for the future. By the time of its formal launch in July 2004, Journal of Experimental & Clinical Assisted Reproduction already had received submissions and inquiries from authors seeking to publish manuscripts in the categories of original research, review, case report, opinion/debate, medical history, and letters. Although column space does not constrain electronic publishing in the same way as traditional print journals, our peer-review process determined that not every submission was appropriate for publication. Hence, from the first instance the journal has pledged to observe a peer-review policy that ensures only the best quality manuscripts are published. Our editorial board membership reflects a depth of expertise required to guide editorial policies of a journal that aspires to join the top tier of selective scientific journals. Journal of Experimental and Clinical Assisted Reproduction is fully committed to the philosophy of Open Access as articulated in the following policies: Free, full-text access for all articles While the availability of published articles on the journal's homepage provides high and unrestricted visibility for accepted articles, contents of Journal of Experimental & Clinical Assisted Reproduction (ISSN 1743-1050) are also accessioned in PubMed Central, maintained by the U.S. National Library of Medicine – the world's largest medical library. Accordingly, all journal publications are available free and without password requirements to anyone with internet access. Furthermore, the platform provided by our publisher, BioMed Central (NLMID b101153627), allows author(s) to track how frequently their manuscript has been accessed and viewed by the public at no charge. Journal scope and article categories The journal reports on important developments impacting the field of human reproductive medicine and surgery. We accept manuscripts describing research in reproductive endocrinology, infertility, bioethics and the advanced reproductive technologies. At the discretion of the Editors, conference proceedings may be considered for publication by special arrangement. Rapid publication After a manuscript is received via the journal's on-line electronic submission system, selection of referees and editorial review is underway within a few business days. Our peer review process evaluates the appropriateness and suitability of all submitted manuscripts, as well as supplying authors with any additional requirements or modifications needed before the article can be published. We seek to reach an editorial decision on a manuscript within three weeks of its initial receipt, although author delays in addressing referees' comments are likely to extend this timeline. Questions regarding the suitability of a proposed submission may be sent to the journal office at , although this is not a requirement Formal peer-review policy The abstract of each manuscript is reviewed by two internal referees, which may include members of the Editorial Board. Authors can expect a response regarding this preliminary review within one week, and manuscripts judged outside the scope of the journal are identified accordingly. Submissions considered to be appropriate for the journal are reviewed by at least one independent referee. Referees' comments are relayed to corresponding authors in a confidential/anonymous manner, and all communication regarding submissions is through the journal's editorial office. The Editorial Office will review commentaries, together with input from an Editorial Board member with the relevant expertise. It may be necessary to enlist additional reviewers and/or statisticians in certain cases to assure the highest quality manuscripts are chosen for publication. Monetary author costs For the first six months, no author charges will apply to any work accepted for publication in Journal of Experimental & Clinical Assisted Reproduction . However, for each manuscript accepted thereafter, our publisher will charge a modest fee comparable to the cost of a single color page charge in a print journal (see ). This fee may be waived in hardship cases or for selected authors without budget support, provided that such a waiver request is made at the time of manuscript submission. These policies assure no publishing bias will exist against residents, fellows, or other investigators with little or no research funding. Copyright retained by authors We do not require authors to assign their copyright claim to the publisher as a condition of publishing any article. This means that our authors keep the copyright to their article with the freedom to include article components in subsequent published work, to submit the article in full to colleagues, or to include the work in their own homepage(s) without the publisher's prior authorization or permission. Conclusion With the beginning of a new century of medical research, library privileges and institutional journal holdings will continue to be eclipsed by electronic publishing and unrestricted access to the internet. The publishing revolution made possible by such technology presages a bright future for continued interdisciplinary collaboration among researchers. Against this background, Journal of Experimental & Clinical Assisted Reproduction exists for the scientific community to facilitate this scholarly dialogue. Author contributions ESS, RMW, and GDP drafted and reviewed the manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524035.xml

526781

Exploiting Thiol Modifications

Molecular oxygen may be necessary for life but with its beneficial properties comes formation of potentially toxic reactive oxygen species. One of the ways in which bacteria protect themselves is explained

As the premier biological electron acceptor, molecular oxygen (O 2 ) serves a vital role in fundamental cellular functions, including the process of aerobic respiration. Nevertheless, with the beneficial properties of O 2 comes the inadvertent formation of reactive oxygen species, including superoxide (O − 2 ), hydrogen peroxide (H 2 O 2 ), and hydroxyl radical (•OH); these differ from O 2 in having one, two, and three additional electrons, respectively ( Figure 1 ). Cells also encounter elevated levels of these reactive oxygen species when they are released by animals, plants, and insects as a defense against detrimental organisms such as microbial pathogens. Reactive oxygen species can damage cells in many ways: by inactivating proteins, damaging nucleic acids, and altering the fatty acids of lipids, which leads in turn to perturbations in membrane structure and function. The accumulation of this oxidative damage underlies the formation of many disease states in humans. It is postulated that tissue injury by these reactive oxygen species accumulates over a long period of time and plays roles in the aging process and the development of heart disease, diabetes, chronic inflammatory diseases, cancer, and several neurodegenerative diseases ( Halliwell 1999 ). Figure 1 Formation of Reactive Oxygen Species The four-electron reduction of molecular O 2 generates two molecules of H 2 O, which is O 2 in its most reduced form. While this reduction normally occurs within the enzyme cytochrome oxidase, one-electron transfers to O 2 also occur outside of cytochrome oxidase via inadvertent reactions with other reduced electron carriers, resulting in partially reduced and reactive forms of O 2 · H 2 O 2 is also produced by the enzymatic or spontaneous dismutation of O 2 − , and •OH is generated by the reaction of iron with H 2 O 2 (the Fenton reaction). In addition, the reactive oxygen intermediates are produced by a variety of organisms as a defense against microbial invasion. (Illustration: Rusty Howson, sososo design) Many organisms have evolved strategies to remove reactive oxygen species and repair damage, which have enabled them to prosper from the tremendous oxidizing potential of O 2 without succumbing to oxidative damage. Bacteria, yeast, and mammalian cells all induce the synthesis of global regulatory responses to survive oxidative insults. The consequences of oxidative stress and the corresponding defense responses have been extensively studied in Escherichia coli . For ease of study in the laboratory, the stress responses are often provoked by the external addition of chemical oxidants that specifically elevate the levels of reactive oxygen species within cells, or by the use of mutant strains that disrupt the normal “homeostatic mechanisms” for removing reactive oxygen species or the damage they do. While this primer focuses on a particular set of protective and regulatory protein modifications induced by oxidative stress in E. coli , it should be noted that many of the same mechanisms are present in other organisms; some specific examples from other species will also be described. The major target of O 2 − damage identified in bacteria is a class of dehydratase enzymes that utilize [4Fe–4S] clusters to bind their substrate ( Imlay 2003 ; Djaman et al. 2004 ). Since some of these enzymes function in the citric acid cycle (also called the Krebs cycle) and in amino acid biosynthesis, high levels of O 2 − lead to a requirement for certain amino acids in growth media ( Imlay and is well known Fridovich 1991 ). H 2 O 2 for its role in oxidizing thiol (SH) groups of cysteinyl amino acid residues in proteins. Elevated levels of H 2 O 2 also are associated with the oxidation of other amino acids, leading to the formation of methionine sulfoxide and a variety of carbonyls. Lastly, because of its extreme reactivity, •OH targets all of the major macromolecules of cells: RNA, DNA, protein, and lipids. The extent to which membrane lipids are targets appears to depend on the presence of polyunsaturated fatty acids in lipids, which are not as prevalent in bacteria as they are in mammals. Many enzymes that protect against oxidative damage have been identified in E. coli ( Imlay 2002 , 2003 ). Three superoxide dismutases, each of which contain a different metal center and show different expression patterns and subcellular localization, catalyze the dismutation of O 2 − to H 2 O 2 . While the superoxide dismutases eliminate O 2 − , they also are a source of endogenously produced H 2 O 2 in E. coli . The major enzymes involved in reducing H 2 O 2 to H 2 O and O 2 in E. coli are catalase and alkyl hydroperoxide reductase. There is no enzymatic mechanism for decreasing levels of •OH, produced from H 2 O 2 . Thus, levels of •OH will be directly proportional to levels of H 2 O 2 , and accordingly, catalase and alkyl hydroperoxide reductase activities are critical to oxidative stress survival. Another component to the oxidative stress response is the reduction of oxidized thiols that arises through one of the mechanisms described below. The tripeptide glutathione and the thiol reductants glutaredoxin and thioredoxin are key to the restoration of thiols to their reduced state (SH) ( Fernandes and Holmgren 2004 ). E. coli contains three glutaredoxins that utilize the reducing power of glutathione to catalyze the reduction of disulfide bonds (–S–S–) in the presence of NADPH and glutathione reductase. There are two thioredoxins in E. coli that also function to reduce disulfide bonds. Reduced thioredoxin is regenerated by thioredoxin reductase and NADPH. The fact that NADPH is required to maintain the reduced state of glutathione and thioredoxin indicates that the response to oxidative stress is coupled to the physiological status of core pathways that generate NADPH. Regulatory Roles of Thiol Modifications As mentioned above, proteins—in particular, the thiols of cysteines—are the major targets of H 2 O 2 . The reaction of cysteinyl thiolates with H 2 O 2 can lead to the formation of different modifications, such as sulfenic acid (–SOH), sulfinic acid (–SO 2 H), and sulfonic acid (–SO3H), as well as disulfide bond formation (–S–S–) and glutathione conjugation (–S–GSH) ( Jacob et al. 2004 ; Poole et al. 2004 ) ( Figure 2 ). These modifications often alter the structure and function of the protein. Recent progress in this field points to a common chemistry in the reaction of H 2 O 2 with thiolates through the initial formation of sulfenic acid. In the case of proteins that have a nearby cysteinyl residue, a disulfide bond forms between the two sulfur atoms. The sulfenated cysteinyl residue also can react with a cysteinyl residue on another protein or with glutathione, leading to a mixed disulfide. If no cysteinyl residue is nearby, the sulfenated cysteine can be further oxidized to sulfinic or sulfonic acid, or it can remain in the sulfenic acid state. All but the sulfinic and sulfonic acid modifications are readily reversible by reduction, using proteins such as thioredoxin or glutaredoxin; though sulfinic acid reductase activities have recently been identified in yeast and mammalian cells (denoted sulfiredoxin and sestrin, respectively) ( Biteau et al. 2003 ; Budanov et al. 2004 ). Figure 2 Thiol Modifications of Proteins Formation of sulfenic acid from the reaction of H 2 O 2 with protein thiolates leads to different protein modifications, depending on the protein. In proteins without a second sulfhydryl, the sulfenic acid (–SOH) may be stabilized (e.g., OhrR) or may react with reactive oxygen species to generate the further oxidized sulfinic (–SO 2 H) (e.g., thiolperoxidase; Tpx) and sulfonic acid (–SO 3 H) derivatives. Alternatively, if a second cysteinyl residue is in proximity within the same polypeptide (e.g., OxyR) or an associated protein (e.g., Yap1 and Orp1), a disulfide bond can form between the two sulfur atoms (–S–S–). Lastly, the sulfenated cysteinyl residue can react with glutathione (GSH), leading to a mixed disulfide (e.g., MetE). (Illustration: Rusty Howson, sososo design) Given the reversible nature of most forms of thiol oxidation, it has been suggested that thiol modifications can play roles in signal transduction that are similar to protein phosphorylation/dephosphorylation ( Sitia and Molteni 2004 ). In support of this model, there are several examples of proteins whose activities are modulated by thiol oxidation and reduction. The first of these examples is the OxyR transcription factor, which upregulates peroxide defenses in E. coli and a variety of other bacteria. OxyR contains two critical cysteines that are oxidized to form an intramolecular disulfide bond when cells encounter peroxide stress ( Zheng et al. 1998 ; Aslund et al. 1999 ). Disulfide bond formation is associated with a conformational change that alters OxyR binding to DNA and allows the protein to activate the transcription of genes encoding enzymes, such as catalase and the alkylhydroperoxide reductase, that destroy H 2 O 2 . Once the H 2 O 2 concentration is decreased, OxyR is reduced and the system is reset. The unusually reactive cysteine in OxyR that is oxidized by H 2 O 2 to form the sulfenic acid intermediate can clearly be nitrosylated and glutathionylated in vitro ( Hausladen et al. 1996 ; Kim et al. 2002 ), but the in vivo relevance of these other modifications is questionable ( Mukhopadhyay et al. 2004 ). Two other examples of redox-regulated proteins are the E. coli chaperone protein Hsp33 ( Jakob et al. 2000 ) and the Streptomyces coelicolor anti-sigma factor, RsrA ( Li et al. 2003 ; Paget and Buttner 2003 ; Bae et al. 2004 ). For these proteins, the cysteine residues, which form intramolecular disulfide bonds, are in a reduced state when coordinated to a zinc ion (Zn 2+ ), and zinc is released upon oxidation of the thiols. For both proteins, oxidation and zinc release are associated with an opening of the protein structure. For Hsp33, this structural change allows for dimerization and activates its chaperone activity ( Graf et al. 2004 ). RsrA, on the other hand, dissociates from a promoter specificity factor of RNA polymerase (an extracytoplasmic-function-type alternative sigma factor) allowing the transcription of genes that permit recovery from the stress ( Li et al. 2003 ; Bae et al. 2004 ). Among the target gene products is a thioredoxin, which reduces the disulfide bonds that form within oxidized RsrA. Presumably, reduction of the disulfide restores the binding of zinc and its inhibitory association with the sigma factor. Thus, the RsrA regulatory circuit provides another example, comparable to OxyR, in which the modification of a regulatory protein thiol group can be linked to a change in the transcriptional output of genes that remediate stress. The peroxide-sensing repressor OhrR from Xanthomonas campestris pv. phaseoli ( Panmanee et al. 2002 ) and Bacillus subtilus ( Fuangthong and Helmann 2002 ) can be inactivated by H 2 O 2 or by organic peroxides (ROOH) formed by the oxidation of a variety of organic molecules in the cell or in the environment. The B. subtilis OhrR transcription regulator contains only a single cysteine that forms a relatively stable sulfenic acid upon its reaction with H 2 O 2 or organic peroxides ( Fuangthong and Helmann 2002 ). Oxidation of the single cysteine leads to the dissociation of OhrR from its DNA binding site and the derepression of the gene encoding an organic hydroperoxidase that eliminates the initial oxidizing insult. In this issue, Hondorp and Matthews (2004) provide an example of a thiol modification that protects an enzyme activity during oxidative stress. Their data suggest that when cells encounter oxidative stress, a key cysteinyl residue near the active site of methionine synthase (MetE) is glutathionylated. This modification blocks access of the substrate and prevents further synthesis of methionine. This finding is significant in that it presents a mechanism to reversibly preserve the function of a protein during oxidative challenge. By glutathionylating a single cysteinyl residue, the protein is protected from further oxidation of that cysteinyl residue to the irreversible sulfinic and sulfonic acid forms. Once the stress is removed, the mixed disulfide bond will be readily reduced, and access to the substrate restored. Prevalence of Regulatory Thiol Modifications? As illustrated by the examples above, an array of chemical modifications obtained by oxidizing cysteinyl residues has been exploited in combating oxidative stress. Yet it is important to note that not all cysteinyl residues of proteins are readily oxidized by oxidants such as H 2 O 2 . We do not currently understand all of the features that determine the reactivity of a particular thiol to H 2 O 2 ( Poole et al. 2004 ). The pKa of the thiolates clearly plays an important role, as thiolates are more reactive than their protonated counterparts. In addition, the contribution of protein environment to the stability of the oxidized products is also known to be a factor, but is not well understood. Given that many of the thiol modifications do not appear to be in equilibrium with the redox state of the cell, the features of the protein that determine the rate at which the modifications are formed are another important parameter. The added complexity of the cysteine targets that compose part of a Zn binding site found for Hsp33 and RsrA raises questions about the function of the zinc. Perhaps Zn binding provides some additional control over the reactivity of the cysteine thiols, or perhaps the loss of the zinc facilitates conformational changes. Recently, the oxidative, stress-induced thioredoxin-2 from E. coli has also been shown to contain a H 2 O 2 -labile zinc site, although the loss of zinc does not change its reductase activity ( Collet et al. 2003 ). Thus, the way this oxidatively labile Zn site affects thioredoxin function has yet to be established. The extent of thiol oxidation within the cell remains another open question. The variety of modifications that arise from treatment with H 2 O 2 and the experimental challenges associated with their detection has made it difficult to catalog all the proteins that are modified and all the types of modifications that exist. In this issue, Leichert and Jakob (2004) report a general method for detecting cellular proteins whose cysteinyl residues were modified after imposing an oxidative stress. Such an approach will greatly enhance our understanding of targets of oxidative stress. The method described by Leichert and Jakob also will be useful in detecting transient cysteine modifications. The importance of monitoring transient changes in cysteines is highlighted by the recent finding that oxidation of the Yap1 activator of antioxidant genes in the yeast Saccharomyces cerevisiae requires a peroxidase denoted Gpx3 or Orp1 ( Delaunay et al. 2002 ). In this case, H 2 O 2 reacts with a cysteine in Orp1, forming an unstable sulfenic acid intermediate that then reacts with a cysteinyl residue of Yap1 to form an intermolecular disulfide. The disulfide undergoes an exchange with a second cysteine within Yap1 to form an intramolecular disulfide that locks Yap1 in a confirmation that masks the nuclear export signal ( Wood et al. 2004 ). Thus, methods that allow the appearance of thiol modifications in cells to be monitored kinetically will greatly enhance our understanding of how cysteine residues become oxidized. The examples mentioned here illustrate the versatile potential of thiol modifications. Given the reversibility of thiol oxidations and the wide range of structural constraints that can be imposed by the formation of a sulfenic or sulfinic acid or a disulfide bond, we predict there will be many more examples of regulation by thiol modification.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526781.xml

529317

Lmo Mutants Reveal a Novel Role for Circadian Pacemaker Neurons in Cocaine-Induced Behaviors

Drosophila has been developed recently as a model system to investigate the molecular and neural mechanisms underlying responses to drugs of abuse. Genetic screens for mutants with altered drug-induced behaviors thus provide an unbiased approach to define novel molecules involved in the process. We identified mutations in the Drosophila LIM-only (LMO) gene, encoding a regulator of LIM-homeodomain proteins, in a genetic screen for mutants with altered cocaine sensitivity. Reduced Lmo function increases behavioral responses to cocaine, while Lmo overexpression causes the opposite effect, reduced cocaine responsiveness. Expression of Lmo in the principal Drosophila circadian pacemaker cells, the PDF-expressing ventral lateral neurons (LN v s), is sufficient to confer normal cocaine sensitivity. Consistent with a role for Lmo in LN v function, Lmo mutants also show defects in circadian rhythms of behavior. However, the role for LN v s in modulating cocaine responses is separable from their role as pacemaker neurons: ablation or functional silencing of the LN v s reduces cocaine sensitivity, while loss of the principal circadian neurotransmitter PDF has no effect. Together, these results reveal a novel role for Lmo in modulating acute cocaine sensitivity and circadian locomotor rhythmicity, and add to growing evidence that these behaviors are regulated by shared molecular mechanisms. The finding that the degree of cocaine responsiveness is controlled by the Drosophila pacemaker neurons provides a neuroanatomical basis for this overlap. We propose that Lmo controls the responsiveness of LN v s to cocaine, which in turn regulate the flies' behavioral sensitivity to the drug.

Introduction Cocaine, a naturally occurring plant alkaloid, is the prototype addictive psychomotor stimulant. It elicits a variety of acute behavioral changes ranging from mood elevation, disinhibition, and motor activation at low doses to compulsive stereotypies and psychosis at higher doses ( Gawin 1991 ). Long-term cocaine use generally results in tolerance to many of its subjective effects, an increased craving towards the drug, and, eventually, drug abuse and addiction. Cocaine's primary mechanism of action is to bind and inhibit plasma membrane monoamine transporters, thereby increasing synaptic monoamine neurotransmitter levels and potentiating their actions. Cocaine's direct role in increasing dopamine (DA) in the nucleus accumbens via inhibition of the DA transporter (DAT) has contributed to the prevalent DA hypothesis of drug addiction, which posits that the shared ability of drugs of abuse to increase DA in the nucleus accumbens underlies their reinforcing properties ( Wise and Bozarth 1985 ; Kuhar et al. 1991 ). More recent animal studies, however, have shown that cocaine still elicits a robust conditioned place preference in mice with genetic deletions of DAT, suggesting that the rewarding properties of cocaine are not solely mediated by its action on DAT ( Sora et al. 1998 ). Additional studies on mice in which the serotonin transporter or norepinephrine transporter has been deleted, together with studies using selective serotonin transporter or norepinephrine transporter inhibitors, have suggested roles for both serotonin and norepinephrine systems in mediating cocaine's rewarding properties ( Uhl et al. 2002 ). These data show that the molecular bases for cocaine's psychostimulant and reinforcing properties are more complicated than once thought, and that a combination of actions at multiple sites may mediate its effects. Indeed, multiple genes and signaling pathways have been implicated in the stimulant and rewarding properties of cocaine in mice ( Laakso et al. 2002 ). The fruit fly Drosophila melanogaster has been advanced as a useful model system for identifying novel genes that regulate behavioral responses to drugs of abuse including cocaine ( Wolf and Heberlein 2003 ). Several of cocaine's most characteristic properties have been recapitulated in flies. First, cocaine induces motor behaviors in flies (see below) that are remarkably similar to those observed in mammals ( McClung and Hirsh 1998 ; Bainton et al. 2000 ). Second, repeated cocaine administration induces behavioral sensitization ( McClung and Hirsh 1998 ), a form of behavioral plasticity believed to underlie certain aspects of addiction ( Robinson and Berridge 1993 ; Schenk and Partridge 1997 ). Finally, a key role for dopaminergic systems in mediating cocaine's effects has been demonstrated through both pharmacologic and genetic methods ( Bainton et al. 2000 ; Li et al. 2000 ). More importantly, Drosophila studies have identified genes and pathways whose role in cocaine responsiveness had not been anticipated ( Hirsh 2001 ; Rothenfluh and Heberlein 2002 ). For instance, the Drosophila circadian gene period was identified as a necessary mediator of cocaine sensitization ( Andretic et al. 1999 ). Subsequently, mice carrying various period mutations were found to have altered cocaine sensitization and conditioned place preference ( Abarca et al. 2002 ). In order to identify novel molecules and pathways involved in behavioral responses to cocaine, we carried out a genetic screen for Drosophila mutants with altered acute responses to cocaine. Here, we report the phenotypic and molecular characterization of mutations in the Drosophila LIM-only gene, Lmo (also called Beadex [Bx] ), isolated due to their increased sensitivity to cocaine-induced loss of negative geotaxis. The products of Drosophila and mammalian Lmo genes modulate the function of LIM-homeodomain (LIM-HD) proteins ( Milan et al. 1998 ; Retaux and Bachy 2002 ), which in turn regulate various aspects of nervous system development, including the specification of neural identity ( Thor and Thomas 1997 ; Hobert and Westphal 2000 ; Shirasaki and Pfaff 2002 ; Tsalik et al. 2003 ). We find that cocaine sensitivity is inversely related to the levels of Lmo function: reduced function causes increased sensitivity, while increased function causes resistance. This bidirectional regulation of cocaine responsiveness is mediated by Lmo function in a small set of neurons, the ventral lateral neurons (LN v s), which are the primary circadian pacemaker cells in Drosophila ( Helfrich-Förster 1997 ; Renn et al. 1999 ). Like other mutants that affect LN v function, Lmo mutants show altered circadian locomotor behaviors. However, the roles of the LN v s in regulating cocaine sensitivity and circadian behavioral rhythms are genetically separable, as mutants lacking the neuropeptide PDF—the only known functional output of these neurons—show normal cocaine sensitivity. Thus, Lmo defines a novel role for the circadian pacemaker cells in regulating behavioral responses to cocaine. Results Loss- and Gain-of-Function Mutations in the Lmo Locus Show Altered Cocaine Sensitivity To identify novel molecules involved in cocaine-related behaviors, we carried out a genetic screen for mutants with altered acute responses to volatilized freebase cocaine using the crackometer, a simple assay that measures cocaine-induced loss of negative geotaxis ( Bainton et al. 2000 ). Screening of 400 first chromosome P-element insertions from the EP collection ( Rørth 1996 ) led to the identification of five mutants with reduced cocaine sensitivity and seven with increased sensitivity. Two of the mutants conferring increased cocaine sensitivity, EP1306 and EP1383 ( Figure 1 A), carry a P-element insertion in the promoter region of the Drosophila Lmo locus ( Milan et al. 1998 ; Zeng et al. 1998 ). Drosophila LMO protein has been shown to inhibit the activity of the LIM-HD transcription factor apterous through its interactions with the LIM-HD activator Chip ( Milan and Cohen 1999 ; Weihe et al. 2001 ). Figure 1 Lmo Loss-of-Function Mutants Show Increased Sensitivity to Cocaine (A) Cocaine phenotypes of various Lmo mutants. Male flies hemizygous for the indicated Lmo alleles (and their appropriate genetic controls) were exposed to 150 μg of cocaine and tested in the crackometer as described in Materials and Methods . Compared to their control (Ctl-1), EP1383 ( p < 0.02) and EP1306 ( p < 0.001) flies show significantly increased sensitivity to cocaine. Similarly, compared to their respective controls, pdrm and hdp flies are significantly more sensitive to cocaine ( p < 0.001). Asterisks denote significant differences from controls (Student's paired t -test assuming equal variance); n = 20 experiments. (B) Cocaine dose–response. EP1306 flies (filled squares) and pdrm flies (filled circles) and their respective controls were exposed to the indicated doses of cocaine. At each dose, the responses of EP1306 and pdrm flies are significantly higher than their controls ( p < 0.001, n = 16–20 experiments). (C) EP1306 flies show alterations in cocaine-induced locomotor patterns of activity. Flies were exposed to 0, 75, or 100 μg of cocaine, as indicated, for 1 min. Representative traces shown correspond to 30 s of recorded activity of about ten flies starting 1 min after the end of cocaine exposure ( n ≥ 4). Top panels show response of control flies to indicated amounts of cocaine; bottom panels show activity of EP1306 flies after cocaine administration. Ctl-1 is EP1631, Ctl-2 is P[GAL4] line 8.142, and Ctl-3 is w 1118 . Independently, we isolated a P[GAL4] insertion in the Lmo locus, dubbed pipedream (pdrm), that also showed increased cocaine sensitivity ( Figure 1 A). Finally, heldup (hdp) mutant flies, which carry well-characterized loss-of-function mutations in Lmo ( Milan et al. 1998 ), displayed increased sensitivity to cocaine that is similar in magnitude to that of the EP1306 line ( Figure 1 A). This increased sensitivity was observed at all cocaine doses tested ( Figure 1 B and 1 C; data not shown). The EP1306 and hdp alleles consistently demonstrated stronger phenotypes than EP1383 and pdrm . All mutants tested performed within the normal range in task-specific baseline behaviors (see Materials and Methods ). Finally, we were able to revert the cocaine phenotype of EP1306 by excision of the P element (data not shown). Taken together, these data demonstrate that loss-of-function mutations in Lmo result in increased sensitivity to cocaine. The dominant Beadex (Bx) mutations were first described in 1925 by Morgan, Bridges, and Sturtevant, and named for their characteristic beaded wing margin ( Morgan et al. 1925 ). More recently, Bx mutants were shown to be overexpression alleles of Lmo ( Milan et al. 1998 ; Shoresh et al. 1998 ; Zeng et al. 1998 ). Overexpression results from the insertion of naturally occurring transposable elements into the 3′ UTR of Lmo, which leads to transcript stabilization ( Shoresh et al. 1998 ). The resultant overexpression of Lmo causes decreased apterous activity in the dorsal compartment of the wing and, consequently, disorganization of the wing margin ( Milan and Cohen 1999 , 2000 ). We tested multiple available Bx alleles, and found that they all caused resistance to the acute effects of cocaine at every dose tested ( Figure 2 A and 2 B). The strength of the cocaine resistance phenotype correlated well with the severity of the beaded wing-margin phenotype, Bx J > Bx 1 = Bx 2 = Bx 3 ( Figure 2 A and 2 B; data not shown). Because in Bx mutants Lmo transcripts are stabilized in their normal spatiotemporal pattern ( Shoresh et al. 1998 ), the observed cocaine phenotypes likely result from overexpression, rather than from misexpression of Lmo . Taken together with the loss-of-function effects described above (see Figure 1 A and 1 B), these data reveal a graded behavioral response to cocaine that is inversely related to Lmo levels. Figure 2 Lmo Gain-of-Function Bx Alleles Show Reduced Sensitivity to Cocaine (A) Cocaine phenotypes of Bx mutants. Male flies hemizygous for Bx alleles Bx 1 or Bx J show significant reductions in sensitivity to cocaine compared to control (Ctl) flies ( p < 0.001, n = 12 experiments). Asterisks denote significant differences from control (Student's paired t -test assuming equal variance). (B) Dose–response. Bx J flies (filled circles) show reduced sensitivity compared to Ctl flies (open circles) at all doses tested ( p < 0.001, n = 16–20 for all doses except for 250 μg, where p = 0.0015, n = 8). Two additional Bx alleles ( Bx 2 and Bx 3 ) had similar phenotypes to Bx 1 (not shown). (C) Bx J flies show alterations in cocaine-induced locomotor patterns of activity. Flies were exposed to 0, 100, or 125 μg of cocaine, as indicated, for 1 min. Representative traces shown are 30 s of recorded activity of about ten flies starting 30 or 60 s after the end of cocaine exposure ( n ≥ 3). Top panels show response of control flies to indicated amounts of cocaine; bottom panels show activity of Bx J flies after cocaine administration. Ctl flies are w 1118 . In order to study Lmo mutant behavior in more detail, we recorded the cocaine-induced patterns of locomotor activity of control and Lmo flies using a movement tracking system ( Bainton et al. 2000 ; Wolf et al. 2002 ). Upon mock exposure, control flies showed robust locomotor activity, generally walking in straight lines ( Figure 1 C, top left panel). At relatively low doses of cocaine (75 and 100 μg), flies engaged in stereotyped circling patterns of activity ( Figure 1 C, top right panels). At higher doses (125–200 μg), flies began to show spasmodic movements (seen as zigzag patterns in movement) and severe hypokinesis (data not shown; Bainton et al. 2000 ). EP1306 flies, while showing a reduced baseline speed, showed normal patterns of activity upon mock exposure ( Figure 1 C, bottom left panel). Upon exposure to cocaine, these mutant flies showed a shift in the dose–response relationship. At low doses (75 μg), EP1306 flies showed a marked increase in stereotyped circling behavior compared to controls ( Figure 1 C, bottom middle panel), while at higher doses (100 μg), the mutant flies were more likely to be spasmodic or akinetic ( Figure 1 C, bottom right panel). Thus, despite the reduced speed observed in the mock exposures, increases in specific cocaine-induced locomotor behaviors demonstrate that EP1306 flies have a shifted dose–response to cocaine: mutant flies exposed to 75 μg of cocaine behaved similarly to control flies exposed to 100 μg of the drug. Bx J flies also showed changes in walking patterns that are consistent with a shift in the cocaine dose–response relationship. After exposure to 100 μg of drug, most control flies showed slow circling behavior and some were akinetic ( Figure 2 C, top middle panel). Bx J flies were much less affected, showing increased locomotion (mostly in straight lines), decreased slow circling, and almost no akinesia ( Figure 2 C, bottom middle panel). At 125 μg, control flies showed very little movement ( Figure 2 C, top right panel), while most Bx J flies continued to show circling behaviors seen in control flies at lower doses ( Figure 2 C, bottom right panel). Bx J flies, like EP1306 flies, showed reduced activity upon mock exposure ( Figure 2 C, bottom left panel). Despite this reduced activity, these two mutants showed very different sensitivities to cocaine: EP1306 flies were more affected and Bx J flies were less affected, suggesting that the reduced activity upon mock exposure is unrelated to the cocaine response. Long-term activity recordings revealed no differences between Bx J , EP1306, and control lines (data not shown; see below). In summary, using two different behavioral assays we show that flies carrying loss-of-function mutations in Lmo display increased sensitivity to the effects of cocaine on locomotor behaviors, while gain-of-function mutations show the converse effect, a reduced response to the drug. This inverse relationship between Lmo gene activity and drug responsiveness suggests that Lmo may regulate the expression of genes that might play a direct role in controlling cocaine responses (see Discussion). Molecular Characterization of Lmo Mutants The Lmo locus produces at least three transcripts by differential promoter use ( Figure 3 A). The RA and RC transcripts differ only in their 5′ UTRs and are predicted to encode identical proteins of 313 amino acids; the RB transcript is predicted to utilize an alternative translational start site in its first exon, which would result in an addition of 71 N-terminal amino acids. The insertions that produce increased cocaine sensitivity all lie within the putative promoter region for the RA transcript, 25–90 bp upstream of its transcriptional start site ( Figure 3 A). In order to identify molecular changes caused by these insertions, we assessed Lmo transcript levels by quantitative RT-PCR in wild-type flies and Lmo mutants. In wild-type flies the RA transcript was about 4-fold more abundant in heads as compared to bodies, suggesting that this transcript is enriched in the nervous system ( Figure 3 B; see also Figure 4 ). Importantly, the RA transcript was reduced by greater than 50% in the heads, but not the bodies, of EP1306 flies ( Figure 3 B); a 40% increase in the RA transcript was observed in the heads of Bx J flies (data not shown). Thus, as predicted by the location of the P-element insertion, the EP1306 mutation causes a reduction in Lmo transcript levels, a finding that is consistent with the observation that the cocaine sensitivity of EP1306 flies is similar to that seen with known loss-of-function alleles of Lmo (the hdp alleles) and opposite to that seen with gain-of-function Bx alleles. In addition, the finding that transcript levels are specifically reduced in the heads of EP1306 flies, and not their bodies, suggests that this mutation affects a nervous-system-enriched (or -specific) Lmo transcript (see below). Finally, none of the P-element insertions in the promoter of the RA transcript leads to a held-up wing phenotype, which is characteristic of loss-of-function hdp alleles ( Shoresh et al. 1998 ; Milan and Cohen 1999 ). This suggests that the P-element insertions isolated in our genetic screen cause either less severe or more spatially restricted changes in Lmo gene expression. Figure 3 Molecular Structure of the Lmo Locus (A) A genomic map of the Lmo locus. Three different first exons can be utilized, forming the basis for three alternative transcripts. Exon RA-1 is separated from the alternative start sites RB-1 and RC-1 by a large (∼30 kb) intron. EP1306, EP1383, and pdrm carry insertions 25, 73, and 91 bp, respectively, upstream of the exon RA-1 transcriptional start site. Arrows within the EP elements refer to the orientation of the insertion and the expected direction of inducible expression via UAS sites contained within the EP element. Bx alleles are insertions of natural transposons into the 3′ UTR of the Lmo gene that have been shown to stabilize Lmo transcript ( Shoresh et al. 1998 ). Protein-coding exons are shaded. (B) Expression of the Lmo RA transcript is enriched in Drosophila heads, and is reduced in the EP1306 mutant. RNA was isolated from heads and bodies; after cDNA synthesis, quantitative RT-PCR was performed using primers specific to the RA transcript of Lmo in addition to primers to a reference transcript, the ribosomal protein rp49 . Relative abundance is expressed as fold increase over control (EP1631) body mRNA. No detectable amplification was seen in RNase-treated controls (data not shown). Error bars represent standard error of the mean. Asterisk denotes significant difference from control (Student's paired t -test assuming equal variance; p < 0.001, n = 3). Figure 4 pdrm 's GAL4 Expression Is Sufficient to Drive Lmo -Transgene-Mediated Rescue of Cocaine Sensitivity (A) pdrm 's GAL4 expression pattern. UAS-GFP reveals GAL4 expression pattern of the pdrm enhancer-trap insertion in MB lobes and calyces, ALs, the large cell bodies of the peptidergic LN v s, and neurons of the pars intercerebralis (PI). (B) Lmo expression restores wild-type cocaine responses. In the absence of a UAS transgene (“no UAS” columns), hemizygous male pdrm flies (hatched bar) are more sensitive than controls (Ctl GAL4 is line 8.142, solid bar), as shown before in Figure 1 A. Male flies hemizygous for pdrm and heterozygous for either of two UAS-Lmo transgenes ( UAS-Lmo 1 and UAS-Lmo 2 ; hatched black bars), show normal cocaine sensitivity when compared to either UAS-Lmo transgene alone (white bars) or UAS-Lmo transgenes in the presence of a control GAL4 line (Ctl GAL4, line 8.142; black bars). To control for non-specific effects of transgene overexpression, UAS-GFP and UAS-lacZ transgenes were also driven by pdrm GAL4. Male flies were hemizygous for pdrm (or heterozygous for the control GAL4 insertion) and heterozygous for the specific UAS transgene. One-way ANOVA revealed a significant effect of genotype in the UAS-lacZ ( F = 17.4, p < 0.001), UAS-GFP ( F = 19.47, p < 0.001), or no UAS transgene ( F = 4.1, p < 0.001) groups, but not in either of the UAS-Lmo transgene groups ( F = 1.58, p = 0.22 and F = 1.21, p = 0.31 for UAS-Lmo 1 and UAS-LMO 2 , respectively); thus, UAS-Lmo expression specifically restores normal cocaine sensitivity to pdrm flies. Post hoc pairwise planned comparisons, with the critical p -value adjusted to 0.025, revealed significant differences between the “non-rescued” pdrm/UAS-GFP flies and the appropriate controls ( UAS-GFP/+ or 8.142/UAS-GFP, p < 0.002); similarly, pdrm/UAS-lacZ flies are significantly different from their controls ( UAS-lacZ/+ and 8.142/UAS-lacZ, p < 0.002). Pairwise comparisons revealed no significant differences between “rescued” pdrm/UAS-Lmo 1 flies and their “normal” controls ( 8.142/UAS-Lmo 1 and UAS-Lmo 1 /+, p = 0.99 and p = 0.14, respectively) or pdrm/UAS-Lmo 2 flies and their controls ( 8.142/UAS-Lmo 2 /+ and UAS-Lmo 2 /+, p = 0.09 and p = 0.99, respectively), indicating full rescue of pdrm cocaine sensitivity. For all genotypes, n = 16–20 experiments. Restricted Expression of Lmo in the Nervous System Is Sufficient for Wild-Type Cocaine Sensitivity The pdrm P[GAL4] insertion, which acts as a promoter/enhancer detector, is expected to reproduce, at least in part, the expression pattern of the Lmo gene. In adult flies, the pdrm line showed extensive GAL4 expression in the brain as visualized with the UAS–green fluorescent protein (GFP) transgene ( Figure 4 A). Prominent expression was observed in the antennal lobes (ALs) and the Kenyon cells of the mushroom bodies (MBs), which are major brain centers involved in olfaction and olfactory conditioning, respectively ( Stocker 1994 ; Zars 2000 ). In addition, GAL4 was expressed in the LN v s, which are the major pacemaker cells regulating circadian locomotor rhythmicity in flies ( Renn et al. 1999 ). Finally, GAL4 expression was seen in neurosecretory cells located in the pars intercerebralis, in glial cells surrounding the optic lobes, and in scattered cells throughout the ventral nerve cord (VNC) ( Figure 4 A; data not shown). In pdrm larvae and adult flies, GAL4 expression was restricted to the nervous system. GAL4 expression was not detected in larval imaginal discs, such as the wing disc (data not shown), where Lmo expression has been shown to play an important patterning role ( Milan and Cohen 1999 ). Thus, pdrm traps a specific subset of Lmo regulatory elements, possibly those controlling nervous-system-restricted expression of the RA transcript. We hypothesized that the expression pattern of the pdrm enhancer trap may identify the cells in which Lmo expression is disrupted in the P-element insertion lines, causing increased cocaine sensitivity. If this is the case, pdrm -driven expression of UAS-Lmo transgenes is expected to restore normal cocaine sensitivity to pdrm flies. Indeed, we found that two UAS-Lmo transgenes rescued the cocaine-sensitivity defect of pdrm mutant flies ( Figure 4 B). This effect was specific to Lmo, as pdrm -driven expression of UAS-tauGFP or UAS-lacZ failed to restore normal behavior. Finally, Lmo mutants that did not contain a GAL4 enhancer trap, such as EP1306, failed to be rescued by the presence of UAS-Lmo transgenes (data not shown). These data demonstrate that the cocaine-sensitivity defect of pdrm flies is due to the loss of Lmo function, and that nervous-system-specific expression of Lmo in cells dictated by the pdrm GAL4 line is sufficient to confer normal cocaine responses. Lmo Expression in PDF Neurons Is Sufficient to Confer Normal Cocaine Responses In order to refine further the spatial requirements for Lmo function, we attempted to rescue the EP1306 phenotype by expression of Lmo in specific brain regions. EP lines carry a P-element insertion containing multiple UAS sites to which GAL4 can bind to drive expression of adjacent genomic sequences ( Rørth 1996 ). Therefore, mutagenic EP insertions, if oriented appropriately, can be used to drive expression of the disrupted gene. It was demonstrated previously that GAL4-driven Lmo expression can be mediated by the EP1306 and EP1383 insertions ( Milan et al. 1998 ; Zeng et al. 1998 ). Consistent with this, we were able to rescue the EP1306 phenotype in the presence of the heat-inducible hs-GAL4 transgene, which drives low levels of ubiquitous GAL4 expression even in the absence of heat shock (data not shown). We then used GAL4 enhancer-trap lines with expression in specific brain regions to drive Lmo expression from the EP1306 insertion. We focused specifically on GAL4 lines that drive expression in the MBs, ALs, pars intercerebralis neurons, LN v s, and glia, sites of expression revealed by the pdrm line ( Figure 4 A). GAL4 lines that drive expression specifically in the MBs and the ALs failed to rescue the cocaine-sensitivity phenotype of EP1306 (data not shown). Similar negative results were obtained with the glial-specific repo-GAL4 driver ( Xiong et al. 1994 ). However, expression of Lmo in the LN v s using the pdf-GAL4 driver, which drives expression in cells that contain the neuropeptide PDF ( Renn et al. 1999 ), restored nearly wild-type cocaine responsiveness to EP1306 flies ( Figure 5 A and 5 C). pdf-GAL4 did not rescue the cocaine phenotype of hdp and pdrm flies, demonstrating dependence on induced Lmo expression provided by the EP1306 insertion. These results show that Lmo expression in the PDF-expressing neurons alone is sufficient to rescue wild-type cocaine responses, implicating these cells in the increased cocaine responses observed in Lmo loss-of-function mutants. Figure 5 Lmo Expression in PDF Neurons Regulates Cocaine Responses (A) pdf-GAL4 -driven expression of Lmo using the EP1306 element rescues the EP1306 insertional phenotype. Lmo mutants EP1306, hdp, and pdrm, as well as a control EP line (Ctl-1 = EP1413 ), were tested in the absence (− pdf-GAL4; white bars) and presence ( + pdf-GAL4; black bars) of pdf-GAL4 . All male flies are hemizygous for the Lmo mutation (or Ctl-1) and heterozygous for pdf-GAL4 (when carrying the transgene). One-way ANOVAs with post hoc planned comparisons (critical p- value adjusted to 0.0125) confirmed that EP1306, pdrm, and hdp flies (in the absence of pdf-GAL4 ) had significantly increased sensitivity to cocaine compared to control flies (Ctl-1 = EP1413 ) ( p ≤ 0.003, n = 15–26 experiments). One-way ANOVA with post hoc planned comparisons (critical p- value adjusted to 0.01) revealed a significant difference between pdrm/pdf-GAL4 and hdp/pdf-GAL4 flies and their controls ( EP1413/pdf-GAL4, pdf-GAL4/+, or EP1413/+, p < 0.003, n = 24–27 experiments), showing that the presence of pdf-GAL4 does not rescue the cocaine sensitivity of pdrm or hdp flies. In contrast, similar comparisons for “rescued” EP1306/pdf-GAL4 flies revealed no significant differences from their “normal” controls ( p ≥ 0.026, n = 27–36). Furthermore, within-group comparisons ( + /− pdf-GAL4 ) using t -tests indicate a significant difference only in the EP1306 group ( p = 0.002). Asterisk denotes significant difference between − pdf -GAL4 and + pdf -GAL4 phenotype. (B) Flies overexpressing Lmo in PDF cells show decreased sensitivity to cocaine. Flies heterozygous for both pdf-GAL4 and either one of two UAS-Lmo transgenes (black bars) were compared to flies carrying UAS-Lmo (white bars) or pdf-GAL4 (gray bar) alone. One-way ANOVA revealed a significant effect of genotype for both UAS-Lmo transgene groups. Post-test planned comparisons, with the critical p -value adjusted to 0.025, showed significant differences between the pdf-GAL4/UAS-Lmo flies and either pdf-GAL4/+ ( p < 0.02) or UAS-Lmo/+ controls ( p < 0.005). Asterisks denote significant differences , n = 16 experiments. (C) Confocal images demonstrate overlap between pdrm and PDF expression in the LN v s. In the left panel, UAS-mCD8GFP reveals the pdrm -driven GAL4 expression pattern (green) in the adult brain, and α-PAP staining (magenta) reveals PDF-expressing LN v s. Right panels are close-ups of the cell bodies of the LN v s; white areas correspond to regions of overlap between GFP (green) and PAP (magenta) expression. In order to test whether LN v s are also the locus for the resistance to cocaine observed in Bx mutants, we overexpressed Lmo in wild-type flies using the pdf-GAL4 driver and UAS-Lmo transgenes. We found that these flies showed a significant decrease in cocaine sensitivity ( Figure 5 B). The magnitude of the induced resistance was lower than that observed in the Bx J strain, but similar to that of weaker Bx alleles. It is possible that overexpression of Lmo with the pdf-GAL4 driver may not be as high as that caused by the Bx J mutation, or, alternatively, that overexpression of Lmo in cells other than the LN v s mediates the remaining resistance to cocaine observed in Bx J . Nonetheless, this experiment supports a role for LN v s as a site where Lmo regulates sensitivity to cocaine. In the adult fly, pdf - GAL4 drives expression in small and large LN v s (s-LN v s and l-LN v s) and in a group of neuroendocrine cells located at the very tip of the VNC (J. H. Park et al. 2000 ; Figure 5 C). The LN v s express many central clock genes and have been demonstrated, through genetic ablations and electrical silencing studies, to play a central role in maintaining circadian locomotor rhythmicity ( Renn et al. 1999 ; Nitabach et al. 2002 ; Peng et al. 2003 ); the function of the pdf-GAL4- expressing cells in the VNC is unknown. The pdrm line drives expression in both the s-LN v s and l-LN v s, as demonstrated by immunohistochemical analysis of pdrm/UAS-GFP flies with antibodies directed against the PDF precursor PAP ( Figure 5 C). pdrm does not, however, drive expression in the PDF-expressing cells located in the VNC (data not shown). This overlap in expression of pdrm and pdf-GAL4 in the LN v s implicates these few neurons as mediators of the altered cocaine sensitivity observed in Lmo mutants. Because of the known role of LMOs as regulators of developmentally important transcription factors ( Hobert and Westphal 2000 ), we asked whether the development of PDF-expressing LN v s is disrupted in Lmo mutants. Projections of the s-LN v s and l-LN v s can be visualized with an antibody that recognizes the PDF precursor PAP ( Renn et al. 1999 ). L-LN v s make an elaborate network of varicosities on the surface of the optic medulla, and project across the midline to the contralateral LN v s, while s-LN v s make a very specific projection to the dorsal central brain ( Figure 5 C). Both sets of LN v neurons also make extensive arborizations in the accessory medulla (reviewed in Hellfrich-Förster 2003). In the EP1306 and pdrm mutants, anti-PAP staining revealed that the number and detailed morphology of LN v neurons were completely normal (data not shown). Thus, Lmo does not appear to play a role in the development of the LN v s, although subtle developmental defects could have been missed. Lmo Is Required for Robust Circadian Locomotor Rhythms The increased sensitivity to cocaine of Lmo mutants suggested that these mutations alter LN v function without grossly affecting LN v development. In order to determine whether Lmo mutants have a more general dysfunction in the LN v s, we tested Lmo mutants in locomotor rhythm assays. The LN v s are required for the maintenance of circadian locomotor rhythms in constant darkness, and are thus the pacemaker neurons of the fly ( Stanewsky 2003 ). Lmo mutants EP1306, EP1383, hdp R26 , and hdp rev83 had less robust locomotor rhythms in constant darkness than their control lines ( Figure 6 A; data not shown), and all alleles had a higher tendency for arrhythmicity ( Figure 6 B). We used the power of the rhythm as an estimate of the degree of rhythmicity and found that EP1306 flies had weaker behavioral rhythms than EP1383 flies, which, in turn, were less robustly rhythmic than their control line (Ctl-2; Figure 6 C); similarly, hdp R26 and hdp rev83 flies had significantly weaker rhythms than their control line (Ctl-1; Figure 6 C). However, there were rhythmic flies in almost all genotypes assayed ( Figure 6 B), and the period lengths of the rhythms were very similar across the genotypes (see legend to Figure 6 for details). Together, these data demonstrate that disruption of Lmo results in weak circadian rhythms of behavior and suggest a generalized dysfunction of the LN v s in Lmo mutants. Figure 6 Wild-Type Lmo Is Required for Robust Circadian Rhythms of Locomotor Activity Locomotor activity of control (Ctl-1 and Ctl-2) and Lmo mutant ( EP1383, EP1306, hdp R26 , and hdp rev83 ) flies was recorded in constant darkness as previously described ( Nitabach et al. 2002 ). (A) Representative actograms of control and Lmo mutants. Control flies show robust circadian rhythms with clear distinctions between activity during the subjective day and inactivity during the subjective night. The pattern of the EP1306 and hdp rev83 mutants was more stochastic. (B) Graph showing the proportion of strongly rhythmic (white), weakly rhythmic (gray), and arrhythmic (black) flies for each genotype. Most control flies had strong rhythms (28/30 for Ctl-1 and 27/29 for Ctl-2) while Lmo mutants formed a series with an increasing fraction of the flies arrhythmic. For example, 13/26 hdp rev83 flies were arrhythmic, with the other 13 all having weak rhythms. The power of the rhythm was used to estimate the strength of the activity rhythm, with a power of 300 or more classed as a strong rhythm and a power between 300 and 170 classed as a weak rhythm; arrhythmics were given a power of 170 (for analysis below). Between 24 and 30 flies were assayed for each genotype. There were no major differences in the period length of the rhythmic flies in each genotype (Ctl-1, 23.6 ± 0.3; Ctl-2, 23.5 ± 0.5; EP1383, 23.2 ± 0.4; EP1306, 23.4 ± 0.3; hdp R26 , 24.2 ± 0.5; and hdp rev83 , 23.4 ± 0.3). (C) Quantitation of the average power of the rhythm with error bars showing standard error of the mean. One-way ANOVA revealed significant differences between genotypes ( p < 0.0001). Post-hoc t -tests using a Bonferroni correction revealed that the power of the rhythm was significantly different between control flies and the Lmo mutants hdp R26 , hdp rev83 , and EP1306 ( p < 0.01). EP1383 flies had a significantly weaker rhythm than Ctl-2 flies ( p < 0.05). Ctl-1, w 1118 , is the appropriate genetic control for the hdp alleles (black columns); Ctl-2, EP1631, is the appropriate control for EP1306 and EP1383 (gray columns). All flies tested are in the same genetic background, that of the w 1118 flies. (D) Quantitation of average activity (beam crossings per minute) with error bars showing standard error of the mean. ANOVA did not reveal significant differences between genotypes at the 0.01 level. Ctl-1, w 1118 , is the appropriate genetic control for the hdp alleles (black columns); Ctl-2, EP1631, is the appropriate control for EP1306 and EP1383 (gray columns). All flies tested are in the same genetic background, that of the w 1118 flies. LN v s Regulate Circadian Rhythmicity and Cocaine Sensitivity Independently The observation that Lmo functions in the PDF neurons to regulate cocaine sensitivity, together with the finding that Lmo mutants show weak circadian locomotor rhythms, suggested that the pathways regulating cocaine sensitivity interact with the circadian clock. Alternatively, Lmo could regulate these two behaviors independently. We addressed these possibilities in two sets of experiments. First, we determined whether cocaine sensitivity was regulated by the circadian clock. For this purpose we entrained flies to light–dark (LD) cycles and then tested them for cocaine sensitivity during the light and dark phases at 3-h intervals over 24 h (see Materials and Methods ). We found that cocaine sensitivity was essentially the same at all times tested ( Figure 7 A), demonstrating that cocaine responsiveness is not a behavioral output of the circadian clock. Figure 7 Cocaine Responses Are Not a Circadian Output and pdf Mutants Show Wild-Type Cocaine Sensitivity (A) Cocaine responses do not vary with the circadian clock. Control (EP1631) flies were raised under LD conditions and assayed for cocaine phenotypes in the crackometer at the indicated Zeitgeber (ZT) times. One-way ANOVA revealed no significant effect of time of day ( F = 0.53, p = 0.82, n = 32). (B) Flies lacking the neuropeptide PDF (pdf 01 ) display normal cocaine sensitivity. pdf 01 homozygotes (pdf 01 /pdf 01 ) and pdf 01 hemizygotes (pdf 01 /Df) showed wild-type responses to cocaine in the crackometer. Individual pairwise comparisons using Student's t -tests revealed no significant differences between control (+/+ and +/ Df ) and pdf mutant genotypes ( p = 0.69, p = 0.97, n = 6–8 experiments) Second, we asked whether PDF, the only known functional output of the LN v s in the context of circadian rhythms, is involved in regulating cocaine sensitivity. pdf mutant flies show a circadian phenotype that has been localized to the LN v s. We found that the pdf 01 mutant flies, which completely lack PDF ( Renn et al. 1999 ), showed normal cocaine responses compared to the control strain ( Figure 7 B). To eliminate the possibility that the absence of a phenotype was caused by genetic modifiers in the background of the pdf 01 strain, we also tested flies carrying the pdf 01 chromosome over a deficiency for the locus (see Materials and Methods ). These pdf 01 hemizygous flies also displayed normal cocaine sensitivity ( Figure 7 B). Taken together, these results show that the altered cocaine sensitivity of Lmo flies is not secondary to their abnormal circadian rhythms. Moreover, our data show that cocaine sensitivity and circadian behaviors, although both localized to the LN v s, are genetically separable. LN v s and Their Synaptic Activity Regulate Acute Cocaine Responsivity The hypersensitivity of Lmo mutants to cocaine could result from either the disruption of an LN v output that acts normally to dampen cocaine sensitivity or from an increase in an output of LN v s that normally enhances cocaine sensitivity. In order to differentiate between these two possibilities, we tested flies that lacked LN v s, generated by targeted expression of the cell death gene head involution defective (hid) using pdf-GAL4 . This approach was previously used to demonstrate that LN v s are the Drosophila pacemaker neurons responsible for rhythmic locomotor activity ( Renn et al. 1999 ). When tested in the crackometer, flies with LN v ablations showed reduced sensitivity to cocaine when compared to controls ( Figure 8 ). These results show that LN v s normally act to increase cocaine responsiveness and that this effect is antagonized by Lmo function in these cells (see Discussion). Figure 8 Silencing or Ablating PDF Cells Induces Resistance to Cocaine Ablation of PDF cells with pdf-GAL4 and UAS-hid reduced sensitivity to cocaine compared to parental lines ( pdf-GAL4/+ and UAS-hid/+ ). Electrical (UAS-Kir2.1 8 or UAS-Kir2.1 7 ) or synaptic (UAS-TeTx) silencing of PDF cells with pdf - GAL4 phenocopied PDF cell ablations. One-way ANOVAs with post hoc planned comparisons revealed a significant effect of genotype for the UAS-hid ( p < 0.003, n = 20), UAS-Kir2.1 7 ( p < 0.008, n = 32), UAS-Kir2.1 8 ( p < 0.002, n = 28), and UAS-TeTx ( p < 0.001, n = 28) groups, but not for UAS-TeTx in ( p > 0.045, n = 28) ( n corresponds to the number of experiments). Critical p -value was adjusted to p = 0.025. Asterisks denote significant differences in both planned comparisons ( pdf-GAL4/+ and UAS- transgene/+ versus pdf-GAL4/UAS -transgene). Variations in phenotype of pdf-GAL4 flies for each set of experiments is caused by day-to-day variability. To confirm that the behavioral resistance observed in LN v -ablated animals results from loss of neuronal signaling from these cells, rather than from developmental compensations induced by their ablation, we functionally silenced the LN v s either electrically, by targeted expression of the mammalian inward rectifying K + channel (Kir2.1), or synaptically, by expression of tetanus toxin light chain (TeTx) ( Sweeney et al. 1995 ; Baines et al. 2001 ). These manipulations do not affect LN v survival or the normal projection patterns of LN v s ( Kaneko et al. 2000 ; Nitabach et al. 2002 ). Expression of either Kir2.1 or TeTx resulted in reduced cocaine sensitivity similar to that seen with LN v ablations ( Figure 8 ). Importantly, expression of an inactive TeTx (TeTx in ) did not significantly alter cocaine responses, confirming that the actions of these transgenes were specific to their ability to silence or ablate the LN v s ( Figure 8 ). These data further demonstrate that circadian phenotype does not predict cocaine phenotype, as targeted expression of either hid or Kir2.1 in LN v s causes arrhythmia, while expression of TetTx does not ( Renn et al. 1999 ; Kaneko et al. 2000 ; Nitabach et al. 2002 ). In summary, these results confirm a novel role for LN v activity and synaptic output in increasing behavioral responses to cocaine in a manner independent from its role in regulating circadian locomotor rhythmicity. Discussion In a genetic screen for Drosophila mutants with altered acute responses to cocaine, we isolated multiple mutations in the Lmo locus. Behavioral characterization of gain- and loss-of-function alleles of Lmo demonstrates an inverse correlation between Lmo expression levels and cocaine sensitivity. Through targeted expression of Lmo, we show that these altered cocaine responses are caused by differential Lmo expression in the circadian pacemaker neurons, the LN v s. Consistent with a dysfunction of these pacemaker neurons in Lmo mutants is our finding that the mutant flies also show altered circadian locomotor rhythms. However, using a variety of genetic methods to ablate or functionally silence these neurons, we reveal a novel role for the LN v s in modulating cocaine's acute locomotor responses that is independent of their pacemaker function. These findings add to mounting data in Drosophila and mice supporting a role for circadian genes in cocaine-related behaviors ( Andretic et al. 1999 ; S. K. Park et al. 2000 ; Abarca et al. 2002 ; Rothenfluh and Heberlein 2002 ). Our discovery that cocaine actions are modulated by neurons critical for normal circadian locomotor rhythmicity suggests a basis for this overlap. Lmo Functions in PDF Neurons to Regulate Cocaine Sensitivity We provide several lines of evidence that levels of Lmo expression in PDF-expressing LN v s regulate acute sensitivity to volatilized cocaine in Drosophila . First, loss-of-function mutations in Lmo show increased cocaine sensitivity, a defect that can be reversed by induced expression of Lmo in the LN v s. Second, Bx mutants, in which Lmo is overexpressed, show reduced cocaine sensitivity; this resistance can be mimicked by overexpression of Lmo in the LN v s. The LN v s are a group of 8–10 neurons in each brain hemisphere that express the neuropeptide PDF. These neurons have previously been identified as the circadian pacemaker neurons of adult Drosophila and are involved in modulating rhythmic locomotor behavior ( Renn et al. 1999 ; Blanchardon et al. 2001 ). Interestingly, expression of a mouse homolog of Lmo, Lmo4, is highly enriched in the suprachiasmatic nucleus (SCN) (A. Lasek, D. Kapfhamer, and U. H., unpublished data), the mammalian central pacemaker. Furthermore, microarray analysis revealed that in many tissues, Lmo4 expression varies with circadian time ( Panda et al. 2002 ). These data suggest an evolutionarily conserved role for Lmo in clock neuron function. LMOs are known to act as regulators of LIM-HD protein activity and stability ( Retaux and Bachy 2002 ). LIM-HD proteins are transcription factors involved in many stages of nervous system development, from neuronal generation and axon guidance to determination of neuronal subtype identity ( Hobert and Westphal 2000 ). In addition, expression of LIM-HD proteins in postmitotic neurons suggests a role in maintaining the differentiated state of these neurons ( Hobert and Westphal 2000 ). The development and structure of the PDF-expressing LN v s, the neurons to which we localize Lmo action, have been studied in detail ( Helfrich-Förster 1997 ). The LN v s can be divided into two groups of 4–5 neurons based on cell body size, projection pattern, and time of development. The s-LN v s, which arise in early larval development, project to the dorsal central brain, terminating near two sets of dorsal neurons that express clock genes at high levels. The l-LN v s arise during pupal stages and project onto the surface of the optic medulla, as well as to the contralateral LN v s through fibers running in the posterior optic tract. Both s- and l-LN v s have dense arborizations in the accessory medulla, a neuropil proposed to be a circadian pacemaker center in cockroaches and crickets ( Helfrich-Förster 1998 ). An assessment of s-LN v and l-LN v numbers and detailed projection patterns revealed no differences between wild-type flies and Lmo mutants. Furthermore, by both quantitative RT-PCR and immunohistochemistry, we determined that PDF levels are normal in Lmo mutants (data not shown). Expression of PDF, the only known LN v output, is restricted to the LN v s and a few tritocerebral neurons. As PDF expression is a specific marker for the differentiated state of the LN v s, it is unlikely that in Lmo mutants these neurons are grossly abnormal in their terminal differentiation. The absence of obvious structural abnormalities of LN v s suggests that Lmo may play an active role in regulating cocaine responses. In fact, evidence is mounting that the expression and activity of LMOs are dynamically regulated within the nervous system. For instance, expression of the murine Lmo homologs Lmo1, Lmo2, and Lmo3 is differentially regulated by seizure activity in specific regions of the hippocampus and forebrain of adult mice ( Hinks et al. 1997 ). In addition, gene array experiments have revealed that expression of mammalian Lmo homologs is under circadian regulation in the SCN and is increased in the cerebral cortex during sleep deprivation ( Cirelli and Tononi 2000 ; Panda et al. 2002 ). Furthermore, Lmo3 was isolated as a transcript upregulated by DA administration in cultured astrocytes ( Shi et al. 2001 ). Lastly, Lmo2 and Lmo4 were recently isolated in a screen for calcium-regulated activators of transcription ( Aizawa et al. 2004 ), suggesting a role for LMOs in regulating gene expression changes induced by neural activity. These data are intriguing in light of the many functional changes that occur in reward pathways in the addicted state. Whether LMO activity and/or expression are regulated by acute cocaine exposure remains to be studied. Separable Roles of LN v s in Regulating Cocaine and Circadian Behaviors We provide evidence that the LN v s regulate cocaine-induced behaviors, in addition to their well-known role in controlling circadian locomotor behaviors. Flies with LN v ablations show reduced sensitivity to cocaine, establishing that these cells normally increase behavioral responses to cocaine. Our experiments also indicate that LN v s drive circadian locomotor and cocaine behaviors in distinct ways. First, we found that sensitivity to cocaine is not modulated in a circadian manner, showing that the cocaine response is not simply an output of the central clock. Second, we showed that Lmo mutants, pdf mutants, and flies in which LN v s have been electrically silenced or ablated—all known to display similar locomotor rhythm deficits—show completely uncorrelated cocaine sensitivities: increased, unchanged, and reduced sensitivity, respectively. Lastly, we provide evidence that the LN v outputs that mediate cocaine and circadian behaviors are divergent. PDF, the only known LN v neurotransmitter, is required for normal locomotor rhythms ( Renn et al. 1999 ). pdf null mutants, however, show normal cocaine responses. Furthermore, while synaptic silencing of PDF neurons does not disrupt circadian locomotor activity ( Kaneko et al. 2000 ), we show here that the same manipulation reduces cocaine responses to the same extent as neuronal ablation. Together, these results imply the existence of an alternate, TeTx-sensitive functional output that mediates LN v modulation of cocaine responses. Interestingly, Blau and colleagues have also hypothesized a PDF-independent LN v output that regulates another rapid behavioral response, larval photophobicity (E. Mazzoni, C. Desplan, and J. Blau, unpublished data). How might LN v s modulate circadian rhythmicity and cocaine sensitivity independently? It is possible that these cells use distinct output mechanisms, PDF to regulate circadian locomotor rhythms and another unknown signal to regulate cocaine sensitivity. Alternatively, these behaviors could be regulated by distinct subsets of LN v s. For example, several recent findings suggest that l-LN v s may play a lesser role in regulating circadian rhythmicity ( Helfrich-Förster 2003 ). First, l-LN v s do not project to the dorsal brain, an area implicated in locomotor rhythmicity ( Helfrich-Förster 1997 ). Second, unlike in s-LN v s, molecular clock cycling is not sustained for long in these cells during free-running conditions ( Kaneko et al. 2000 ; Yang and Sehgal 2001 ; Shafer et al. 2002 ). Lastly, PDF expression and release is modulated by the molecular clock in s-LN v s, but not in l-LN v s ( Blau and Young 1999 ; J. H. Park et al. 2000 ). However, another study found that normal rhythmicity can be obtained in flies lacking s-LN v s ( Helfrich-Förster 1998 ). Moreover, the projections of l-LN v s connect the l- and s-LN v s from both brain hemispheres through projections in the posterior optic tract, suggesting a functional link between the two groups of cells. Whether Lmo functions in the s- and/or l-LN v s to regulate cocaine sensitivity and circadian rhythmicity cannot be established with currently available tools. There is growing evidence for a functional link between circadian neurons and the modulation of cocaine-related behaviors. In mammals, the peptidergic neurons of the SCN have been shown by a variety of studies to be the central pacemakers controlling circadian rhythms ( van Esseveldt et al. 2000 ). Interestingly, the fetal mammalian SCN contains DA D1 receptors through which cocaine can influence entrainment of fetal biological rhythms ( Simonik et al. 1994 ; Viswanathan et al. 1994 ; Bender et al. 1997 ). In addition, disruption of another major component of the mammalian circadian system, the pineal gland, or its secretory product melatonin, results in altered cocaine responses ( Uz et al. 2002 , 2003 ; Zhdanova and Giorgetti 2002 ). Hirsh and colleagues showed that mutants in the Drosophila clock genes period, clock, and cycle, but not timeless fail to sensitize to repeated cocaine exposures ( Andretic et al. 1999 ). These genes, which are expressed at high levels in LN v s, have been to shown to act in these cells to modulate circadian behavior ( Kaneko et al. 1997 ). It is not known whether circadian gene function in LN v s also regulates behavioral sensitization to cocaine. How Might LN v s Modulate Cocaine Responses We have demonstrated that LN v electrical activity and synaptic output contribute to cocaine-induced behavioral responses, raising a number of questions regarding the interaction of cocaine with these neurons and their output. We propose a simple model whereby the activity of some or all LN v s directly increases upon cocaine administration, which in turn results in cocaine-induced changes in locomotion ( Figure 9 ). Interestingly, a recent report showed that cultured LN v neurons can respond to either DA or acetylcholine, but not to glutamate, serotonin, octopamine, or histamine ( Wegener et al. 2004 ). The inferred presence of DA receptors on a subset of LN v s provides a potential mechanism by which LN v activity could be directly increased by cocaine administration, as cocaine's primary mechanism of action is to inhibit the reuptake of DA by DAT. In this model, when the LN v s are ablated or silenced, one site of cocaine action would be eliminated, thus reducing cocaine's effect ( Figure 9 B). Figure 9 A Model for LN v and LMO Regulation of Cocaine Sensitivity (A) In wild type, LN v s modulate locomotor responses via electrical activity and synaptic transmission. We propose a model in which cocaine acts to directly increase LN v activity. Upon cocaine administration, synaptic DA concentrations are increased (via cocaine's inhibition of the plasma membrane DA transporter). Activation of presumed DA receptors on the LN v (dark arrowheads) stimulates electrical activity and subsequent synaptic output. This activity contributes to the behavioral response of the fly to cocaine. (B) LN v ablations eliminate LN v contribution to the cocaine response, reducing cocaine sensitivity. (C) In our model, Lmo loss-of-function mutants (Lmo LOF ), which have increased cocaine sensitivity, have increased activity/output during the cocaine response. This increased activity may be mediated by increases in receptor content on the LN v or by recruitment of other LN v s that normally do not participate in the cocaine response. (D) Lmo gain-of-function mutants (Lmo GOF ) mutants have reduced LN v output and reduced cocaine sensitivity. This could also result from a reduction in receptor density. How would Lmo fit into this model? We propose that in Lmo loss-of-function mutants, LN v output is boosted, possibly because of increased expression of DA receptors ( Figure 9 C), leading to enhanced cocaine sensitivity. Conversely, increased Lmo expression (in Bx mutants or in flies specifically overexpressing Lmo in PDF cells) would result in reduced receptor content and, consequently, in dampened cocaine sensitivity ( Figure 9 D). Further studies are needed to identify the putative DA receptor (or other molecules) that functions in LN v s to regulate cocaine-induced behaviors. LMO-induced changes in receptor expression are not inconceivable given LMO's interaction with LIM-HD proteins. Changes in LIM-HD protein function have been shown to affect aspects of neuronal subtype identity, including neurotransmitter and receptor expression profiles. For instance, mutations in the Drosophila LIM-HD gene islet cause a loss of DA and serotonin synthesis, while ectopic expression leads to ectopic expression of tyrosine hydroxylase, an enzyme required for DA biosynthesis ( Thor and Thomas 1997 ). In addition, expression of a Drosophila DA receptor in larval neurons requires the function of the LIM-HD gene apterous ( Park et al. 2004 ). The observation that DA receptor expression is also altered in various Caenorhabditis elegans LIM-HD mutants ( Tsalik et al. 2003 ) suggests an evolutionarily conserved role of LIM-HD proteins and possibly LMOs in the regulation of neurotransmitter identity and responsiveness. Consistent with this idea is the finding that Lmo homologs are highly expressed in the mesolimbic DA system of mice (A. Lasek, D. Kapfhamer, and U. H., unpublished observations)—a neural pathway involved in the acute stimulant and rewarding properties of abused drugs. We posit that LMOs are ideally suited to modulate in a subtle manner the neurochemical identity and sensitivity of the nervous system to various stimuli, including drugs of abuse. Materials and Methods Drosophila culture and strains. All flies were maintained on standard cornmeal molasses agar at 25 °C and 70% humidity under constant weak light. The Rørth EP collection was obtained from G. M. Rubin (University of California, Berkeley, California, United States) ( Rørth 1996 ). The pdrm allele was originally isolated in a genetic screen of P[GAL4] insertions (carrying the GawB element) for altered response to the locomotor-activating effects of ethanol (F. Wolf, unpublished data) and consequently tested for cocaine sensitivity. The location of the insertion was determined by inverse PCR ( http://www.fruitfly.org/about/methods/inverse.pcr.html ). The hdp R590 and hdp R26 loss-of-function alleles were provided by S. Cohen (European Molecular Biology Laboratory, Heidelberg, Germany) ( Milan et al. 1998 ). UAS-Lmo lines were provided by C. Zeng (University of Wisconsin, Milwaukee, Wisconsin, United States) ( Zeng et al. 1998 ). The hdp rev83 strain was generated by N. Justice by imprecise excision of the EP1306 line, screening for the hdp phenotype (unpublished data). Bx alleles and pdf deficiencies (Df(3R)T1-X and Df(3R)T1-P) were obtained from the Drosophila Stock Center (Bloomington, Indiana, United States). w33, pdf 01 and pdf-GAL4 flies were obtained from P. Taghert (Washington University, St. Louis, Missouri, United States) ( Renn et al. 1999 ). UAS-Kir2.1 lines were obtained from S. Sweeney (University of California, San Francisco, California, United States) ( Baines et al. 2001 ). UAS-TeTx lines were generated as previously described ( Sweeney et al. 1995 ; Scholz et al. 2000 ). UAS-hid flies were provided by R. S. Stowers (NASA Ames Research Center, Moffett Field, California, United States) ( Zhou et al. 1997 ). UAS-tauGFP and UAS-lacZ lines were obtained from Y.-N. Jan (University of California, San Francisco). All lines used for behavioral experiments, unless noted below , were out-crossed for five generations to a w 1118 stock isogenic for Chromosomes II and III. Because pdf 01 and its background control strain (w33) are unmarked, only Chromosomes I and II were replaced via crosses to balancer stocks in the w 1118 genetic background. Genetic screen and selection of controls The approximately 400 X-linked EP lines were initially screened as hemizygotes crossed to a w, X^X/Y tester strain in the crackometer as described below ( n = 6–9). These lines distributed normally, and 30 lines with phenotypes greater than 1.5 standard deviation from the mean were out-crossed into our w 1118 background and retested. An additional 70 lines were selected to be out-crossed in order to provide a population distribution from which to select control lines. These 100 out-crossed lines were rescreened for acute responses to cocaine. A group of five control EP lines were chosen based on having a normal acute response (near the mode of the distribution). In all behavioral experiments involving EP lines, at least three of these control lines were tested to ensure that the control shown was representative of the EP population. The specific control EP line shown for each experiment is noted in the figure legends. As a control for the pdrm line we used a P[GAL4] insertion (line 8.142 ) that shows a normal response to multiple drugs, including cocaine (A. Rothenfluh, D. Guarnieri, A. Rodan, and F. Wolf, unpublished data). For data shown in Figures 1 and 3 , all lines were crossed to a w, X^X/Y tester strain to reduce possible effects of recessive autosomal modifiers. In addition to the genotypes shown, we tested three P-element lines in the Lmo locus that did not produce cocaine phenotypes: MS1096, which lies 20 bp downstream of the first RA transcript exon, and two additional EP lines, EP0443 and EP1394, inserted 5′ to EP1383; the EP lines have normal wings. MS1096 has been shown to have a very weak wing venation phenotype, but has otherwise not been phenotypically or molecularly characterized ( Milan et al. 1998 ). Behavioral assays and statistics For cocaine-sensitivity assays, for all behavioral experiments, 15 male flies were collected under CO 2 anesthesia 0–2 d post eclosion (day 12), and tested 2–3 d later. Flies were equilibrated to room temperature for at least 1 h before being exposed to cocaine. Cocaine exposures were performed as described previously ( McClung and Hirsh 1998 ; Bainton et al. 2000 ), and startle-induced negative geotaxis was assayed in a glass cylinder as described previously ( Bainton et al. 2000 ). A drug effect score was determined as the average (measured every minute over 5 min) number of flies that remained on the bottom of the cylinder, expressed as percent of the total number of flies. Significance was established for each experiment as described in figure legends. Generally, either Student's paired t -tests assuming equal variance or one-way ANOVAs with post hoc Bonferroni planned comparisons or Tukey-Kramer multiple comparisons were performed in GraphPad Prism 4 (GraphPad, San Diego, California, United States). Error bars in all experiments represent the standard error of the mean. To maintain an experiment-wide error rate of α = 0.05, the adjusted error rates were p = 0.05/ n for the n subsequent planned pairwise comparisons in each experiment. To observe cocaine locomotor activity patterns, ten flies were exposed to volatilized cocaine for 1 min, transferred to a 7.5 cm × 10 cm × 0.5 cm acrylic box, and images were captured on an Apple G4 computer (Apple, Cupertino, California, United States) using Adobe Premiere (Adobe Systems, San Jose, California, United States) at 10 frames/s for 5 min. Fly locomotion was tracked using the dynamic image analysis system software (Solltech, Oakdale, Iowa, United States). All genotypes were tested at each dose multiple times ( n ≥ 3), and representative data were selected for the figures. Baseline negative geotaxis was measured by mock exposures of flies and subsequent assay in the crackometer as above; all genotypes displayed behavioral scores of less than 10%, with no significant differences between genotypes ( n > 8). For circadian experiments, locomotor activity of individual flies was measured using the TriKinetics (Waltham, Massachusetts, United States) infrared beam-crossing system recording total crosses in 10-min bins. Raw activity histograms were analyzed for circadian rhythms using Actimetrics (Wilmette, Illinois, United States) Clocklab software. Chi-square periodograms were constructed according to Sokolove and Bushell (1978) , and significant circadian rhythmicity was defined as presence of a peak in periodogram power that extends above the 0.01 Chi-prosquare significance line. Since this line is equal to a power of approximately 175 at a period of 24 h, flies with no periodogram peak crossing the significance line were assigned a circadian power of 170. This would tend to overestimate the circadian power of these flies, and thus is conservative with regard to assessing statistical differences in power between genotypes exhibiting frequent arrhythmicity and those that are predominately rhythmic. For the circadian cocaine-sensitivity experiment, flies were set up and raised in LD, collected under CO 2 anesthesia during the light phase 1–2 d after eclosion, and placed back in LD for 2–3 d. At the indicated Zeitgeber times, flies were tested in the crackometer, under lights, within 5 min of being removed from LD conditions. Histology Expression pattern of the pdrm P[GAL4] insertion was examined by crossing to UAS-GFPT2 and UAS-mCD8GFP . All adult and larval preparations were dissected in PBS, fixed in 4% formaldehyde/PEM for 40 min, washed in PBS, and dehydrated in 50% glycerol/PBS for 1 h. Tissue was mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, California, United States) and analyzed with a Bio-Rad confocal microscope with Bio-Rad Lasersharp 2000 software (Bio-Rad, Hercules, California, United States). PAP antibody staining (provided by P. Taghert) was performed on CNS preparations that were dissected, fixed, and washed as above. Specimens were incubated in 1:2,000 dilution of anti-PAP in PBT, and with a Texas Red–coupled goat anti–guinea pig secondary antibody, diluted 1:200 (Jackson Laboratory, Bar Harbor, Maine, United States). Real-time quantitative RT-PCR Flies 2- to 4-d-old were collected and frozen immediately at −80 °C. Heads were removed from bodies by vortexing, and separated in a sieve. RNA was extracted from heads and bodies by homogenizing the flies in hot phenol and NTES. Complementary DNA (cDNA) was prepared using TaqMan Reverse Transcription Reagents (Applied Biosystems, Foster City, California, United States) according to the manufacturer's specifications, with the addition of random sequence hexamers to 2.5 μM. cDNA was analyzed by quantitative, real-time PCR using the ABI PRISM 7700 Sequence Detection System (Applied Biosystems). The following probes and primers were designed using ABI PrimerExpress software: Lmo RA-For, GAAGAGAAACAACAGCAGCAACA; Lmo RA-Rev, ATTTGCATATTTCGCACTTGTTTAGCT; and Lmo RA-Probe, CTGCTGCCGTTGCTG. rp49 probe and primers (CT 6405) were obtained from Applied Biosystems. TaqMan PCR reactions consisted of 50 ng of cDNA, 0.9 μM each diagnostic primer, 0.25 μM diagnostic probe, and 1x final of TaqMan Universal PCR Mastermix (Applied Biosystems) in a reaction volume of 25 μl. The TaqMan PCR conditions used were as described in TaqMan guidelines. Each sample was analyzed in triplicate. As negative controls, we used both no-template and DNase-treated non-reverse-transcribed mRNA samples; no significant amplification was observed in these samples. rp49 transcript levels were used as an endogenous normalization control for RNA samples, and relative mRNA abundance was calculated using the comparative delta-Ct method. Reference mRNA is noted in figure legends. Quantitative RT-PCR analysis was performed on at least two independent RNA preparations, with similar results.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529317.xml

539064

The Evolution of Self-Fertile Hermaphroditism: The Fog Is Clearing

The nematode Caenorhabditis elegans is a little less lonely than the rest of us—it is a self-fertile hermaphrodite, which as a larva makes and stores sperm before switching to egg production for the remainder of its lifespan. ( C. elegans also maintains some males at a low frequency, about 1 in 500, and the hermaphrodite's eggs can be fertilized by sperm either from males or themselves.) A sister species, C. briggsae , is also hermaphroditic, but phylogenetic evidence suggests the last common ancestor of the two species had a female/male mode of reproduction. This raises the question of how the sex determination mechanisms, which must have evolved independently, differ between the two species. In this issue, Sudhir Nayak, Johnathan Goree, and Tim Schedl show that a crucial difference lies in the activities of two genes. In C. elegans , the early period of sperm production is controlled by multiple proteins, two of which are the focus of this study, the RNA-binding protein GLD-1 (encoded by the gene gld-1 ) and the F-box-containing protein FOG-2 (encoded by the gene fog-2 ). Together, they repress translation of a gene, tra-2 , by binding to its messenger RNA. This allows another gene, fem-3 , to transiently masculinize the larval germline to produce sperm. Wild-type C. elegans hermaphrodite stained to highlight the nuclei of all cells Comparing the genomes of C. elegans and C. briggsae , Schedl and colleagues found they share 30 out of 31 sex determination genes, but not fog-2 . More surprisingly, they found that the role of gld-1 in sex determination is opposite in the two species. When C. elegans is deprived of gld-1 , would-be hermaphrodites produce only oocytes. But when C. briggsae is deprived of gld-1 , would-be hermaphrodites produce only sperm. Thus, the authors conclude, the control of hermaphrodite spermatogenesis is fundamentally different in the two species. By further examining the C. elegans genome, the authors showed that fog-2 arose from a gene duplication event after the C. elegans – C. briggsae split, which occurred approximately 100 million years ago. Since then, its final exon, which codes for the C-terminal end of the protein, has undergone rapid evolution. The authors also show that this is the “business end” of the protein for its interaction with GLD-1, suggesting that the divergence of C. elegans and C. briggsae sex determination pathways resulted, in part, from FOG-2's new interaction with GLD-1. Exactly what the role of fog-2 is in C. elegans is still unclear. The authors speculate that it may recruit additional factors onto the gld-1/tra-2 mRNA complex, increasing efficiency of translation repression. Much remains to be discovered about C. briggsae sex determination as well. The authors suggest that additional genetic differences promoting self-fertility are likely to have accumulated since the two species diverged, which may act to strengthen the male–female germline switching signal. Investigation of this possibility may shed more light on how hermaphroditism operates in these two species, and how a developmental pathway controlling sex determination can evolve.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539064.xml

544879

Polymorphic segmental duplications at 8p23.1 challenge the determination of individual defensin gene repertoires and the assembly of a contiguous human reference sequence

Background Defensins are important components of innate immunity to combat bacterial and viral infections, and can even elicit antitumor responses. Clusters of defensin (DEF) genes are located in a 2 Mb range of the human chromosome 8p23.1. This DEF locus, however, represents one of the regions in the euchromatic part of the final human genome sequence which contains segmental duplications, and recalcitrant gaps indicating high structural dynamics. Results We find that inter- and intraindividual genetic variations within this locus prevent a correct automatic assembly of the human reference genome (NCBI Build 34) which currently even contains misassemblies. Manual clone-by-clone alignment and gene annotation as well as repeat and SNP/haplotype analyses result in an alternative alignment significantly improving the DEF locus representation. Our assembly better reflects the experimentally verified variability of DEF gene and DEF cluster copy numbers. It contains an additional DEF cluster which we propose to reside between two already known clusters. Furthermore, manual annotation revealed a novel DEF gene and several pseudogenes expanding the hitherto known DEF repertoire. Analyses of BAC and working draft sequences of the chimpanzee indicates that its DEF region is also complex as in humans and DEF genes and a cluster are multiplied. Comparative analysis of human and chimpanzee DEF genes identified differences affecting the protein structure. Whether this might contribute to differences in disease susceptibility between man and ape remains to be solved. For the determination of individual DEF gene repertoires we provide a molecular approach based on DEF haplotypes. Conclusions Complexity and variability seem to be essential genomic features of the human DEF locus at 8p23.1 and provides an ongoing challenge for the best possible representation in the human reference sequence. Dissection of paralogous sequence variations, duplicon SNPs ans multisite variations as well as haplotypes by sequencing based methods is the way for future studies of interindividual DEF locus variability and its disease association.

Background Despite the tremendous efforts and successful completion of the Human Genome Project by April 14 th 2003, a set of recalcitrant gaps remain in the euchromatic part of the final human genome sequence. One obvious reason for these gaps is that the appropriate regions are enriched in sequences that are not tolerated by the cloning systems. The second possibility is that even if clones are available and amenable for sequencing, their sequences cannot be unambiguously aligned due to gap flanking segmental duplications. Generally, those duplicons are defined by >90% sequence identity and lengths of >1 kb and about 87% of all human ones are longer than 50 kb [ 1 ]. In these regions with nucleotide identities up to >99% over several kb it is nearly impossible to decide whether very similar sequences represent distinct loci or different alleles of a single locus. Here, sequencing of a single chromosomal haplotype is a straightforward approach to achieve a „consistent“ assembly. It was successfully applied to decipher intrachromosomal duplications of the human Y [ 2 ]. If, however, duplications are located on autosomes and their copy numbers vary interindividually, as shown for regions in 15q11-q13 [ 3 ], the situation becomes even more complicated and requires the extra effort of resolving haplotype differences that result from the diploid nature of the underlying BAC library. In the Williams-Beuren syndrome (WBS) region on human chromosome 7, only extensive redundant sequencing from a single BAC library led to a representative sequence [ 4 ]. Alternatively, monospermic complete hydatidiform moles [ 5 , 6 ] and hamster somatic cell hybrids [ 7 ] provide access to fully homozygous genomes or individual autosomes, respectively. It is a fact that structural variations between chromosomal haplotypes complicate the sequence assembly and lead to the formation of de facto gaps [ 1 , 8 ]. The more haplotypes are represented by BAC clones, the more de facto gaps may be formed. In the case of unresolved segmental duplications, usually a large number of clones has been sequenced with high accuracy [ 9 ] and the clone coverage of the loci is well above-average of the entire human genome. However, no contiguous tiling path can be build and gaps remain. Nevertheless, the available data are an invaluable resource for the investigation of individual genetic variations in duplicated regions and of their association with diseases. One of those complex regions is located in 8p23.1 at 6.3 – 8.3 Mb of the July 2003 human reference sequence (NCBI Build 34; UCSC version hg16, Fig. 1A ). In the Golden Path assembly [ 10 ], there are 22 finished clones from five different libraries and 20 working draft or predraft clones (<4x coverage shotgun; four different libraries) grouped on both sides of a recalcitrant gap at 7.5 Mb. Another 10 finished clones from four different libraries are not included in the hg16 assembly but map to the 8p23.1 locus. Several attempts to close this gap have failed due to the highly repetitive structure of the flanking sequences. The gap flanking regions harbor defensin (DEF) genes, encoding a group of small cationic peptides with characteristic three intramolecular disulfide bonds. These peptides play a prominent role in innate immunity to defend bacterial and viral infections in animals, plants and insects [ 11 ]. Furthermore, in humans, loss or down regulation of DEF genes is shown to be related with cancer, such as renal cell carcinoma [ 12 - 14 ], prostate cancer [ 14 ] and bladder tumors [ 15 ]. Two different DEF gene clusters can be distinguished: DEF cluster a contains the genes DEFB1 , DEFA6 , DEFA4 , DEFA1 , DEFT1 , DEFA3 and DEFA5 ; DEF cluster b comprises the genes DEFB109p , DEFB108 , DEFB4 , DEFB103 , SPAG11 , DEFB104 , DEFB106 , DEFB105 , and DEFB107 (Fig. 2 ). DEF cluster b is duplicated in reverse complementary orientation on either side of the gap, forming the distal cluster b1 and the proximal cluster b2 . Figure 1 Alternative alignments of the 8p23.1 DEF locus. ( A ) July 2003 UCSC version hg16 [10] chr8:6,258,283-8,262,034. Only finished clones are shown and arranged by libraries, which are indicated by background colors: RP11 = gray, bottom; SCb = white, middle; other (CTB, CTD, GS, RP13) = gray, top. Defensin gene clusters are shown as arrows, repeat blocks are indicated as striped boxes, (+) strand is above the black line, (-) strand is below the black line, same stripe patterns indicate similar structures. The light blue background indicates the distal repeat region for chromosomal rearrangements [16]. ( B ) Revised alignment of the 8p23.1 DEF locus, containing an additional 360-kb-contig and five clones which cannot be aligned; colors: black = aligned as in Fig. 1A; orange = clones not present in the UCSC browser; red = clones with different positions in both alignments, blue = clones presented in the UCSC browser but excluded in the revised assembly. The yellow vertical bar in DEF a illustrates the widening of the DEF cluster a as a result of the alternative alignment of [GenBank:AF200455] / [GenBank:AF238378] (see text). Clone (number) GenBank accession.version / library: (1) [GenBank:AC018398] / RP11, (2) [GenBank:AF287957] / CTD, (3) [GenBank:AF233439] / GS, CTD, (4) [GenBank:AF200455] / SCb, (5) [GenBank:AF238378] / SCb, (6) [GenBank:AF228730] / SCb, CTB, (7) [GenBank:AF215847] / CTB, (8) [GenBank:AC130339] / RP13, (9) [GenBank:AC130360] / RP11, (10) [GenBank:AC130367] / RP11, (11) [GenBank:AC134395] / RP11, (12) [GenBank:AC134683] / RP11, (13) [GenBank:AC285443] / SCb, (14) [GenBank:AC202031] / SCb, (15) [GenBank:AC134684] / RP11, (16) [GenBank:AC084121] / RP11, (17) [GenBank:AC144950] / RP11, (18) [GenBank:AC130365] / RP11, (19) [GenBank:AC131269] / RP11, (20) [GenBank:AC105233] / RP11, (21) [GenBank:AC068020] / RP11, (22) [GenBank:AC068353] / RP11, (23) [GenBank:AF298854] / SCb, (24) [GenBank:AF205406] / SCb, (25) [GenBank:AF314060] / GS, (26) [GenBank:AF314059] / SCb, (27) [GenBank:AF252831] / SCb, (28) [GenBank:AF189745] / SCb, (29) [GenBank:AF252830] / SCb, (30) [GenBank:AC148106] / RP11, (31) [GenBank:AC105214] / RP11, (32) [GenBank:AC092766] / RP11 Figure 2 Genes and pseudogenes in DEF clusters a and b . Names correspond to the Vertebrate Genome Annotation, intergenic distances are scaled 1:10. Defensin and defensin like genes and pseudogenes are written in black, novel defensin genes and pseudogenes are underlined, other genes / transcripts are indicated in gray. DEF cluster a : The presence of four copies of the DEFA1/DEFTP tandem and the DEFA3 gene in [GenBank:AF200455] requires the illustrated clone alignment, resulting in a "widening" of the hg16 assembly, pictured by the striped gray box (corresponding to the yellow bar in Fig.1). Analysis of the intergenic distances (data not shown) suggests, that [GenBank:AF238378] harbors copies 2 and 4 of the DEFA1/DEFTP tandem whereas copy 3 is missing (dotted line). Since both clones are derived from the same library (SCb), either copy 3 is lost during the cloning process or the clones represent different alleles. DEF cluster b : The DEF cluster b is illustrated in the orientation of DEF cluster b1 . Interestingly, the DEF cluster region was identified as the distal breakpoint (REPD) of a 4.7-Mb segment inversion, identified as a common polymorphism with frequencies of 39% and 26% in the Japanese population and in Europeans, respectively [ 16 - 18 ]. Although the inversion itself apparently do not have any pathological effects, in heterozygous female carriers unequal recombinations can occur, leading to three macrorearrangements – inv dup del (8p); +der(8)(8p23.1pter) and del (8)(p23.1p23.2) – related to severe disease phenotypes. The fact that low copy repeats (LCR) flanking the DEF clusters represent the essential sites for such recombination events is a strong argument to resolve the structure of the LCRs themselves as well as the genomic organization of the entire region. Also, genes of both DEF cluster a and b vary interindividually in their copy numbers. This was shown by somatic cell hybrid mapping for DEFA1 , DEFA3 ( DEF cluster a ; 2–3 copies each) [ 19 ] and by a combination of multiplex amplifiable probe hybridisation and semiquantitative fluorescence in situ hybridization for DEFB4 , DEFB103 , DEFB104 ( DEF cluster b ; 2–12 copies each) [ 20 ]. It is generally assumed that this variability crucially contributes to the differences in the innate immunity network between individuals and influences predisposition and susceptibility for diseases. The polymorphic nature of this locus suggested to us that the pool of clones presented in the hg16 assembly should be aligned in a different way. Our alternative assembly creates more DEF cluster copies and better reflects the individual variability of the locus. In addition, comparative sequence analysis of DEF genes in our closest relative, the chimpanzee ( Pan troglodytes ), both revealed the differences in the defensin protein panel of both species and showed that DEF clusters are also multiplied in the ape. Moreover, extraction of single nucleotide polymorphisms (SNPs) from overlapping regions of clones harboring DEF genes provided haplotypes which were analyzed for their ratio in individuals and used for the determination of individual gene copy numbers. Results Revision of the hg16 assembly In the framework of the International Human Genome Sequencing Consortium we sequenced 19 out of 32 BAC clones mapping to the 8p23.1 DEF region (for linking clone numbers with accession numbers see legend of Fig. 1 ; accession numbers and corresponding clone names are given in Additional file 1 ). In addition to the existing clone alignments of DEF clusters b1 and b2 , in our sequence assembly a consistent 360-kb-contig comprising five clones was built that also contains a DEF cluster b . Interestingly, this additional contig could neither be joined unambiguously to cluster b1 nor to cluster b2 . This is in contradiction to the hg16 assembly where clones 13 and 14 of the novel contig are positioned in DEF cluster b1 . This disagreement prompted us to evaluate carefully the alignment of the entire region. For this, we used all 32 finished clones which map in the region. To circumvent the problems of an automatic assembly in repeat rich, duplicated regions, the clones were manually joined according to the following criteria: only single base exchanges and insertions/deletions in repeat stretches were tolerated and joins were not performed if the total ratio of single base differences in the overlapping clone portions was >0.8%. The result is an alternative alignment shown in Fig. 1B , which differs from the hg16 assembly in three major points and is supported by a detailed repeat and SNP/ haplotype analysis: (1) Clones 23–27 not present in the hg16 assembly were located distally of the gap, whereas clones 11, 12 and 17 were excluded since they cannot be aligned according to our criteria. Clones 13 and 14 were moved to the new 360-kb-contig (see below). (2) In the framework of the human and vertebrate analysis and annotation initiative (HAVANA) [ 21 ], where gene structures are annotated on the basis of human interpretation of combined supportive evidence generated during sequence analysis, we manually annotated seven DEF gene containing clones, five of them located in DEF cluster a . We found that clone 4 contains four copies of a DEFA1 / DEFTP tandem and one copy of DEFA3 , whereas clone 5 harbors only two DEFA1 / DEFTP tandems and the DEFA3 gene (Fig. 2 ). Consequently, in our assembly, clone 4 is aligned to clone 5 in a way that results in four copies of the DEFA1 / DEFTP tandem instead of three copies in the hg16 assembly. Thus, this region will be "widened" by shifting all proximally adjacent clones for 23,270 bp. Analysis of the intergenic distances between the DEF genes suggests that clone 5 harbors copies 2 and 4 of the four DEFA1 / DEFTP tandems. Furthermore, hitherto undiscovered, additional DEF genes and pseudogenes could be annotated: clones 3–5 harbor the DEFA7 gene [GenBank:A98570], [GenBank:A98571] coding for a novel protein, similar to DEFA4 , as well as pseudogenes DEFAP1 , DEFAP2 and DEFAP3 similar to DEFA (Fig. 2 ). Our annotation of the entire DEF cluster region is submitted to the Vertebrate Genome Annotation database [ 22 ]. (3) Our assembly of the 32 finished clones mapping to the locus created an additional 360-kb-contig which consists of five clones representing an additional, third DEF cluster copy named b3. Since the repeat in clone 30 is located on the (-)-strand, the contig cannot be joined to any of the repeats at either side of the gap but instead has to be located within the gap in unknown orientation. Low copy repeat analysis The DEF locus contains several genomewide low copy repeats (LCR) (Fig. 3 ). The repeats form clusters of up to 165 kb length and paralogs exist on other chromosomes as well as on chromosome 8 at about 12 Mb. In silico identification of these paralogs was in good accordance with fluorescence in situ hybridization (FISH) experiments using clones from the 8p23.1 DEF locus as probes ( Additional files 3 and 4 ). Five types of LCRs can be distinguished: Figure 3 DEF gene cluster flanking repeat blocks in 8p23.1 and their paralogs on other genomic loci. Repeats and repeat clusters are drawn as striped bars as in Fig. 1. Inverted repeat pairs of the paralogs are highlighted in gray. Type I The 89.6 kb cluster consists of 4.7 subunits of a 19.2 kb repeat. Each complete subunit contains DEFA1 and DEFTP . The incomplete fifth copy harbors DEFA3 . The content of interspersed repeats is low (20%); the nucleotide identity between the subunits is 99%. The repeat subunits of cluster I are unique to this locus and do not show significant homologies to other regions of the human genome. Type II Repeats of type II flank DEF clusters b , do not show substructures and do not contain genes. In contrast to repeat units of type I they differ in length and orientation, contain inversions and show a high degree of interspersed repeats (55%). The considerable length differences are caused by an initial partial duplication and/or subsequent deletions. The nucleotide identity of the repeats is 96–99% except for II.1, located adjacent to DEF cluster a , which exhibits an identity of ~92% to repeats II.2-6. Blat search of the longest repeat (II.2, 82 kb) revealed that repeats of type II are also present at about 12 Mb on chromosome 8p23.1 as well as on other human chromosomes, e.g. 3, 4, 7, 11, 12 and 16. The paralogs vary in length (25–133 kb) and show nucleotide identities of 93–95% compared to II.2 on 8p23.1. On chromosomes 3, 4 and 11, they are arranged as inverted repeats. The distances in between "repeat pairs" vary remarkably: Whereas in 11q13.2-11q13.4 and 4p16.2-16.1 the parts are separated by 4 and 5 Mb, respectively, the repeats on chromosome 3 are located in 3p12.3 and 3q13.31, framing 51 Mb on both arms of the chromosome. Type III Located adjacent to DEF clusters b , they differ in length and orientation and consist of various numbers of 7.7 kb-subunits showing a low content of interspersed repeats (17%). The nucleotide identity of two subunits of a single type III cluster is >99%, whereas the identity of two subunits of different clusters is in the range of 96–98%. Each of the 7.7 kb-subunits harbors one copy of a gene for the hypothetical protein FLJ10408 [GenBank:NM_018088]. The coding sequences of the gene copies differ and in some cases the reading frame contains premature termination codons. Blat search of the type III repeat subunit revealed three paralogs in the human genome. Two of these are located on chromosome 8p23.1 at about 12 Mb (flanked by repeats of type II), the third at 12p13.31. The paralogs are slightly rearranged in comparison to the subunits of III.1-5. Type IV Repeats of this type partially cover the DEF clusters b (containing DEFB109P and DEFB108 , Fig. 2 ) and do not show substructures. The content of interspersed repeats is about 30%; the nucleotide identity between repeats IV.1-3 is >98%. Paralogs of these repeats exist on chromosome 8 at about 12 Mb and on chromosome 12 with identities of about 96% to IV1-3. Type V These repeats are unique to the 8p23.1 DEF locus, show no substructures and contain 35–38% interspersed repeats as well as the major part of the DEF cluster b genes (all genes downstream of DEFB108 , Fig. 2 ). The nucleotide identities between V.1, V.2 and V.3 are >98%. SNPs and haplotypes We manually inspected seven clones covering DEF cluster a and 16 clones covering DEF clusters b1 , b2 and b3 for SNPs in exons and introns of all DEF genes. In total we found 270 overlap SNPs: 25 are located in coding sequences, comprising 16 nonsynonymous and nine synonymous changes. 36 SNPs were identified in untranslated regions, and 209 are located in introns. With respect to the coding SNPs in DEF clusters b and regarding 11 of 16 clones, six distinct haplotypes H1-6 can be defined (Table 1 ). One clone each supports haplotypes H3, H4 and H6, whereas haplotypes H1, H2 and H5 are found in either two or three clones. The remaining five clones harbor only parts of DEF genes rendering the unambiguous identification of coding SNP based haplotypes impossible. Examination of all SNPs leading to amino acid (aa) changes in defensins indicates that diversity in the peptides is not restricted to residues outside and in between the cystein motif, but also occurs in the vicinity of cysteins, or even a cystein itself is changed (rs1800968 in DEFB1 ; C67S; data not shown). Table 1 SNPs and haplotypes H1-H6 extracted from DEF cluster b covering clones Gene Pos. mRNA 1 Haplotypes 2 Change H1 H2 H3 H4 H5 H6 DEFB107 13 G / / T G T F5V DEFB105 107 C / C T C C P36L DEFB106 125 T / C T T T Silent DEFB104 42 A A A A A G V10I DEFB4 78 T C C C C C Silent 120 T T T T C T Silent 275 C C C C T C 3'UTR 335 C C C C G C 3'UTR DEFB108 138 A A A A T / Silent 111 C C C C T / Silent 97 G G A A G / G33S DEFB109p 133 A A A / G / V45I 132 C C A / A / K44N 119 C C G / G / R40T 104 T T C / C / S35F 41 C C / / A / S14ochre 1 mRNA positions are referred to the following human mRNAs: DEFB107 = [GenBank:AY122467]; DEFB105 = [GenBank:NM_152250]; DEFB106 = [GenBank:NM_152251]; DEFB104 = [GenBank:NM_080389], DEFB4 = [GenBank:NM_004942]; DEFB108 = [GenBank:AF540980]; DEFB109p = [GenBank:AF540981]. 2 Haplotypes are derived from the following clones (libraries): H1 = 9, 10 (both RP11); H2 = 8 (RP13), 18, 19 (both RP11); H3 = 11 (RP11); H4 = 12 (RP11); H5 = 13, 14, 29 (all SCb); H6 = 28 (SCb). Chimpanzee defensin loci In order to compare the human chromosome 8p23.1 DEF region to the orthologous locus in our closest relative, we both employed the chimpanzee ( Pan troglodytes , ptr) whole genome shotgun (WGS) working draft (WD, [ 23 ]) and high quality chimpanzee BAC sequences. Close inspection of the chimpanzee WD scaffold 32935 (chain ID 462) revealed all orthologs of the genes in human DEF cluster a except DEFA3 . In contrast to the human organization, ptrDEFA1 and ptrDEFTP were found as single copies. In order to check whether ptrDEFA1 and ptrDEFTP are also multiplied in the ape, but misassembled into a single locus, we inspected the NCBI trace archive [ 24 ] for chimpanzee WGS sequences covering the ptrDEFA1 locus. For a region of about 500 bp spanning exon 1 of ptrDEFA1 , there are shotgun reads representing six different haplotypes. Since the sequences derived from one chimpanzee this is a clear indication that ptrDEFA1 is also multiplied in the ape. No evidence was found for the presence of ptrDEFA3 . Concerning DEF cluster b we encountered the same problem: the cluster is represented only once in the chimpanzee WD (chain ID 900), but trace data inspection indicates the presence of several different haplotypes. Additionally, ptrDEFB108 and ptrDEFB109p are not covered by any chimpanzee WD sequences. As an alternative to the WD approach, we sequenced for ptrDEF cluster b three BAC clones. Examination of SNPs in overlapping regions (104 kb) of the three clones [GenBank:AC150655], [GenBank:AC150656], [GenBank:AC150657] revealed three different haplotypes originating from one chimpanzee. The detected aa changes in human and chimpanzee defensins are illustrated in Additional file 5 . Interestingly, in the ptrDEFA5 protein, one of the disulfide bridging cysteins is changed into serine (C54S). In ptrDEFB108 the canonical cystein motif is truncated (R53opal) which suggests that ptrDEFB108 is a pseudogene, since also the start codon ATG is changed into GTG. We also used the ptrDEF sequences for the detection of ancestral alleles of all 19 nonsynonymous and synonymous human DEF coding SNPs ( Additional file 2 ). Individual DEF haplotypes and copy numbers In order to determine individual DEF copy numbers we PCR-amplified a 500 bp fragment of DEFB104 which contains 4 SNPs, and a 511 bp region of DEFB4 containing 5 SNPs in four individuals (Table 2 ). Three of the DEFB4 SNPs (Table 2 , SNPs 5,6 and 7) were previously described in a haplotype study in different ethnic populations [ 25 ]. The PCR products were cloned, individual clones were sequenced and respective haplotypes were determined according to the base composition at polymorphic positions. Each individual tested bears between three and four different haplotypes of DEFB104 and two to four haplotypes of DEFB4 . In different individuals the ratios of the single haplotypes vary remarkably. For instance, at DEFB104 the haplotype GAGC is found in all four individuals but compared to all other haplotypes at ratios of 2:3 (proband 2) to 1:7 (proband 1). Interestingly, this haplotype is also found in the trace archives of chimpanzee and baboon. Furthermore, ratios of the individual haplotypes of DEFB104 as well as of DEFB4 indicate different DEF cluster b numbers in the four individuals. While proband 3 bears five copies, proband 4 most probably harbors eight copies or multiples thereof. Table 2 Haplotype based estimation of gene and cluster copy numbers Haplotypes SNP 1 Proband DEFB104 1 2 3 4 1 2 3 4 1 C A A T 50 19 - 5 2 G A A T - - 16 - 3 G A G C 15 18 32 9 4 G G G C 65 10 16 58 5 G A A C - - 14 - Ratios of single haplotypes 3:1:4 2:2:1 1:2:1:1 1:1:6 Minimal gene copy number 8 5 5 8 DEFB4 5 6 7 8 9 1 2 3 4 1 C C C G A 14 - 8 - 2 C T G G A - 17 9 5 3 T C C A G 8 - 8 - 4 T C C G G 24 68 11 30 Ratios of single haplotypes 2:1:3 1:4 1:1:1:1 1:6 Minimal Gene copy number 6 5 4 7 Minimal DEF cluster b copy number 8 5 5 8 1 SNP1: ss28489415, 2: rs2680507 = rs11774031, 3: ss28489416, 4: rs4259430, 5: rs2740090, 6: rs2740091, 7: rs2737912, 8: rs2737913, 9: rs2737531. Discussion The manual clone-by-clone alignment and gene annotation as well as detailed repeat and SNP/haplotype analyses significantly improved the assembly of the human DEF 8p23.1 locus. Eventhough the revised alignment (Fig. 1B ) does not represent a gap-free version of the locus and in fact introduces a second de facto gap, it better reflects the region in the sense of a „human genome reference“, since the clones harboring copies of DEF clusters b derive from three libraries (RP11, RP13, SCb) and may therefore represent up to five alleles (library RP13 is represented by only one clone). Our assembly also reflects better the diversity of all available sequence data of this chromosomal region: 27 out of 32 finished clones are incorporated into the tiling path. The remaining five clones cannot be included in the assembly according to our quality criteria and therefore must be regarded as parts of additional copies or alleles. Furthermore we point out that the identification of a third copy of the DEF cluster b in the 360-kb-contig does not represent an allele of clusters b1 or b2 derived from an alternative library / donor, since besides four SCb clones one RP11 clone is incorporated. With respect to the RP11 library of which most of the DEF cluster b covering clones derive we conclude that at present sequence information of at least five variants of the cluster is available from a single individual: cluster b1 (clones 9, 10, 15); b2 (clones 16 – 20); b3 (clone 30); b4 (clone 11) and b5 (clone 12). All these results are in agreement with the reported interindividual variability of DEF cluster b genes [ 20 ]. Alignment and analysis of the intergenic distances of clones 4 and 5 ( DEF cluster a ; Fig. 2 ) show that clone 5 harbors copies 2 and 4 of the four DEFA1 / DEFTP tandems present in clone 4. Since both clones derive from the same library (SCb), we conclude that either copy 3 of the tandem was lost during the cloning process of clone 5 or the clones represent two different alleles of the same chromosomal locus. This perfectly agrees with the variation in the copy number of DEFA1 reported by Mars [ 19 ]. Moreover, the identification of new DEF genes and pseudogenes demonstrates the advantages of a curated manual annotation over automatic approaches. The LCR analysis (Fig. 3 ) allows to draw conclusions about the role of these repeats in chromosomal rearrangement processes: The difference in nucleotide identities between LCR II.1 at one hand and II.2-6 on the other hand indicates that repeats II.2-6 might be involved in rearrangement events of DEF clusters b , whereas repeat II.1, separating DEF cluster a from the clusters b , has evolved independently from its paralogs. Aditionally, for repeats on chromosome Y, a similar genomic structure as for inverted LCR type II is described in the literature: a 300-kb inverted repeat flanks a 3.5 Mb region that occurs in opposite orientations in different individuals [ 2 ]. This supports the assumption that also type II repeats may be involved in homologous recombination events resulting in chromosomal macrorearrangements including inversions, even pericentromeric ones. In particular, this may hold true for the polymorphic 4.7-Mb inv dup del (8p) reported in the literature [ 16 - 18 ]: LCRs of type II are located on chromosome 8p both at 6.9–8.0 Mb and 11.9–12.6 Mb and therefore separated from each other in a range of 3.9–5.7 Mb. Thus, they can be supposed to be inversion breakpoints in REPP and REPD. The function of the protein encoded by FLJ10408 located in LCR type III is unknown, but the genomic arrangement facilitates proteome plasticity by multiple copies of the same gene. SNP detection and correct assignment to regions with segmental duplications is not trivial and hampered by paralogous sequence variations [ 26 ], duplicon SNPs and multisite variations [ 6 ]. Moreover, there is considerable evidence that gene conversion [ 27 , 28 ] promotes allele plasticity in duplicated regions. This is illustrated by the fact that in the UCSC browser ~2800 SNPs from dbSNP [ 29 ] are assigned to DEF cluster a (224 kb) and the two DEF clusters b (196 kb each) resulting in a SNP density of 1 SNP per 220 bp. Close examination indicates that the SNPs are arbitrarily allocated to the two DEF cluster b loci present in the hg16 assembly. Therefore, in such regions, only manual clone-by-clone inspection as performed during our assembly process provides a reliable set of SNPs for the determination of haplotypes. Human SNPs such as rs1800968 in DEFB1 , affecting cysteins (C67S) might be of functional relevance, since the cystein connectivity is assumed to determine the correct fold of the defensins which is essential to elicit chemotactic responses as shown for DEFB103 [ 30 ]. In order to answer the question whether the extraordinary complexity of the DEF locus is human specific, we closely inspected the orthologous region of the chimpanzee WD. In contrast to the human organization, ptrDEFA1 and ptrDEF cluster b were found as single copies. Familiar with the drawbacks of the WGS automatic assembly [ 31 , 32 ], we suspected these loci are also multiplied in the ape, but assembled wrongly into a single locus due to the high nucleotide identity. In accordance with this assumption, we identified WGS reads representing more than two haplotypes in one chimpanzee. In order to overcome the WD problems we sequenced BAC clones containing ptrDEF cluster b according to a high quality standard and conclude that it is at least duplicated. This suggests that the DEF locus of the chimpanzee is probably as complex as in humans. The exceptional genomic complexity and heterogeneity of the human 8p23.1 DEF locus and the prominent position of defensins in the innate immunity framework raise the question whether individual patterns of haplotypes together with their variable copy number affect the functionality of the defensin system. A similar situation is found for a chemokine gene cluster where an individually variable gene copy number of CCL3-L1 regulates the gene's expression and is supposed to affect the susceptibility to and progression of inflammatory diseases [ 33 ]. Systematic typing of physically linked SNPs should allow to detect interindividual differences in haplotypes and locus copy numbers. Sequencing provides a robust method for the determination of haplotypes and their frequencies scalable to large numbers as required for association studies. As outlined above, SNP genotyping in duplicated regions is demanding and in addition to very careful initial data mining and laboratory practice requires methods allowing the quantitative assessment of allele ratios like dynamic allele-specific hybridization [ 34 ] and pyrosequencing [ 35 ] as well as of copy-number-variation like multiplex ligation-dependent probe amplification [ 36 , 37 ] and representational oligonucleotide microarray analysis [ 38 , 39 ]. In order to differentiate between valid SNPs, duplicon SNPs, paralogous sequence variations and multisite variations, complete hydatidiform moles or haploid genomes have to be included in upstream assay validation [ 6 ]. Nevertheless, for highly complex and polymorphic regions the significance of single SNP based assays may be insufficient. As shown in our approach, systematic typing of linked SNPs can overcome this limitation. An estimation of haplotype ratios provides information about the copy number, however, it depends on the number of individual clones analyzed. Our results confirm that the easy-to-handle „classical sequencing approach“ is a valuable tool for the determination of DEF gene variants, DEF haplotypes and DEF cluster copy numbers. More detailed analyses will give a catalog of haplotype combinations associated with different phenotypes and diseases. Finally, the presented work provides a set of SNPs and haplotypes suitable for future studies of interindividual DEF locus variability and its disease association. Conclusions Complexity and variability seem to be essential genomic features of the major human DEF locus and of – yet unknown – functional significance for the innate immunity framework. This is supported by our human-chimpanzee genomic comparison. In conclusion of the presented repeat analyses we propose a model of the repeat and DEF cluster organization ( Additional file 6 ) that is consistent with the available sequence information and explains the observed extensive variability of the locus. Since all proposed structural elements of the highly complex locus are available at least once as finished sequence, no fosmid-end mapping problems have been observed (International Human Genome Sequencing Consortium, unpublished results). This is a strong indication that despite there are at least two de facto gaps no essential elements of the DEF locus are missing in the human reference sequence. In comparison to its actual representation, our revised clone alignment clearly better represents the 8p23.1 complexity and improves the human reference sequence as an invaluable resource for the investigation of individual genetic variations. Finally, the presented work provides a set of SNPs and haplotypes as well as a robust sequencing based method suitable for future studies of interindividual DEF locus variability and its disease association. Methods Alignment revision Clone alignments were performed using the GAP4 assembly program, version 6 [ 40 ], using the sequences of the GenBank versions listed in the legend of Fig. 1 and Additional file 1 . Additionally, for all clones sequenced at the IMB Jena the original GAP4 projects including the trace data were used. The clones were joined allowing only single base exchanges and insertions/deletions in repeat stretches. Joins were not performed if the total ratio of single base differences in the overlapping clone portions exceeded 0.8%. Repeat analysis LCRs were identified by application of the Miropeats program [ 41 ] to the revised alignment. The repeat sequences were extracted from the revised alignment of joined clones following the positions of the Miropeat's output and checked for their nucleotide identity using sim2 [ 42 ] and Blast 2.0 [ 43 , 44 ]. Interspersed repeats in the repeat blocks were identified by RepeatMasker (Version: 20040306-web; [ 45 ]. Paralogs of the repeat blocks were identified by Blat [ 46 ] to the July 2003 UCSC version h16 [ 10 ]. The DNA sequences in between the Blat match limits were fetched from the browser and also analyzed for their similarity to the DEF cluster repeats by Blast and sim. Chimpanzee BAC clones The clones were identified by Blast of the revised alignment consensus to the chimpanzee BAC end sequence database [ 47 ]. Subcloning was performed into pUC18 followed by sequencing using dye terminator chemistry and ABI 3730/3700 technology. Base calling and assembly were performed by Phred/Phrap and GAP4 was used for editing and finishing in accordance to the Human Genome Project standards [ 48 ]. Sequence annotation The gene annotation was performed by using the automated sequence annotation system RUMMAGE [ 49 ] and ANA_NOTES, SPANDIT and LACE as a local client of the HAVANA pipeline at the Sanger Institute (Hinxton, UK; [ 21 ]. Detailed descriptions of the analysis tools are given by [ 50 ] and [ 51 ]. Haplotypes and copy numbers Genomic DNA was extracted from the blood of four male volunteers using the QIAamp DNA Blood Kit (Qiagen). PCR was performed in a total volume of 25 μl using ReadyToGo PCR beads (Amersham) with 5 pmoles of each primer and 100 ng of DNA. Cycling conditions were 94°C for 30 sec followed by 35 cycles with 94°C for 20 sec, 58°C for 30 sec and 72°C for 60 sec, plus a final 72°C extension for 10 min. Oligos used were for DEFB4 GGCGATACTGACACAGGGTT (sense) and ATGGGGAAGGTCAAGGAATC (antisense) and for DEFB104 TTCTGTAGCCCCAACACCTC (sense) and GGTGCCAAGGACATCTAGGA (antisense), respectively. PCR products were cloned into PCRTopo2.1 (Invitrogen) and individual clones were sequenced as described above. List of abbreviations BAC Bacterial artificial chromosome DEF Defensin FISH Fluorescence in situ hybridization GAP Genome assembly program HAVANA Human and vertebrate analysis and annotation initiative LCR Low copy repeat NCBI National Center for Biotechnology Information PCR Polymerase chain reaction ptr Pan troglodytes (chimpanzee) SNP Single nucleotide polymorphism UCSC University of California Santa Cruz WD Working draft WGS Whole genome shotgun Author's contributions ST performed the assembly revision, the gene annotations, the LCR, SNP and chimpanzee sequence analyses and drafted the manuscript. PG performed mapping and repeat analyses in the 8p23.1 region, completed by IFL's FISH experiments. KR contributed to the sequencing process of clones in the region. AS, SA, NS, KS and MS accounted for the assembly revision and the LCR analyses. AF and RS supported the gene annotations and the DEF structure and function discussions. KH and KR were in charge for the individual haplotype and copy number determinations. MP conceived, designed and coordinated the project. Supplementary Material Additional File 1 Accession numbers, libraries and clone names of all clones shown in main text, Fig. 1 . Click here for file Additional File 3 LCR of clones SCb-561b17 [GenBank:AF238378]; green/yellow signals) and CTB-415D8 [GenBank:AF228730]; red signals) visualized by FISH on metaphase chromosomes according to standard protocols [ 52 , 53 ]. Metaphase spread after DAPI counter stain (A) and color inversion (B). Targeted chromosomes are numbered in B. Any FISH signal is shown quadruplicated within the metaphase spread on four chromatides from two homologue chromosomes. Probe CTB-415D8 generated strong FISH signals with declining intensity at 8p23, 4p16, 11q13.3 and 3q21, corresponding to the in silico identified LCR type II paralogs (see Fig. 3 , main text). Additional weaker signals at 3p12-13, 7q21, 11p15, 12p13 and 16p13.3 are also an indication for LCR type II and III paralogs. In contrast, the single locus of SCb-561b17 at 8p23 highlighted by open triangles corresponds to the unique LCR type I. Double signals, resulting from two close located targets at 4p16, 8p23 and 3q21 are marked by asterisks (compare also Additional_file_4). Click here for file Additional File 4 Resolving LCR type II and III "pairs" on chromosomes at approx. 900 band stage. Probe CTB-415D8 [GenBank:AF228730]; LCR type II and III) generates two clearly separated FISH signals at 4p16 and 8p23, respectively (red signals marked by double arrows): In contrast, probe SCb-561b17 [GenBank:AF238378]; LCR type I) yield a single signal at 8p23, solely (green/yellow signal, open triangle), that is co-localized with the telomeric signal of probe CTB-415D8 [GenBank:AF228730]. Signals with lower intensity are indeterminable in this picture. Click here for file Additional File 5 DEF aa sequences with highlighted residues (bold) different between human and chimpanzee. Boxes: human aa – human position – chimpanzee aa. All aa positions refer to the following human protein accessions: DEFA6 = [GenBank:NP_001917]; DEFA4 = [GenBank:NP_001916]; DEFA1 = [GenBank:NP_004075]; DEFA5 = [GenBank:NP_066290]; DEFB1 = [GenBank:NP_005209]; DEFB107 = [GenBank:AAM93909]; DEFB105 = [GenBank:NP_689463]; DEFB103 = [GenBank:NP_061131]; DEFB4 = [GenBank:NP_004933]; DEFB108 = [GenBank:AAN33116]. Aa for the chimpanzee orthologs ptrDEFA6, ptrDEFA4, ptrDEFA1, ptr novel defensin similar to DEFA4, ptrDEFA5 and ptrDEFB1 ( ptrDEF cluster a ) are deduced from the chimpanzee WD and might therefore include sequencing errors. Those for ptrDEFB1, ptrDEFB107, ptrDEFB105, ptrDEFB103, ptrDEFB4 and ptrDEFB108 ( DEF cluster b ) are derived from high quality BAC sequences and the appropriate traces were visually inspected. The gray shadow indicates the motif of six cystein residues (except for DEFB107 with only five cysteins). Click here for file Additional File 2 Synonymous and non synonymous changes by SNPs in human DEF genes and their ancestral alleles by comparison to chimpanzee sequences. Click here for file Additional File 6 Predicted genomic organization of the human 8p23.1 DEF locus. For simplicity only the DEF clusters (arrows) as well as LCRs type II (rectangles) are shown. Black : high quality sequence available; Gray : hypothetical structures, no finished sequence available. In addition to a 'minimal' DEF locus consisting of one a and two b clusters (middle), individual loci may have incorporated variable numbers ( F , R ) of additional b clusters in either orientation. The proposed duplicon consists of two inverted LCRs flanking a DEF cluster b (top/bottom). The orientation of any DEF cluster b can change either by inverted duplication/crossover ( i ) or homologous recombination within inverted LCRs ( x , right). Moreover, the proposed genomic structure indicates that even in a 'minimal' DEF locus one or both DEF clusters may be deleted due to homologous recombination between direct LCR copies (Δ). Sequence features of the most distal LCR (II.1; see text and Fig. 3 ) suggest, that it may be less often involved in recombination or gene conversion events. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544879.xml

549583

The Courage to Change the Rules: A Proposal for an Essential Health R&D Treaty

The medical needs of many of the world's population go unmet. A new treaty on essential health R&D could provide a binding framework to redirect today's scientific expertise to priority needs

Biomedical science and technology are developing at a more rapid pace than ever. Investments in health research and development (R&D) have never been higher—global spending on health research increased from US$30 billion in 1990 to US$105.9 billion in 2001. But despite advances in technology and unparalleled research spending, the medical needs of many of the world's population go unmet. For example, only 1% of new drugs approved between 1975 and 1999 were specifically developed for tropical diseases and tuberculosis—diseases that account for over 10% of the global disease burden ( Figure 1 ) [ 1 ]. Figure 1 Clinical Officer Preparing Sodium Stibogluconate Solution Injection for a Patient with Visceral Leishmaniasis Sodium stibogluconate solution is administered by intramuscular injection for 30 days. The injection is painful and can cause toxic reactions. Developed in 1934, resistance of up to 65% has been documented in India. Around 50,000 people die from visceral leishmaniasis each year. New, effective drugs and diagnostics are urgently needed. (Photograph: Copyright Espen Rasmussen/MSF, Somalia, 2004) In recent years, some important steps have been taken to improve access to existing treatments in the developing world by increasing generic competition. Yet there continues to be a tension between promoting access to lifesaving medicines as a human right and maintaining a global trade regime that seeks to finance health R&D by allowing monopolies to charge high prices [ 2 ]. There is a growing demand from many quarters for a new international policy framework [ 3 ]. A new international treaty on essential health R&D could provide a binding framework to redirect today's knowledge and scientific expertise to priority health needs. The treaty could help to cement new political commitments and coordinate complementary partnerships aimed at generating and rewarding health innovation as a global public good. The Patent System: An Unhealthy Motive for Medical Innovation? Until recently, providing patent protection for pharmaceuticals was a choice made by individual governments according to their level of industrial development. Today, pharmaceutical patents are globalized through the World Trade Organization's Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS Agreement) [ 4 ], and then further reinforced through bilateral and regional arrangements (the so-called TRIPS-Plus agreements [ 5 ]). But the patent system stimulates innovation only where industry sees the opportunity for increasing sales and market share; much of the resulting “innovation” is in fact imitation, producing “me-too” drugs that offer little, if anything, in the way of therapeutic benefit over existing drugs ( Box 1 ). Box 1. How Innovative Is the Profit Motive? A study published in the Lancet in 2002 showed that 68% of all new chemical entities marketed worldwide in the last 25 years were me-too products, representing little or no therapeutic gain [ 1 ]. According to the United Nations Development Programme, less than 5% of drugs introduced by the top 25 pharmaceutical companies in the US represented true therapeutic advances; of these, 70% were developed with government involvement ([ 20 ], p. 69). Studies of drug development over the last decade in the US [ 21 ] and the last two decades in France [ 23 ] show that around two-thirds of medicines are me-too products. 92% of medicines approved in 2002 by the US Federal Drug Administration were me-too drugs [ 23 ]. The poorest are hardest hit. While R&D of new therapies against tropical diseases has ground to a standstill, 14 million people die from infectious diseases each year, predominantly in developing countries [ 1 ]. Most of the world's 40 million people with HIV/AIDS, including 2.2 million children under 15, live in the developing world ( www.unaids.org ). The poor also dominate non-communicable disease tables, accounting for 59% of the 56.5 million annual global deaths [ 6 ]. The patent system is also promoting new inequalities in high-income countries. Americans now spend a staggering $200 billion a year on prescription drugs. This figure is growing at a rate of about 12% per year [ 7 ]. The average price of the fifty drugs most used by senior citizens in America was nearly $1,500 for a year's supply in 2002. Prescription drugs have become inaccessible even to many people in the rich world. Patents, with their focus on maximizing profits, have at least three negative consequences. First, it has been argued that the patent system causes substantial welfare losses because consumers who would buy the product if it were priced at somewhere nearer production cost do not buy it at the monopoly price [ 8 ]. Second, the system encourages counterfeiting—counterfeit drugs may represent up to 10% of the global market for pharmaceuticals [ 9 ]. Third, patented drugs are promoted through excessive marketing—on average, twice as much is spent on marketing a drug as on its R&D [ 10 ]. Across industries, it is becoming increasingly apparent that the patent system isn't working well [ 11 ], leading some in industry to express public concern that the blockbuster business model is “irreparably broken” [ 12 ]. A new approach is needed, for all our sakes. Prescriptions for an Innovative Approach: A New Treaty for Essential Health R&D The only major international policy instrument that exists today to stimulate and finance health R&D is the TRIPS Agreement [ 4 ]. The TRIPS Agreement provides 20 years of patent protection on pharmaceuticals in the hope of stimulating the development of new medicines. Beyond that, governments try to stimulate R&D in neglected areas by providing industry with incentives such as tax breaks and patent extensions. However, the effectiveness of these policies is hardly known. In 2001, the Doha Declaration on TRIPS and Public Health affirmed the sovereign right of WTO members to take measures to protect public health by overcoming patents whenever needed [ 4 ]. The last few years have seen increased attention to the fact that patents keep drug costs high and limit access to medicines. However, there has been no movement in international policy to address the crisis in pharmaceutical innovation. Health R&D must be treated as an international problem that requires an international solution. It should be treated like other strategic sectors, as happens today for defense and space discovery—sectors that both benefit from very strong government support for innovation. When global public goods do correspond to national needs, governments should step in to mobilize and enforce the collective action required. For example, global cooperation in the sharing of infectious disease monitoring from 1890 onwards set a valuable precedent [ 13 ]. The recent epidemic of severe acute respiratory syndrome—SARS—clearly shows that biomedical knowledge and the pharmaceutical sciences can be mobilized to achieve rapid advances relevant to social needs if sufficient resources and political will can be mustered. The SARS virus was completely sequenced in just six days, and a diagnostic test was developed in only three months. The public-sector funded, collaborative “public-goods model” used for the Human Genome Project shows that public collaborative research can be more efficient than the closed, monopolistic, private sector approach. An international treaty ( Box 2 ) would promote a health-needs-driven approach to drug discovery. The elaboration of such a treaty would have to meet the two crucial requirements for an effective system of funding innovation in pharmaceuticals. First, the reward for innovation should be proportional to the social (that is, therapeutic) value. Second, prices should be near average production cost. Box 2. Key Concepts of an R&D Treaty A global, needs-driven R&D agenda: allowing policy makers, funding agencies, and the research community to set priorities for developing safe, effective, and affordable medicines according to health needs. Prioritization for neglected diseases: to ensure that immediate efforts are made toward finding new tools for lethal diseases that are currently difficult or impossible to diagnose and treat. Adequate international financing of health R&D: a new funding mechanism is urgently needed to support R&D on an ongoing basis, particularly for neglected diseases. All governments will need to participate according to their means. Equitable pricing: governments should ensure that the poor also have access to innovations resulting from government-funded or university research. Open access: governments should require access to the compounds and tools that result from public research in order to stimulate follow-on innovation elsewhere. International exchange: strengthening openness and transfer of technologies on a global basis will greatly help developing countries by improving access to information and ideas and accelerating the development of science and technology. The idea is to shift the discourse from trade to health. The treaty—focussed directly on R&D rather than patent rights or drug prices—would address the global management of publicly funded health R&D. Priorities for R&D would be defined through public-sector leadership and based on public health needs. R&D opportunities would be aimed at new lead compounds, new types of health tools ( Figure 2 ) and new treatment approaches. As the only legally mandated international government agency responsible for global health, the World Health Organization should work toward establishing this essential R&D agenda. Individual states would need to periodically evaluate targets for priority research and make adequate recommendations toward needs-driven R&D. Figure 2 Detection of African Sleeping Sickness by Lumbar Puncture Lumbar puncture in patients with African sleeping sickness can be a painful and potentially dangerous maneuver that is the only way to determine if the disease has progressed to the second stage. New diagnostic tools are urgently needed, as are new treatments; current medicines are old, toxic, difficult to use, and their production is not guaranteed. Around 60,000 people die from Afican sleeping sickness every year. (Photograph: copyright Serge Sibert/MSF, Uganda, 1998.) How Will the Treaty Work? One of the main objectives of the treaty would be to encourage the broad dissemination of information and knowledge-sharing, and to support diversity, competition, and collaboration among researchers from developed and developing countries. There are already precedents for the free, public sharing of innovations with the aim of developing new drugs. The Tropical Diseases Initiative ( www.tropicaldisease.org ), for example, is a new, Internet-based, community-wide effort to develop new drugs for tropical diseases [ 14 ]. The BioBricks project ( http://parts.mit.edu/ ) at the Massachusetts Institute of Technology is exploring standardized tools and processes for DNA work, largely by computer. The Bios Initiative (Biological Innovation for Open Society), launched by the Australian non-profit organization Cambia (the Center for the Application of Molecular Biology to International Agriculture; www.cambia.org ), is an effort to develop new innovation systems for market failures and for neglected priorities [ 15 ]. Among other incentives, technology exchange frameworks could include licensing agreements with developing countries, or affirmative commitments of research funds for collaborative projects with these countries. Such collaboration is currently being implemented in Europe, for example, through the European and Developing Countries Clinical Trial Partnership ( www.edctp.org ) [ 16 ]. The partnership is a new funding body established to fund research in developing countries, particularly in Africa, which contributes to the development of affordable prophylactics and drugs for HIV/AIDS, tuberculosis, and malaria. The treaty on health R&D should also promote partnerships between countries in the developing world and encourage the creation of regional technology networks in developing countries. Much of today's drug development know-how exists within the private sector. Further work is needed to define obligations and incentives in the treaty that maximize industry contributions to publicly funded R&D by providing in-kind contributions in areas where industry has the skills that public groups need. The treaty should also provide an expanded use of government rights against patent abuse on drugs developed with public support. This would include the right of a government to intervene if an invention is not made available to the public on reasonable terms, such as is included in the march-in rights clause of the United States's 1980 Bayh-Dole Act (which enabled public universities to license inventions for commercial development [ 17 ]). Making the Treaty Happen There are a number of obvious difficulties in moving the treaty forward, and these should not be underestimated. A delicate issue is the treaty's relation with other binding agreements, particularly the TRIPS Agreement. Governments that join the treaty should be granted patent exceptions and should not be accused of “free-riding,” since they would be contributing to R&D through a different juridical avenue. Substantial government resources would need to be mobilized to finance the highest priority medical research. All governments should participate according to their means. Countries already contribute significantly to global R&D through the purchase of costly patented drugs. Among other measures, not-for-profit initiatives working to develop new drugs, vaccines, and diagnostic tools for neglected diseases should be funded at levels that enable them to reach their objectives. Recent examples clearly show that when political will is mobilized, resources are rapidly made available to generate R&D in a particular area. In 2001, the anthrax scare in the US led to increases in biodefense research spending at the US National Institutes of Health from US$53 million in 2001 to US$1.6 billion in 2004. A treaty on health R&D is certainly a feasible proposal—the successful adoption of a treaty on plant genetic resources shows that it can be done. After seven years of negotiations, the Food and Agriculture Organization of the United Nations adopted the International Treaty on Plant Genetic Resources for Food and Agriculture in November 2001 [ 18 ]. This legally binding treaty covers all plant genetic resources relevant for food and agriculture. Through the treaty, countries agree to establish an efficient, effective, and transparent multilateral system to facilitate access to plant genetic resources for food and agriculture, and to share the benefits in a fair and equitable way. While there has been a growing consensus in development circles that more international public goods need to be supplied as part of the development strategy, increasing their provision will be influenced by the extent to which inspirational groups of individuals step in to play a leadership role to meet the collective need. The Ottawa Convention to Ban Anti-Personnel Landmines and the 2003 Framework Convention on Tobacco Control show that international frameworks are essential to the regulation of the private sector for the good of global health. The World Health Organization, together with other relevant United Nations agencies, has full legitimacy to work with member states toward crafting challenging proposals, and provoking policy action. One lesson from these treaties is that support will have to be built from a strong coalition of like-minded countries that would steer the process internationally. While the development of the treaty is still at an early stage of discussion, the concept is already being aggressively opposed. As with tobacco, landmines, and more recently, sugar, the involvement of civil society will be crucial to defend these health improvement strategies where these may conflict with powerful vested interests in the private sector [ 19 ]. It takes courage to change the rules. If governments are indeed persuaded to face up to their responsibilities in the coming years, it may very well be because of the many voluntary organisations that seek to promote the global public interest.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549583.xml

539058

Transcription-Based Prediction of Response to IFNβ Using Supervised Computational Methods

Changes in cellular functions in response to drug therapy are mediated by specific transcriptional profiles resulting from the induction or repression in the activity of a number of genes, thereby modifying the preexisting gene activity pattern of the drug-targeted cell(s). Recombinant human interferon beta (rIFNβ) is routinely used to control exacerbations in multiple sclerosis patients with only partial success, mainly because of adverse effects and a relatively large proportion of nonresponders. We applied advanced data-mining and predictive modeling tools to a longitudinal 70-gene expression dataset generated by kinetic reverse-transcription PCR from 52 multiple sclerosis patients treated with rIFNβ to discover higher-order predictive patterns associated with treatment outcome and to define the molecular footprint that rIFNβ engraves on peripheral blood mononuclear cells. We identified nine sets of gene triplets whose expression, when tested before the initiation of therapy, can predict the response to interferon beta with up to 86% accuracy. In addition, time-series analysis revealed potential key players involved in a good or poor response to interferon beta. Statistical testing of a random outcome class and tolerance to noise was carried out to establish the robustness of the predictive models. Large-scale kinetic reverse-transcription PCR, coupled with advanced data-mining efforts, can effectively reveal preexisting and drug-induced gene expression signatures associated with therapeutic effects.

Introduction Interferons are small, inducible proteins secreted by nucleated cells in response to viral infection and other stimuli. They act in a paracrine fashion on other cells in their immediate vicinity, triggering a state of growth arrest, so that infected cells cannot be forced to produce viral proteins, and activating the process of programmed cell death, so that infected cells can be removed [ 1 ]. Interferons are important not only in the defense against a wide range of viruses but also in the regulation of immune responses and hematopoietic cell development [ 2 , 3 ]. Recombinant human interferon beta (rIFNβ) is routinely used to control exacerbations in relapsing-remitting multiple sclerosis (MS) [ 4 , 5 ]. Although effective in reducing the number of exacerbations and brain disease activity in some patients, rIFNβ produces no benefit in almost one-half of these patients [ 6 , 7 ]. Furthermore, it is not at all certain how significant its long-term effects on disease progression are. Therapy has been associated with a number of adverse reactions, including flu-like symptoms, transient laboratory abnormalities, menstrual disorders, increased spasticity, and dermal reactions [ 8 ]. We generated and analyzed longitudinal patterns of gene expression from interferon beta (IFNβ)–treated patients suffering from MS with the aim of identifying preexisting and drug-induced signatures that would predict or explain the clinical response to the drug. Results/Discussion Fifty-two patients with relapsing-remitting MS were followed for at least 2 y after initiation of therapy with IFNβ. Clinical follow-up included a neurological examination every 3 mo and at the time of relapse. At each visit, a blood sample was obtained by venipuncture. After the 2-y endpoint, patients were classified as either good or poor responders based on strict criteria, as described in Materials and Methods . We measured the expression profile of 70 carefully selected genes from peripheral blood mononuclear cells isolated from each patient at each time point, using one-step kinetic reverse-transcription PCR ( Dataset S1 ). This process offers remarkable sensitivity and specificity and a dynamic range of several orders of magnitude, allowing the comparison of expressed transcripts from many different genes without compromising accuracy. Targets for analysis were selected on the basis of their presumed biological action and included genes coding for type I and II IFN-responsive molecules, cytokine receptors, members of the interferon (IFN) signaling and apoptosis pathways, and several transcription factors involved in immune regulation ( Table S1 ). Altogether, more than 70,000 reactions were carried out. A common inherent prediction performance limitation of most high-throughput gene-expression profiling projects arises from the largely asymmetric expression data matrix obtained as a result of measuring far more genes than samples [ 9 ]. Such ill-conditioned data matrices inevitably lead to overfitting of predictive models (among other difficulties), some effects of which can be mitigated by judicious application of various established inverse and regularization schemes [ 10 ]. The undesirable properties (i.e., overfitting) of such massively under-determined datasets are largely avoided in this study design because the number of genes measured is commensurable with the numbers of samples. Using linear discriminant analysis–based integrated Bayesian inference system (IBIS), we were able to detect the gene MX1 as the single best discriminating variable between samples obtained at baseline ( T = 0) and at 3 mo after initiation of therapy ( T = 3) with a classification accuracy of 79% (data not shown). Given that MX1 is a known marker of IFN bioavailability [ 11 ], this result validates our experimental approach as well as our sample handling and processing. To search for expression signatures associated with therapeutic outcome (good or poor responder), we conducted clustering of samples using normalized data for all 70 genes at each time point [ 12 ]. Despite applying several different similarity measures and clustering algorithms [ 13 ], we did not observe concomitant segregation of samples according to their responder status, with the exception of a few local clusters of small size ( Figure 1 ). This result may suggest that overall differences in gene expression in the two groups of patients, as assessed by conventional similarity measures, are small or negligible. The clustering null results with respect to concomitant class segregation, however, do not rule out the possibility of discovering outcome-predictive combinatorial and nonlinear relationships. To investigate this possibility further, we used quadratic discriminant analysis–based IBIS, implemented for three-dimensional (3D) searches, in the search for highly predictive sets of three genes whose expression at T = 0 correlated with a good or poor outcome of therapy at the 2-y endpoint. This process exhaustively carried out sample classification by searching through all 54,740 possible three-gene combinations of 70 genes. Higher-order combinatorial searches beyond combinations of three genes are possible using IBIS and are currently under investigation. However, higher-order predictive variable combinations do require the support of many more samples to prevent overfitting of the model. Figure 1 Nonsupervised Two-Way Hierarchical Clustering of Samples at T = 0 A clear aggregation of samples cannot be seen by this technique. The first column indicates the type of responder to which each sample belongs (red, good; blue, poor). We implemented a stringent method for examining the statistical validity of our classification results that consisted in testing the obtained classifier in an independent set of samples not previously “seen” by the program. All of the following statistical analyses were thus performed on split datasets, namely, training (75% of the samples) and test (25%), each reflecting the same relative proportion of classes (63% good and 37% poor responders). We started by conducting 3D IBIS searches using expression data from only the training set and selected the top nine scoring triplets (on the basis of their high prediction accuracy and low mean squared error [MSE] values). For each gene triplet, and using the training data only, a committee of classifiers was built based on an internal cross-validation scheme. Subsequently, the classifiers were used to predict the outcome of an independent test set of samples. Gene triplets were ranked on the basis of the prediction accuracies of the classifiers on this independent test set. We identified nine gene triplets with a predictive accuracy of at least 80% ( Table 1 ). We considered it essential to empirically rule out the chances of fortuitous data splits in the accuracy results obtained from the top-scoring gene triplets. Table 1 Best-Scoring Predictors of Response to IFNβ at T = 0 For each triplet, the mean prediction accuracy for 100 and 500 splits of the data are shown. Prediction was recalculated after the addition of Gaussian noise to each expression value a The percentage drop in prediction accuracy after the addition of noise Consequently, for the nine top-scoring gene triplets and their corresponding classifiers, we generated 100 random splits and built classifiers for each new resulting training set. Next, we tested how well the classifiers predicted therapeutic outcome in the corresponding test datasets. Figure 2 illustrates the distribution in the prediction accuracies obtained for the triplet composed of Caspase 2, Caspase 10, and FLIP, which yielded a predictive accuracy of 86% in the original split. The bell-shaped distribution resulting from 100 tests for this triplet displayed a mean accuracy of 87.8% and a tenth percentile of 78.6%, meaning that if the prediction were performed multiple times, in 90% of these instances an accuracy of almost 79% or better would be obtained. This histogram only reflects the range of accuracies obtained, should the initial data split be different. Notably, the genes in the top-scoring triplet were Caspase 2, Caspase 10, and FLIP —three apoptosis-related molecules. The second-highest-scoring triplet was that of Caspase 2, Caspase 3, and IRF4 (86.8% mean accuracy after 100 splits). Other high-scoring triplets included IL4Ra and MAP3K1, in addition to other apoptotic molecules ( Table 1 ). When we repeated this experiment with the top three scoring genes, using F -test, the obtained mean accuracy was 64% (tenth percentile at 50%) ( Figure 3 ). Figure 2 Accuracy Ranges of the Three-Gene Predictive Model of IFNβ Response After the initial data split into training and test sets, using IBIS on the training set only, nine best-performing triplets were identified. The triplet of Caspase 2, Caspase 10, and FLIP resulted in an accuracy rate of 86% correct prediction on the blind test set resulting from the original split. To minimize the effect of fortuitous initial data division in the accuracy outcome, an extra 100 data splits were performed as a coarse approximation of the possible ranges of accuracies in which this gene triplet could result. A histogram of prediction accuracy over the 100 trials for the gene triplet composed of Caspase 2, Caspase 10, and FLIP is shown as an example of classification and prediction of response to IFNβ at T = 0. A red Gaussian curve encompasses the distribution, where the mean prediction accuracy was 87.9%, with a maximum of 100% (in 11 cases) and a minimum of 64.3% (in two cases). The broken blue line indicates the tenth percentile (78.6%). No major differences were found when we performed the same classification/prediction strategy in 500 random splits of the data. Figure 3 Best-Scoring Gene Triplet by F -Test Analysis Notably, as observed with IBIS, Caspase 10 was also the single best discriminant ( p = 1.87 × 10 −4 ) variable, but the second and third best scoring genes by F -test (IL12Rb2, IL4Ra) did not seem to add any significant predictive power. The mean prediction accuracy for the test set of samples was 65.6% (tenth percentile, 57.1%), well below that observed for the triplet derived from IBIS (Caspase 2, Caspase 10, and FLIP) shown in Figure 2 . This suggests that F -test could efficiently capture individual linear separators but cannot identify and prioritize the nonlinear combinations of genes discovered by IBIS that ultimately provide the most predictive accuracy and robustness. In Figure 4 , the predictive capability of the best-scoring triplet ( Caspase 2, Caspase 10, and FLIP; 3D model) was compared with those obtained with the single-gene (1D) and gene-pair (2D) models. We observed that the classification accuracy improves as more genes are added to the classifier. We next plotted the samples of a test dataset (25% of samples) on the predictive probability model shown in Figure 4 G and compared the performance of the 3D IBIS model to those of the individual 2D models ( Figure 5 ). Overall, the 2D projections of the 3D predictive model show that the Caspase 2 / Caspase 10 and Caspase 10/FLIP gene pairs show significant predictive capability, but that all three genes are required to provide the highest level of model accuracy and robustness. Figure 4 Training Dataset Performance of the Three Genes from the Top Predictive Model of IFNβ Response One-, two-, and three-dimensional IBIS searches were conducted independently on the same training dataset. Each chart shows a two-colored background, corresponding to regions predictive of good response (red) and poor response (blue). Each colored dot corresponds to an individual sample (red, good responder; blue, poor responder). (A–C) One-dimensional IBIS predictive models. High values of Caspase 10 are associated with poor response according to a linear relationship. In contrast, Caspase 2 levels are associated with poor response at intermediate values, suggesting a nonlinear relationship. FLIP expression is associated with good responders at low values, again depicting a linear relationship. The highest cross-validation accuracy score for a single gene predictor was 73% (Caspase 10) . (D–F) Two-dimensional IBIS predictive models. Each of the three possible pairs of this classifier was tested. Linear and nonlinear combinatorial predictive relationships were revealed, specifically, a nonlinear predictive relationship associating poor response with high values of Caspase 10 and intermediate values of Caspase 2, a nonlinear relationship associating good response with high values of FLIP and either low or high (but notintermediate) values of Caspase 2, and a linear relationship associating poor response with low values of FLIP and high values of Caspase 10 . The highest cross-validation score was obtained for the Caspase 2/Caspase 10 pair according to a nonlinear, quadratic distribution (85% accuracy). (G) Three-dimensional IBIS predictive model. The shapes identified in the 1D and 2D distributions were optimized by the 3D model, providing a better separation of good and poor responders. Figure 5 Test Dataset Performance of the Top Three-Gene Predictive Model of IFNβ Response The same probability model generated from the training dataset (see Figure 4 G) provides the background shading of volumes predictive of good response (red) and poor response (blue). Three samples are identified with arrows and followed along different graphical representations. (A and B) The two rotations of the full 3D model show that all good responder samples are correctly classified. (C) Projection of full model onto one of the possible 2D surfaces is provided as an aid to visualization. (D–F) Two-dimensional IBIS predictive models. Three samples are identified with arrows and followed along different graphical representations. If prediction was performed in only two dimensions, a higher number of misclassifications would have occurred. For example, the 2D model built using only Caspase 2/FLIP (D), could not resolve the good responding sample identified by a cyan arrow, whereas it correctly resolves the good responding sample shown by the orange arrow. The model built using Caspase 10/FLIP (E), in contrast, acts oppositely and can resolve the good responding sample shown by the cyan arrow and not the sample shown by the orange arrow. Both these sample are correctly resolved the 2D model built using Caspase 2/Caspase 10 (F); however, this model is unable to resolve the poor responding sample identified by the yellow arrow, whereas one of the previous models (E) was able to do this. As demonstrated in the full 3D model view from (A) and (B), as well as the projection of model (C), all the labeled poor and good responding patients are correctly classified. Although 2D models show high predictive capabilities, all three genes are needed to increase the classification accuracy of the IBIS model. To validate the specificity and predictive capability of the top-scoring gene triplet (for the good and poor responding classes) and its associated classifiers, we examined the model exhibiting the best performance on a “default” expression dataset. This null dataset was built keeping the original gene expression data and randomly permuting the class labels of the outcomes 1,000 times (keeping the same counts of good and poor responding patients as were in the original dataset). The prediction accuracies for all the gene triplets obtained with this dataset dropped dramatically as the means ranged from 49.2% to 53.6% (data not shown), emphasizing the specificity of the classifiers. In addition, for the top-scoring gene triplet (for good and poor responder classes), we calculated the probability of achieving, under the null hypothesis, an equal or better accuracy than that obtained in the original prediction (86%), as previously described [ 14 ]. This achieved significance level was 0.009, suggesting that it is very unlikely that the prediction accuracies observed for this classifier are caused by chance. Finally, we tested the robustness of each of these gene sets as predictors of IFNβ response by simulating experimental measurement error. To accomplish this, we first calculated the standard deviation of the expression measurements for all genes as an estimation of the overall experimental error. We then added a fixed amount of Gaussian noise corresponding to one standard deviation (taken from 20 random deviations) to each expression value and repeated the classification/prediction in 30 different splits of the data (a total of 600 tests). Notably, the mean drop in predictive accuracy after the addition of noise was less than 10%, denoting a significant tolerance to reasonable measurement errors ( Table 1 ). Because all the patients in this study were systematically followed up for a period of 2 y, we were able to perform a longitudinal analysis. Using a repeated-measures analysis of variance (ANOVA), we searched for genes with significantly different expression patterns based on models that tested for responder effect, time effect, and interaction effect (time × response). Significant responder effect for 20 genes ( Figure 6 ) and significant time effect for 13 genes were detected ( Figure 7 ). Interestingly, six of the genes that showed statistically significant differences between good and poor responders, IRF4 ( p = 0.03), IL4Ra ( p = 0.01), Caspase 10 ( p = 0.0008), Caspase 7 ( p = 0.01), IRF2 ( p = 0.02), and IRF6 ( p = 0.03) are among the 12 genes that best predict response at T = 0 (shown in bold in Figure 7 B). A pattern consistent with increased apoptosis (five members of the Caspase family of proteins, TRADD, and BAX ) was observed for the poor responders. Figure 6 Characteristic Gene Expression Profiles of Good and Poor Responders to IFNβ over Time (A) An unsupervised hierarchical clustering representation of the weighted difference between the average expression of good and poor responders. For each gene, the obtained differences were log normalized and multiplied by the F -statistic from an ANOVA (responder effect) run previously (shown in [B]). The “heat” colored bar represents the absolute value of this difference. With the exception of MX1 (indicated by an arrow), all genes showing a significant difference in expression between the two groups of patients were automatically arranged in only two clusters (framed in blue). (B) List of all genes showing a significant responder effect along with their F -statistic and p -values. Genes that were part of any triplet showing more than 80% prediction accuracy at T = 0 are shown in bold. (C) A continuous representation of the longitudinal average expression of two representative genes for good (^) and poor (•) responders. TRADD shows two widely parallel curves, indicative of a significant difference in the expression averages, correlating with its profile (#) observed in the clustering shown in (A). In contrast, GATA 3 displays two almost overlapping curves, consistent with its shading (*) in the clustering in (A). Figure 7 IFNβ-Induced Changes in Gene Expression over Time (A) An unsupervised hierarchical clustering representation of the weighted difference in gene expression at each time point versus baseline. For each gene, the obtained differences were log normalized and multiplied by the F -statistic from an ANOVA (time effect) run previously (shown in [B]). The “heat” colored bar represents the absolute value of this difference. With the exception of IFNAR1 (arrow), all genes showing a significantly different expression in at least one time point with respect to baseline were arranged in the same cluster (framed in blue). (B) List of all genes showing a significant time effect along with their F -statistic and p -values. Genes that were part of any triplet showing more than 80% prediction accuracy at T = 0 are in bold. (C) A continuous representation of the longitudinal average expression of two representative genes over all samples. MX1 (^) shows a marked departure from T = 0 and remains elevated for the rest of the observed period. This correlates well with the shading (#) displayed in the clustering shown in (A). In contrast, IRF6 (•) displays an almost flat curve, consistent with its color in the clustering (*). Although we successfully identified informative, combinatorial relationships, establishing the causality of the association between gene expression and outcome to therapy is beyond of the scope of this work, and these genes are therefore considered surrogate markers. Moreover, although extensive in vivo and in vitro experiments have been conducted, the full mechanism of action of IFNβ in MS remains unknown. Transcription profiling experiments have involved IFNβ in the regulation of apoptosis in both cancer and MS [ 15 , 16 , 17 , 18 ]. Induction of programmed cell death could lead to a reduction in the number of activated lymphocytes, macrophages, and monocyte-derived dendritic cells—all key components of the pathogenic process leading to tissue damage in MS [ 19 , 20 , 21 ]. However, increased levels of some anti-apoptotic molecules have also been observed in IFNβ-treated MS patients, possibly reflecting a compensatory mechanism [ 16 ]. Furthermore, even the inhibition of activated T cell apoptosis in response to IFNα and IFNβ has been reported [ 22 ]. Our finding of increased apoptosis in poor responders does not support the hypothesis of programmed cell death as a primary therapeutic mechanism for IFNβ. We hypothesize that a net increase in pro-apoptotic transcripts in peripheral blood mononuclear cells from poor responders could be reflecting undesired elimination of certain regulatory cell populations that are much needed to maintain a homeostatic balance. Other differentially expressed transcripts included IRF4, a gene essential for mature T and B lymphocyte function and homeostasis, and a transcription factor with dual function (activator/repressor) that regulates transcription of IL4 through physical interaction with NFATc2 [ 23 ]. Remarkably, IRF4 is a repressor of other IFN-induced genes [ 24 ], an observation consistent with the elevated expression of IRF4 observed in the poor responders before initiation of therapy. As expected, the gene MX1 showed a significant time effect independent of clinical response ( p = 0.01). This result is in agreement with previous findings indicating substantial MX1 upregulation in response to type I IFNs [ 25 ]. Interestingly, upregulation of MX1, which occurs minutes after IFN stimulation [ 26 ], is sustained over at least 2 y, spanning several orders of magnitude of time units. This also correlates well with our results identifying MX1 as the best single classifier for samples from patients before ( T = 0) and after ( T = 3) initiation of therapy. In fact, as Figure 7 illustrates, most of the significance for the time effect in MX1 comes from the difference between T = 0 and T = 3. Also of interest, a significant time effect (but not responder effect) was observed for IFNAR1 and STAT2 ( Figure 7 B). Because IFNAR1 is a subunit of the heterodimeric type I IFN receptor and STAT2 is a critical component of the DNA binding complex ISGF3a (which regulates the expression of IFN-responsive genes), their upregulation on administration of rIFNβ is likely related to mechanistic aspects of IFN signaling. Our results suggest that poor response is associated with downstream signaling events rather than deficient recognition or metabolism of the drug. Our previous finding that IFN receptor polymorphisms do not affect therapeutic response in this same set of patients partially supports this hypothesis [ 27 ]. Only two genes with significant time effects, Caspase 10 ( p = 0.01) and MAP3K1 ( p = 0.01) were part of any predictor set ( Figure 7 B). In addition, MAP3K1 also showed a significant interaction effect ( p = 0.05; data not shown). These results highlight the involvement of these genes in the response to IFN both at T = 0 and once therapy has started. Here we combined large-scale, function-oriented gene expression with advanced data mining to identify a set of markers that accurately and robustly predict the response to rIFNβ therapy. Although larger, prospective studies are needed to confirm these findings, our results suggest that the underlying gene activity profile of an individual at the verge of therapy harbors sufficient information to allow investigators to estimate the chances of experiencing satisfactory therapeutic effects. As analytical tools to predict clinical outcomes based on molecular evidence evolve, these types of studies are likely to become a substantial aid to the physician, taking the paradigm of personalized medicine one step further. Materials and Methods Patients and samples All studies were approved by the respective Committees of Human Research at Hospital Vall d'Hebron, Barcelona, Spain, and the University of California, San Francisco, United States. Informed consent was obtained for all study participants. All patients were examined by a trained neurologist at the CNI Unit, Vall d'Hebron Hospital. Inclusion criteria for this study were clinically definite MS (Poser's criteria), disease in relapsing-remitting phase, age between 18 and 65 y, recorded history of at least two clearly identified relapses within the preceding 24 mo, and expanded disability status scale between zero and 5.5 (inclusive). Detailed information about clinical aspects of these patients has been recently reported elsewhere [ 6 ]. Patients were categorized as good responders ( n = 33) if they experienced a total suppression of relapses and no increase in the expanded disability status scale after a 2-y follow-up period. Poor responders ( n = 19) were defined as having suffered two or more relapses or having a confirmed increase of one point in the expanded disability status scale score. Patients with intermediate phenotypes were excluded from this study. Blood specimens were taken following institutional guidelines at approximately the same time of the day just before the administration of the first dose of rIFNβ and every 3 mo thereafter during the neurological examination, with the exception of T = 15 and T = 21 mo. Altogether, 336 samples were tested (an average of 6.5 time points along 2 y for each individual). Immediately after being drawn, all blood samples were spun over Ficoll-Paque (Amersham Biosciences, Piscataway, New Jersey, United States) gradients to enrich the sample for mononuclear cells. After three washes with PBS, aliquots of 5 × 10 6 cells were frozen in RPMI1640 containing 20% DMSO and 20% fetal calf serum. RNA purification, quantitation, and handling Peripheral blood mononuclear cells were thawed at 37 °C for 1 min, and RNA lysis buffer was added immediately. RNA was purified with the Strataprep kit (Stratagene, La Jolla, California, United States) and finally resuspended in nuclease-free water (Promega, Madison, Wisconsin, United States). One-microliter RNA aliquots were subjected to fluorescence-based quantitation (in duplicate) using the Ribogreen reagent (Molecular Probes, Eugene, Oregon, United States) in a Spectra Max Gemini fluorometer (Molecular Devices, Sunnyvale, California, United States). Samples were diluted to 1 ng μl −1 using nuclease-free water, and 5 μl was aliquoted in triplicates into 384-well plates using a Multiprobe II liquid-handling instrument (PerkinElmer Life and Analytical Sciences, Boston, Massachusetts, United States). Plates were kept frozen at −70 °C until needed. One-step kinetic reverse-transcription PCR A master mix was prepared essentially as described previously [ 28 ], with the addition of 200 μM ROX (Sigma, St. Louis, Missouri, United States), and overlaid on top of each well of a freshly thawed 384-well plate containing 5 ng of RNA in each well. Reactions were performed in triplicate using an ABI 7900 Sequence Detection System (Applied Biosystems, Foster City, California, United States). Positive and negative controls, as well as calibration curves, all in triplicate, were also included in each reaction plate. Total reaction volume was 10 μl. All expression values were calculated by interpolation in a calibration curve spanning five orders of magnitude constructed with an in vitro transcribed clone of GAPDH. The average of each expression measurement was then divided by that of one of the positive controls (thymus RNA) to account for plate-to-plate variability. On the basis of reports addressing the limited utility of normalization [ 29 , 30 ] and of our unpublished observations, we avoided housekeeping gene normalization and used instead RNA content, thus relying on repeated precise fluorescence-based quantitation and highly accurate liquid-handling procedures. Data collection and analysis A custom Microsoft Excel (Microsoft, Redmond, Washington, United States) worksheet was prepared for handling reaction data import and performing initial statistics. Normalized data were imported as a .csv file into GeneLinker Platinum (Predictive Patterns Software, Kingston, Ontario, Canada) for preprocessing and clustering analysis. Quadratic discriminant analysis–based IBIS implemented for 3D searches was carried out at the School of Computing, Queen's University, Kingston, Ontario, Canada, and at Biosystemix, Sydenham, Ontario, Canada. IBIS is a data-mining algorithm that searches through the gene space for a single gene (or group of genes) that can predict the outcome class (in this case, good and poor response to rIFNβ therapy). This algorithm incorporates a complete 10-fold cross-validation method with several independently trained classifiers to predict an outcome on the basis of a voting scheme (see below). We used MSE and classification accuracy to assess how well the classification predictions matched the true response of the patients to therapy. The top-performing gene triplets were selected on the basis of a mixed threshold for low MSE levels and high accuracy rates. The algorithm IBIS identifies genes (or gene pairs or groups of genes) that are highly predictive of the outcome based on probability distributions of those genes in different outcome classes. For example, for a given gene g i , two Gaussian functions are fitted to the distributions of the observed expression levels in good responding and poor responding patients (let us call these fitted distributions D g and D p for good and poor responding patients, respectively). Our fitted distribution, D g ( x ), denotes the probability of a patient having an expression level of g i = x, given that this patient is a good responder. The fundamental question we are aiming to answer using data-mining methods (here using IBIS particularly) is as follows: what is the probability of a patient being a good responder given the observed expression level of a gene for that patient? Taking advantage of the fitted distributions, a classifier applies Bayes' formula to answer the fundamental question. According to this formula: where P 1 = P (( g i = x ) |Pa good_responder ) P ( Pa good_responder ). In fact, P (( g i = x ) |Pa good_responder ) is the distribution function fitted to the observed gene expression values of g i above (D g ), and P ( Pa good_responder ), the probability of a patient being a good responder, can be easily calculated using the total and good responding patient counts. The term P 2 is strictly analogous to P 1 but applies to poor responders. Therefore, according to equation 1 , for a given gene, a comprehensive model is built that predicts the probability of a patient being a good responder for different values of observed expressions of that gene. IBIS searches through all the genes and calculates such models for a single gene or combination of genes, resulting in singular or combinatorial mining of most relevant genes. The probability of a patient being a poor responder can also be calculated in a similar fashion. To obtain a reliable classifier that is generalizable to all patient samples obtained under similar conditions, the Gaussian distributions and the classifier were only trained on a subsample of the patient data (training set). The results of the classification (i.e., probability of a patient being a good or poor responder), however, were tested on patient samples never seen by the classifier before (test set). This ensures limitation of the classifier overfitting. Further, a complete 10-fold cross-validation scheme was built into the training phase. In IBIS, linear and quadratic classifiers correspond to classifiers built using Gaussian distributions with equal or different covariance matrices, respectively. The prediction results of IBIS are visualized graphically within the observed gene expression space by presenting the probabilities of a patient being a good or poor responder as a background color (see Figures 4 and 5 ). The red background in the gene space represents a high probability of a patient sample being a good responder if the observed gene expression values are in that range in the gene space. The blue background, similarly, represents a high probability of a patient sample being a poor responder if the observed gene expression values are in that range in the gene space. Several measures were used to assess how well the calculated probabilities matched the true patient responses to therapy. Two of these measures were MSE and classification accuracy. MSE was calculated as the sum of (response i observed − response i expected ) 2 averaged over all patients. For a given patient, the clinical response determined by the end of the 2-y monitoring period is denoted by response i observed and response i expected and represents the probability of that patient being a good responder to rIFNβ therapy, using the Bayes' formula above. Classification accuracy simply expresses the percentage of patients that were correctly predicted as being good or poor responders. Classification and prediction procedure The initial dataset of patients was divided into two parts; namely, a training set with 75% of the samples and a test set with 25% of the samples, each reflecting the same proportion of classes (63% good and 37% poor responders). Only the training set was used for identifying the best predictive gene triplets with the IBIS method, as well as for building the classifier. A committee of classifiers was then generated using a 10-fold cross-validation scheme during training. The training data themselves were divided into ten parts, and each time, a classifier was built using only nine parts of the data. That classifier's predictive capability was determined by its accuracy over the one-tenth of the data withheld. A committee of ten classifiers was assembled from the results of this training stage; this committee was then applied to the test data (which have thus far been hidden from the classifiers). For a patient sample in the test data, each classifier in the committee made a prediction. A majority voting scheme then decided as to which class the sample would be assigned. Given the initial data split into training and test sets, it was important to rule out the role of fortuitous idiosyncrasies in this split and the resulting accuracy rates. To address this point, we created 100 random splits of the data into training and test subgroups. A committee of classifiers was trained on the training set for each data split, and the accuracies were calculated over the blind test set. A histogram of the test set accuracies was then built, representing the expected ranges of accuracies had the initial data split been different. This histogram is not representative of the estimated or idealized distribution of the accuracies for a gene triplet in a machine learning sense. Rather, it is a coarse approximation of the possible range of gene triplet outcome–prediction accuracies that could be expected. Controlling for false discoveries To assess the significance and specificity of the top-scoring gene triplets and their corresponding trained committee of classifiers, a null dataset was created by keeping the same expression levels of genes in the dataset and randomly permuting the class labels of the patients 1,000 times (the total count of poor and good responding patients was unchanged). Classifiers were built using the training null data, and accuracies were calculated on the corresponding test sets. The mean of these accuracies for all the top-performing gene triplets was around 50%. The achieved significance level, which represents the probability of achieving accuracy levels better than or equal to that of the nonpermuted classification, was calculated to be 0.009. This value can be considered a significance level, or p -value, and indicates the number of times in 1,000 trials for which accuracies of 86% or higher can be achieved under the “no predictive capability” null hypothesis. Time-series analysis was performed using SAS version 8.0 (SAS Institute, Cary, North Carolina, United States). Permutation analysis and histogram graphic outputs were produced with Matlab (The Mathworks, Natick, Massachusetts, United States). Supporting Information Dataset S1 Raw Expression Dataset Gene expression values for all samples at all time points. This is the raw file from which all analyses were performed. (491 KB XLS). Click here for additional data file. Table S1 Target Information Gene names, symbols, and LocusLink and GenBank accession numbers, as well as primer sequences, are listed for all targets. (160 KB DOC). Click here for additional data file. Accession Numbers The LocusLink ( http://www.ncbi.nlm.nih.gov/projects/LocusLink/ ) accession numbers for the genes discussed in this paper are BAX (LLID 581), Caspase 10 (LLID 843), Caspase 2 (LLID 835), Caspase 3 (LLID 836), Caspase 7 (LLID 840), FLIP (LLID 8837), IFNAR1 (LLID 3454), IL4 (LLID 3565), IL4Ra (LLID 3566), IRF2 (LLID 3660), IRF4 (LLID 3662), IRF6 (LLID 3664), MAP3K1 (LLID 4214), MX1 (LLID 4599), NFATc2 (LLID 4773), STAT2 (LLID 6773), and TRADD (LLID 8717).

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539058.xml

555753

Gene expression signature of estrogen receptor α status in breast cancer

Background Estrogens are known to regulate the proliferation of breast cancer cells and to modify their phenotypic properties. Identification of estrogen-regulated genes in human breast tumors is an essential step toward understanding the molecular mechanisms of estrogen action in cancer. To this end we generated and compared the Serial Analysis of Gene Expression (SAGE) profiles of 26 human breast carcinomas based on their estrogen receptor α (ER) status. Thus, producing a breast cancer SAGE database of almost 2.5 million tags, representing over 50,000 transcripts. Results We identified 520 transcripts differentially expressed between ERα-positive (+) and ERα-negative (-) primary breast tumors (Fold change ≥ 2; p < 0.05). Furthermore, we identified 220 high-affinity Estrogen Responsive Elements (EREs) distributed on the promoter regions of 163 out of the 473 up-modulated genes in ERα (+) breast tumors. In brief, we observed predominantly up-regulation of cell growth related genes, DNA binding and transcription factor activity related genes based on Gene Ontology (GO) biological functional annotation. GO terms over-representation analysis showed a statistically significant enrichment of various transcript families including: metal ion binding related transcripts (p = 0.011), calcium ion binding related transcripts (p = 0.033) and steroid hormone receptor activity related transcripts (p = 0.031). SAGE data associated with ERα status was compared with reported information from breast cancer DNA microarrays studies. A significant proportion of ERα associated gene expression changes was validated by this cross-platform comparison. However, our SAGE study also identified novel sets of genes as highly expressed in ERα (+) invasive breast tumors not previously reported. These observations were further validated in an independent set of human breast tumors by means of real time RT-PCR. Conclusion The integration of the breast cancer comparative transcriptome analysis based on ERα status coupled to the genome-wide identification of high-affinity EREs and GO over-representation analysis, provide useful information for validation and discovery of signaling networks related to estrogen response in this malignancy.

Background Estrogen plays essential roles in the development, growth control and differentiation of the normal mammary gland. However, it is well documented that endogenous estrogens are powerful mitogens critical for the initiation and progression of human breast and gynecological cancers [ 1 ]. This cell proliferation signal is mediated by the estrogen receptors (ER), members of the nuclear receptor family that function both as signal transducers and transcription factors to modulate expression of target genes [ 2 ]. There are two main subtypes of estrogen receptors: ERα and ERβ that generally can form homo- and heterodimers before binding to DNA. Although the DNA binding domains of these receptors are very similar, the overall degree of homology is low [ 3 ]. Transcriptional regulation of target genes in response to 17β-estradiol (E 2 ) is mediated by two main mechanisms. In one, the E 2 -ER complex binds to a specific DNA sequence called the estrogen response element (ERE), this receptor-ligand DNA bounded complex interacts with co-regulatory proteins, promoting chromatin remodeling and bridging with the general gene transcription machinery thus resulting in transcription initiation [ 4 ]. Alternatively, the ligand-ER complex can interact with other DNA-bound transcription factors that in turn bind DNA sequences (e.g. via AP1, SP1 complexes) [ 5 , 6 ]. ERα and ERβ have different affinities for different response elements and exhibit distinct transcriptional properties. Additionally, E 2 also exerts rapid, non-genomic effects attributed to cell membrane-initiated signaling [ 7 ]. Approximately two-thirds of all breast cancers are ERα (+) at the time of diagnosis and expression of this receptor is determinant of a tumor phenotype that is associated with hormone-responsiveness. Patients with tumors that express ERα have a longer disease-free interval and overall survival than patients with tumors that lack ERα expression [ 8 ]. However, the association between ERα expression and hormonal responsiveness is not perfect: approximately 30% of ERα-positive tumors are not hormone-responsive while 5–15% of ERα-negative tumors respond to hormonal therapy [ 9 ]. The molecular basis for the association between ERα expression, hormonal responsiveness and breast cancer prognosis remains unclear. Several studies have been carried out using cDNA and oligonucleotide microarrays identifying breast cancer subclasses possessing distinct biological and clinical properties [ 10 - 13 ]. Among the distinctions made to date, the clearest separation was observed between ERα (+) and ERα (-) tumors [ 10 - 15 ]. It has been suggested that there are sets of genes expressed in association with ERα that could play an important role in determining the hormone-responsive breast cancer phenotype [ 16 ]. ERα is obviously likely to be important for the E 2 induced proliferative response predominantly via the regulation of estradiol-responsive genes. Nevertheless, the expression of additional subsets of genes not necessarily directly regulated by estrogen may also be fundamental in defining the breast cancer hormone-responsive phenotype. To further elucidate the molecular basis of estrogen-dependent breast carcinogenesis, we here report a comparative transcriptome profiling of invasive breast tumors based on ERα status obtained by SAGE. The SAGE method provides a statistical description of the mRNA population present in a cell without prior selection of the genes to be studied, and this constitutes a major advantage [ 17 ]. The breast cancer SAGE comparative analysis was combined with promoter sequence analysis of genes of interest using high-throughput methods of high-affinity ERE identification. In order to have an even more comprehensive picture we also performed a cross-platform comparison between SAGE and DNA microarray studies. Results and discussion Biomarkers of ERα status in breast carcinomas The primary goal of our study was to identify the most commonly deregulated genes in invasive breast carcinomas related to ERα status. To this end SAGE data was obtained from a set of primary breast carcinomas. Thus, a breast cancer SAGE database of almost 2.5 million tags was analyzed, representing over 50,000 tag species. We performed a comprehensive evaluation and comparison of gene expression profiles using a recently developed supervised method [ 18 ], to identify the most representative differentially expressed transcripts between tumors groups, i.e. ERα (+) vs. ERα (-) breast tumors. This statistical analysis revealed 520 genes differentially expressed (Fold change ≥ 2; p < 0.05) between ERα (+) and ERα (-) primary breast carcinomas (see additional data file 1 ). Among the 520 transcripts, 473 were up-modulated and 47 were down-modulated transcripts in ERα (+) tumors. The most commonly over-expressed transcripts in ERα (+) tumors were: trefoil factor 1 ( TFF1/pS2 ), synaptotagmin-like 4 ( SYTL4 ), regulating synaptic membrane exocytosis 4 (RIMS4), dual specificity phosphatase 4 ( DUSP4 ), chromosome 1 open reading frame 34 ( C1orf34 ), necdin homolog ( NDN ), n-acetyltransferase 1 ( NAT1 ) and caspase recruitment domain family 10 ( CARD10 ) (Table 1 and additional data file 1 ). Table 1 Most highly up-modulated transcripts in ERα (+) breast carcinomas identified by SAGE. Gene name Tag Locus Link Fold change ( p value) Frequency # Cell proliferation related TFF1* (trefoil factor 1) CTGGCCCTCG 7031 51.4 (0.0016) 15/18 (83%) DUSP4 (dual specificity phosphatase 4) CGGGCAGAAA 1846 14.7 (0.0016) 14/18 (78%) NDN* (necdin homolog) ACCTTGCTGG 4692 13.3 (0.0026) 11/18 (61%) HDGFRP3 (hepatoma-derived growth factor) TGTAAAGTTT 50810 9.8 (0.0019) 12/18 (67%) TSPAN1* (tetraspan 1) GGAACTGTGA 10103 9.5 (0.0017) 15/18 (83%) SEP6 (septin 6) TCAATTTTCA 23157 7.6 (0.0044) 12/18 (67%) DHX34* (DEAH box polypeptide 34) GTTGCTCACT 9704 7.1 (0.0129) 9/18 (50%) Apoptosis related CARD10* (caspase recruitment domain family) AGAATGTACG 29775 11.1 (0.0030) 15/18 (83%) Signal transduction related SYTL4* (synaptotagmin-like 4) TATGTGTGCT 94121 28.0 (0.0003) 15/18 (83%) ECM1* (extracellular matrix protein 1) ACTGCCCGCT 1893 10.1 (0.0175) 13/18 (72%) LEPR* (leptin receptor) AAAGTTTGAG 3953 9.8 (0.0302) 10/18 (55%) PTGES (prostaglandin E synthase) TGAGTCCCTG 9536 8.0 (0.0168) 8/18 (44%) SCUBE2 (signal peptide, CUB domain EGF-like 2) TCAGCACAGT 57758 7.5 (0.0024) 14/18 (78%) ADORA2A* (adenosine A2a receptor) TGCTGAGTAG 135 7.1 (0.0460) 11/18 (61%) ITGBL1 (integrin beta-like 1) CATATTCACA 9358 7.1 (0.0159) 8/18 (44%) Regulation of transcription related ESR1 (estrogen receptor 1) AGCAGGTGCC 2099 9.8 (0.0000) 18/18 (100%) TCEAL1 (transcriptional elongation factor A) AAAGATGTAC 9338 9.8 (0.0014) 13/18 (72%) ZNF14 (zinc finger protein 14) TAAACAGCCC 7561 8.4 (0.0023) 13/18 (72%) ZNF38* (zinc finger protein 38) CCAGCATTAC 7589 7.6 (0.0051) 10/18 (55%) HIF1AN* (hypoxia-inducible factor 1α subunit inhibitor) CCTGAGTGCG 55662 7.1 (0.0094) 10/18 (55%) HOXC13 (homeo box C13) TTTTTAAAAT 3229 7.1 (0.0157) 9/18 (50%) Cytoskeleton MAPT (microtubule-associated protein tau) GTAGACTCGC 4137 9.8 (0.0085) 9/18 (50%) MYLIP (myosin regulatory light chain interacting) TTTTCCACTC 29116 9.3 (0.0036) 11/18 (61%) Metabolism and Miscelaneous RIMS4 (regulating synaptic membrane exocytosis) TTGAAATTAA 140730 24.9 (0.0378) 8/18 (44%) NAT1 (N-acetyltransferase 1) TATCTTCTGT 9 11.7 (0.0385) 15/18 (83%) ATP6V1B1* (ATPase, H+ transporting) CCTCCCCCTC 525 10.7 (0.0111) 10/18 (55%) JDP1 (J domain containing protein 1) TCTGTGAATT 56521 10.0 (0.0035) 12/18 (67%) CHST11 (carbohydrate sulfotransferase 11) AACCTTCCTC 50515 9.8 (0.0009) 13/18 (72%) CILP (nucleotide pyrophosphohydrolase) GTTTTGCCCA 8483 9.3 (0.0054) 14/18 (78%) ABCA3 (ATP-binding cassette sub-family A) GTAGTCACCG 21 8.9 (0.0149) 10/18 (55%) SEC14L2 GGAAGGCGGC 23541 8.7 (0.0487) 9/18 (50%) ANXA9* (annexin A9) ACATCCGAGG 8416 8.4 (0.0145) 10/18 (55%) KCTD3 (K channel tetramerisation domain 3) ATAATTAAAT 51133 8.4 (0.0001) 17/18 (94%) SFRS7 (splicing factor) TAGCTAATAT 6432 8.0 (0.0031) 12/18 (67%) SNRPA* (small nuclear ribonucleoprot. polypep. A) AAGATCTCCT 6626 7.6 (0.0009) 15/18 (83%) NNMT (nicotinamide N-methyltransferase) CCTGCAATTC 4837 7.6 (0.0120) 10/18 (55%) SLC1A4 (solute carrier family 1 member 4) GACTCACAGG 6509 7.6 (0.0254) 9/18 (50%) TIPARP (TCDD-inducible polymerase) AAATGGCCAA 25976 7.6 (0.0051) 10/18 (55%) SLC7A2 (solute carrier family 7 member 2) CACTGACAGC 6542 7.3 (0.0190) 11/18 (61%) GA* (liver mitochondrial glutaminase) CTGCTGCTAC 27165 7.1 (0.0126) 9/18 (50%) Function unknown C1orf34 AGGATGTACA 22996 13.3 (0.0025) 14/18 (78%) SMILE (hypothetical protein FLJ90492) TAGAGAGTTT 160418 11.1 (0.0004) 15/18 (83%) RHBDL4 (rhomboid, veinlet-like 4) TTGTTTCTAA 162494 10.7 (0.0099) 9/18 (50%) KIAA0882 GTCTCATTTC 23158 10.1 (0.0007) 18/18 (100%) C20orf103* TTTAGTGATT 24141 9.3 (0.0277) 10/18 (55%) FLJ33387 GCAGGGAGAG 161145 9.3 (0.0118) 10/18 (55%) TRALPUSH GTTTCCAGAG 116931 8.9 (0.0458) 9/18 (50%) KIAA0980* TGGTGCTTCC 22981 7.6 (0.0096) 11/18 (61%) C10orf32 AGTCTGTTGT 119032 7.3 (0.0002) 15/18 (83%) FLJ13611 TAATCACACT 80006 7.1 (0.0069) 10/18 (55%) * Genes with known or putative high-affinity EREs mapping in the vicinity of the TSS. # Transcripts tags changing > 2-fold when compared with the average expression of ER (-) tumors in at least 8 of 18 (44%) ERα (+) invasive carcinomas SAGE libraries. For the whole list of ERα associated transcripts see additional data file 1 . To validate novel ERα associated genes detected by SAGE not reported in other studies, we performed Real Time RT-PCR analysis of representative transcripts in an independent set of 36 invasive ductal breast carcinomas. In agreement with our SAGE analysis, we detected statistical differences in the over-expression of 8 out of 9 evaluated transcripts in ERα (+) breast tumors including: signal peptide CUB domain EGF-like 2 ( SCUBE2 ) (p = 0.0001), SYTL4 (p = 0.0005), KIAA0882 protein (p = 0.0005), tetraspan 1 ( TSPAN1 ) (p = 0.001), myeloblastosis viral oncogene homolog ( C-MYB ) (p = 0.002), epidermal growth factor-like 2 ( CELSR2 ) (p = 0.011), nuclear receptor subfamily 4 ( NR4A1 ) (p = 0.029), and enolase 2 ( ENO2 ) (p = 0.033) (Figure 1 ). A trend of borderline significance was detected for the lectin galactoside-binding protein ( LGALS3BP ) (p = 0.079) transcript (Figure 1 ). Figure 1 Real time RT-PCR validation of nine over-expressed genes in 36 invasive breast carcinomas. a) SCUBE2 (p = 0.0001); b) SYTL4 (p = 0.0005); c) KIAA0882 (p = 0.0005); d) TSPAN1 (p = 0.001); e) CMYB (p = 0.002); f) CELSR2 (p = 0.011); g) NR4A1 (p = 0.029); h) ENO2 (p = 0.033); i) LGALS3BP (p = 0.079). Mean ± 2 Standard Error based on Log2 transformation of real time RT-PCR values of the assayed gene relative to 18S rRNA used as normalizing control. SCUBE2 (also known as EGF-like 2 or CEGP1 ) encodes a secreted and cell-surface protein containing EGF and CUB domains that defines a novel gene family [ 19 ]. The epidermal growth factor (EGF) motif is found in many extracellular proteins that play an important role during development, functioning as secreted growth factors, transmembrane receptors, signaling molecules, and important components of the extracellular matrix. The CUB domain is found in several proteins implicated in the regulation of extracellular process such as cell-cell communication and adhesion [ 20 ]. Expression of SCUBE2 has been detected in vascular endothelium and may play important roles in development, inflammation and perhaps carcinogenesis [ 19 ]. The CELSR2 gene (also known EGFL2 ) encodes a protein member of the nonclassic-type cadherins (flamingo subfamily). These 7-pass transmembrane proteins have nine cadherin domains, seven-epidermal growth factor-like repeats and two laminin A G-type repeats [ 21 ]. It is postulated that these proteins are receptors involved in cell adhesion and receptor-ligand interactions [ 21 ] playing a role in developmental processes and cell growth/ maintenance in epithelial and neuronal cells [ 22 , 23 ]. SYTL4 (also known as granuphilin-a or SLP4 ) contains an N-terminal Slp homology domain (SHD) than can specifically and directly bind the GTP-bound form of Rab27A, a small GTP-binding protein involved in granule exocytosis in cytotoxic T lymphocytes [ 24 ]. We determined that over-expression of SYTL4 is associated with ERα (+) tumors (Figure 1b ). However, the potential role of this gene in breast carcinogenesis remains unknown. ENO2 (also known as NSE/neuron-specific gamma enolase ) encodes one of three enolase isoenzymes found in mammals. This isoenzyme was described to be expressed in cells of neuronal origin. Interestingly, in a recent report Hao et al . (2004) showed high expression of ENO2 transcripts in breast cancer lymph node metastases when compared with primary breast tumors [ 25 ]. The TSPAN1 gene (also known as tetraspanin or NET1 ) encodes a cell-surface protein member of the transmembrane 4 superfamily ( TM4SF ), involved in the regulation of cell development, activation, growth and motility. A number of tetraspanins were described as tumor-specific antigens, and it was suggested that the function of some TM4SF proteins may be particularly relevant to tumor cell metastasis [ 26 ]. Sugiura and Berditchevski (1999) observed that TSPAN1 protein complexes may control the invasive migration of tumor cells and contribute to ECM-induced production of MMP2 in breast cancer cell line [ 27 ]. NR4A1 , a nuclear receptor subfamily 4, group A gene (also known as steroid receptor TR3 or NUR77 ) encodes an orphan member of the steroid-thyroid hormone-retinoid receptor superfamily whose members mainly act as transcriptional factors to positively or negatively regulate gene expression and play roles in regulating growth and apoptosis [ 28 , 29 ]. A role for NR4A1 in cell proliferation has been previously reported. It was shown that its expression is rapidly induced by various mitogenic stimuli such as: serum growth factor, epidermal growth factor and fibroblast growth factor [ 28 ]. Taken together, the genes that we identified and validated appear to be involved in signaling pathways related to cell proliferation, invasion and metastatic processes, but their exact role in breast carcinogenesis remains to be elucidated. Gene Ontology analysis Classification of genes based on Gene Ontology (GO) terms is a powerful bioinformatics tool suited for the analysis of DNA microarray and SAGE data. Analysis of GO annotation allows one to identify families of genes that may play significant roles related to specific molecular or biological processes in expression profiles [ 30 ]. We used the Expression Analysis Systematic Explorer software ( EASE ) [ 31 ] to annotate the 520 deregulated genes according to the information provided by the GO Consortium [ 30 ]. The GO database provided annotation for 80% (419 out of 520) of the genes identified by SAGE. Results of this analysis are shown in Figure 2 and in detail in additional data file 2 . Figure 2 GO classification of the ERα associated genes identified by SAGE. Percent of coverage representing the percentage of genes annotated with a specific GO term related to Biological Processes (blue bars) and Molecular Function (yellow bars). We observed that 31% of ERα associated transcripts are involved in biological processes related to cell growth and/or maintenance , 21% are related to cell communication , and 16% are related to regulation of transcription . Approximately 16% of these deregulated genes are related to molecular functions associated with DNA binding and more specifically with transcription factor activity (10%) (Figure 2 ). Interestingly, using the enrichment GO terms analysis, we identified statistical significant over-representation of specific groups of proteins including: metal ion binding proteins (54 hits out of 419 annotated genes; p = 0.011), calcium ion binding proteins (27 hits out of 419; p = 0.032) and steroid hormone receptor activity related proteins (6 hits out of 419; p = 0.031) ( additional data file 2 ). The GO cluster related to steroid hormone receptor activity proteins includes: estrogen receptor 1 ( ESR1, i.e. ERα ), androgen receptor ( AR ), hydroxysteroid 17-β dehydrogenase 4 ( HSD17β4 ), glucocorticoid receptor ( NR3C1 ), oxysterol binding protein ( OSBP ), and retinoic acid receptor α ( RARA ). The observation of functionally related groups of genes identified in the SAGE dataset via GO over representation analysis allows the identification of distinct biological pathways directly or indirectly associated to estrogen response related processes and provides the basis for future mechanistic studies. Identification of high-affinity Estrogen Response Elements We used a recently reported genome-wide high-affinity ERE database [ 32 ] to identify putative EREs in the promoter regions of the SAGE-identified 473 up-modulated genes in ERα (+) breast tumors. We identified 220 EREs distributed on 163 out of the 473 genes (35%) (see additional data file 3 ). Seventy-two percent of these genes contain one high affinity ERE (117 out of 163) and 28% of them contain two or more EREs in proximity to the transcriptional start sites (TSS) (46 out of 163) (Figure 3a ). These EREs can be located in both coding and non-coding sequences such as was described by Bourdeau et al . [ 32 ]. Figure 3 High-affinity EREs in ERα (+) up-modulated genes (n = 163). a) Percentage of genes according to number of EREs. b) Distribution of EREs in 5' (blue bars) and 3' (aquamarine bars) regions relative to the TSS (-10 to + 5 kb). Each bar represents an interval width of 500 bp. The observed frequency of these elements in our study was 220 EREs in 3260 kb (considering a DNA window of 20 kb for each one of the 163 up-modulated genes with EREs). Compared with the expected frequency from random distribution of high-affinity EREs found in the genome (732 EREs in 3,069334 kb 0.8 ERE in 3260 kb) (see material and methods) [ 32 ], the number of individual EREs was 270 fold higher than expected by chance (p < 0.00001). Fifty percent (110 out of 220) of the detected EREs mapped within a 10 kb region 5' of the TSS, while the rest mapped to 3' regions (Figure 3b ). Approximately 68% of EREs mapped within the region between -5 to +5 kb from the TSS; in agreement with those observations of Bourdeau et al . [ 32 ]. However, it remains to be determined whether distantly located EREs (e.g. -10 kb from the TSS) are functional E 2 -ER binding sites related to transcriptional activation. Of the validated transcripts previously discussed (Figure 1 ), we detected high-affinity EREs on the upstream or downstream regions related to the TSS of SYTL4 (-8384 bp from the TSS: tggacatcatgacct), TSPAN1 (+974 bp and +9384 bp from the TSS: tggtctgaatgaccc and aggtcatttccacct respectively), CELSR2 (+173 bp and +3607 bp from the TSS: tgctcagggtgaccc and aggtcaccatgaccg respectively), and NR4A1 (-3478 bp and +4217 bp from the TSS: tgttcactctgacct). It is interesting to note that we were unable to identify high-affinity EREs on the majority of deregulated genes (65%) associated with a positive ER α status. The possibility exists that many of these genes are transcriptionally regulated by non-ERE mediated mechanisms such as those involving ER binding to the AP1 or SP1 transcription factors [ 33 ]. The AP1 transcription factor is a heterodimer formed by Jun and Fos family member proteins that binds to the phorbol diester (TPA) response element as well as to the AP1 consensus DNA sequence. In this pathway, ER plays a co-activator role for AP1 [ 6 ]. The ER/AP-1 complex can confer estrogen responsiveness to additional subset of genes found in our dataset such as: ovalbumin (Fold change: 3; p = 0.033) and c-fos (Fold change: 2.1; p = 0.033); two transcripts detected as over-expressed in ERα (+) breast tumors by SAGE ( additional data file 1 ). Similarly the ER/SP1 complex confers estrogen responsiveness to genes such as: retinoic acid receptor α ( RARA ) (Fold change: 6.7; p = 0.038), vascular endothelial growth factor ( VEGFC ) (Fold change: 2.6; p = 0.037), insulin-like growth factor binding protein-4 ( IGFBP4 ) (Fold change: 2; p = 0.01) and heat shock protein 27 ( HSPB1 ) (Fold change: 2; p = 0.045); four transcripts detected as over-expressed in ERα (+) tumors in our study ( additional data file 1 ). An additional pathway of transcription regulation by estrogen involves the ER-related receptors (ERR), nuclear orphan receptors with significant homology to ERs, which do not bind estrogen and have unknown physiological ligands. ERRs are known to bind to the steroidogenic factor 1 response element (SFRE) and also bind to classic EREs, by means of which they exert constitutive transcriptional activity [ 34 ]. We detected over-expression of the nuclear orphan receptor NR4A1 by SAGE and subsequently validated this observation by real time RT-PCR (Figure 1g ). Interestingly, and as previously mentioned, the genomic region 5' and 3' to the TSS of NR4A1 contain high-affinity EREs. Interaction between ERs and ERRs has been observed in the transcriptional regulation of certain genes such as the human breast cancer related gene TFF1/pS2, the promoter of which is not only activated by ERs but also by ERRs [ 35 ]. As described, ERα can mediate estrogenic response through multiple genomic and non-genomic mechanisms, many of which affect proteins and pathways not necessarily directly or exclusively associated with ERα. Thus it is worth stressing that it will the totality of deregulated proteins the ones that ultimately define the phenotype of ERα (+) breast carcinomas regardless of whether a "direct association" with ER transcriptional regulation exists or not. In vivo versus in vitro estrogen induced global gene expression findings The SAGE profiles for E 2 -responsive genes in MCF-7 cell line, previously reported by us [ 36 ], was compared with the ER status genes expression profile found in primary breast carcinomas. Briefly, we detected 199 transcripts differentially expressed (p < 0.01) in MCF-7 treated cells, 124 were up-regulated and 75 were down-regulated transcripts. Basically and as reported Charpentier et al , we observed a general up-regulation cell cycle progression-related genes including: CCT2 , CCND1 , PES1 , RAN/TC4 , CALM1 , CALM2 ; and tumor-associated genes such as: RFP , D52L1 , TFF1/PS2 , CAV1 , and NDKA among others [ 36 ]. These together could contribute to the stimulation of proliferation and the suppression of apoptosis by E 2 -ER transcriptional regulation. By comparing the in vitro (199 differentially expressed transcripts) and in vivo (520 differentially expressed transcripts) gene expression profiles, to our surprise we detect that only few transcripts: TFF1, CCND1, H19, SREBF1 and WWP1 behaved similarly (i.e. up-regulation) in both studies. This is similar to observations made previously by Meltzer and co-workers whom showed that the majority of genes regulated in cell culture do not predict ER status in breast carcinomas [ 11 , 37 ]. This result suggests that the estrogen-responsive pathways affected in vitro represent only a minor portion of the global gene expression profiles characteristic of ERα (+) breast tumors. This maybe in great part the result of the heterogenous nature of bulk tumor tissue but in addition, the in vitro response of a single cell line to E 2 , in this particular case the widely used MCF-7 cells, may not faithfully reproduce the physiological effects of ER signaling in vivo . Cross-platform gene expression profiling comparison In order to identify and validate the most reliable set of genes able to discriminate breast carcinomas based on their ERα status, we performed a cross-platform comparison between the described SAGE dataset with two previously reported breast cancer studies based on DNA microarray methods [ 12 , 13 ]. van't Veer et al. [ 12 ] reported the gene expression profile of 97 primary breast tumors based on oligonucleotide microarrays containing 24,479 elements (Agilent Technologies, Palo Alto, CA, USA). In another study, Sotiriou et al. [ 13 ] reported the gene expression profile of 99 primary breast tumors using a cDNA microarray containing 7650 elements. Only files containing differentially expressed genes associated to ERα status tumors from both microarrays studies were obtained for cross-platform comparison (see material and methods). Among the three platforms, a total of 1686 transcripts were identified as over-expressed in ERα (+) breast tumors. One hundred and eighty-three genes were identified by more than one method (Figure 4 ; additional data file 4 ). Eleven of these 183 genes were identified by all three methods displaying over-expression in ERα (+) breast carcinomas: estrogen receptor 1 ( ESR1 ), GATA-binding protein 3 ( GATA3 ), mucin 1 ( MUC1 ), v-myb-myeloblastosis viral oncogene homolog ( C-MYB ) , X-box-binding protein 1 ( XBP1 ), hydroxysteroid 17-β dehydrogenase 4 ( HSD17B4 ), BTG family member 2 ( BTG2 ), transforming growth factor β-3 ( TGFB3 ), member RAS oncogene family ( RAB31 ), START domain containing 10 ( STARD10 ), and KIAA0089 (Table 2 ). Figure 4 Cross-platform comparisons of the up-modulated transcripts in ERα (+) breast carcinomas. One hundred and eighty-three genes were identified by more than one study, eleven of which were commonly identified across the three platforms. a) Comparison between SAGE and oligonucleotide microarray platforms [12] showing a highly significant number of overlapping genes (p < 0.001) (see table 2). b) Comparison between SAGE and cDNA microarray platforms [13] (p > 0.05). c) Statistically significant number of overlapping genes identified by both DNA microarrays platforms (p < 0.01). Table 2 Transcripts identified as over-expressed in ERα (+) breast cancers commonly detected by cross-platforms comparison (SAGE and oligonucleotide microarrays). Gene name Locus Link ID Fold change Frequency Gene name Locus Link Fold change Frequency # TFF1* 7031 51.4 15/18 (83%) SULF2 55959 2.9 11/18 (61%) SYTL4* 94121 28.0 15/18 (83%) THBS4 7060 2.9 8/18 (44%) DUSP4 1846 14.7 14/18 (78%) AZGP1 563 2.8 9/18 (50%) NAT1 9 11.7 15/18 (83%) BBC3* 27113 2.8 12/18 (67%) ECM1* 1893 10.1 13/18 (72%) NET7* 23555 2.8 10/18 (55%) KIAA0882 23158 10.1 18/18 (100%) NET6 27075 2.8 12/18 (67%) JDP1 56521 10.0 12/18 (67%) TRAF5 7188 2.8 9/18 (50%) ESR1 2099 9.8 18/18 (100%) BTG2 7832 2.7 9/18 (50%) HDGFRP3 50810 9.8 12/18 (67%) RNF123* 63891 2.7 11/18 (61%) TCEAL1 9338 9.8 13/18 (72%) CHAD* 1101 2.6 12/18 (67%) TSPAN1* 10103 9.5 15/18 (83%) CSNK1A1 1452 2.6 14/18 (78%) C20orf103* 24141 9.3 10/18 (55%) EVL 51466 2.6 12/18 (67%) MYLIP 29116 9.3 11/18 (61%) HIST1H2BD 3017 2.6 10/18 (55%) ABCA3 21 8.9 10/18 (55%) SUSD3 203328 2.6 9/18 (50%) SEC14L2 23541 8.7 9/18 (50%) PLAT* 5327 2.6 8/18 (44%) ANXA9* 8416 8.4 10/18 (55%) RARRES3* 5920 2.6 11/18 (61%) KCTD3 51133 8.4 17/18 (94%) SH3BGRL* 6451 2.6 8/18 (44%) SCUBE2 57758 7.5 14/18 (78%) TPBG* 7162 2.6 9/18 (50%) ITGBL1 9358 7.1 8/18 (44%) UGCG 7357 2.6 11/18 (61%) C14orf168 83544 6.7 6/18 (33%) CELSR2* 1952 2.5 8/18 (44%) FBP1 2203 6.7 14/18 (78%) CRIM1 51232 2.5 11/18 (61%) MYB 4602 6.7 14/18 (78%) FLJ90798* 219654 2.5 9/18 (50%) RARA* 5914 6.7 12/18 (67%) KIF12 113220 2.5 7/18 (39%) CaMKIINα 55450 6.3 18/18 (100%) LRIG1 26018 2.5 9/18 (50%) AR* 367 6.2 10/18 (55%) LRP2* 4036 2.5 10/18 (55%) ZNF552 79818 6.2 16/18 (89%) PHF15* 23338 2.5 12/18 (67%) MIPEP* 4285 6.0 14/18 (78%) HSMNP1 55861 2.4 8/18 (44%) BAI2 576 5.3 15/18 (83%) LOC123169 123169 2.4 12/18 (67%) DP1L1 92840 5.3 15/18 (83%) PINK1* 65018 2.4 11/18 (61%) VAV3 10451 5.3 12/18 (67%) PRKAR2B 5577 2.4 7/18 (39%) KIAA0089 23171 5.2 17/18 (94%) TJP3* 27134 2.4 11/18 (61%) GATA3 2625 5.1 15/18 (83%) CCND1 595 2.3 9/18 (50%) QDPR 5860 5.1 11/18 (61%) CYBRD1 79901 2.3 10/18 (55%) C1orf21 81563 4.9 11/18 (61%) KRT18 3875 2.3 10/18 (55%) KIAA1143 57456 4.9 7/18 (39%) PURA 5813 2.3 9/18 (50%) OIP106 22906 4.9 16/18 (89%) SREBF1* 6720 2.3 10/18 (55%) AGR2 10551 4.6 10/18 (55%) CYB5R1 51706 2.2 6/18 (33%) MGC4251 84336 4.6 13/18 (72%) DLG3* 1741 2.2 9/18 (50%) FER1L3 26509 4.4 10/18 (55%) EEF1A2 1917 2.2 11/18 (61%) C4A 720 4.1 11/18 (61%) GSTZ1 2954 2.2 9/18 (50%) CRIP2 1397 4.0 15/18 (83%) LOC159090 159090 2.2 6/18 (33%) NTN4 59277 4.0 10/18 (55%) MGC11242* 79170 2.2 10/18 (55%) GJA1 2697 3.8 11/18 (61%) MGC18216* 145815 2.2 8/18 (44%) CGI-111* 51015 3.7 14/18 (78%) NEIL1 79661 2.2 6/18 (33%) CROT* 54677 3.6 15/18 (83%) XBP1* 7494 2.2 8/18 (44%) DACH 1602 3.6 13/18 (72%) IRX5 10265 2.1 8/18 (44%) DKFZP564D172 83989 3.6 10/18 (55%) RAB31 11031 2.1 9/18 (50%) FGD3 89846 3.6 10/18 (55%) SSBP2 23635 2.1 7/18 (39%) RNASE4* 6038 3.6 12/18 (67%) TGFB3 7043 2.1 8/18 (44%) GLUL* 2752 3.3 11/18 (61%) BMPR1B 658 2.0 7/18 (39%) FOXA1 3169 3.2 10/18 (55%) FLJ21174 79921 2.0 6/18 (33%) MGC7036 196383 3.2 14/18 (78%) FLJ22386 79641 2.0 7/18 (39%) MUC1* 4582 3.2 12/18 (67%) HSPB1* 3315 2.0 6/18 (33%) NAV1 89796 3.1 13/18 (72%) IGFBP4* 3487 2.0 8/18 (44%) RPLP1* 6176 3.1 12/18 (67%) MGC15737* 85012 2.0 8/18 (44%) ALCAM 214 2.9 9/18 (50%) SPARCL1 8404 2.0 9/18 (50%) HSD17B4* 3295 2.9 13/18 (72%) STARD10 * 10809 2.0 7/18 (39%) * Genes with known or putative high-affinity EREs mapping in the vicinity of the TSS. # Transcripts tags changing > 2-fold when compared with the average expression of ERα (-) tumors. Underlined genes correspond to the transcripts cross-validated among all three compared platforms. One hundred and fourteen genes were identified as over-expressed by oligonucleotide microarrays [ 12 ] and SAGE in ERα (+) tumors, representing a non-random significant number of overlapping genes based on normal approximation to the binomial distribution (p < 0.001) (Figure 4 and Table 2 ). Sixty-six genes were identified as over-expressed in ERα (+) tumors by both DNA microarrays platforms (p < 0.01). The set of 25 genes overlapping between cDNA microarrays [ 13 ] and SAGE were not statistical significant (p > 0.05). Interestingly, we found a higher number of overlapping genes between the oligonucleotide microarray and SAGE platforms (114 genes), while only 66 genes were observed overlapping when comparing both microarray platforms. It is worth noting that 96% of the 470 genes (Figure 4 ) identified as overexpressed by the cDNA microarray method [ 13 ] were included within the total set of elements in the oligonucleotide microarray platform [ 12 ]. In other words, it appears that a better correlation was observed between SAGE and oligonucleotide arrays, than between both DNA microarray methods. Conclusion In summary, our comprehensive comparison of overlapping genes across different gene expression platforms provides validation for a significant number of transcripts identified as highly expressed in ERα (+) breast tumors. More importantly this analysis identifies the most promising biomarkers for further evaluation as ERα associated genes in breast cancer. Furthermore, the identified proteins may be of value as breast cancer prognostic indicators analyzed either as a group or individually. It is also likely that groups of co-regulated genes in ERα (+) breast cancers may be associated to the hormonal control of mammary epithelial cells growth and differentiation. Finally, a better understanding of the signaling networks controlled or associated with the estrogen response may lead to the identification of novel breast cancer therapeutic targets. Methods SAGE libraries To perform the comparative breast cancer SAGE analysis based on ERα status, we analyzed 26 Stage I – Stage II invasive breast carcinomas (8 ERα-negative tumors and 18 ERα-positive tumors). To this end, we generated and sequenced 24 breast cancer SAGE libraries at an approximate resolution of 100,000 tags per library, combined with 2 additional breast cancer libraries (ERα-negative tumors) downloaded from the Cancer Genome Anatomy Project – SAGE Genie database (SAGE_Breast_Carcinoma_B_95-259 and B_IDC_4) . For the generation of our SAGE libraries, snap frozen samples were obtained from the M.D. Anderson breast cancer tumor bank, and SAGE analysis was performed as previously described [ 36 , 38 ]. Data processing and statistical analysis of SAGE libraries SAGE tag extraction from sequencing files was performed by using the SAGE2000 software version 4.0 (a kind gift of Dr. K. Kinzler, John Hopkins University). SAGE data management, tag to gene matching as well as additional gene annotations and links to publicly available resources such as GO, UniGene, LocusLink, were performed using a suite of web-based SAGE library tools developed by us . In our analyses we only considered tags with single tag-to-gene reliable matches. To compare these SAGE libraries, we utilized a modified t-test recently developed by us [ 18 ]. This test is based on a beta binomial sampling model that takes into account both, the intra-library and the inter-library variability, thus identifying 'common patterns' of SAGE transcript tag changes systematically occurring across samples [ 18 ]. All raw SAGE data reported as Supplementary tables in this manuscript is publicly available at . Real Time RT-PCR analysis Template cDNAs were synthesized on mRNAs isolated from an independent set of 36 Stage I – Stage II human breast carcinomas (13 ERα-negative tumors and 23 ERα-positive tumors) obtained from our tumor bank. Primers and probes were obtained from the TaqMan Assays-on-Demand™ Gene Expression Products (Applied Biosystems, Foster City, CA, USA). All the PCR reactions were performed using the TaqMan PCR Core Reagents kit and the ABI Prism ® 7700 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Experiments were performed in duplicate for each data point and 18s rRNA was used as control. Results were expressed as mean ± 2 Standard Error based on Log2 transformation of normalized real time RT-PCR values of the assayed genes. We used t-test to compare the gene expression levels of validated genes between ERα (+) and ERα (-) breast tumors (p < 0.05). Immunohistochemical determination of ER status IHC staining and ER status determination was performed by the Pathology Department, MDACC following routine immunohistochemical procedures. Briefly, five micrometer sections of invasive breast carcinomas paraffin embedded tissues were used. Endogenous peroxidase activity was blocked with 3% H 2 O 2 in methanol for 10 min. After pretreatment with Tris-EDTA buffer, in order to block non-specific antibody binding, the slides were incubated with 10% goat serum in PBS for 30 min. Primary monoclonal ERα antibody (ER-6F11, Novocastra, Newcastle, UK) was used at 1:50 dilution and detected following standard immunohistochemical techniques. DAB was used as chromogen and Mayers hematoxylin is used as counterstain. Scoring was performed by breast pathologist (AS). Cuttoff for positivity was determined at 5% of tumor cells staining positively for ER (i.e. < 5% of cells the tumor was considered negative for ERα). Bioinformatics analysis For automated functional annotation and classification of genes of interest based on GO terms, we used the EASE [ 31 ] available at the Database for Annotation, Visualization and Integrated Discovery ( DAVID ) at [ 39 ]. The EASE software calculates over-representation of specific GO terms with respect to the total number of genes assayed and annotated. Statistical measures of specific enrichment of GO terms are determined by means of an EASE score (p < 0.05). The EASE score is a conservative adjustment of the Fisher exact probability that weights significance in favor of biological themes supported by more genes and is calculated using the Gaussian hypergeometric probability distribution that describes sampling without replacement from a finite population [ 31 ]. This allows one to identify biological themes within a specific list of EASE analyzed genes. High-affinity Estrogen Response Elements (ERE) analysis To identify the occurrence of EREs within the promoter regions of up-modulated genes in ERα (+) breast tumors, we used a human genome-wide high-affinity ERE database [ 32 ]. This public available database contains 71,119 EREs identified across the human genome (related to 17,353 transcriptional start sites), representing the consensus ERE (5'-Pu-GGTCA-NNN-TGACC-Py-3'), and equivalent sequences with only one or two nucleotide variations from such consensus. Based on these restrictions the expected random frequency was calculated as the total number of base pairs in the human genome divided by the frequency of occurrence of a sequence with specified base pairs at 10 positions and two base pair choices at two positions (3,069334246/4 11 = 732 high-affinity EREs) [ 32 ]. Comparison of gene expression patterns identified by different methodologies ERα status associated genes identified in previous breast cancer studies [ 12 , 13 ] using DNA microarray methods were compared with our SAGE findings. All over-expressed genes in ERα (+) breast tumors obtained from these studies were downloaded from the corresponding web sites ( and ) [ 12 , 13 ]. These datasets were annotated by LocusLink ID using the EASE software [ 26 ], and then compiled into one Excel spreadsheet pivotTable for comparison of overlapping genes between platforms, i.e. SAGE, Oligonucleotide and cDNA arrays. Anonymous ESTs from the microarrays platforms were excluded due to the inability to cross validate the identities between different gene expression profiles. Any combination of two lists was compared for matching gene-identity. The number and identity of genes commonly affected in two platforms ( e.g. SAGE study vs. DNA microarray) was determined. We used the normal approximation to the binomial distribution as previously described [ 40 ] to calculate whether the number of matching genes derived from each cross-platform comparison was of statistical significance (p < 0.05). Authors' contributions M.C.A. conceived the study idea and carried out the real time RT-PCR validations, the biostatistical/ bioinformatics analysis and writing the manuscript. Y.H. , H.S. and J.A.D. carried out the breast cancer SAGE libraries and provided practical feedback on aspects of the manuscript. K.B. and S.G. developed the biostatistical and web-page base methodology. A.S. provides the tissue samples and clinical information. C.M.A. is the principal investigator and was involved in the conceptualization, design and writing of the manuscript. All authors read and approved the final manuscript. Competing interests The author(s) declare that they have no competing interests. Supplementary Material Additional File 1 Differentially expressed genes between ERα (+) vs. ER α (-) breast carcinomas (Fold change ≥2; p < 0.05). Click here for file Additional File 2 Gene Ontology overrepresentation analysis. Click here for file Additional File 3 High-affinity EREs identified in ERα (+) up-modulated genes. Click here for file Additional File 4 Cross-platform comparison of the up-modulated transcripts in ERα (+) breast carcinomas. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555753.xml

521171

Sensitivity to Oxidative Stress in DJ-1-Deficient Dopamine Neurons: An ES- Derived Cell Model of Primary Parkinsonism

The hallmark of Parkinson's disease (PD) is the selective loss of dopamine neurons in the ventral midbrain. Although the cause of neurodegeneration in PD is unknown, a Mendelian inheritance pattern is observed in rare cases, indicating a genetic factor. Furthermore, pathological analyses of PD substantia nigra have correlated cellular oxidative stress and altered proteasomal function with PD. Homozygous mutations in DJ-1 were recently described in two families with autosomal recessive Parkinsonism, one of which is a large deletion that is likely to lead to loss of function. Here we show that embryonic stem cells deficient in DJ-1 display increased sensitivity to oxidative stress and proteasomal inhibition. The accumulation of reactive oxygen species in toxin-treated DJ-1-deficient cells initially appears normal, but these cells are unable to cope with the consequent damage that ultimately leads to apoptotic death. Furthermore, we find that dopamine neurons derived from in vitro–differentiated DJ-1-deficient embryonic stem cells display decreased survival and increased sensitivity to oxidative stress. These data are consistent with a protective role for DJ-1, and demonstrate the utility of genetically modified embryonic stem cell–derived neurons as cellular models of neuronal disorders.

Introduction Parkinson's disease (PD) is a progressive neurodegenerative disorder characterized by rigidity, slowed movement, gait difficulty, and tremor at rest ( Dauer and Przedborski 2003 ). The pathological hallmark of PD is the relatively selective loss of dopamine neurons (DNs) in the substantia nigra pars compacta in the ventral midbrain. Although the cause of neurodegeneration in PD is unknown, a Mendelian inheritance pattern is observed in approximately 5% of patients, suggesting a genetic factor. Extremely rare cases of PD have been associated with the toxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine, which is taken up specifically by DNs through the dopamine transporter and is thought to induce cellular oxidative stress. Population-based epidemiological studies have further supported roles for genetic and environmental mechanisms in the etiology of PD ( Dauer and Przedborski 2003 ; Jenner 2003 ). The identification of several genes that underlie familial forms of PD has allowed for the molecular dissection of mechanisms of DN survival. Autosomal dominant mutations in α-synuclein lead to a rare familial form of PD ( Polymeropoulos et al. 1997 ), and there is evidence that these mutations generate toxic, abnormal protein aggregates ( Goldberg and Lansbury 2000 ) and cause proteasomal dysfunction ( Rideout et al. 2001 ). A majority of patients with sporadic PD harbor prominent intracytoplasmic inclusions, termed Lewy bodies, enriched for α-synuclein ( Spillantini et al. 1998 ), as well as neurofilament protein ( Trojanowski and Lee 1998 ). Mutations in a second gene, Parkin, lead to autosomal recessive PD ( Hattori et al. 2000 ). Parkin is a ubiquitin ligase that appears to participate in the proteasome-mediated degradation of several substrates ( Staropoli et al. 2003 ). Homozygous mutations in a third gene, DJ-1, were recently associated with autosomal recessive primary parkinsonism ( Bonifati et al. 2003 ). DJ-1 encodes a ThiJ domain protein of 189 amino acids that is broadly expressed in mammalian tissues ( Nagakubo et al. 1997 ). Interestingly, DJ-1 was independently identified in a screen for human endothelial cell proteins that are modified with respect to isoelectric point in response to sublethal doses of paraquat ( Mitsumoto and Nakagawa 2001 ; Mitsumoto et al. 2001 ), a toxin that generates reactive oxygen species (ROS) within cells and has been associated with DN toxicity ( McCormack et al. 2002 ). Gene expression of a yeast homolog of DJ-1, YDR533C, is upregulated in response to sorbic acid ( de Nobel et al. 2001 ), an inducer of cellular oxidative stress. These results suggest a causal role for DJ-1 in the cellular oxidative stress response. Surprisingly, animal models that harbor genetic lesions that mimic inherited forms of human PD, such as homozygous deletions in Parkin ( Goldberg et al. 2003 ; Itier et al. 2003 ) or overexpression of α-synuclein ( Masliah et al. 2000 ; Giasson et al. 2002 ; Lee et al. 2002 ), have failed to recapitulate the loss of dopamine cells. An alternative approach, the genetic modification of midbrain DNs in vitro ( Staropoli et al. 2003 ), is potentially useful but limited by the difficulty and variability in culturing primary postmitotic midbrain neurons. Other studies have focused on immortalized tumor cell lines, such as neuroblastoma cells, but these may not accurately model the survival of postmitotic midbrain neurons. Here we show that DJ-1-deficient cells display increased sensitivity to oxidative stress. DNs appear to be particularly sensitive to the loss of DJ-1. The initial accumulation of ROS is normal in DJ-1-deficient cells, but subsequent cellular defenses to ROS are impaired, leading to increased apoptosis. Results Generation of DJ-1-Deficient ES Cells To investigate the normal cellular function of DJ-1 and the pathogenic mechanism of the PD mutations, we generated cells deficient in DJ-1. A murine embryonic stem (ES) cell clone, F063A04, that harbors a retroviral integration at the DJ-1 locus was obtained through the German Gene Trap Consortium ( http://tikus.gsf.de ) ( Figure 1 A; Floss and Wurst 2002 ). This integration is predicted to disrupt the normal splicing of DJ-1, leading to the generation of a truncated protein that lacks the carboxy-terminal domain required for dimerization and stability (unpublished data). Of note, a mutation that encodes a similarly truncated protein (at the human DJ-1 exon 7 splice acceptor) has been described in a patient with early-onset PD ( Hague et al. 2003 ). Figure 1 DJ-1-Deficient ES Cells Are Sensitized to Oxidative Stress (A) Schematic map of the murine DJ-1 gene in clone F063A04. The retroviral insertion places the engrailed-2 (En2) intron, the splice acceptor (SA), and the β-galactosidase/neomycin resistance gene fusion (β-geo) between exons 6 and 7. (B) Southern blot analysis of KpnI-digested genomic DNA from DJ-1 homozygous mutant (–/–), WT (+/+), and heterozygous (+/–)cells, probed with murine DJ-1 cDNA. WT DNA shows a predicted 14-kb band (WT), whereas the mutant allele migrates as a 9-kb band (insertion). (C) Western blot (WB) of ES cell lysates from WT (+/+), DJ-1 heterozygous (+/–), and mutant homozygous (–/–) clones with antibodies to murine DJ-1 (α-DJ-1) or β-actin (α-β-actin). DJ-1 migrates at 20 kDa, β-actin at 45 kDa. (D) ES cells were exposed to 0, 5, 10, and 20 μM H 2 O 2 for 15 h and viability was assayed by MTT. Responses of DJ-1 heterozygous cells (diamonds) and DJ-1 knockout clones 9 (open circles), 16 (solid circles), 23 (squares), and 32 (triangles) are shown. ** p ≤ 0.01; *** p ≤ 0.0001. (E and F) Cell death of DJ-1 heterozygous and DJ-1-deficient cells (clone 32) after exposure to H 2 O 2 (10 μM) was quantified by staining with PI and an antibody to AV with subsequent FACS analysis. AV staining marks cells undergoing apoptosis, whereas PI staining indicates dead cells. * p ≤ 0.05. (G) DJ-1 heterozygous (+/–) and knockout (clone 32; –/–) cells were assayed at 1, 6, and 24 h after treatment with 10 μM H 2 O 2 by Western blotting for cleaved PARP (89 kDa), which indicates apoptosis. No band is seen for cleaved PARP or β-actin for the DJ-1-deficient cells at 24 h due to cell death. Data represent means ± standard error of the mean (SEM) and were analyzed by ANOVA with Fisher's post-hoc test. To generate ES cell subclones homozygous for the trapped DJ-1 allele, clone F063A04 was exposed to a high dose of the antibiotic G418, which selects cells that are homozygous for the neomycin resistance gene insertion ( Mizushima et al. 2001 ). Several homozygous mutant ES cell subclones (that had undergone gene conversion at the DJ-1 locus) were identified by Southern blotting ( Figure 1 B). To confirm that the trapped allele leads to the loss of wild-type (WT) DJ-1 protein, cell lysates from homozygous DJ-1-deficient (also termed “knockout” here) ES cell clones as well as the parental heterozygous clone were analyzed by Western blotting using polyclonal antibodies to the amino-terminal region of DJ-1 (amino acids 64–82) or full-length DJ-1 protein (unpublished data). Neither full-length nor truncated DJ-1 protein products were detected in knockout clones ( Figure 1 C), consistent with instability of the predicted truncated DJ-1 product. In addition, no full-length DJ-1 RNA was detected in cultures of knockout cells ( Figure S1 ). In contrast, heterozygous and WT ES cells express high levels of DJ-1. Initial phenotypic analysis of knockout subclones indicated that DJ-1 is not essential to the growth rate of ES cells in culture, consistent with the viability of humans homozygous for DJ-1 mutations. DJ-1 Protects Cells from Oxidative Stress and Proteasomal Inhibition DJ-1 has been hypothesized to function in the cellular response to oxidative stress. To investigate the role of DJ-1 in the oxidative stress response in vivo, DJ-1-deficient knockout and heterozygous ES cell clones were analyzed for cell viability in the context of increasing concentrations of H 2 O 2 . Heterozygous cells were used as controls because the knockout subclones were derived from these. Cell viability was initially determined by MTT assay (which detects reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide [MTT] by metabolic enzymes) in triplicate ( Fezoui et al. 2000 ). Exposure to H 2 O 2 led to significantly greater toxicity in the DJ-1-deficient cells; similar results were obtained with multiple knockout subclones in independent experiments ( Figures 1 D and 2 A). In contrast, in the absence of toxin, heterozygous and knockout cells displayed comparable viability in the MTT assay ( Figure S2 ). Consistent with the MTT assay, fluorescence-activated cell sorting (FACS) analysis of cells stained with annexin V (AV) and propidium iodide (PI) revealed increased death of knockout cells compared to heterozygous cells in the context of H 2 O 2 exposure ( Figure 1 E). The increase in AV-positive cells implicated an apoptotic mechanism of cell death ( Figure 1 F). Furthermore, when exposed to H 2 O 2 , knockout cells displayed potentiated cleavage of poly(ADP-ribose)polymerase-1 (PARP) in a pattern indicative of an apoptotic death program ( Gobeil et al. 2001 ) ( Figure 1 G). Figure 2 Specificity and Mechanism of Altered Toxin Sensitivity in DJ-1-Deficient Cells (A–C) Cell viability of DJ-1 heterozygous cells (solid bar) and DJ-1-deficient knockout clone 32 cells (open bar) after 15 h exposure to H 2 O 2 (A), lactacystin (B), or tunicamycin (C) as assayed by MTT reduction. * p ≤ 0.05. (D) DJ-1-deficient knockout cells (clone 32) were transiently transfected with plasmids containing WT human DJ-1 vector (solid bar) and PD-associated L166P mutant DJ-1 vector (gray bar); as a control, knockout cells were also transfected with vector alone (open bar). 48 h after transfection, cells were exposed to 10 μM H 2 O 2 for 15 h and then assayed by MTT reduction. WT human DJ-1 significantly enhanced survival of the knockout cells, whereas the L166P mutant did not. Similar results were obtained at 20 μM H 2 O 2 and with a second DJ-1-deficient clone (unpublished data). Transfection efficiency exceeded 90% in all cases and protein expression level was comparable for human WT and L166P mutant DJ-1 as determined by Western blotting ( Figure S1 ). * p ≤ 0.05. (E) DJ-1-deficient cells (clone 32; open bar) and control heterozygous cells (solid bar) were assayed for intracellular formation of ROS in response to H 2 O 2 treatment (15 min, 1 or 10 μM) using DHR and FACS analysis. (F) Protein carbonyl levels were measured by spectrophotometric analysis of DNP-conjugated lysates from DJ-1-deficient (clone 32; solid red line) and control heterozygous cells (dashed blue line). Data are shown as the mean ± SEM and were analyzed by ANOVA with Fisher's post-hoc test. Additional toxin exposure studies demonstrated that DJ-1-deficient cells were sensitized to the proteasomal inhibitor lactacystin ( Figure 2 B), as well as to copper ( Figure S2 ), which catalyzes the production of ROS. We did not observe altered sensitivity to several other toxins, including tunicamycin (an inducer of the unfolded protein response in the endoplasmic reticulum; Figure 2 C), staurosporine (a general kinase inhibitor that induces apoptosis) ( Figure S2 ), or cycloheximide (an inhibitor of protein translation) (unpublished data). WT but Not PD-Associated L166P Mutant DJ-1 Protects Cells from Oxidative Stress To confirm that altered sensitivity to oxidative stress is a consequence of the loss of DJ-1, we performed rescue experiments by overexpressing WT or mutant human DJ-1 in knockout ES cells. Plasmids encoding human Flag-tagged WT DJ-1, Flag-tagged PD-associated L166P mutant DJ-1, or vector alone, were transiently transfected into DJ-1-deficient clones, and these were subsequently assayed for sensitivity to H 2 O 2 using the MTT viability assay. DJ-1-deficient cells transfected with a vector encoding Flag-WT human DJ-1 were effectively rescued in terms of viability in the presence of H 2 O 2 ( Figure 2 D); Thus, viability in rescued knockout cells mimicked the viability of untransfected heterozygous cells in the context of H 2 O 2 treatment ( Figure 2 A and 2 D). In contrast, transfection of knockout cells with a vector encoding the PD-associated L166P mutant DJ-1 did not significantly increase the viability of H 2 O 2 -treated knockout cells ( Figure 2 D). Baseline cell viability in the absence of toxin exposure was not altered by DJ-1 overexpression, and Western blotting of lysates from transfected cells with an antibody specific to human DJ-1 demonstrated that transfected Flag-WT DJ-1 and Flag-L166P mutant DJ-1 accumulated comparably ( Figure S2 ). DJ-1 Deficiency Does Not Alter the H 2 O 2 -Induced Intracellular ROS Burst We hypothesized that DJ-1 either alters the initial accumulation of intracellular ROS in response to H 2 O 2 exposure, or that it functions downstream of the ROS burst and protects cells from consequent damage. Therefore, we quantified the accumulation of ROS in response to H 2 O 2 treatment in knockout and heterozygous cells using the ROS-sensitive fluorescent indicator dye dihydrorhodamine-123 (DHR) and FACS analysis. Initial ROS accumulation (at 15 min after stimulation) appeared unaltered in the DJ-1-deficient cells in comparison to control heterozygous cells ( Figure 2 E). Consistent with this, accumulation of protein carbonyls, an index of oxidative damage to proteins ( Sherer et al. 2002 ), appeared normal initially (at 1 h after toxin exposure; Figure 2 F). However, at 6 h after toxin exposure, a point at which knockout cells already display increased apoptosis (as indicated by PARP cleavage; see Figure 1 G), protein carbonyl accumulation was robustly increased in the DJ-1-deficient cells. These data suggest that initial ROS accumulation is not altered by DJ-1 deficiency, but that the mutant cells are unable to appropriately cope with the consequent damage. Consistent with this result, no antioxidant or peroxiredoxin activity with purified DJ-1 protein in vitro was detected (S.S. and A.A., personal communication). DJ-1 Is Required for Survival of ES Cell–Derived DNs Several methods have been established for the differentiation of ES cells into DNs in vitro ( Morizane et al. 2002 ). To extend our analysis of DJ-1 function to DNs, we differentiated DJ-1-deficient ES cells or control heterozygous cells into DNs in vitro by coculture with stromal cell–derived inducing activity (SDIA; Figure 3 A) ( Morizane et al. 2002 ; Barberi et al. 2003 ). DNs were quantified by immunohistochemistry for tyrosine hydroxylase (TH; a marker for DNs and other catecholaminergic cells), or by analysis of dopamine transporter uptake activity (a quantitative DN marker) ( Han et al. 2003 ). Production of DNs appeared to be significantly reduced in knockout ES cell cultures compared to parental heterozygous cultures at 18 days in vitro (DIV) as determined by both dopamine uptake and TH immunoreactivity ( Figures 3 B and 3 C; 4 A– 4 L). In contrast, general neuronal production did not appear altered in this assay in terms of the postmitotic neuronal marker TuJ1 (a monoclonal antibody specific to neuronal, not glial, class III β-tubulin) ( Figures 3 E and 4 A– 4 L′); other neuronal subtypes also appeared normal, including GABAergic ( Figures 3 D and 4 A′– 4 L′) and motor neurons (HB9-positive; Figure S3 ). To investigate whether the reduction in DNs in DJ-1-deficient cultures is due to defective generation or survival, a time course analysis was performed. At early time points (8 and 12 DIV), dopamine uptake activity was comparable in WT and DJ-1-deficient cultures, whereas subsequently the DJ-1-deficient cultures appeared defective ( Figure 3 F). Consistent with this, intracellular dopamine accumulation (as quantified using high-performance liquid chromatography) was significantly reduced in DJ-1-deficient cultures (6.4 ± 1.5 ng dopamine/mg protein) relative to control heterozygous cultures (66.0 ± 17.4 ng/mg) at 35 DIV. These data strongly suggest that DJ-1 deficiency leads to loss of DNs, rather than simply to downregulation of cell marker expression. Figure 3 DJ-1-Deficient ES Cell Cultures Display Reduced DN Production (A) The SDIA coculture method. DJ-1 knockout or control heterozygous ES cells are cocultured with mouse stromal cells (MS5) in the absence of serum and leukemia inhibitory factor for 18 DIV. (B) DN production was quantified at 18 DIV by 3 H-dopamine uptake assay. DJ-1-deficient ES cell cultures were defective relative to heterozygous control cultures. (C–D) Neuron production was quantified by immunohistochemical analysis as a percent of TuJ1-positive colonies that express TH (C) or GABA (D). Quantification of TH and GABA immunostaining was performed on all colonies in each of three independent wells. Colonies were scored as positive if any immunostained cells were present. * p ≤ 0.05. (E) The absolute number of TuJ1-positive colonies was not significantly different between the two genotypes. (F) Kinetic analysis of DN differentiation in DJ-1-deficient cultures (clone 32, solid square) and heterozygous controls (open circle) as quantified by 3 H-dopamine uptake assay. * p ≤ 0.05. (G) DJ-1-deficient (open bar) and heterozygous control (closed bar) cultures differentiated for 9 DIV and then exposed to 6-OHDA at the indicated dose for 72 h. DNs were quantified by 3 H-dopamine uptake assay. Data represent the means ± SEM and were analyzed by ANOVA followed by Fisher's post-hoc test. * p ≤ 0.05. Figure 4 Neuronal Differentiation of DJ-1-Deficient and Control Heterozygous ES Cell Cultures (A–L) DJ-1 heterozygous (+/–; A–F) and knockout (–/– [clone 32]; G–L) cultures were differentiated by SDIA for 18 DIV and immunostained with antibodies to TH (green) and TuJ1 (red). Images of both (Merge) are also shown. (A′–L′) Immunostaining of DJ-1 heterozygous (+/–, A′–F′) and deficient (–/–, G′–L′) cultures with antibodies for GABA (green) and TuJ1 (red). Scale bar, 50 μm. Images of both (Merge) are also shown. We hypothesized that DJ-1-deficient DNs may be sensitized to oxidative stress, akin to DJ-1-deficient undifferentiated ES cells. To test this, DN cultures from DJ-1-deficient or heterozygous control ES cell cultures at 9 DIV were exposed to oxidative stress in the form of 6-hydroxydopamine (6-OHDA), a DN-specific toxin that enters DNs through the dopamine transporter and leads to oxidative stress and apoptotic death ( Dauer and Przedborski 2003 ). DJ-1-deficient DNs displayed an increased sensitivity to oxidative stress in this assay ( Figure 3 G). Post-hoc analysis of the data indicates that the difference among genotypes is maximal at an intermediate dose of toxin (50 μM); at the highest dose of 6-OHDA employed (100 μM), the difference is lessened (because the heterozygote is increasingly affected as well), indicating that DJ-1-mediated protection is limited. Although we cannot exclude a role for DJ-1 in the late-stage differentiation of DNs, these data suggest that DJ-1 deficiency leads to reduced DN survival and predisposes these cells to endogenous and exogenous toxic insults. RNAi “Knockdown” of DJ-1 in Midbrain Embryonic DNs Leads to Increased Sensitivity to Oxidative Stress To confirm the role of DJ-1 in primary midbrain DNs, DJ-1 expression was inhibited by RNA interference (RNAi) in embryonic day 13.5 (E13.5) murine primary midbrain cultures by lentiviral transduction of short hairpin RNAs (shRNAs) ( Figure 5 ) ( Rubinson et al. 2003 ). E13.5 midbrain cultures ( Staropoli et al. 2003 ) were transduced with a lentiviral vector that includes a gene encoding the green fluorescent protein marker eGFP, along with shRNAs homologous to murine DJ-1. DJ1 -shRNA virus-infected cells displayed efficient silencing of DJ-1 gene expression to 10%–20% of control vector-infected cultures (as determined by Western blotting [ Figure 5 Q]). Transduction efficiency, as assessed by visualization of the fluorescent eGFP marker, exceeded 95% in all cases ( Figure 5 I and unpublished data). After 7 DIV, cultures were exposed to H 2 O 2 for 24 h and then evaluated for DN survival as quantified by immunostaining for TH and dopamine transporter (DAT). Figure 5 RNAi “Knockdown” of DJ-1 in Primary Embryonic Midbrain DNs Display Increased Sensitivity to Oxidative Stress (A–P) Primary midbrain cultures from E13.5 embryos were infected with lentiviral vectors encoding DJ-1 shRNA (or vector alone) under the regulation of the control vector (A–H) or the U6 promoter (I–P). Cells were cultured for 1 wk after infection and then exposed to H 2 O 2 (5 μM; E–H and M–P) for 24 h. Cultures were immunostained for TH (B, F, J, and N) or DAT (C, G, K, or O) and visualized by confocal microscopy. Images containing all stains are included (Merge; D, H, L, and P). Scale bar, 100 μm. (Q) Cell lysates prepared from midbrain primary cultures infected with DJ-1 shRNA lentivirus (or control vector) were analyzed by Western blotting for murine DJ-1 or β-actin. (R–T) Quantification of TH, DAT, and GFP signal was performed on ten randomly selected fields in each of three wells for each condition. Red triangles, DJ-1 shRNA treated; black circles, control vector. Data represent the means ± SEM and were analyzed by ANOVA followed by Fisher's post-hoc test. * p ≤ 0.05. Midbrain cultures transduced with DJ-1 shRNA virus and with control vector displayed similar numbers of TH-positive neurons in the absence of exposure to H 2 O 2 ( Figure 5 A–D, 5 I–L, and 5 R–S). In contrast, in the presence of H 2 O 2 , DJ-1-deficient cultures displayed significantly reduced DN survival as quantified by immunohistochemistry for TH or DAT ( Figure 5 E–H, 5 M–P, 5 R–S). These studies were repeated three times with similar results. The reduction in DAT immunoreactivity appears to be more robust than the reduction in TH-positive cell number in the context of H 2 O 2 ; this may reflect the differential localization of DAT to DN processes, whereas TH is primarily in the cell body. As we described in a previous manuscript, nondopaminergic cells in the E13.5 primary midbrain cultures are predominantly GABAergic neurons (90%–95%) ( Staropoli et al. 2003 ). Total embryonic midbrain neurons transduced with either DJ-1 shRNA or vector displayed comparable survival in the context of toxin exposure, suggesting that DJ-1 deficiency leads to a relatively specific alteration in DN survival ( Figure 5 T). These data are consistent with the analyses of ES cell–derived DNs above and indicate that DJ-1 is required for the normal survival of midbrain DNs in the context of toxin exposure. Discussion In this study we present evidence that DJ-1 is an essential component of the oxidative stress response of DNs. DJ-1-deficient cells display an apparently normal initial burst of ROS in response to H 2 O 2 , but they are unable to cope with the consequent toxicity, culminating in apoptosis. Additionally, we find that DJ-1 deficiency sensitizes cells to the proteasomal inhibitor lactacystin but not other toxic stimuli such as tunicamycin. Proteasomal inhibition induces the accumulation of short-lived and misfolded cytoplasmic proteins, leading to oxidative stress and apoptosis ( Demasi and Davies 2003 ). ROS and proteasomal inhibition have previously been correlated with PD pathology ( Dauer and Przedborski 2003 ), and it is therefore tempting to hypothesize that DJ-1 mutations lead to PD because of an increased sensitivity to such stressors. The apparent cell-type specificity of DN impairment in patients with the Parkinsonism-associated DJ-1 mutation is not predicted by the ubiquitous expression of DJ-1 ( Nagakubo et al. 1997 ). In this study, we find that DJ-1 protects both dopaminergic and nondopaminergic cells from oxidative insult. However, DJ-1-deficient DNs appear to be especially sensitive to oxidative insult, suggesting relative cell-type specificity to the consequences of DJ-1 deficiency. Similar results are observed in DJ-1-knockout ES cell–derived DNs (which are devoid of any detectable DJ-1) and in primary DNs with DJ-1 levels reduced by RNAi “knockdown.” However, we find that even in the absence of exogenous toxin exposure, the knockout ES cell–derived DNs display reduced survival, whereas survival of the primary embryonic midbrain RNAi knockdown DNs appears to be similar to WT cells. We hypothesize that this discrepancy reflects the activity of residual DJ-1 (approximately 10%–20%) in the RNAi knockdown cultures. Alternatively, the knockout ES cell–derived DNs may be exposed to a greater degree of oxidative stress in vitro than are the knockdown-derived DNs even in the absence of added toxin. The mechanism by which DNs are preferentially targeted for destruction in the absence of DJ-1 is unclear. It has been proposed that DNs are subject to high levels of endogenous oxidative stress that may relate to dopamine metabolism ( Jenner and Olanow 1998 ). DJ-1 is structurally modified in the context of cellular oxidative stress ( Mitsumoto and Nakagawa 2001 ), suggesting a possible function. Two recent studies ( Yokota et al. 2003 ; Taira et al. 2004 ) investigated the role of DJ-1 in the oxidative stress response of neuroblastoma tumor cells. Both studies used RNAi to perturb the expression of DJ-1 in neuroblastoma tumor cell lines, and suggested that DJ-1 deficiency sensitizes cells to oxidative stress; these results are consistent with our data. Taira et al. (2004) further reported that overexpression of DJ-1 in neuroblastoma cells leads to a reduction in ROS accumulation and hypothesized that DJ-1 may harbor antioxidant activity in vivo. In contrast, we find that ES cells that are deficient in DJ-1 display a normal initial burst of ROS in the context of H 2 O 2 . Consistent with this, we fail to detect DJ-1 antioxidant activity in vitro ( Shendelman et al. 2004 ). Finally, this study presents a novel, ES cell-based genetic approach to the study of neurodegenerative disorders. Mouse genetic models of disease are often limited by the inherent variability of animal experiments, the limited mouse life span, and the difficulties in manipulating whole animals. For instance, genetic rescue experiments and toxicological dose-response studies are impractical in whole animals. Furthermore, genetic cell models are more readily amenable to molecular dissection of disease mechanisms than are whole animals. Thus, genetically altered, ES cell–derived neurons are likely to be generally useful as cellular models of neurodegenerative disorders. Future studies may also utilize available human ES cells to investigate species differences. Materials and Methods Cell culture. Undifferentiated ES cells were cultured using standard techniques ( Abeliovich et al. 2000 ). SDIA differentiation of ES cell cultures to DNs was performed as described in Kawasaki et al. (2000) , except that ES cells were plated at a density of 500 cells/cm 2 rather than approximatively 1,000 cells/cm 2 , and were cocultured with the MS5 mouse stromal cell line ( Barberi et al. 2003 ). For rescue experiments, cells were plated at a cell density of 1.4 × 10 6 cells/well. Transfections with plasmids encoding human Flag-WT DJ-1, PD-associated L166P mutant DJ-1, or vector alone, were performed using Lipofectamine 2000 (Life Technologies) for 18–36 h according to the manufacturer's instructions ( Staropoli et al. 2003 ). 24 hours post-transfection, cells were split into 96-well plates and treated as described below. Primary cultures and lentiviral transductions were performed as described in Staropoli et al. (2003) . Generation of knockout ES cell clones The pT1ATGβgeo gene trap vector, which includes βgeo, a fusion of the genes for β-galactosidase and neomycin resistance, is present between exons 6 and 7 of the murine DJ-1 gene, as determined by cDNA sequencing of trapped transcripts and genomic analysis ( Figure 1 A). To select for ES cell subclones homozygous for the trapped DJ-1 allele, we treated clone F063A04 with 4 mg/ml G418. Several subclones that were homozygous for the mutant DJ-1 allele were identified by Southern blotting (see Figure 1 B), and three were chosen for further experimentation: clones 9, 23, and 32. To confirm that the trapped allele leads to the loss of wild-type (WT) DJ-1 protein, cell lysates from these clones, as well as from the parental heterozygous clone, were analyzed by Western blotting using polyclonal antibodies to the amino-terminal region of DJ-1 or the full-length DJ-1 protein. For Western blotting, cells were resuspended in 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.2% Triton X-100, and incubated at 4 °C, rotating for 20 min. Cleared lysate was prepared by centrifuging the lysate at 13,000 rpm for 10 min at 4 °C. Antibodies. A rabbit polyclonal antibody to DJ-1 was generated against the synthetic polypeptide QNLSESPMVKEILKEQESR, which corresponds to amino acids 64–82 of the mouse DJ-1 protein. Antiserum was produced by the Polyquick polyclonal antibody production service of Zymed Laboratories (South San Francisco, California, United States). The antiserum was affinity-purified on a peptide-coupled Sulfolink column (Pierce Biotechnology, Rockford, Illinois, United States) according to the manufacturer's instructions. Antibody was used at a dilution of 1:200 for immunohistochemistry and Western blotting as described ( Staropoli et al. 2003 ). Immunohistochemistry was performed with a rabbit polyclonal antibody to TH (PelFreez, Rogers, Arizona, United States; dilution 1:1000), the mouse monoclonal antibody to neuronal class III β-tubulin TuJ1 (Covance, Princeton, New Jersey, United States; dilution 1:500), and a rabbit polyclonal antibody to GABA (Sigma, St. Louis, Missouri, United States; dilution 1:1000). Western blotting was performed using a polyclonal antibody to cleaved PARP (Cell Signaling Technology, Beverly, Massachusetts, United States; dilution 1:500), a monoclonal antibody to DJ-1 (Stressgen Biotechnologies, San Diego, California, United States; dilution 1:1000), and a mouse monoclonal antibody to β-actin (Sigma, 1:500). In vivo assays. ES cells plated in 96-well format (2.3 × 10 4 cells/well) were treated for 15 h with H 2 O 2 in ES cell medium deficient in β-mercaptoethanol ( Abeliovich et al. 2000 ). Cell viability (as a percent of untreated control) was determined by MTT assay in triplicate ( Fezoui et al. 2000 ). AV/PI (Molecular Probes, Eugene, Oregon, United States) staining was performed according to the manufacturer's instructions. For DHR staining (Molecular Probes) ( Walrand et al. 2003 ), cells were preincubated for 30 min with DHR (5 μM), washed with PBS, then treated with H 2 O 2 in ES cell medium deficient in β-mercaptoethanol for 15 min at 37 °C. The FACS analysis was performed using a FACSTAR sorter (Becton-Dickinson, Palo Alto, California, United States). Dopamine uptake assays were performed as described ( Farrer et al. 1998 ). Reported values represent specific uptake from which nonspecific uptake, determined in the presence of mazindol, was subtracted. Uptake values were normalized for protein content with the BCA kit (Pierce). For 6- hydroxydopamine (6-OHDA, Sigma) treatment, the drug was diluted in the differentiation medium ( Kawasaki et al. 2000 ) and medium was changed every day for 72 h. Primary midbrain embryonic cultures were prepared and transduced with lentiviral vectors as described in Staropoli et al. (2003) . The DJ-1 shRNA vector was generated by insertion of annealed oligonucleotides 5′-TGTCACTGTTGCAGGCTTGGTTCAAGAGACCAAGCCTGCAACAGTGACTTTTTTC-3′ and 5′-ACAGTGACAACGTCCGAACCAAGTTCTCTGGTTCGGACGTTGTCACTGAAAAAAGAGCT-3′ into the LentiLox vector ( Rubinson et al. 2003 ). For cellular dopamine quantification, cultures were incubated in standard differentiation medium supplemented with L-DOPA (50 μM) for 1 h to amplify dopamine production, as described in Pothos et al. (1996) . Subsequently, cells were washed in PBS and then lysed in 0.2 M perchloric acid. Dopamine levels were quantified by HPLC ( Yang et al. 1998 ) and normalized for protein content as above. Expression vectors. The cDNA for human DJ-1 was PCR-amplified from a human liver cDNA library (Clontech, Palo Alto, California, United States). For expression of DJ-1 in ES cell rescue experiments, DJ-1 was cloned into the expression vector pcDNA3.1 (Invitrogen, Carlsbad, California, United States) containing a Flag peptide sequence at the N-terminus using standard cloning techniques. Flag-L166P DJ-1 (pcDNA3) was generated by PCR-mediated mutagenesis. Protein carbonyl analysis. For protein carbonyl quantitation ( Bian et al. 2003 ), cells were plated (1.4 × 10 5 cells per well), grown for 24 h, and then treated with 10 μM H 2 O 2 as indicated. Cells were lysed in 200 μl lysis buffer and cleared lysate was prepared as described above. An aliquot of 40 μl from each time point was added to 2 M HCl (120 μl) with or without 10 μM 2,4-dinitrophenyl-hydrazine and incubated for 1 h at 24 °C with shaking. Proteins were then TCA-precipitated and resuspended in 200 μl of 6 M guanidinium chloride. Absorbance was measured at 360 nm, and DNP-conjugated samples were normalized for protein concentration with the underivitized control samples. Supporting Information Figure S1 Quantitative Real-Time PCR for DJ-1 Gene Expression (A) Real-time PCR analyses of DJ-1 cDNA in WT (+/+), heterozygous (+/–), and knockout (–/–) cultures. Each expression value was normalized to that of β-actin and expressed relative to the respective value of the WT (+/+) control. These gene expression patterns were replicated in at least three independent PCR experiments. Total RNA from ES cells differentiated with the SDIA method for 18 days was isolated using the Absolutely RNA Miniprep kit (Stratagene, La Jolla, California, United States). Synthesis of cDNA was performed using the SuperScript first strand synthesis system for RT-PCR (Invitrogen). Real-time PCR reactions were optimized to determine the linear amplification range. Quantitative real-time RT-PCRs were performed (Stratagene MX3000P) using the QuantiTect SYBR Green PCR Master Mix (Qiagen, Valencia, California, United States) according to the manufacturer's instructions. DJ-1 primer sequences were 5′-CGAAGAAATTCGATGGCTTCCAAAAGAGCTCTGGT-3′ and 5′-CAGACTCGAGCTGCTTCACATACTACTGCTGAGGT-3′; primers used for β-actin were 5′-TTTTGGATGCAAGGTCACAA-3′ and 5′-CTCCACAATGGCTAGTGCAA-3′. For quantitative analyses, PCR product levels were measured in real time during the annealing step, and values were normalized to those of β-actin. (B) Ethidium bromide staining of the PCR products obtained after 29 cycles for DJ-1 (625 bp) and β -actin (350 bp). (704 KB EPS). Click here for additional data file. Figure S2 Analysis of DJ-1-Deficient ES Cells (A and B) Cell viability of DJ-1 heterozygous cells (solid bar) and DJ-1-deficient knockout clone 32 (open bar) after exposure to CuCl 2 or staurosporine at the doses indicated. (C) MTT values of untreated DJ-1-deficient ES cell clones and the control heterozygous cells. Assays were performed exactly as in Figure 2 , but in the absence of toxin. (D) MTT values of untreated DJ-1-deficient ES cells transfected with vector alone or various DJ-1-encoding plasmids. Transfection and expression of WT DJ-1 or mutant forms of DJ-1 does not alter the basal metabolic activity or viability of the cells. (E) Western blotting of extracts from ES cells transfected with vectors harboring WT human DJ-1 or the L166P mutant. (545 KB EPS). Click here for additional data file. Figure S3 Immunocytochemistry for HB9 and GABA Neurons in DJ-1-Deficient and Control Heterozygous ES Cells Both cell cultures were differentiated by SDIA for 18 DIV. Cells were fixed with 4% paraformaldehyde and stained with mouse monoclonal antibodies against HB9 (gift from T. Jessell, dilution 1:50) and rabbit polyclonal antibodies against GABA (Sigma, dilution 1:1000) as in Figure 5 . Scale bar, 50 μM. (5.5 MB TIF). Click here for additional data file. Accession Numbers The GenBank ( http://www.ncbi.nlm.nih.gov/ ) accession numbers of the genes discussed in this paper are α-synuclein (NM_000345), Parkin (AB009973), and DJ-1 (AB073864).

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521171.xml

535892

Odontogenic tumors in Nigerian children and adolescents- a retrospective study of 92 cases

Background Tumours arising from odontogenic tissues are rare and constitute a heterogenous group of interesting lesions. The aim of this study was to determine the relative frequency of odontogenic tumors (OT) among Nigerian children and adolescents 19 years or younger. Patients and methods The histopathology records were retrospectively reviewed for all the tumors and tumor-like lesions of the oral cavity and the jaws seen in children and adolescents ≤ 19 years seen between January 1980 and December 2003. Hematoxylin and eosin-stained sections were re-evaluated and the diagnosis in each case was confirmed or modified according to World Health Organization (WHO) classification, 1992; and were subjected to analysis of age, sex, site of tumor and histopathologic type. Results A total of 477 tumors and tumor-like lesions were seen in patients ≤ 19 years during the period of the study. Of these, 92 (19.3%) were odontogenic tumors. Benign odontogenic tumors constituted 98.9% of the cases seen, while only 1 case (1.1%) of malignant variety was seen during the period. The mean (SD) age of patients was 14.9 (± 3.1) years (range, 4–19 years). Male-to-female ratio was 1:1; and mandible-to-maxilla ratio was 2.7:1. OT's were most frequently seen in patients aged 16–19 years (46.7%) and the least number (2.2%) were found in patients aged 0–5 years. Among nine histologic types of OT seen, ameloblastoma (48.9%), adenomatoid odontogenic tumor (19.6%) and odontogenic myxoma (8.7%) were predominant. Multicystic/solid and unicystic variants of ameloblastoma were diagnosed in 40 (89%) and 5 (11%) cases respectively. Conclusions Odontogenic tumors are relatively common in children and adolescents in Nigeria. One out of every 5 children and adolescents with tumors and tumor-like lesions of oral cavity and the jaws seen in this study had a diagnosis of odontogenic tumor.

Background Tumors and tumor-like growths arising from the odontogenic tissues constitute a heterogenous group of particularly interesting lesions, as they display the various inductive interactions that normally occur among the embryologic components of the developing tooth germ [ 1 ]. In humans, tumors of odontogenic tissues are comparatively rare, comprising of about 1% of all jaw tumors [ 2 ]. In children and adolescents, neoplastic lesions are often benign and are of mesenchymal origin [ 3 , 4 ]. Choung and Kaban [ 5 ] were of the opinion that tumor histology in this age group did not correspond to their clinical behaviour. There are few reports especially from Africa, which have specifically reported the high frequency of odontogenic tumors (OT) in children and adolescents in the literature. In most previous African reports, odontogenic tumours in children were presented as part of orofacial [ 3 ] or oral tumors in this age group [ 3 , 6 ]; or presented as specific tumours e.g. ameloblastoma or adenomatoid odontogenic tumor [ 7 , 8 ]. We could find only a single report on odontogenic tumors in children and adolescent in African environment [ 9 ]. The aim of this study was to determine the relative frequency of odontogenic tumors among children and adolescent ≤ 19 years seen over a period of 24 years (1980–2003). Patients and methods The histopathology records of the Department of Oral Pathology and Biology, College of Medicine, University of Lagos, Lagos, Nigeria, were reviewed for all the tumors and tumor-like lesions of the oral cavity and the jaws seen in children and adolescent ≤ 19 years from January 1980 to December 2003. Hematoxylin and eosin-stained sections was re-evaluated and the diagnosis in each case was confirmed or modified according to World Health Organization (WHO) 1992 classification [ 10 ]. The data was subjected to analysis of age, sex, site of tumor and histopathologic type. The age of the patients was divided in four groups: Group 1 (0–5 years); group 2 (6–10 years); group 3 (11–15 years); group 4 (16–19 years). Data was analyzed using SPSS for Window (version 11.0; SPSS Inc., Chicago, IL) and frequency tables and cross tables were prepared. Results A total of 477 tumors and tumor-like lesions were seen in patients ≤ 19 years during the period of the study. Of these, 92 (19.3%) were odontogenic tumors. Benign odontogenic tumors constituted 98.9% of the cases seen, while only 1 case (1.1%) of malignant variety was seen during the period. Table 1 shows the various histologic types and their relative frequency. There were 47 males and 45 females; a male-to-female ratio of 1:1 was observed. The mean (SD) age of patients was 14.9 (± 3.1) years (range, 4–19 years). Most of the patients (46.7%) were in age group 4 and the least number of patients were found in age group 1 (2.2%) Table 2 . Table 1 Relative frequency of odontogenic tumors in children and adolescents (≤ 19 years) Histological Types Frequency (%) Ameloblastoma 45 (48.9) Calcifying epithelial odentogenic tumor 1(1.1) Ameloblastic fibroma 5 (5.4) Adenomatoid odontogenic tumor (AOT) 18 (19.6%) Calcifying odontogenic cyst (COC) 3 (3.3) Odontoma 4 (4.3) Odontogenic fibroma 7 (7.6) Myxoma 8 (8.7) Ameloblastic carcinoma 1 (1.1) Total 92 (100) Table 2 Age distribution of patients with odontogenic tumors (years) Histological Types Group 1 Group 2 Group 3 Group 4 Total Age 0–5 Age 6–10 Age 11–15 Age 15–19 Ameloblastoma 1 2 22 20 45 Calcifying epithelial odontogenic tumour 0 0 1 0 1 Ameloblastic fibroma 0 1 1 3 5 Adenomatoid odontogenic tumor (AOT) 0 3 6 9 18 Calcifying odontogenic cyst 0 2 1 0 3 Odontoma 0 1 1 2 4 Odontogenic fibroma 1 0 2 4 7 Myxoma 0 1 4 3 8 Ameloblastic carcinoma 0 0 0 1 1 Total 2 10 38 42 92 Table 3 shows the gender and site distribution of various histologic typing of odontogenic tumors. The tumors occurred more often in the mandible (67) than in the maxilla (25) giving a maxilla-to-mandible ratio of 1:2.7. Table 3 Distribution of histologic types of odontogenic tumors according to gender and site of tumor Histological type Number (%) Gender Site Male Female Mandible Maxilla Ameloblastoma 45 (48.9) 28 17 40 5 Calcifying epithelial odontogenic tumor 1 (1.1) 1 0 0 1 Ameloblastic fibroma 5 (5.4) 1 4 5 0 Adenomatoid odontogenic tumor 18 (19.6) 6 12 8 10 Calcifying odontogenic cyst 3 (3.3) 2 1 1 2 Odontoma 4 (4.3) 2 2 0 4 Odontogenic fibroma 7 (7.6) 5 2 5 2 Myxoma 8 (8.7) 2 6 7 1 Ameloblastic carcinoma 1 (1.1) 0 1 1 0 Ameloblastoma constituted almost half (48.9%) of the odontogenic tumors with female-to-male and maxilla-to-mandible ratios of 1:1.7 and 1:8 respectively. The mean (SD) age of patients in this group was 15.1 (± 3.0) years (range, 4–19 years) with most patients (49%) in age group 3. Multicystic/solid and unicystic variants were diagnosed in 40 (89%) and 5 (11%) cases respectively. Adenomatoid odontogenic tumor (AOT) was the second most common tumor in this series (Table 1 ) accounting for 19.6% of odontogenic tumors in this population. All the patients were between 8 and 19 years (Mean ± SD; 14.6 ± 3.2) with most of them (50%) in age group 4. More females (12) were affected than males (6); a male-to-female ratio of 1:2 was observed. Maxillary lesions (10) were commoner than mandibular lesions (8). Odontogenic myxoma accounted for 8 cases (8.7%) of odontogenic tumors with a male-to-female ratio of 1:3. Only 1 case occurred in the maxilla, the rest (7) were found in the mandible. The patients were found between 10 and 19 years. Odontogenic fibroma accounted for 7 cases (7.6%) of tumors in this series. More cases were found in males and in the mandible. Patients were seen between 5 and 19 years in this group. Ameloblastic fibroma accounted for 5.4% of cases seen. The lesion occurred exclusively in the mandible with a male-to-female ratio of 1:4. Odontomas accounted for 4.3% of OT seen. They were exclusively found in the maxilla with equal sex predilection. Table 3 shows the gender and site distribution of other less common odontogenic tumors in children and adolescent ≤ 19 years of age. Discussion We present a report of odontogenic tumours in children and adolescents aged ≤ 19 years. This report represents the largest series of odontogenic tumors in children and adolescents in Africa. We found that 19.3% of tumors and tumor-like lesions of the oral cavity and the jaws in children and adolescent in this study were odontogenic tumors. This is similar to the findings of Arotiba [ 6 ]. However, other authors have reported higher [ 9 , 11 - 13 ] or lower [ 4 , 14 ] frequency of these heterogenous group of lesions in children and adolescents. A major problem in comparing our report with previous studies is the lack of uniformty in the age group studied in those reports. Some studies were restricted to children under the age of 14 years [ 3 , 15 ], or 15 years [ 4 , 6 , 11 - 13 , 16 ], while others included higher age groups [ 9 , 14 ]. Odontogenic tumors were most frequently seen in patients in group 4 in this study (> 15 years) in agreement with Adebayo et al [ 9 ]. Al-Khateeb et al [ 14 ] reported that odontogenic tumors were most commonly (72%) seen in patients aged 12–18 years in their report. Other authors have reported that odontogenic tumors were most frequently seen in patients aged 11–15 years [ 4 , 11 , 13 ]; these authors, however considered patients aged 0–15 years in their studies. Odontogenic tumors were less frequently seen in patients aged 0–5 years in this study in agreement with most reports in the literature [ 4 , 9 , 11 - 14 , 17 ]. About 98% of OT in the present series was found in patients older than 5 years. Many odontogenic tumors are thought to arise from the tooth germ [ 18 ]. In most permanent teeth, crown formation is completed by the age of 4 or 5 years; odontogenic tumors seemed to develop after crown formation [ 12 , 13 ]. This strengthens the impression that the majority of odontogenic tumors arise from quiescent remnants of the tooth germ [ 14 ]. Odontogenic tumors in children are known to have predilection for the mandible [ 11 - 13 ]. This is also corroborated by our findings with 73% of the tumors in this series found in the mandible. Al-Khateeb et al [ 14 ] however found 64% of OT in the maxilla. Odontomas were exclusively found in the maxilla in the present series. Males were slightly affected by OT in our study in agreement with Adebayo et al [ 11 ]; whereas Ulmansky et al [ 4 ] reported female preponderance in their study. Ameloblastoma was the most common tumor in this study in agreement with reports from Africa [ 6 , 11 ]. Other authors reported odontoma as the most frequently seen OT in children [ 12 - 14 ]. Ulmansky et al [ 4 ] found odontogenic myxoma as the most common OT in children in their study. The gender and site distribution of ameloblastoma in this study is in agreement with other reports [ 6 , 11 - 14 ]. Ameloblastoma was found in all the age groups considered in this study, unlike other histologic types of OT. Few cases of unicystic variant of ameloblastoma (11%) were seen in the present series. Unicystic ameloblastoma has been reported to be more common in Western children than African children [ 19 ]. Previous data from Africa as corroborated by the present series, have shown a low percentage of unicystic ameloblastomas in their patient population compared to other parts of the World [ 20 , 21 ]. Unicystic tumors have a different prognosis to the multi cystic type and are said to be more common in children [ 19 , 22 , 23 ]. Adenomatoid odontogenic tumor (AOT) was the second most common OT in this study and 50% of this tumor was found in patients >15 years. Asamoa [ 3 ] reported AOT as the most frequent pediatric odontogenic tumor in Nigeria; whereas Arotiba [ 6 ] found AOT as the second most common OT after ameloblastoma in Nigerian children. The relative frequency of 19.6% found in this study is higher than in other reports [ 6 , 9 , 13 , 14 ]. More of this lesion was found in the maxilla in concordance with previous reports [ 6 , 7 , 9 , 24 , 25 ]. OT is reported to be more common in females [ 1 , 7 , 17 , 26 ]; and this is also confirmed by our findings. In children, the reported incidence of myxoma ranges from 1.2% to 39% [ 4 , 6 , 9 , 13 ]. Myxoma was the third most common OT in this study with an incidence of 8.7%. Ulmansky et al [ 4 ] reported that odontogenic myxoma was the most common OT in Israeli children. The gender predilection favors females in previous publications [ 4 , 9 , 24 , 26 , 27 ]. This study found that the male-to-female ratio was 1:3, in agreement with previous reports. Previous publications [ 9 , 24 , 27 ] reported that both jaws were equally affected in their reports; a ratio of 1:7 was found in the maxilla and mandible respectively in the present study. Odukoya [ 26 ] also found maxilla-to-mandible ratio of 1:3. Odontogenic fibroma was reported to be rare in children with incidence ranging from 0% to 1.3% of OT [ 4 , 9 , 11 , 13 ]. An incidence of 7.8% found in our study was however, similar to that of Al-Khateeb et al [ 14 ]. Arotiba [ 6 ] reported that 12.5% of odontogenic tumors in his study were odontogenic fibroma. More cases were found in males and in the mandible in agreement with the report of Lu et al [ 17 ]; whereas others [ 1 , 26 ] have reported females and maxillary predilection. Ameloblastic fibroma was exclusively seen in the mandible, accounting for 5.4% of OT seen in this study. Females were also affected more than males in the ratio of 1:4. Other authors [ 17 , 26 ] have also reported predilection of ameloblastic fibroma for the mandible but with both jaws affected equally. Odontomas are often regarded as dental hamartomas, rather than odontogenic neoplasms [ 14 , 28 , 29 ]. Odontoma is relatively rare in Nigerian children as confirmed in our report and others from Nigeria [ 6 , 9 , 11 ], accounting for 4.3% of OT in this study. This contrasts the findings of Al-Khateeb et al [ 14 ], Tanaka et al [ 12 ] and Sato et al [ 13 ] that reported odontoma as the most frequently seen OT in North Jordanian and Japanese children respectively. Most odontomas are discovered on routine radiograph and do not produce clinical symptoms [ 1 ]. This may be responsible for the low incidence observed in African population, because most patients in our environment do not seek medical consultation unless there are symptoms suggesting an obvious pathology. While some authors reported that odontomas commonly affect the mandible [ 12 , 17 ], others have reported predilection for the maxilla [ 1 , 13 , 14 ] and some authors have reported equal distribution in both jaws [ 9 , 11 , 25 ]. Odontomas were found exclusively in the maxilla in the present series. Malignant odontogenic tumors are rare, most especially in children [ 1 , 4 , 6 , 9 , 11 , 12 , 15 , 30 ]. No cases of malignant odontogenic tumors were found in African children and adolescents [ 6 , 9 , 11 , 30 ], Israeli children [ 4 ], North Jordanian children and adolescents [ 14 ] and Japanese children [ 12 , 13 ]. A case ameloblastic carcinoma of the mandible in a 16-year old girl was, however seen in this study. Conclusions Odontogenic tumors are relatively common in children and adolescents in Nigeria. One out of every 5 children and adolescents with tumor and tumor-like lesions of oral cavity and the jaws seen in this study had a diagnosis of odontogenic tumor. Competing interests The author(s) declare that they have no competing interests. Authors' contribution OFA conceived the study, coordinated the write-up and submission of the article, and also reviewed the slides. ALL , WLA and MOO did the literature search and participated in the writing of the manuscript. All the authors read and approved the final manuscript. Funding Source None declared

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535892.xml

176547

Borneo Elephants: A High Priority for Conservation

A new study settles a long-standing dispute about the genesis of an endangered species. With scant fossil evidence supporting a prehistoric presence, scientists could not say for sure where Borneo's elephants came from. Did they descend from ancient prototypes of the Pleistocene era or from modern relatives introduced just 300–500 years ago? That question, as Fernando et al. report in this issue, is no longer subject to debate. Applying DNA analysis and dating techniques to investigate the elephants' evolutionary path, researchers from the United States, India, and Malaysia, led by Don Melnick of the Center for Environmental Research and Conservation at Columbia, demonstrate that Borneo's elephants are not recent arrivals. They are genetically distinct from other Asian elephants and may have parted ways with their closest Asian cousins when Borneo separated from the mainland, effectively isolating the Borneo elephants some 300,000 years ago. In the 1950s, Borneo elephants had been classified as a subspecies of Asian elephants (either Indian or Sumatran) based on anatomical differences, such as smaller skull size and tusk variations. This classification was later changed, partly because of the popular view that these animals had descended from imported domesticated elephants. Until now, there was no solid evidence to refute this belief and no reason to prioritize the conservation of Borneo elephants. Their new status, as revealed by this study, has profound implications for the fate of Borneo's largest mammals. Wild Asian elephant populations are disappearing as expanding human development disrupts their migration routes, depletes their food sources, and destroys their habitat. Recognizing these elephants as native to Borneo makes their conservation a high priority and gives biologists important clues about how to manage them. Borneo elephant

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC176547.xml

529263

Traumatic-event headaches

Background Chronic headaches from head trauma and whiplash injury are well-known and common, but chronic headaches from other sorts of physical traumas are not recognized. Methods Specific information was obtained from the medical records of 15 consecutive patients with chronic headaches related to physically injurious traumatic events that did not include either head trauma or whiplash injury. The events and the physical injuries produced by them were noted. The headaches' development, characteristics, duration, frequency, and accompaniments were recorded, as were the patients' use of pain-alleviative drugs. From this latter information, the headaches were classified by the diagnostic criteria of the International Headache Society as though they were naturally-occurring headaches. The presence of other post-traumatic symptoms and litigation were also recorded. Results The intervals between the events and the onset of the headaches resembled those between head traumas or whiplash injuries and their subsequent headaches. The headaches themselves were, as a group, similar to those after head trauma and whiplash injury. Thirteen of the patients had chronic tension-type headache, two had migraine. The sustained bodily injuries were trivial or unidentifiable in nine patients. Fabrication of symptoms for financial remuneration was not evident in these patients of whom seven were not even seeking payments of any kind. Conclusions This study suggests that these hitherto unrecognized post-traumatic headaches constitute a class of headaches characterized by a relation to traumatic events affecting the body but not including head or whiplash traumas. The bodily injuries per se can be discounted as the cause of the headaches. So can fabrication of symptoms for financial remuneration. Altered mental states, not systematically evaluated here, were a possible cause of the headaches. The overall resemblance of these headaches to the headaches after head or whiplash traumas implies that these latter two headache types may likewise not be products of structural injuries.

Background Chronic headaches from head trauma [ 1 - 4 ] and whiplash injury [ 5 - 8 ] are well-known and common. Together with their accompanying symptoms they are usually referred to as "postconcussive (or post-traumatic) syndrome" and "whiplash syndrome," respectively. Chronic headaches from other sorts of physical traumas are not recognized, but a few authors have mentioned them [ 9 - 11 ]. Parker [ 9 ] reported that, among 750 consecutive litigants seen by him for industrial and motor-vehicle accidents, 53% of those who had sustained neither a head nor whiplash injury complained of headache. Duckro et al. [ 10 ] were perplexed about their patients with such headaches: "Even more puzzling, from a diagnostic standpoint, are those patients who suffer persistent exacerbation of headache following physical trauma not involving the head or neck." They also noted that "...in our experience at a university-based clinic for chronic head and neck pain, the problem is not uncommon." The present paper details a series of chronic headaches that began soon after and in apparent relation to various physical traumas that did not include head trauma or whiplash injury. It describes the traumatic events, their temporal relation to the headaches, the features of the headaches, and the symptoms associated with the headaches. It also analyses the relation of the headaches to the events and compares the headaches with those following head trauma and whiplash injuries. Methods The 15 patients in this series were all those seen by the author from February 1997 through December 2001 for chronic headaches apparently related to traumatic events that involved the body but did not cause either head trauma or whiplash injury (a term denoting a painful cervical injury, typically occurring during a motor-vehicle collision and generally considered to be a cervical sprain) [ 7 , 12 , 13 ]. The patients themselves attributed their headaches to the events and so did most of their referring physicians. Detailed accounts of the events and the headaches had been obtained from the patients, family members sometimes, and medical records sent by the referring physicians. Before their events, none of the patients had more than minor occasional headaches. Detailed information had been systematically collected on the headaches' characteristics (intensity, location, quality, and response to physical activities), duration and frequency (when episodic), and accompaniments (nausea, vomiting, hypersensitivity to light and noise), and on the patients' use of pain-alleviative drugs. With this information, the headaches were classified by the 1988 diagnostic criteria of the International Headache Society (IHS) [ 14 ], extant during this review, as though they were naturally-occurring headaches [ 15 ]. Information on the presence of other post-traumatic symptoms and litigation had also been recorded, but less systematically than that for the headaches. Results The physical traumas The physical traumas were very diverse (Table 1 ). None occurred in motor-vehicle accidents. One patient had his face cut by a metal sign as he fell, but he was not stunned by this contact. Two patients lost consciousness briefly, one from syncope after giving blood, and the other from a high-voltage electrical current. Six events produced identifiable damage to a part of the body (patients 1, 2, 9, 10, 13, 15). All but two of the events (patients 8, 11) were sudden and unexpected. Table 1 Patients with headaches from diverse traumatic events. Patient G/Age Traumatic event Headache onset Headache class* 1 M/38 He slid down a roof into a wall and broke a foot and back bone. 4 weeks 2.2 2 F/50 A falling mass of snow buried her and fractured 3 back vertebrae. 2 weeks 2.2 3 F/36 She fainted after donating blood. 1 day 2.2 4 M/41 He was knocked into his car when another car hit his shopping cart. Immediate 2.2 5 M/46 He hurt his neck while yanking a wrench on a rusted bolt. 5 days 2.2 6 M/52 He heard his neck "pop" while lifting a cargo door. Hours 2.2 7 M/28 He slipped off a plank while carrying buckets but landed on his feet. Days 2.2 8 M/43 He had a bone-marrow transplant for a lymphoma. Days 2.2 9 M/48 His face was badly cut on a metal sign when he fell while walking. Hours 2.2 10 F/56 Her scalp was burned by a hair-curling chemical at a beauty parlor. 1 day 1.1 11 M/40 He became angry at the urologist right after his cystoscopy. 1 hour 2.2 12 F/47 Her snowmobile was mistakenly backed up 5 feet into a tree. Days 2.2 13 M/53 He fell sideways in a beachchair and fractured his ribs on a rock. 2 days 2.2 14 F/33 She was exposed to an acute Freon-gas leak at work. Minutes 1.1 15 M/43 A high-voltage electrical injury led to amputations of both arms. 4 weeks 2.2 *IHS codes: 2.2 = chronic tension-type headache, 1.1 = migraine without aura Intervals between traumatic events and headaches The reported intervals between the traumatic events and the headaches' onsets are listed in Table 1 . These intervals can be placed into four groups: within minutes (patients 4, 14), within hours (patients 6, 9, 11), within days (patients 3, 5, 7, 8, 10, 12, 13), and within weeks (patients 1, 2, 15). The interval for the one patient who did not have an abrupt traumatic event (patient 8) was arbitrarily set as the interval between his hospital discharge and headache onset. The headaches Thirteen of the 15 patients had serious continuous headaches of varying intensity. Among these, 11 patients had headaches that met the IHS's diagnostic criteria for chronic tension-type headache [ 14 ] fully, while the other 2 patients' headaches met all but one of the criteria. The 2 patients without continuous headaches had frequent headaches that met the criteria for migraine without aura , and one of them also had headaches fitting the criteria for migraine with aura [ 14 ]. Table 1 lists each patient's headache class. After the completion of this study, the IHS added a new primary headache class called new daily-persistent headache , which has the same features as chronic tension-type headache, but is distinguished from it by becoming daily and unremitting within three days of its onset [ 16 ]. Many, if not all, of the 13 patients in this study with continuous headaches may have had their headaches develop in this way, but as information about this was not specifically sought and as this distinction is of unproven merit, chronic tension-type headache is used herein. The IHS's criteria for chronic tension-type headache and migraine without aura [ 14 ] are listed below in abbreviated form. Chronic tension-type headache Headache frequency more than 15 days/month. At least 2 of the following pain characteristics: 1. Pressing quality 2. Mild or moderate severity 3. Bilateral location 4. No aggravation by routine physical activity Both of the following: 1. No vomiting 2. No more than one of the following: Nausea, photophobia or phonophobia Migraine without aura Headaches last 4 to 72 hours. Headache has at least 2 of the following characteristics: 1. Unilateral location 2. Pulsating quality 3. Moderate or severe intensity 4. Aggravation by routine physical activity During headache at least one of the following: 1. Nausea and/or vomiting 2. Photophobia and phonophobia Examples of chronic tension-type headache after traumatic events Patient 5 This 46-year-old man was seen in 1999 for headache that began five days after an "injury" at work five months earlier. Ten minutes after forcefully yanking a wrench to loosen a rusted bolt, he felt pain in his neck and right shoulder. An urgent-care facility prescribed analgesics after taking (unremarkable) cervical radiographs. This pain disappeared before I saw him. The headache was a continuous dull to moderate ache mostly in the right cranium. It was unaffected by physical activities or neck movements, and was not nauseating. Neurological examinations were normal. Pressure on his posterolateral neck was not painful. A cranial CT was normal. His only other symptom was insomnia. Analgesics, taken just a few days per week, had little effect. Amitriptyline lessened the headache's intensity enough for him to return to work. Patient 9 This 48-year-old man with a neurologic impairment of gait was seen in 1999 for continuous headache that began a few hours after he fell while walking two months earlier and cut his face on the edge of a metal sign. His head was not struck and he was not stunned. He bled profusely from the laceration, which was closed with 26 stitches in the emergency room. His neurologist detected no new neurologic findings and a cranial CT was normal. His headache was a non-nauseating steady ache of mild to moderate intensity in his forehead, unaffected by exercises, brightness, or noise. Non-prescription analgesics and opioids had been ineffective and discontinued. Patient 11 This 40-year-old man was seen in 2000 for a continuous headache that began 10 months earlier soon after he awoke from anesthesia for a cystoscopy. When the urologist did not report the (negative) result of the procedure to him in the recovery suite, the patient became visibly angry and soon complained about his treatment to the health-care facility. His anger persisted and was expressed at his consultation. The headache was a steady non-nauseating pain that fluctuated from dull to moderate intensity at the vertex, temples, and posterior neck, and was unaffected by physical activities, brightness, or noise. He worked despite it and no longer took analgesics. A cranial MRI had been normal. A trial of amitriptyline had been unsuccessful. At his last report a month later he reported improvement on buspirone. Patient 12 This 47-year-old woman was seen in 2001 for symptoms that she attributed to a snowmobile accident seven weeks earlier. She was seated behind the driver when he mistakenly shifted into reverse sending the machine backwards five feet into a tree. At impact, he fell on top of her without hurting her or himself. She recalled no impact of her helmet against the tree or the snowmobile. She felt no pain, but was upset and asked to be taken home. On the next day, her neck ached. Cervical radiographs taken eight days after the accident were normal. Two days later, she developed severe headache, nausea, dizziness, and confusion. A cranial CT taken later that day was normal. Subsequent MRIs of her head and neck were normal. Headache soon became her most prominent pain. It was a continuous pressing ache of mild to moderate intensity in her temples and orbits, unaffected by physical activities. It was sometimes nauseating, without emesis. Brightness and noise bothered her. She also complained of dizziness and impaired thinking and memory. Infrequent doses of analgesics were not beneficial. She was unable to work. Neurologic examinations were normal. Preventive medications were refused. When seen next, by a colleague, two months after her visit with me, the headache and dizziness had lessened considerably, but her thinking difficulties remained disabling. Four months later, she reported continuing improvement. Patient 13 This 53-year-old man was seen in 2001 for a headache of six-months duration that began two days after he struck the right side of his chest against a rock without striking his head when he toppled over in a beach chair. His chest pain was extreme. He obtained a prescription for hydrocodone/acetaminophen tablets that day, but discontinued taking them after several doses because of side effects. Coughing, sneezing, and lying on his right side were excruciating. His physician diagnosed a fractured rib. Two days after the injury, he returned to work despite his chest pain and new headache. The chest pain disappeared in a few weeks, but the headache persisted. It was a continuous, non-nauseating, bifrontal "tightness" of dull to moderate intensity, unaffected by mild physical activities, brightness or noise. Amitriptyline and propranolol had been ineffective and produced side effects. When I saw him, he was taking only occasional doses of non-prescription analgesics. Neurological examinations and cranial MRI were normal. He declined other medications. Seven months later, he reported that his headache persisted. Patient 15 This 43-year-old man was seen in 2001 for a continuous headache that he first became aware of soon after discharge from hospital, in 1997, where he had undergone 25 days of intensive treatment for a high-voltage electrical injury that had necessitated amputation of his arms. Unconsciousness had been instantaneous, but brief, and post-traumatic amnesia lasted about ten minutes. His headache was a bi-occipital, non-pressing ache, usually of mild to moderate intensity, and only occasionally severe enough to force him to cease physical activities. Then it was nauseating, without emesis. Some loud noises, but not brightness, seemed to intensify it. He had been getting slight relief from a few doses of ibuprofen per week, but had received no preventive medications. His cognitive and emotional states and cranial MRI were unremarkable. He had been provided with prosthetic upper limbs with grasping hands. Amitriptyline decreased the headache slightly. The addition of progressively larger doses of dextroamphetamine limited the headache to only a few days per month. Other post-traumatic symptoms Thirteen of the fifteen patients had, besides headache, at least one other post-traumatic symptom of the type commonly seen after head trauma [ 3 , 17 , 18 ]. Seven patients had either three or four symptoms. The number of patients having each of the symptoms is listed in Table 2 . Only patient 2 had any of the symptoms included in the syndrome of post-traumatic stress disorder [ 19 ]. Table 2 Other post-traumatic symptoms. Symptoms Number of patients with symptom Insomnia 14 Decreased concentration 7 Decreased memory 5 Dizziness 5 Mild depression 3 Anxiety 3 Tiredness 3 Litigation/compensation Seven patients (numbers 3, 5, 9, 10, 11, 12, 13) were not seeking and could not seek financial compensation for their headaches. Three patients (numbers 6, 7, 14) were receiving Workers' Compensation (WC) payments for their headaches and other symptoms. One patient (number 15) was receiving both WC and Social Security (SS) disability payments. One other receiving WC payments was also seeking SS disability payments (number 1). One patient (number 8) was seeking SS payments only. One patient (number 2) was litigating for payment of medical bills. Only one patient (number 4) had pursued a lawsuit for monetary compensation for the symptoms, but he continued to experience headaches even after receiving the award. Discussion This report has presented evidence suggesting that traumatic accidents and other events without head trauma or whiplash injury but with other physical effects can induce chronic headaches. Such headaches have barely been alluded to before (see Background). The 15 patients presented here blamed their headaches and other symptoms on their traumas and are supported in this by the juxtaposition of their headaches to the traumas. Twelve of the 15 patients developed their headaches minutes, hours, or a few days afterwards (Table 1 ). These intervals are well within those deemed necessary by the IHS for connecting both head trauma and whiplash injury to subsequent headaches. Their 1988 classification [ 14 ] lists "less than 14 days" but their 2004 revision [ 16 ] lists "within 7 days" as the maximum interval acceptable for relating chronic headaches to these injuries. The other 3 patients reported headache onsets of two weeks, four weeks, and four weeks. Nevertheless, they are included in this series because the apparent connection of their headaches to their serious accidents outweighs the still non-evidential latent-period requirement chosen by the IHS. In support of this inclusion is evidence suggesting that genuine post-traumatic symptoms can develop as late as three months after head trauma [ 11 , 20 ] and that chronic headaches after whiplash injuries may not appear for weeks or even months [ 21 ]. Although the headaches in this study were temporally related to events involving the body, they were unlikely to have been caused by the bodily injuries themselves, for there is no conceptual link between injuries such as a facial laceration or a fractured rib and chronic headaches. Moreover, nine patients had either trivial or unidentifiable injuries (Table 1 , patients 3, 4, 5, 6, 7, 8, 11, 12, 14). Fabrication of symptoms for financial remuneration was not evident in this series, in which seven patients were not even seeking payments of any kind. If these headaches can not be attributed to bodily injury, then the symptoms would appear to be of psychological origin, as they are after some other sorts of traumatic events, such as those that set off chronic "post-traumatic stress disorder" (PTSD) [ 19 ]. Traumatic psychological symptoms can include headache. It has been reported, for example, to accompany the characteristic symptoms of PTSD [ 22 ], and it was the most prominent post-traumatic symptom in a "mass psychogenic illness" induced by false perceptions of exposure to toxic fumes at a school [ 23 ]. In this study the psychological states of the patients were not investigated (though they were not ignored), because the study's purpose was to analyze the traumatic events, the headaches following them, and the relationship between the two. Thus, this study can not present positive evidence for a psychological basis of the headaches. The headaches were not, however, related to PTSD, since only one patient had (some) symptoms of this condition. Both head-trauma and whiplash headaches form clinical classes based on their preceding traumas, but the headaches of the present series have no single trauma to join them. They were, however, all related to acute and unexpected (with one exception) traumatic events that affected the patients physically. Hence, they could be designated traumatic-event headaches . Recognition of this class of post-traumatic headaches would link heretofore puzzling individual phenomena (see Background) and thereby foster their investigation. The headaches themselves were serious chronic headaches. They had the features of continuous chronic tension-type headaches in 13 patients and frequent migraines in the other two [ 14 , 16 ]. This distribution of headache types is similar to that of the chronic post-traumatic headaches after head-trauma and whiplash injury (excepting the advocated cervical syndrome from some whiplash injuries [ 24 ]). In studies using the 1988 IHS diagnostic criteria, head-trauma headaches were 75% chronic tension-type, 21% migraine, and 4% unclassifiable [ 15 ], whiplash headaches were 74% tension-type, 15% migraine, and 11% cervicogenic [ 6 ], and "cranio-cervical acceleration/deceleration trauma" headaches (22% with head trauma, 78% with whiplash) were 37% tension-type, 27% migraine, 18% cervicogenic, and 18% unclassifiable [ 25 ]. In addition to the headaches, the present series of patients suffered other symptoms like those in the post-concussion [ 3 , 17 , 18 ] and whiplash syndromes [ 5 , 26 ] (Table 2 ). The apparent overall likeness of the traumatic-event headaches to those after head and whiplash traumas suggests a pathogenic link between them. Such a link would be discredited if, as many believe, headaches after head and whiplash traumas are due to structural injuries to the brain and neck, respectively [ 21 , 24 , 27 - 29 ]. But such injuries have not been seen in cranial or cervical MRIs in the great majority of cases and whatever injuries may be present have an uncertain relation to the chronic headaches [ 7 , 28 , 30 , 31 ]. Moreover, certain evidence seems incompatible with the presence of symptomatic injuries. Firstly, the incidence of chronic symptoms after head trauma [ 2 , 3 , 28 ] or car crashes [ 32 , 33 ] does not increase in step with increasing degrees of trauma. Secondly, the development of chronic symptoms is dependent on the circumstances in which the trauma occurs. For example, whereas chronic headaches are common after head traumas at work, they are rare after head blows during sports [ 34 - 36 ]. Likewise, chronic whiplash symptoms have not been produced in volunteers subjected to rear-end car crashes [ 31 , 33 ] and they do not occur in demolition-derby drivers [ 37 ] and drivers or passengers subjected to rear-end crashes in certain countries where whiplash is not common knowledge [ 8 ]. This and additional evidence has led some authors to discount a role for cranial or cervical injuries in the development of chronic post-traumatic symptoms, offering instead explanations based on altered mental states [ 2 , 8 , 31 , 38 , 39 ]. These include, among others, neurotic reactions and culturally-related expectations of symptom development. The present study supports a mental origin from another perspective by showing that chronic post-traumatic headaches can be sequelae of traumatic events having neither head trauma nor whiplash injury. Conclusions This study indicates that chronic headaches similar to those after head trauma and whiplash injury can follow other types of acute and unexpected physically injurious traumatic events. The evidence suggests that these traumatic-event headaches are psychogenic. The occurrence of such headaches supports the concept that the chronic headaches after head trauma or whiplash injury may likewise not be products of structural injuries. Competing interests The author(s) declare that they have no competing interests. Authors' contributions DCH was the sole investigator and author. Pre-publication history The pre-publication history for this paper can be accessed here: http://www.biomedcentral.com/1471-2377/4/17/prepub

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529263.xml

548138

Sister Joseph's nodule in a liver transplant recipient: Case report and mini-review of literature

Background Umbilical metastasis is one of the main characteristic signs of extensive neoplastic disease and is universally referred to as Sister Mary Joseph's nodule. Case presentation A 59-years-old Caucasian female underwent liver transplant for end stage liver disease due to hepatitis C with whole graft from cadaveric donor in 2003. After transplantation the patient developed multiple subcutaneous nodules in the umbilical region and bilateral inguinal lymphadenopathy. The excision biopsy of the umbilical mass showed the features of a poorly differentiated papillary serous cystadenocarcinoma. Computed tomographic scan and transvaginal ultrasonography were unable to demonstrate any primary lesion. Chemotherapy was start and the dosage of the immunosuppressive drugs was reduced. To date the patient is doing well and liver function is normal. Conclusions The umbilical metastasis can arise from many sites. In some cases, primary tumor may be not identified; nonetheless chemotherapy must be administrated based on patient's history, anatomical and histological findings.

Background Metastases to the umbilicus are universally referred to as Sister Joseph's (or Sister Mary Joseph's) nodule. Etiology is related to the presence of primary malignant disease in the abdominal cavity or occasionally in the chest and/or breast [ 1 - 5 ]. Historically, Sister Mary Joseph (1856–1939) was a surgical assistant under the guidance of Dr. William Mayo. She was the first one to note the connection between the umbilical nodule and intra-abdominal cancer. The first case reporting the presence of Sister Mary Joseph's nodule was in 1864 by Storer, however; Hamilton Bailey was the first one to use the term "Sister Mary Joseph's nodule. Although skin metastasis is rare and range between 5% and 9%, it is estimated that 1% to 3% of abdomino-pelvic tumors metastasize to the umbilicus [ 2 , 3 , 5 ]. The most common primary neoplasm is adenocarcinoma (75%), more rarely squamous cell carcinoma followed by undifferentiated tumors or carcinoid can metastasize to umbilicus. In men, gastrointestinal tract (55%) is the most common location of the primary neoplasm that metastasizes to the umbilicus, followed by stomach, colon, rectum, small bowel, and pancreas in a decreasing order [ 1 - 5 ]. In women, on the other hand, gynecological neoplasms particularly ovarian cancer is the most common primary site, of which serous papillary cystadenocarcinoma (34%) is the most frequent. Further, endometrial carcinoma, cervix, vagina and vulva may also be responsible for the metastasis to the umbilicus. In addition to these the literature reports other but rare sites that could form potential grounds for metastasis to the umbilicus, namely gallbladder, liver, breast, lung, prostate, penis, peritoneum, lymphoma, bladder and kidney. In about 11% of the umbilical metastasis, the origin of the metastasis is unknown [ 4 , 5 ]. Upon recognition of the positive nodules in the umbilicus, physician must consider other potential sources such as endometriosis, melanocytic nevi, fibroma, epithelial inclusion cysts, seborrheic keratosis, pilonidal sinus, keloid, foreign body granulomas, myxoma, omphalitis, abscesses, umbilical hernia and of course primary malignant umbilical tumor (melanoma, squamous and basal cell carcinoma, sarcoma and adenocarcinoma) [ 6 ]. In many instances, Sister Joseph's nodule is the first and often the only indication of an underlying or occult cancer and/or may indicate recurrence of previously treated malignancy. The presence of umbilical metastasis usually suggests already advanced metastatic process characterized by poor prognosis [ 4 , 5 ]. Therefore, a biopsy of umbilical nodule is recommended which is safe and easy. Obtaining histological description can further direct the clinician to dictate a prompt treatment. In the case of recurrence of the neoplasm, fine needle aspiration cytology may be sufficient to provide definite diagnosis; however, sometimes clinical, cytological, histological, radiological and/or surgical investigations may not be sufficient to identify primary site of the metastasis [ 7 ]. Serous papillary cystadenocarcinoma has been documented previously as a cancer responsible for Sister Joseph's nodule because of the metastasis to the regional lymph nodes [ 6 - 9 ]. Herein, we report, to the best of our knowledge, the first case of Sister Joseph's nodule in a patient after liver transplantation complicated by serous papillary cystadenocarcinoma. In addition, we review the potential routes by which tumor spread may have occurred and the prognostic significance associated with umbilicus metastasis. Case presentation A 59-year-old Caucasian woman was admitted to our hospital for progressively enlarging but painless mass in the region of the umbilicus. The umbilical mass of approximately 5 cm in diameter and has developed two months after liver transplantation. The nodule was firm with irregular borders, wine-red shade and fixed to the surrounding tissues. The overlying skin was not ulcerated. She never smoked, consumed alcohol or used oral contraceptive pills. The gynecological history was unremarkable. The patient underwent a liver transplantation for end stage liver disease (ESLD) secondary to hepatitis C in mid of 2003. Prior to the transplant, her serum alpha-fetoprotein (alpha-FP) and CEA concentration were measured and both were within normal limits. The transplant procedure and immediate post-transplant period were unremarkable. The post-transplant immunosuppressive regimen consisted of Tacrolimus (6 mg/day) and steroids (prednisone 10 mg/day). Two months after transplantation, during routine follow-up, physical examination revealed multiple subcutaneous satellite nodules in close proximity to the umbilical region and lower abdominal wall. There was bilateral inguinal lymphadenopathy. A plain anterior/posterior chest x-ray was normal. Ultrasound of the abdominal cavity showed a 4 × 4 cm mass below the umbilicus and additional smaller nodules confined to the lower abdominal wall. Paraaortic lymph nodes were not identified. The echogenicity of the mass consisted of alternative hyper- and hypoechoic with poorly defined edges. However, the general condition of the patient was good and the laboratory test showed: hemoglobin level 9.6 g/dl, total bilirubin 0.6 mg/dl, conjugated bilirubin 0.1 mg/dl, Alkaline Phosphates 273 U/l, GGT 40 U/L, AST 36 U/L, and ALT 56 U/L. There was serological evidence of past hepatitis C virus infection. The biopsy of the umbilical mass showed features characteristic for a poorly differentiated serous papillary cystadenocarcinoma. An extended resection of the umbilical and paraumbilical metastasis was performed. In addition, extensive investigation to detect the primary neoplastic origin was initiated. Tumor marker evaluation consisted of: CEA 0.22 [normal 0–3.7 ng/ml], alpha feto protein 1.49 (normal: 0.8–9.4 UI/ml), CA125 27.73 (normal: 0–30 UI/ml), CA 19.9 was 8.61 (normal: 0–36.2 UI/ml), CA 15.3 was 22.48 (normal: 0–35 UI/ml). Multiple tumor marker measurements were obtained and each time similar results were reported. A computed tomographic (CT) scan was unable to demonstrate any lesion that could be considered a primary source of the metastasis to the umbilicus. Trans-vaginal ultrasonography showed a normal uterus and ovaries; ultrasonographic examination of thyroid and breast did not show any pathological changes as well. In spite of laborious efforts, the tumor's primary site was not found at that time. However, considering the histological results, the umbilicus metastasis could be attributed almost entirely to the ovarian cancer stage 4, since the main primary site in females are the ovaries. In order to prevent further tumor extension and expansion, tacrolimus was reduced from 6 mg/day to 1.5 mg/day. In addition, chemotherapy with Taxol was initiated (50 mg/m 2 I.V) once a week for 16 weeks. To date, after 6 months of follow-up, the patient is doing well, the liver function is normal and tacrolimus serum level is 1 ng/ml. A recent whole body CT-scan showed regression of iliac and inguinal lymph nodes involvement. Discussion Metastases may appear virtually anywhere in the body; however, certain sites are more common then others and umbilical metastasis are unusual sites [ 9 , 10 ]. In majority of cases, the metastatic lesions accompany the symptoms and signs of the primary tumor; however, it can be the first and often the only sign of a carcinoma, although this happens rarely [ 10 ]. A variation in vascularity and embryological development makes the umbilicus easy target for metastasis from an intra-abdominal tumor. In fact, during fetal development, the ductus venous (ductus venous Arantii) connects the umbilical portion of the left branch of the portal vein to the inferior vena cava, thus shunting oxygenated umbilical cord blood away form the liver. After birth the duct obliterates and persists as ligamentum venosum or Arantius' ligament. Because of its functional role, it is commonly believed that the ligament runs from the left branch of the portal vein to the vena cava itself. However, attention to anatomical detail demonstrates that the fibers of the ligament insert either on the left hepatic vein or at the junction between the left hepatic and the middle hepatic vein [ 1 , 11 ]. We know that in cirrhotic patients, the umbilical vein remains open due to portal hypertension. There are numerous potential routes by which a carcinoma can metastasize to the umbilicus in a liver transplant recipient. Malignant carcinomatous cells in the portal venous system may reach the umbilicus by way of a patent umbilical vein [ 12 - 15 ]. This would be more likely to occur in the presence of portal hypertension and the resulting portal-systemic shunting of blood [ 16 ]. Since our patient received liver transplant for cirrhosis secondary to hepatitis C virus infection accompanied by portal hypertension, this could be a possible mode of spread of papillary serous cystadenocarcinoma in our patient. In addition, presence of severe esophageal varicies and large umbilical vein that was viewed by angiography examination before the liver transplantation could have contributed to the spread of tumor. Therefore, in this case, malignant cells could have reached the umbilicus via the umbilical vein. In addition, a pelvic carcinoma of unknown origin may spread via lymph nodes to the umbilical region [ 9 , 17 , 18 ]. Furthermore, dermal lymphatics are a potential route of spread to the umbilicus and should be considered, since the umbilicus can be reached by retrograde lymphatic flow [ 4 ]. Nevertheless, the spread of malignant cells could have occurred via embolization thru the arterial blood supply to the umbilicus. Furthermore, direct implantation of tumor cells may have occurred via direct extension of the neoplasm from the anterior peritoneal surface or via redistribution of the peritoneal fluid flow [ 19 , 20 ]. Finally, cutaneous metastasis to the anterior abdominal wall, may perhaps cause malignant cells to spread in a retrograde direction, namely from the metastatic lesions along the subcutaneous lymphatic vessels to the umbilicus. It is known that the immunosuppression promotes tumor spreading and potentates its expansion, thus immunosuppressive regimen plays a crucial role in the tumor expansion and metastasis. Umbilical metastasis is one of many characteristic signs of extensive neoplastic disease. It suggests advanced distant metastasis and is associated with poor prognosis; mean survival is approximately 10–12 months, although long-term survival has been reported, but only in the presence of solitary metastatic umbilicus nodule [ 5 ]. Conclusions Presence of umbilical nodule or nodules requires extensive, meticulous and laborious evaluations since the differential diagnosis for umbilical nodule are many. In our case although, the primary tumor was not found, we treated it as ovarian tumor considering the histological findings and the iliac-inguinal nodes involvement. Competing interests The author(s) declare that they have no competing interests. Authors' contributions FP : conception and design, interpretation of data, drafting the article and revising EA : interpretation of data, drafting the article and revising SDD : acquisition of data and revising NM : interpretation of data, design and coordination and helped to draft the manuscript GB : collection of data, design and coordination and helped to draft the manuscript RM : collection of data, design and coordination MM : acquisition of data and drafting the article TJ : drafting the article and interpretation FR : acquisition of data and design MC : interpretation of data, drafting the article UV : drafting the article, revising and supervision Figure 1 Preoperative Sister Mary Joseph's nodule ultrasonography: 4 × 4 cm mass confined below the umbilicus (arrows). The main lesion is partly hyperechoic and partly hypoechoic with a poorly defined edge. Figure 2 Intraoperative specimen of the umbilical region: Sister Mary Joseph's nodule and umbilicus.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548138.xml

535925

Evolutionary relationships of Fusobacterium nucleatum based on phylogenetic analysis and comparative genomics

Background The phylogenetic position and evolutionary relationships of Fusobacteria remain uncertain. Especially intriguing is their relatedness to low G+C Gram positive bacteria (Firmicutes) by ribosomal molecular phylogenies, but their possession of a typical gram negative outer membrane. Taking advantage of the recent completion of the Fusobacterium nucleatum genome sequence we have examined the evolutionary relationships of Fusobacterium genes by phylogenetic analysis and comparative genomics tools. Results The data indicate that Fusobacterium has a core genome of a very different nature to other bacterial lineages, and branches out at the base of Firmicutes. However, depending on the method used, 35–56% of Fusobacterium genes appear to have a xenologous origin from bacteroidetes, proteobacteria, spirochaetes and the Firmicutes themselves. A high number of hypothetical ORFs with unusual codon usage and short lengths were found and hypothesized to be remnants of transferred genes that were discarded. Some proteins and operons are also hypothesized to be of mixed ancestry. A large portion of the Gram-negative cell wall-related genes seems to have been transferred from proteobacteria. Conclusions Many instances of similarity to other inhabitants of the dental plaque that have been sequenced were found. This suggests that the close physical contact found in this environment might facilitate horizontal gene transfer, supporting the idea of niche-specific gene pools. We hypothesize that at a point in time, probably associated to the rise of mammals, a strong selective pressure might have existed for a cell with a Clostridia-like metabolic apparatus but with the adhesive and immune camouflage features of Proteobacteria.

Background The genus Fusobacterium , together with some close relatives such as Leptotrichia , forms an ecologically and physiologically coherent group [ 1 ]. They seem to be inhabitants of the mammal gastrointestinal tract probably specialized in the oral cavity. Specifically, they are components of the dental plaque, a highly complex habitat that has received considerable attention in recent years due to its involvement in dental pathology [ 2 ]. They are all fermentative anaerobes that use mostly peptides as their energy source (see, for example, [ 3 ]). The species Fusobacterium nucleatum has received particular attention being a key component of the human dental plaque that also has considerable pathogenic potential. In fact after Bacteroides , Fusobacterium is responsible for most human anaerobic infections, producing abscesses at different locations and aspiration pneumonia among other serious conditions [ 4 , 5 ]. Phylogenetically speaking the fusobacteria have become somewhat of a puzzle [ 6 ]. Originally classified with Bacteroides and other Gram negative anaerobes, their association became conflicting when, after the extensive gene sequencing carried out by the mid 80's, it became clear that Bacteroides showed a clear relationship to other aerobic Gram negatives such as Flavobacterium or Cytophaga [ 7 - 9 ] while on the grounds of the 16S rRNA sequence Fusobacterium appeared as a separate cluster only distantly associated to the low G+C Gram positives [ 10 , 11 ]. However, this association is methodology sensitive, and different algorithms or genes associate them with other groups such as the Proteobacteria, the Cyanobacteria, the Thermotogales, or within the Firmicutes (see for example [ 12 - 14 ]). The publication of the Fusobacterium nucleatum genome [ 3 ] did not solve the problem since although most BLAST top-hits appeared as Clostridium species (low G+C Gram positives) genomic analysis showed also a strong proximity to Proteobacteria. Based on the ERGO chromosomal clustering tool, F. nucleatum had more "clusters" of genes with the same gene order in common with Escherichia coli than with Enterococcus or Staphylococcus , although less than with Clostridium or Bacillus [ 3 ]. As expected, most elements typical of a Gram negative cell wall were found in the genome including porins, outer membrane transport systems, lipid A synthesis pathways and LPS core compounds. It may be argued that the Gram negative cell wall is the ancestral situation and the Gram positives have lost the outer membrane. However, this scenario requires a remarkable stability in the components of the fusobacterial cell wall to remain so similar to other distant bacterial phyla [ 15 ]. On the other hand, there is the possibility that large portions of the fusobacterial genome could be the result of horizontal gene transfer (HGT). The oral cavity environment where F. nucleatum thrives is an ecosystem with a large bacterial biodiversity. In a recent survey using 16S rDNA sequences from sub gingival plaque samples, 347 species or phylogroups were identified, and the best estimate of the total species diversity in the oral cavity is approximately 500 species [ 16 ]. These 347 species belonged to 9 different bacterial taxa and F. nucleatum interacts with a great deal of them, because it plays a crucial "bridge" role between early and late colonizers of the tooth surface [ 17 ] and forms carbohydrate-mediated coaggregations with other species [ 18 - 21 ]. Because of the many species with which F. nucleatum interacts and aggregates (including spirochaetes, proteobacteria, bacteroidetes, firmicutes, and even fungi) there is a great potential for HGT. We have reanalysed the fully sequenced genome of F. nucleatum , using a variety of bioinformatics tools, in an attempt to clarify the phylogenetic position of the Fusobacteria and the relative contributions of vertical descent and horizontal transfer in shaping the genome of this highly specialized organism. In addition, our study aims at providing material for further discussions on evolution of the gram-negative cell wall, and on the evolution of bacterial communities in micro-environments. Results and Discussion Phylogenetic position of core fusobacterial genes It is generally assumed that in every genome there is, at least, a basic core of genes that are inherited vertically and may be used to infer relationships among prokaryotes [ 22 ]. Although most often the relationships obtained with the core genes are consistent with that of the 16S rRNA gene we have extended this type of analysis to include as many genes as possible. Firstly, a Bayesian tree using the combined 16S-23S rRNA sequences was constructed (Figure 1 ). A neighbor-joining tree based on the concatenated alignments from 44 ribosomal proteins gave a similar result [see additional file 1 ]. In both cases, the fusobacteria appear as a clearly defined and distinct group that branches out at the base of the Firmicutes but as an independent phylum. Finally, the 23 proteins conserved across all sequenced Bacteria were selected [ 23 ] and trees were constructed based on their sequences. Many of them gave results consistent with the previous two trees [see additional file 1 ]. However, some typical core genes hinted of a mixed ancestry. The ribosome-associated protein prlA, for example, produced a tree that associated Fusobacterium with the cyanobacteria and the elongation factor tufA with the proteobacteria. Other cases are also unclear. DNA pol III is a complex holoenzyme formed by 10 subunits in E. coli [ 24 ]. Interestingly, subunits α and β of the polymerase III seem related to Firmicutes while the gene for subunits γ and τ to Thermotoga and Aquifex . The RNA-directed DNA polymerase and a RNA helicase seem related to archaeal counterparts. In Clostridium , for example, to which most informational genes of Fusobacterium have a best match, all subunits of both polymerases cluster clearly into the Firmicutes (data not shown). Summarizing, the fusobacteria appear as an independent taxon with remote relatedness to the Firmicutes. Although the phylogenetic signal for many genes was considerably weak, the rRNA genes and the ribosomal proteins were very congruent and therefore their trees (Figure 1 and additional file 1 ) are likely to represent a reliable phylogenetic reconstruction of the core genome. Heretofore we will refer to this affiliation as ribosomal phylogeny and consider it as the reference for vertically inherited genes. We have assumed that genes showing a close relationship to other bacterial taxa (including the Firmicutes) are possible candidates for HGT origin, particularly when a close association has been proved by more than one approach. Figure 1 Phylogenetic tree (Bayesian method) using the combined sequence of the 16S-23S rRNA of representative bacterial species. The Fusobacteria are a coherent and taxonomically independent group, that branches out at the base of the lineage leading to Firmicutes. Numbers represent bootstrap values. In the case of the branching of Fusobacteria/Firmicutes, the numbers represent the values obtained by four different methods: BA: Bayesian; NJ: Neighbor-joining; MP: Parsimony; ML: Maximum likelihood. GC-skew plots Figure 2 shows the GC-skew plot for F. nucleatum compared to Clostridium tetani (the sequenced species to which it shows the highest number of homologous genes) and Bacteroides thetaiotaomicron (the Gram-negative species to which it shows the highest number of homologs). Due to differences in mutational biases between the leading and lagging strands, it is common to find the GC skew value (G-C/G+C) with opposite signs on each replichore, the change in sign indicating the origin or terminus of DNA replication [ 25 , 26 ]. This skew is independent of GC content [ 27 ]. As the figure shows, both the Gram-positive and Gram-negative bacteria which appear to have the most similar gene content to F. nucleatum display a standard plot, with mainly positive values over the right replichore and negative values on the left replichore. Fusobacterium , however, does not show a clear pattern, with constant shifts in GC-skew values across the genome. This situation could be caused by horizontal gene transfer (HGT) incorporating xenologous sequences with a different GC-skew across the recipient chromosome, distorting a clear-cut plot. In favour of this, many GC skew oscillations coincide with clusters of putative xenologous origin (see thin arrows in Figure 2 , top). It is interesting to note that an oscillating GC skew plot is also observed in other genomes that appear to have undergone massive HGT episodes (see, for example [ 28 ]). Figure 2 (G-C)/(G+C) values (GC-skew) plotted every 5000 bp for Fusobacterium nucleatum , the low-GC Gram positive Clostridium tetani and the bacteroidete Bacteroides thetaiotaomicron . Red wide arrows represent replication origin (bottom) and terminus (top). Orange thin arrows indicate the 36 "clusters" of four or more contiguous genes that are potentially transferred from species outside the Firmicutes. Note that some of F. nucleatum shifts in GC skew coincide with putative HGT regions. The observed GC-skew could also arise from chromosomal inversions (see, for example, the genome of Yersinia pestis -[ 29 ]). However, F. nucleatum should have undergone massive events of genomic scrambling to account for the effect, including numerous non-symmetric inversions around the replication origin and terminus, which are rarely observed [ 30 , 31 ] and are assumed to be detrimental [ 32 ]. Moreover, homologous genes present in the long DNA fragments sequenced in the close relative F. nucleatum subsp vincentii [ 33 ] show an almost perfect sinteny: In all 6 sequenced segments larger than 30 kb in vincentii , gene order was conserved without a single chromosomal inversion (data not shown). Although other related genomes are not available for comparison and the potential inversions could have happened prior to the split of both subspecies, the suggestion is that the oscillating GC-skew plot is not due to multiple inversions. Finally, the GC-skew plot of F. nucleatum could be partly due to multiple replication origins constantly shifting the values, but this situation has not been observed in any bacterial species. Genome sequence similarity analysis A sequence similarity search performed by BLASTP [ 34 ] against the whole available database reveals homology to over 150 bacterial and archaeal species. More than a quarter of the genes had no significant hit or a hit to a eukaryotic species. 64.6% of the hits went to Firmicutes species and 35.4% to other bacterial species (Table 1 ). These results seem congruent with the ribosomal phylogeny. However, from the hits to Firmicutes, 267 ORFs (representing 12.9% of the total genome) had a hit in only one genus within this bacterial group together with hits in another taxa, a feature suggestive of HGT to or from these bacteria, that are very numerous and diverse in the dental plaque [ 16 ]. Table 1 General function of F. nucleatum genes, divided by group of best BLAST hit 1 . Function Category Archaea CFB group Low GC Gram pos α, β, γ Proteo δ, ε Proteo Other eubact Spiro-chaetes Eukarya/No hit Total Aa biosynthesis 2 (6.5,3.9) 22 (71,2.3) 3 (9.7,2.0) 2 (6.5,1.9) 1 (3.2,1.4) 1 (3.2,0.2) 31 Cofactors and carriers biosynth. 1 (1.4,2.0) 59 (83,6.2) 4 (5.6,2.6) 2 (2.8,1.9) 2 (2.8,2.7) 2 (2.8,2.2) 1 (1.4,0.2) 71 Cell envelope 1 (0.6,2.0) 6 (3.6,11.3) 45 (27.1,4.7) 24 (14.5,16) 12 (7.2,11.4) 7 (4.2,9.5) 12 (7.2,13.3) 59 (35.5,10) 166 Cellular processes 2 (3.1,3.9) 2 (3.1,3.8) 35 (54.7,3.7) 4 (6.2,2.6) 2 (3.1,1.9) 2 (3.1,2.2) 17 (26.6,2.9) 64 Central intermed. metab. 3 (7.0,5.9) 3 (7.0,5.7) 24 (55.8,2.5) 6 (14.0,4.0) 2 (4.7,1.9) 3 (7.0,3.3) 2 (4.7,0.3) 43 DNA metab. 4 (5.5,7.8) 45 (61.6,4.7) 5 (6.8,3.3) 3 (4.1,2.9) 4 (5.5,5.4) 1 (1.4,1.1) 11 (15.1,1.9) 73 Energy metab. 1 (0.9,2.0) 9 (7.8,17.0) 82 (70.7,8.6) 6 (5.2,4.0) 6 (5.2,5.7) 4 (3.4,5.4) 3 (2.6,3.3) 5 (4.3,0.9) 116 Lipid metab. 21 (75.0,2.2) 3 (10.7,2.0) 3 (10.7,2.9) 1 (3.6,0.2) 28 Hypothetical prots. 3 (2.1,5.7) 43 (29.9,4.5) 4 (2.8,2.6) 2 (1.4,1.9) 3 (2.1,4.1) 9 (6.2,10.0) 80 (55.6,14) 144 Other categories 1 (5.3,1.9) 15 (79,1.6) 1 (5.3,0.7) 1 (5.3,1.0) 1 (5.3,0.2) 19 Protein fate 4 (7.0,7.5) 33 (58,3.5) 7 (12,4.6) 4 (7.0,3.8) 4 (7.0,5.4) 1 (1.8,1.1) 4 (7.0,0.7) 57 Protein synthesis 1 (0.9,2.0) 1 (0.9,1.9) 88 (78,9.2) 5 (4.4,3.3) 7 (6.2,6.7) 8 (7,10.8) 3 (2.7,0.5) 113 Nucleotides metab. 1 (2.9,2.0) 1 (2.9,1.9) 28 (80,2.9) 2 (5.7,1.3) 1 (2.9,1.0) 2 (5.7,2.7) 35 Regulat. functions 1 (2.0,2.0) 2 (3.9,3.8) 29 (57,3.0) 3 (5.9,2.0) 3 (5.9,2.9) 2 (3.9,2.7) 5 (9.8,5.6) 6 (11.8,1) 51 Signal transduction 3 (60.0,0.3) 2 (40,0.3) 5 Transcription 1 (5.0,2.0) 1 (5.0,1.9) 15 (75,1.6) 1 (5.0,1.4) 1 (5.0,1.1) 1 (5.0,0.2) 20 Transport/binding proteins 11 (5.7,21.6) 2 (1.0,3.8) 91 (47.4,9.5) 23 (12,15.2) 15 (7.8,14.3) 14 (7.3,18.9) 18 (9.4,20.0) 18 (9.4,3.1) 192 Unclassified 11 (7.0,21.6) 3 (1.9,5.7) 72 (45.6,7.5) 16 (10,10.6) 9 (5.7,8.6) 1 (0.6,1.4) 6 (3.8,6.7) 40 (25.3,6.8) 158 Unknown function 3 (3.4,5.9) 3 (3.4,5.7) 50 (57.5,5.2) 4 (4.6,2.6) 10 (11.5,9.5) 5 (5.7,6.8) 3 (3.4,3.3) 9 (10.3,1.5) 87 Hipothetical function 8 (1.3,15.7) 12 (2.0,22.6) 155 (26.1,16.2) 31 (5.2,20.5) 21 (3.5,20.0) 16 (2.7,21.6) 24 (4.0,26.7) 327 (55,55.6) 594 Total 51 53 955 151 105 74 90 588 2067 1 First number between brackets indicates the percentage of genes with a best match in a given taxon that have the function indicated on the row heading. Second number between brackets indicate the percentage of genes with a given function that have a best match in the taxon indicated on the column heading. When the top hits to the different groups are plotted, the matches outside the Firmicutes are scattered along the Fusobacterium chromosome and, in many cases, they are clustered (Figure 3 ). There are 36 cases of clusters of 4 or more contiguous genes that have a best hit outside the Firmicutes, many of which still preserve some gene order compared to their phylogenetically unrelated counterparts. Five of these cases are shown in more detail in Figure 4 . Gene arrangement conservation with distantly related groups is a strong indication of HGT events. Figure 3 Gene-position plot with a reconstruction of vertical descent and potentially-transferred genes across F. nucleatum genome. Plus signs indicate genes whose phylogenetic affiliation to a certain group on the left is supported by BLAST analysis only. Crosses indicate genes whose phylogenetic origin is supported by BLAST and by one or two other methods (phylogenetic tree reconstruction and gene order conservation). Thirty-six clusters are indicated containing four or more consecutive genes that appear to have a xenologous origin (i.e. a phylogenetic affiliation outside the Firmicutes). Details of these clusters are explained in Table 2. Arrowheads at the top indicate the position of transposases. Arrowheads at the bottom indicate position of phage-related genes. Plus signs and crosses indicate potential transfers to/from Firmicutes (Firm), Spirochaetes (Sp), alpha-beta-gamma Proteobacteria (αβγ-Pr), delta-epsilon Proteobacteria (δε-Pr), Cytophaga-Flexibacter-Bacteroides Group (CFB), other bacterial groups (Others), Archaea, or genes that are consistent with the phylogeny (CWP) shown in Figure 1 and additional file 1. Figure 4 Gene order conservation in some representative cases of potentially-transferred clusters. Homologous genes are shown with the same colour in F. nucleatum and the species with which it is compared. Small black boxes represent short orphan genes. Non-contiguous genes are separated by an interrupted line. Genes are not drawn to scale. It is interesting to note that there are 40 transposase ORFs in the F. nucleatum genome and 73 assignments of possible IS elements [ 3 ]. Thirty-four of the transposase sequences are at the flanks of putative transferred genes, whereas 6 were between core genes (Figure 3 ). There are also two integrase genes, both at the edge of putative HGTs. In addition, Kapatral and collaborators [ 3 ] described that active and remnant IS-elements are flanking many genes with high similarity to proteobacteria. Among these there are outer membrane proteins, hemolysin precursors and activators, pyrophosphate synthesis genes and others. Another possibility for the insertion of xenologous sequences would be through the action of bacteriophages. In F. nucleatum , 31 genes were found to have homologs in phage regions of other bacteria (Figure 3 ) and 13 on plasmids. Small cryptic plasmids containing mobile elements are frequently found in F. nucleatum strains [ 35 ]. In addition, six phage contigs encoding 110 ORFs have been identified in its sister subspecies vincentii . In this bacterium, the phage genes have homology to Gram positive and Gram negative phages, with an average GC content of 28% and a similar codon usage to the chromosome [ 33 ]. Thus, it is possible that an old phage infection is partly responsible for the mosaic genome of F. nucleatum . For example, a region with 6 ORFs presents homology to the proteobacterial bacteriophage P2, a phage that has been shown to be responsible for HGT episodes in some E. coli strains [ 36 ]. Looking into more detail at the gene clusters with a best hit outside Firmicutes it was found that many genes were involved in typical gram-negative features, mainly membrane-associated functions (Table 2 ). For example, the segment of genes FN1893-FN1897 includes 3 homologs found in Salmonella typhimurium , coding for surface-exposed virulence proteins and a membrane-associated gene involved in D-ribose transport. Twenty-eight of the 36 clusters of 4 or more contiguous genes had a function related to the outer membrane, the periplasm or pathogenecity typical of proteobacteria, CFB group species or spirochaetes. Although this would have to be expected since those organisms posses also Gram negative cell walls, the similarity was always to the same groups and much higher than what would be expected based on a distant common origin of the corresponding Gram negative feature. The conservation of gene order and relatively high similarity to groups present in the dental plaque (see below) also hints at a secondary acquisition by HGT. Therefore, the interpretation of F. nucleatum as a gram-positive bacterium with gram-negative clothing [ 6 ] appears quite realistic, with the xenologous sequences being especially relevant in membrane-associated functions associated to a gram-negative cell wall. Table 2 Clusters of 4 or more consecutive genes with a best match outside the Firmicutes 5 . # Putative function Observations 1 Sequence similarity 2 Genes Sinteny 3 1 Transposase + 4 hypothetical proteins of similar sequence Flanked by 3 short orphans 4 One of proteins is a short ORF 24–32 FN1511 to FN1515 2 KDO (LPS core synthesis) + endonuclease and DNA pol III Includes a short orphan 31–58 FN1561 to FN1576 3 Peptide ABC transporter It includes two long (>1500 bp) hypothetical proteins 30–56 FN1650 to FN1656 4 sysnthesis of LPS (O chain) + phosphatidylcholine synthesis Split by a hypothetical protein and 3 short ORFs 37–61 FN1661 to FN1668 5 carbohydrate trasnport-pot operon (periplasmic binding prot dependent transport) Split by long spacer 22–55 FN1792 to FN1800 Thermotoga maritima 6 periplasmic binding protein dependent cation (Mn2+, Zn2+) transport posibly Co2+ Flanked by transposase and archaeal best-match ORF 24–56 FN1807 to FN1814 Pseudomonas putida 7 DNA pol III gamma and tau subunits and TonB OM export system Flanked by hypothetical orphans 25–36 FN1830 to FN1834 Helicobacter pylori 8 Periplasmic amilase and ribose ABC trasnporter Short orphan in the middle 23–32 FN1893 to FN1897 9 LPS synthesis and/or decoration and outer membarne stabilization Flanked by 3528 bp hypothet. protein with eukaryotic best-match followed by long spacer 25–77 FN1908 to FN1911 Geobacter sulfurreducens 10 capsule biosynthesis Includes 2 short ORFs (possible HIPA pseudogenes) 23–46 FN1997 to FN2003 Bordetella bronchiseptica Yersinia pestis 11 Slow porin homologous to OmpA ( Bacteroides ) or Opr ( Pseudomonas ) Split by a long spacer with some homology to membrane proteins. Includes 2 short ORF 23–49 FN2056 to FN2062 12 Hypothetical exported 24-amino acid repeat protein Includes 4 short ORFs (one of them with homology to subunit δ of DNA Polym. III) 34–45 FN2110 to FN2122 13 24 aa repeat protein like in cluster 23 Protein match to Helycobacter hepaticus 31–53 FN0023 to FN0028 14 Endonuclease + 3 genes implicated in porfirinic siderophore synthesis Flanked by short orphan 24–65 FN0185 to FN0188 Haemophilus influenzae 15 DNA helicase + peptide transporters High gene order conservation in an archaeal species 28–42 FN0191 to FN0197 Methanosarcina acetivorans 16 Sugar ABC transporter Short spacers/overlapping genes 31–48 FN0217 to FN0220 Escherichia coli 17 Large cluster of hemolysin/ hemagglutinin containing hemagglutinin FhaB Largest bacterial protein. Some degraded hemolysin copies found throughout genome 23–26 FN0290 to FN0293 Escherichia coli 18 ABC iron/haemin transporter with periplasmic binding protein Flanked by long spacer 27–47 FN0300 to FN0303 Methanosarcina acetivorans 19 Periplasmic binding protein dependent iron transport system Physically linked to other iron transport genes of Gram positive and Archaeal match 34–49 FN0309 to FN0312 Bordetella bronchiseptica 20 NA+/H+ antiporter + 3 genes of unknown function Split by a tRNA gene. Includes 2 short orphans 33–53 FN0350 to FN0354 Treponema denticola 21 Two clusters of genes implicated in drug efflux (detoxification) extrusion out of OM Flanked by two orphans of 402 and 618 bp 21–37 FN0515 to FN0519 Vibrio cholerae 22 Mixed functions cluster 30–44 FN0524 to FN0527 23 LPS synthesis and/or decoration and outer membarne stabilization Includes recA and recX proteins with best match to Caulobacter and Vibrio 29–100 FN0538 to FN0548 Haemophilus ducreyi 24 Structural lipoprotein with release and mureine anchoring components Flanked by short ORF 30–46 FN0579 to FN0582 Helicobacter hepaticus 25 Membrane-related functions + Fe-S oxidoreductase Includes a short hypothetical protein with biased codon use 32–55 FN0734 to FN0739 26 Haemin uptake with periplasmic binding protein iron acquisition Haemin genes tightly-linked, probable operon 24–59 FN0766 to FN0771 Campylobacter jejuni 27 Biotin biosynthesis Most spacers are short, possible cotranscription 31–55 FN0846 to FN0852 Campylobacter jejuni 28 Hydrolase + protease + aromatic compound synthesis Mixed function cluster 30–47 FN0869 to FN0873 29 Iron ABC transporter Flanked by a short orphan with biased codon usage 45–71 FN0879 to FN0882 Treponema denticola 30 Membrane proteins 1 st and 2 nd genes probably permeases 22–37 FN1030 to FN1033 Photorhabdus luminescens 31 Lipase B componet of type II secretion system + 24 aa repeat protein+ bacterioferritin All proteins of short length 26–34 FN1075 to FN1079 32 KDO (cetodeoxyoctulonic acid biosynthetic operon) KDO is a component of LPS core in Fusobacterium and many Gram negatives. 31–100 FN1221 to FN1224 33 Eps synthesis + EpsF (secretion of proteins/large biomolecules) Possible tandem duplication 30–47 FN1242 to FN1245 Ralstonia solanacearum 34 LOS choline decoration + Ton B (biopolymer transport through Outer Membrane) Includes a short ORF (a degraded copy of a biopolymer transporter) 29–40 FN1306 to FN1312 Pseudomonas aeruginosa 35 ABC transporter system Flanked by short orphan followed by a transposase 30–69 FN1346 to FN1355 Treponema denticola 36 ABC amino acid transport system liv G-M operon; biased and homogeneous codon usage 50–62 FN1428 to FN1431 Bifidobacterium longum 1 "Split" indicates a cluster separated by a long intergenic spacer, the two parts of the cluster generally coding for different functions. 2 Range of sequence similarity among the genes from each cluster compared to their BLAST top hits. 3 Representative species with similar gene order. 4 An "orphan" gene is defined as an ORF with an unknown function and no BLAST similarity in the current database. Short orphans (<500 bp) are likely to be pseudogene remnants or other non-functional regions (Mira et al . 2002, Davies et al . 2004). 5 Only protein coding genes are included in the analysis. Phylogenetic, gene-order and compositional analyses BLAST analysis has the advantage of giving a closest similarity match for almost every gene. It is however a crude method that can give as the top sequence similarity hit a species that is not the closest from a phylogenetic point of view [ 37 ] and it could also be much influenced by the undersampling of certain poorly-sequenced groups, such as the Fusobacteria. For example, when the BLAST top 10 best matches are considered, less than 70% of the hits fall on the same bacterial taxa as the top hit. In top hits to archaea, a domain from which fewer sequences are available, 58% of the top ten best matches hit groups other than Archaea. Since we used the top hit to designate potential phylogenetic origin, some degree of inaccuracy is expected. Thus, to complement the BLAST analysis we used a phylogenetic and a gene-order analysis, indicating in Figure 3 whether their results do or do not support the BLAST results. In the phylogenetic analysis, trees were constructed based on the sequence of each individual gene with sufficient homologs in the database (see experimental procedures section for details). Over 1200 trees were generated and analysed to detect a phylogeny either congruent with the ribosomal one, or suggestive of HGT. Almost two thirds of the trees could be resolved, and corroborated the high degree of gene transfer, with at least 25% of the genome being of xenologous origin (Table 3 ). Only 8.4% were consistent with the ribosomal phylogeny, and over 25% indicated a potential HGT to or from Firmicutes species. Part of the latter could also be due to multiple losses of the genes in most Firmicutes genera. However, it is not unreasonable to think that they could have been transferred between Fusobacteria and typical Firmicutes (particularly clostridiales and streptococci ), which share the mouth and dental plaque ecosystem [ 16 ]. The phylogenetic analysis method, therefore, suggests a 25–50% of gene transfer. Although only trees with a bootstrap value over 500 were considered (see methods section), these numbers must be taken with caution. Given the distant relationship of F. nucleatum with most sequenced genomes, a weak phylogenetic signal may remain for many trees. The branching pattern in trees is also influenced by other variables like different rates of evolution for different genes, method of alignment or number of species included. Like in the sequence similarity method, the data presented are inferences based upon the data, and the limitations of each method should be kept in mind. Table 3 Percentage of F. nucleatum ORFs classified by the taxa of potential origin. Sequence similarity method (BLAST) Phylogenetic trees method Gene order conservation Number of genes analyzed 2067 1236 738 Root of Firmicutes 1 33.28 % 8.41 % 35.1 % Inside Firmicutes 2 12.92 % 25.8 % 15.45 % CFB group 2.56 % 2.27 % 4.06 % α, β, γ Proteobacteria 7.34 % 10.3 % 21.0 % δ, ε Proteobacteria 5.07 % 3.4 % 5.7 % Spirochaetes 4.35 % 4.32 6.37 % Other eubacteria 3.58 % 4.53 4.2 % Archaea 2.46 % 1.13 % 0.95 % No hit, hit to eukaryotes, uncertain/unresolved 28.4 % 38.7 % 7.45 % 1 Genes consistent with the 16S-23S and ribosomal proteins phylogeny. 2 It indicates possible HGT to/from Firmicutes. Another method used was based on the conservation of gene order among certain gene clusters, a character that can be used in phylogenetic reconstructions [ 38 , 39 ]. Only 738 F. nucleatum protein-coding genes belonged to clusters of 2 or more genes that had some order conservation in other bacteria. From these, 35% had the same order as most Firmicutes (Table 3 ), suggesting vertical inheritance. Over 15% of the genes belonged to clusters whose gene order was more consistent with HGT from this group (i.e. same order as only one of the Firmicutes genomes). The extent of HGT from Firmicutes could be overestimated if the genes are ancestral but subsequently lost in most Gram-positive lineages. This being the case, the addition of vertically-inherited genes and genes inside Firmicutes in Table 3 would indicate an upper limit of genes consistent with the ribosomal phylogeny. Even if HGT from Firmicutes is not considered, 42% of the genes were assigned as HGT from other bacterial taxa based on the gene-order method. These dramatic figures suggest again that the genome of F. nucleatum could be an amalgamation of genes from different groups, particularly those of species that inhabit mammalian hosts in general and the mouth niche in particular. A summary figure showing the outcome of the three methods is published as supplementary material [see additional file 1 ]. The discrepancies between the three methods can be partly influenced by the different sample sizes used (Table 3 ). In addition, it must be noted that most of the discrepancy appears in the Phylogenetic Trees method, where a very low percentage of vertical inheritance was detected. In this analysis, over 38% of the trees were unresolved, introducing an important degree of variation. It is therefore possible that many of the genes giving uncertain phylogenies are consistent with vertical inheritance, but the phylogenetic signal is too weak to give a clear-cut tree. The gene-order method could give higher numbers of horizontal transfers if operons are more likely to be transferred than single genes [ 40 ]. Thus, all three methods have its limitations, and although the importance of HGT is clear, the numbers obtained may be subject to certain bias imposed by the methodology [ 41 ]. Deviations from genomic GC content and codon usage have been used to infer potential gene transfers across bacteria [ 42 , 43 ]. However, only 40 genes with significantly extraneous DNA composition were found in F. nucleatum [ 44 ] suggesting that many transfers could come from low-GC species or that many of the transfers occurred long ago, allowing the xenologous genes to ameliorate and homogenize its characteristics with those of the recipient genome [ 45 ]. In addition, the extremely low GC content of F. nucleatum could make this method less discriminatory [ 46 , 47 ]. A few potential transfers were identified this way, including a cluster spanning two iron-sulphur binding proteins and two arsenic pump-driving ATPases. Another interesting case was a glutamate fermentation cluster with closest similarity and gene order conservation to the clostridial species Acidaminococcus fermentans . This represents a typical case of potential HGT from the Firmicutes that could be masked in a BLAST analysis as a vertically inherited cluster. As the tree and gene-order methods show, the amount of HGT from/to the Firmicutes species could be as high as 15–25%, assuming that the percentages are maintained among the genes that we could not analyse because the trees were unresolved or because they were not part of conserved-order clusters. Chimeric enzymes and operons To explore the possibility that the chimeric nature of Fusobacterium may apply not only to its genome but also to some of its metabolic pathways and enzymes, some specific cases were looked at in more detail. A potential example includes the RNA polymerase, where the β' subunit has a best BLAST hit to spirochaetes as well as the RNA polymerase sigma-E factor. This is confirmed by comparative analysis of domain architecture across bacteria [ 48 ]. An interesting instance is given by the phenylalanyl-tRNA synthetase, in which the α and β chains have a Clostridium and Geobacter (delta-proteobacteria) best sequence similarity match, respectively. The tree analysis confirms that the β chain is likely to have a proteobacterial origin. Interestingly, although the β chain is located in a proteobacterial cluster (at the edge of cluster 12), it is contiguous to the Firmicutes related α chain gene, separated by a very short spacer without a promoter. This exemplifies how selection may have put together two functionally related genes, presumably to ease cotranscription, even though their phylogenetic origin appears to be different. Another example is given by an iron ABC transporter operon formed by a periplasmic binding protein followed by two iron permeases. A similar structure is repeated two other times in the subsequent genes (Figure 5a ), forming a long iron transport system. Remarkably, the first operon is found in identical gene order in the archaeon Methanosarcina acetivorans , to which it presents the highest sequence similarity, whereas the second and third operon appear to be a blend of genes with relatedness to firmicutes, proteobacteria, spirochaetes and Thermotoga . These genes are present in many Gram negatives including Helicobacter and other Proteobacteria [ 49 , 50 ]. Thus, assuming that some of these genes have a xenologous origin, they must have been selected to occupy a precise gene order to maximise its function within the iron transport system. Another fascinating case of a potential chimeric gene system is that of transport of dipeptides (Figure 5b ). There are as many as five dipeptide transport operons in F. nucleatum , this time dispersed along the genome. Although one of the sets has best matches to Firmicutes species, another one appears to be of spirochaete source (also present in the same gene order in Methanosarcina ), whereas the other three are, according to sequence similarity, gene order and tree reconstruction a mixture of genes with archaeal, Firmicutes and proteobacterial origin. A third case can be seen in the three copies of a hemin transport system located away from each other and formed by a hemin receptor and the genes hmuT , hmuU and hmuV . Although the different taxonomical origin analysis methods are not always consistent, two of the hemin operons are probably of proteobacterial origin (Figure 5c ). The other one has a closest gene order to the spirochaete Treponema denticola , also an inhabitant of the dental plaque, but is absent in its close relative T. pallidum , suggesting again that gene transfer is facilitated across the bacteria that occupy this specialised niche. Figure 5 Chimeric operons (metabolic pathways of putatively mixed origin) in F. nucleatum . Arrowed boxes represent gene orientation, coloured by BLAST top hit. White boxes: top hit to Firmicutes; grey boxes: top hit to Archaeal species; black boxes: top hit in Gram negative species. Numbers below boxes indicate the percent of top ten hits that have matches in Firmicutes. Names above indicate gene names (I-BP: Iron binding protein; NIP: Nitrogenase iron protein; Oxdtase: Oxidoreductase; B, C, D, F: dipeptide permeases B, C, D, F; BP: dipeptide binding protein; Tr: ABC transporter; unk: unknown function gene; Rec: Hemin receptor). Best match taxa by the phylogenetic tree and gene order methods are also indicated (A: Archaea; Pr: Proteobacteria; Sp: Spirochaetes; CP: consistent with (ribosomal) phylogeny; O: other eubacteria; x: unresolved; --: not analysed. Plus signs indicate unusual DNA composition by the method of García-Vallvé et al . 2003. Remnants of HGT An indication of massive gene transfer events comes from looking at intergenic spacer regions of F. nucleatum . Although average spacer length in this species is 115 bp, there are many long spacers of 500 bp and higher scattered across the genome. It was found that 21 of these long spacers were located at positions flanked by a "core" gene (that with a low-GC Gram-positive best match) and a potential transferred gene, whereas only 8 appeared between core genes. Since intergenic spacer regions are known to increase in length as a result of genomic rearrangements and pseudogene formation [ 51 ], many of these long spacers might be signatures of ancient HGT events. In agreement with this view, another 17 long spacers were located inside gene clusters of a putative Gram negative or archaeal origin. We hypothesize that these long non-coding regions are remnants of transferred genes that were not selected for and have been mostly erased. When DNA sequence similarity searches are done with these long spacers located inside xenologous clusters, some significant matches are found to other regions of the genome. For example, the long spacers inside clusters 4, 8 and 11 all have some sequence similarity (more than 85% sequence identity over 125 bp or more using BLAST analysis, E-value <10 -5 ) to one another and to other five long spacers scattered throughout the genome. In all cases except one, these long spacers are flanked by outer membrane proteins of Gram-negative origin, suggesting that they may represent remnants of old membrane-associated genes. A similar case is that of the long spacer located after the hemolysin activator protein precursor (FN1818), which shows high sequence similarity to a hemolysin activator located someplace else in the genome (and to another spacer and a short ORF with unknown function). Another potential signature of ancient transfers subsequently erased is the high number of ORFs without a match in the complete, non-redundant NCBI database (including its closely related subspecies vincentii ), spanning 450 sequences that represent 20% of the genome. On average, these orphan genes are extremely short (440 bp, versus 1040 bp for the rest of genes with a match on the DNA database, see Figure 6 ), suggesting that they do not represent real genes [ 52 , 31 ]. Overannotation of short ORFs that are not functional is more common on GC-rich genomes due to a lower probability of stop codons [ 53 ], but F. nucleatum is just 27% G+C. It is therefore possible that many of these short ORFs are eroded pseudogenes or remnants of fragmented genes, as it has been demonstrated in Rickettsia [ 54 ], where many genes appear to be under low selection coefficients in its intracellular environment. In Fusobacterium , it is likely that many transferred genes were not useful and got eliminated, a process known to happen very rapidly [ 55 ]. This would explain the high number of short ORFs without significant BLAST matches on the database, as small fragments of genes may have accumulated enough mutations to make them frequently unrecognisable by sequence similarity. For many of these small ORFs with no significant matches, some low sequence similarity is found to gram-negative outer membrane proteins (e.g. tolA ), glycine permease, periplasmic-like proteins, etc. Figure 6 Length frequency distributions for F. nucleatum proteins. Genes are divided by the group with closest sequence similarity match. ORFs without sequence similarity on the non-redundant NCBI database (orphan genes) are significantly shorter than the rest. In addition, some short ORFs appear to be degraded fragments of bigger genes. For example, there are 3 sequences with similarity to HIPA proteins, one of which is less than half the length of the other two. As it also has a very biased codon usage, it is likely that it represents a degraded remnant of this protein. The 3 copies of integrases scattered across the genome show another case. Two of them are around 900 bp long and have a normal codon usage. The third copy (FN0402) is only 177 bp long, is flanked by a long spacer and has a very skewed codon usage. In general, the codon usage of these orphans is very biased (mean corrected χ 2 values of 0.47 versus 0.22 for the rest of the genome). As it is unlikely that all these short ORFs are highly expressed, we believe that this biased codon utilization is reflecting very divergent pseudogene fragments. Thus, the picture that emerges is that of massive gene transfer leaving many non-coding segments that are remnants of unnecessary genes and genomic rearrangements. Conclusions The genome of F. nucleatum possesses a remarkable amount of patchiness with any kind of phylogenetic analyses used. This can be said to a certain degree of some other genomes (see for example [ 56 ]). One possible explanation for this kind of results is an undersampling of the group considered what gives only very distant and hence uncertain similarities to a variety of prokaryotic groups. This might be the case for part of the Fusobacterium genome that gives very weak and uncertain phylogenetic signal. However, the observation that certain genes and operons are shared by distantly related species that inhabit the dental plaque (for example, the spirochaete T. denticola , the proteobacteria Campylobacter and the CFB P. gingivalis ) points to HGT as the most likely origin of these genes. Even less apparent, our work suggests multiple episodes of gene transfer to or from phylogenetically-related bacteria, like certain Firmicutes species (such as the cariogenic bacterium S. mutans or some Clostridia), that might be confounded with vertically inherited traits. The origin of the Gram-negative cell wall found in Fusobacterium requires special consideration. Some type of Gram-negative cell wall seems to be the default phenotype in Bacteria (see, for example [ 57 ]), being found in most deeply branching groups. Moreover, even some deep branches of the Firmicutes contain organisms (such as Sporomusa and Desulfotomaculum ) with Gram-negative cell wall structures [ 58 , 59 ]. On the other hand, it has also been proposed that the Gram-positive cell wall is the default structure [ 60 ]. It might be argued that Fusobacterium is a remnant of the ancestral cells predating the bacterial radiation that originated either Gram-positive or Gram-negative cell walls. This is supported by phylogenetic inferences based on conserved indels, which place Fusobacteria at an intermediate position between Gram-positive and Gram-negative taxa [ 61 ]. However, in light of our results this explanation does not seem likely. Fusobacterium does not show any primitive trait and its outer membrane and transport mechanisms show all the characteristics of any sophisticated Gram-negative cell wall. In addition, many of the outer membrane proteins are closest to specific taxa (mainly to proteobacterial species) and not equally dispersed among species with a Gram-negative cell wall. Thus, many of the genes involved in the construction of the Gram-negative outer membrane have probably been horizontally transferred. The extent of this transfer deserves further examination. If we assume that Fusobacterium evolved after the Gram-positive/negative divergence on the low-GC Gram-positive lineage, massive HGT is the most likely explanation for the formation of the outer membrane. On the other end of possible explanations, most genes of the outer membrane would already be present in the common ancestor of fusobacteria and Firmicutes, where a massive loss would be responsible for the differences observed today. Recently, the idea of gene pools that are characteristic of certain environments has been advanced to explain the large number of common genes among groups of thermoacidophiles distantly related by ribosomal phylogeny [ 62 ]. The presence of a common pool of dental plaque genes is not unlikely in light of the results described here. However, the time scale of the adaptation to the latter habitat is much shorter that that of thermoacidophiles and can be probably estimated around the origin of mammals (about 120 million years). Even going backwards to the origin of the vertebrate's intestine it would put the selective pressure for these gene combinations to originate no earlier than 400 Myr ago. Former chimeric genomes have been explained as selected by strong environmental pressure. The case of Thermotoga is paradigmatic, a hyperthermophilic bacteria that is assumed to have recruited genes from the archaeal hyperthermophiles to reach its unusual (for bacteria) thermotolerance. Here (as in the case of Methanosarcina , a mesophilic anaerobe) there is not such an obvious explanation. F. nucleatum natural habitat seems to be the dental plaque of mammals, a rather unique and special environment that probably requires very special features to survive. Strong adhesion mechanisms, such as those found often in the Proteobacteria, probably represent an essential ability for survival in the early stages of plaque formation, particularly for non-motile cells. Also the mucose-associated immune system that prevails in the mouth of mammals could have acted as a strong selective pressure favoring the Gram-negative envelopes that are often less immunogenic and easier to disguise thanks to the LPS polysaccharide O chain [ 63 ]. Thus, it is not difficult to envisage that at a point in time, probably associated to the rise of mammals, a strong selective pressure might have existed for a cell with the metabolic apparatus of Clostridia for amino acid fermentation but with the adhesive and immune camouflage paraphernalia of the Proteobacteria. It is remarkable to note that many of the genes that determine the lifestyle of Fusobacterium and its interaction with the environment, such as peptide transport systems, cell adhesins and outer membrane components have probably been acquired by gene transfer. It is therefore not only the number of horizontal transfers but also their contribution to niche adaptation that makes the HGT mechanism of dramatic impact on genomes. It is interesting that some of these genes are shared by different organisms inhabiting the dental plaque. From an applied point of view, some of these highly transferred genes are likely to provide a critical advantage in the establishment and adaptation of the bacteria to their niche, and could be used as potential targets for antimicrobial agents. Methods Phylogenetic trees rRNA and evolutionary conserved proteins trees The different rRNA and conserved protein data sets were analyzed with Bayesian methods using the program MrBAYES 3 [ 64 ]. For the fusion of 16S+23S rRNA sequences, the GTR model with a Γ law (8 rate categories) and a proportion of invariant sites to take among-site rate variation into account was used. A similar procedure was used to construct the trees based on evolutionary conserved proteins (a mixed substitution model and a Γ law with 8 rate categories and a proportion of invariant sites were applied). The evolutionary conserved proteins were defined as those found in all sequenced species of Bacteria and assumed to form part of the minimal genome necessary for life [ 65 , 66 ]. The list was extracted from [ 23 ] but removing the genes for which paralogous ORFs were found. In all cases, the Markov chain Monte Carlo searches were run with 4 chains for 1,000,000 generations, with trees being sampled every 100 generations (the first 2,500 trees were discarded as "burnin"). Concatenated ribosomal proteins tree The amino-acid sequences of ribosomal genes S1–S20 and L1–L35, excluding S1, S14, L24, L25, L30, L31, L32 and L33, were retrieved from the KEGG website from a total of 60 different bacteria. The bacteria chosen were all those represented in the KEGG ribosomal genes ortholog table [ 67 ], except Rickettsia prowazekii , Rickettsia conorii , Wigglesworthia brevipalpis , and Buchnera aphidicola , and with the addition of Bacteroides thetaiotaomicron and and Desulfovibrio vulgaris . An alignment was generated for each ribosomal gene, using the Clustalw software with default parameters [ 68 ]. When two or more paralogs were found in a species, the most divergent of the paralogs was removed from the alignment. A concatenated alignment including the species for which all of the selected ribosomal genes were present was generated. A neighbor-joining tree with 1000 bootstrap replicates was produced from the alignment using Clustalw [ 68 ], excluding positions with gaps, and correcting for multiple amino-acid substitutions (Kimura correction). The tree was visualized with NJPLOT [ 69 ]. Exclusion of ribosomal proteins was based on the following: S14 has been shown to be subject to horizontal transfer [ 70 ], L24 is truncated and split in Fusobacterium nucleatum , S1 is absent/truncated in the Mollicutes subgroup of the low-GC gram-positives, L25, L30, L31, L32, L33 contained a high number of paralogs and/or were absent in several key species. Methods for detecting HGT Blast method The protein sequences of Fusobacterium nucleatum subsp. nucleatum ATCC 25586 were retrieved from . Peptide sequence database of all non-redundant GenBank CDS translations + PDB + SwissProt + PIR was retrieved from . We performed an all against all BLASTP [ 34 ] search of each protein in Fusobacterium nucleatum subsp. nucleatum ATCC 255586 against peptide sequence database. We then recorded the top hit for each protein sequence with an E-value of 10 -5 , filtering the hits whose sequence identity and length was lower than 30 and 50%, respectively. We categorized all the hits into 8 categories as belonging to the CFB group, Firmicutes bacteria, α,β,γ-Proteobacteria, δ,ε-Proteobacteria, Spirochaetes, other Bacteria, Archaea and Eukaryotes/no hit. Hits to Firmicutes (the group to which Fusobacterium appear to be more closely-related) were refined by further BlastP analysis between F. nucleatum and the 31 sequenced bacteria available from this group. If the gene had a homolog in only one genus from all the available low-GC gram-positive species, it was considered a HGT event from/to this group. If it was present in more than one genus it was considered vertically inherited and consistent with the ribosomal phylogeny. There were 61 cases of genes found in more than one genera from a single subgroup of this taxon (i.e. present only in the Clostridiales, the Bacillales, the Mollicutes or the Lactobacillales). These can be equally explained by HGT or by common descent and were conservatively assigned to the vertical inherited category. Phylogenetic trees method For each F. nucleatum gene, the protein sequences of up to 50 best blast hits with e-value lower than e -5 were retrieved (the hits were identified by the "Blast method" described above). All sequences were then automatically clustered with the Clustalw alignment tool with default parameters. A neighbor-joining tree with 1000 bootstrap replicates was generated from the resulting alignment, using Clustalw with default parameters. The trees were visualized with NJPLOT [ 69 ]. In all cases, the bootstrap values at the nodes chosen for a decision on taxonomical assignment had to be over 500. Assignment of the F. nucleatum genes to a taxonomic group was done using the following criteria: Low-GC gram-positives A F. nucleatum gene was determined to originate from the firmicutes if it was found in the tree most closely associated with at least 5 different species from that group, or with at least 3 species from 2 different subgroups (where the subgroups were: mollicutes, bacillales, lactobacillales, clostridiales). If the F. nucleatum gene branched at the base of the firmicutes, the gene was assigned as being consistent with phylogeny; otherwise, it was assigned as a potential horizontal gene transfer (HGT) from the firmicutes. Proteobacteria Same as described above (low-GC gram-positives), the subgroups in this case were:alpha-, beta-, gamma-, gamma-entero-, delta-, and epsilon-proteobacteria. In the case of the proteobacteria, all F. nucleatum genes with trees fulfilling this criteria were assigned as HGTs from proteobacteria. Note that a distinction was made between the grouping of the alpha-, beta-, and gamma- proteobacteria, and the grouping of the epsilon- and delta- proteobacteria whenever possible. Archaea To be assigned as originating from the archaeales, the F. nucleatum had to be closest to at least 3 species, and there had to be a clear association between the two groups, i.e. the branches were relatively short, and the tree topology did not resemble a "star phylogeny". High-GC gram-positives, Cyanobacteria, Chlamydiales To be assigned to these groups, the F. nucleatum gene had to be found closest in the tree to at least 3 different species, there had to be a clear association (see above in "Archaeales") and there should have been no obvious evidence of gene transfer from the Fusobacteria. Evidence of transfer from the Fusobacteria would be when, apart from the association to some species of the high-GC gram positives (or Cyanobacteriales, or Chlamydiales), the tree placed the F. nucleatum gene in agreement with the accepted species phylogeny (just outside of the firmicutes). Aquifales, Deinococcales The F. nucleatum gene had to be found closest to at least 1 species from that group (Aquifales, or Deinococcales). A clear association was necessary, as well as no evidence of transfer from Fusobacteria (see above). Spirochaetes, CFB group The F. nucleatum gene had to be closest to at least 2 species from that group, or 1 species with a clear association, and no evidence of transfer from Fusobacteria. Unknown If the tree contained less than four hits other than eukaryotes and other Fusobacteria, the gene was not considered for further analysis. In cases where it was not possible to clearly associate a taxonomic group to the F. nucleatum gene, it was then assigned as "unknown/not resolved". Gene-order method In order to identify clusters of at least two genes with conserved order between F. nucleatum and other genomes, all available amino-acid protein sequences sets for all replicons of all published bacterial and archaeal genomes at the time (may 1 st 2004) were downloaded from the NCBI ftp website . All Orthologs of F. nucleatum genes (reciprocal best blast hit) were detected between the two replicons ( F. nucleatum and replicon X). Clusters of consecutive orthologs were found (consecutive orthologs are determined in terms of the numbered position in both F. nucleatum (exactly consecutive) and replicon X (possible gap of 2 genes in between the orthologs), and for each cluster, a score was assigned as follows: (1-(totalCost/numberOfGenesInCluster))*(1-(deletions/numberOfGenesInCluster)) *Mean(Identity%)*Mean(Length%)*numberOfGenesInCluster/10000 Where "totalCost" was determined by the program derange2 [ 71 , 72 ] with the following command-line: "derange2 -U -L $inputFile 5 1 1 1", i.e. the direction of the gene was ignored, and the cost for an inversion, a transition or translocation within the cluster was the same: 1."Deletions" is the number of "gene gaps" found in replicon X for that cluster, whatever their size, "Mean(Identity%)" is the mean of the %identity of all blast results for the orthologs of the cluster. Mean(Length%) is the mean of the length of all blast results for the orthologs of the cluster, where length is defined as the minimum of length of (blast hit/ length of query sequence) and (length of blast hit/length of subject sequence). Each ortholog in the cluster was assigned the same score, and following completion of the procedure for all replicons in the database, for each F. nucleatum gene the orthologs that were part of clusters were ordered by their score, and an excel table was generated for manual investigation. If the gene order of a given gene cluster was not preserved in Firmicutes species but maintained in another procaryotic group, the genes were assigned as HGT from/to the group with the highest score (highest gene-order conservation). If the order was preserved in at least one species from two or more groups of low-GC gram-positives (Clostridiales, Bacillales, Lactobacillales and Mollicutes) the cluster was assumed to be ancestral to the divergence of fusobacteria and Firmicutes, and consistent with the ribosomal phylogeny. If gene order was preserved in one or more species from only one of the low-GC gram-positive groups the cluster was classified as HGT from low-GC gram-positive bacteria. GC-skew plots and gene classification Classical GC-skew plots were done using the formula (G-C)/(G+C) in 5000 bp windows, following Lobry's methods [ 25 , 26 ]. The functional classification of F. nucleatum genes by function was based on the TIGR Gene Attribute Annotation [ 73 ]. Authors' contributions AM carried out the gene-order analysis of gene transfer, and the study of chimeric operons and short ORFs. FRV conceived the study, and FRV and AM drafted the manuscript. RP and BAL did the bioinformatics work. RP and AM carried out the BLAST analysis of gene transfer and made the figures, except the phylogenetic trees. BAL developed the phylogenetic and gene-order method and did the ribosomal proteins tree. DM did all Bayesian trees. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Supplementary material published as additional information in the manuscript Mira et al. 2004. Evolutionary relationships of Fusobacterium nucleatum based on phylogenetic analysis and comparative genomics. The file contains three additional figures. It is available in pdf format and includes figure legends. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535925.xml

526178

Widespread A-to-I RNA Editing of Alu-Containing mRNAs in the Human Transcriptome

RNA editing by adenosine deamination generates RNA and protein diversity through the posttranscriptional modification of single nucleotides in RNA sequences. Few mammalian A-to-I edited genes have been identified despite evidence that many more should exist. Here we identify intramolecular pairs of Alu elements as a major target for editing in the human transcriptome. An experimental demonstration in 43 genes was extended by a broader computational analysis of more than 100,000 human mRNAs. We find that 1,445 human mRNAs (1.4%) are subject to RNA editing at more than 14,500 sites, and our data further suggest that the vast majority of pre-mRNAs (greater than 85%) are targeted in introns by the editing machinery. The editing levels of Alu-containing mRNAs correlate with distance and homology between inverted repeats and vary in different tissues. Alu-mediated RNA duplexes targeted by RNA editing are formed intramolecularly, whereas editing due to intermolecular base-pairing appears to be negligible. We present evidence that these editing events can lead to the posttranscriptional creation or elimination of splice signals affecting alternatively spliced Alu-derived exons. The analysis suggests that modification of repetitive elements is a predominant activity for RNA editing with significant implications for cellular gene expression.

Introduction On the molecular level, the complexity of higher organisms is based on the number of different gene products available for structural, enzymatic, and regulatory functions. Posttranscriptional and/or posttranslational mechanisms have an important role in generating RNA and protein diversity ( Baltimore 2001 ). One posttranscriptional processing pathway present in higher eukaryotes is RNA editing by adenosine deamination involving modification of individual adenosine bases to inosine in RNA by adenosine deaminase acting on RNA (ADARs; reviewed in Bass 2002 ; Schaub and Keller 2002 ; Maas et al. 2003 ). Since inosine acts as guanosine during translation, A-to-I conversion in coding sequences leads to amino acid changes and often entails changes in protein function ( Seeburg et al. 1998 ; Bass 2002 ; Schmauss and Howe 2002 ). The power of RNA editing in generating protein diversity lies in the fact that usually both the edited and unedited versions of the RNA and/or protein coexist in the same cell, and the ratio between the unedited and multiple edited variants can be regulated in a cell type-specific or time-dependent manner. Crucial functional properties of neurotransmitter receptors are regulated by A-to-I editing in the central nervous system ( Seeburg et al. 1998 ; Schmauss and Howe 2002 ), and inactivation of editing enzymes in mice ( Higuchi et al. 2000 ) and in the fruit fly ( Palladino et al. 2000 ) have resulted in profound neurological phenotypes. In addition to amino acid changes, A-to-I RNA editing can theoretically lead to the alteration of transcriptional start and stop codons, as well as that of RNA splice sites. In only one case though has the creation of a splice acceptor site through intronic RNA editing been described ( Rueter et al. 1999 ). Currently it is not known if the recoding of mRNAs at single codon positions is the main function of A-to-I RNA editing or if other types of editing events with as yet unknown roles in the regulation of gene expression are more widespread. The recently reported embryonic lethality in mice with ADAR1 deficiency indicates that additional substrates for this enzyme exist that function during early embryonic development ( Wang et al. 2000 , 2004 ; Hartner et al. 2004 ). Furthermore, a role for ADAR1 in the immune system is widely accepted, as one of its isoforms is interferon induced ( Patterson and Samuel 1995 ) and upregulated in immune cells during chronic inflammation ( Yang et al. 2003 ). The ablation of editing enzymes in Caenorhabditis elegans resulted in transgene silencing, suggesting that the RNA editing and RNA interference (RNAi) pathways intersect ( Knight and Bass 2002 ). This notion was recently confirmed by findings that the behavioral phenotype of ADAR-deficient worms could be rescued by inactivation of the RNAi pathway ( Tonkin and Bass 2003 ). Since both RNAi and RNA editing target double-stranded RNA (dsRNA) molecules, RNA editing could suppress gene silencing by preventing the formation of small interfering RNAs (siRNAs). A recurring theme of edited sequences is the involvement of an imperfectly dsRNA foldback structure ( Higuchi et al. 1993 ). The importance of base-paired RNA elements for site-selective editing to occur is also mirrored in the presence of dsRNA binding domains in ADAR enzymes ( Bass 2002 ). At present, though, it is not possible to predict if and to what extent a given RNA molecule is a substrate for A-to-I RNA editing in vivo. Despite recent progress in identifying additional genes that undergo RNA editing ( Morse and Bass 1997 ; Morse et al. 2002 ; Hoopengardner et al. 2003 ), the total number of currently known A-to-I edited genes in mammals is still small ( Bass 2002 ). However, the activity of the mammalian editing machinery, as measured by inosine content in mRNA fractions ( Paul and Bass 1998 ), is much higher than expected based on the current number of known substrates. Furthermore, ADARs are ubiquitously expressed in mammalian tissues, but almost all ADAR targets identified to date reside in the brain ( Bass 2002 ; Maas et al. 2003 ). This discrepancy between signs that A-to-I editing is omnipresent and the scarcity of identified targets has puzzled researchers in the field for some time, wondering where all the edited transcripts are. In this study we identify a minimum of 1,445 edited human mRNAs present in existing databases. Clusters of adenosine-to-guanosine (AtoG) discrepancies in these cDNAs are the result of RNA editing involving intramolecular pairs of inverted Alu repeat sequences, repetitive elements that represent approximately 10% of the human genome and are concentrated in and around genes ( Batzer and Deininger 2002 ). We also characterize functional consequences of the observed editing events and the factors that determine editing levels in Alu repeats and their modification patterns. The prevalence of Alu elements in primate genes, together with our experimental and computational analysis, suggests that the vast majority of primary human gene transcripts (greater than 85% of RNAs with average structure) are subject to A-to-I RNA editing. We show how editing might influence the alternative splicing of exonized Alu elements and discuss the implications of this extensive modification of mRNAs bearing repetitive elements for the regulation of gene expression. Results/Discussion Clusters of AtoG Discrepancies between Genomic and cDNA Sequences Are Due to A-to-I RNA Editing and They Are Located in Alu Repeat Elements A hallmark of an A-to-I RNA editing event is an AtoG transition when comparing genomic and cDNA sequences of the affected gene since inosine base-pairs with cytosine and therefore is replaced by guanosine during reverse transcription and PCR amplification. However, AtoG discrepancies between genomic and cDNA sequences can also be due to single-nucleotide polymorphisms (SNPs) or errors in databases. Therefore the search for edited sequences on a genome-wide basis is not feasible solely based on this single feature. However, in some cases of editing, not a single, but a cluster of AtoG discrepancies between genomic and cDNA sequences is evident within a stretch of a few hundred nucleotides ( Patton et al. 1997 ; Morse et al. 2002 ; Rosenthal and Bezanilla 2002 ). Therefore, we decided to inquire whether clusters of AtoG transitions seen in cDNA/genomic DNA (gDNA) sequence comparisons might represent bona fide editing events, since multiple base changes, all being of the AtoG type, are not likely accounted for by cosegregating SNPs or sequencing errors. In an initial screen for candidate genes, we used the Human Unidentified Gene-Encoded (HUGE) database of ca. 3,000 human cDNAs derived from the Kazusa cDNA sequencing project ( Kikuno et al. 2002 ). Several examples of cDNA sequences were found that within a window of 200–300 nt differ at several positions from the genomic sequence, such that the cDNA harbors a G where the genomic counterpart specifies an A. AtoG differences that coincide with an annotated A/G SNP were filtered out. Table 1 shows a list of all 26 genes from the HUGE database showing greater than two AtoG transitions in the exonic regions. Remarkably, we found that in all cases except one (KIAA0001) the location of the AtoG cluster coincides with the position of an Alu repeat element in the cDNA. As with Alu elements, most AtoG transition clusters are localized in 5′-UTR and 3′-UTR sequences and few in coding regions. Table 1 A/G Discrepancy Clusters in Human Brain cDNAs a Not all cDNA regions with A/G discrepancies were analyzed b SNPs according to National Center for Biotechnology Information SNP database c Clone hh15303 Alu elements are short interspersed elements found in all primates, which are approximately 300 nt in length ( Batzer and Deininger 2002 ). There are about 1.4 million copies of Alus from several closely related subfamilies present in the human genome, comprising approximately 10% of its mass ( Lander et al. 2001 ). The enrichment of Alu repeats in gene-rich regions of the genome ( Chen et al. 2002 ) makes their prevalence in transcribed sequences even more pronounced. Their high CpG dinucleotide content renders Alu sequences targets for methylation and implicates them in the regulation of gene expression ( Rubin et al. 1994 ). Clusters of A/G discrepancies that mapped to Alu repeats had been noted before in the HUGE database ( Kikuno et al. 2002 ). Furthermore, of ten newly identified editing targets in C. elegans ( Morse and Bass 1999 ; Morse et al. 2002 ) and 19 in human brain ( Morse et al. 2002 ), most were located in repeat elements. These findings suggested that repetitive elements, such as Alus, might be frequent targets for A-to-I RNA editing. In order to better understand the connection of Alu's with the observed AtoG clusters, we analyzed experimentally the cDNAs from all 25 candidate genes for RNA editing in human brain. Total RNA and gDNA were isolated from the same human brain specimen to eliminate false positives from unmapped A/G SNPs. For all 25 genes in vivo RNA editing was detected by single-run sequencing of gene-specific RT-PCR products, and for five of them the editing efficiency was quantitatively evaluated through repeated experiments. Extents of editing ranged from less than 2% to 90% at individual sites ( Table 1 ; Figures 1 – 3 ). Figure 1 RNA Editing of an Alternatively Spliced Alu-Exon in a G-Protein Coupled Receptor (A) Schematic representation of LUSTR (GPR107, KIAA1624) gene structure around edited exon 15a. The AluSx repeat element in intron 15 and the exonic, inversely oriented AluJo are predicted to form an intramolecular foldback structure as depicted below (MFold software). TM, exonic regions predicted to encode transmembrane domains; *, editing sites. (B) Editing analysis of exon 15a (sequence in capital letters) and flanking regions. The two major editing sites predicted to change amino acids (H/R and Q/R) are indicated. Editing levels in brain (filled column) and lung (open column) are shown above each edited nucleotide. The splice acceptor site subject to editing is underlined. Figure 3 RNA Editing of Alternative Exon 22a in Inhibitor BTKI and the 5′-UTR of KIAA1497 (A) The alternatively spliced exon 22a and surrounding region of the BTKI (KIAA1417) gene with two Alu elements and its computer-predicted foldback structure. (B) Editing analysis of the AluSx- element with the exonic sequence in capital letters and edited A's in bold. The alternative splice acceptor site is underlined with a dashed line; the additional alternative consensus splice acceptor site, which undergoes editing, is underlined with a solid line. (C) Gene architecture and Alu foldback structure of KIAA1497. The brain-derived cDNA of KIAA1497, also known as LRRN1; ( Taguchi et al. 1996 ), has a total of 15 nonpolymorphic AtoG discrepancies to gDNA, 14 being located within the 5′-UTR of the gene and one within the coding region. We analyzed PCR products covering all 14 potential editing sites in the 5′-UTR for editing in cDNA from human brain and could confirm in vivo editing to an extend clearly above the detection limit of our method for most of these positions and also at additional adenosines (data not shown). ORF, open reading frame; *, editing sites. Figure 2 RNA Editing of KIAA500 Alu Inverted Repeat KIAA0500 is a cDNA of 6,577 nt in length cloned from human brain (AB007969) with a predicted open reading frame of 213 amino acids. Four AtoG discrepancies were present within the coding region of which two lead to an amino acid change (Q/R and S/G, respectively). (A) Structure of the KIAA500 mRNA with location of Alu elements indicated and the predicted RNA secondary structure according to the MFOLD algorithm. Large open box indicates predicted open reading frame. *, editing sites. (B) Editing analysis of an exonic Alu element in KIAA0500. Editing sites predicted to change amino acids are indicated. Our analysis revealed a significant percentage of editing (%G) at the nucleotide positions 3518 (27% ± 3%), 3522 (20% ± 3%) and 3625 (6% ± 1%) and additional editing sites with less than 5% editing, whereas parallel analysis of human gDNA confirmed the presence of adenosine at these positions. Editing levels in brain (filled column) and lung (open column, where detectable) are shown above each edited nucleotide. Intramolecular Pairs of Oppositely Oriented Alus Are Responsible for Alu Element Editing Since a prerequisite for A-to-I RNA editing is the presence of a partially base-paired RNA foldback structure ( Higuchi et al. 1993 ; Bass 2002 ), the observed modifications in Alu repeats might be the result of two oppositely oriented, base-pairing repeat elements located within the same RNA molecule. For each of the 25 genes with edited, exonic Alu elements we find such oppositely oriented Alu repeats in the same pre-mRNA, many of which are located in intronic sequences. To determine if the predicted Alu pairs and the calculated foldback structures ( Figures 1 A, 3 , and 4 ) actually form in vivo, we analyzed experimentally the predicted Alu partners from the pre-mRNA for four of the identified editing targets (LUSTR, KIAA0500, Bruton's tyrosine kinase [BTKI], and KIAA1497). In each case we found that the closest, oppositely oriented Alu repeat undergoes A-to-I RNA editing as well. Figure 4 Alu-Mediated RNA Editing in p53, SIRT2, NFκB, and Paraplegin Pre-mRNAs Schematic presentation of the gene structures from (A) P53, (B) SIRT2, (C) NFκB, and (D) SPG7. Edited repeat elements are marked by asterisks. RNA folds appear as calculated with MFOLD. The AluJb− in p53 is located in the 3′-UTR (A); all others are intronic. *, editing sites. Because of the abundance of Alu elements in human pre-mRNAs, most primary transcripts contain one or more pairs of oppositely oriented Alus. If a majority of them is indeed subject to A-to-I RNA editing in vivo, it should be possible to predict RNA edited genes by identifying inverted pairs of Alu repeats in pre-mRNA transcripts. As a proof of principle, the analysis was extended to four arbitrary chosen genes (p53, SIRT2, NFκB, and paraplegin (SPG7) containing pairs of Alu repeats as seen schematically in Figure 4 B– 4 E. In all four cases, editing in the Alu elements that are predicted to form a dsRNA foldback structure is readily detectable. Many primary gene transcripts allow several energetically favorable foldback structures to be predicted for a given Alu that involve different combinations of Alu pairs. Do all these alternative Alu-pair foldback structures exist in vivo and are therefore subject to RNA editing? To address this question we examined the editing pattern of the G-protein-coupled receptor 81 (GPR81; identified through a computational search as described below). GPR81 contains four Alu elements, one sense and three antisense oriented, in the 3.6-kb pre-mRNA and was selected based on Alu repeat configuration and transcript length. If the alternative foldback structures depicted in Figure 5 coexist in vivo, all four Alu elements should show signs of editing with the level of editing indicating how prominent each of the alternative structures is. According to the analysis of GPR81 pre-mRNA, all three configurations form in vivo with variant II possibly being the dominant one since AluSp and AluJo show the highest levels of editing ( Figure 5 ). Figure 5 Editing of Alternative Foldback Structures of GPR81 Pre-mRNA (A) The position and orientation of all four Alu elements in GPR81 pre-mRNA is indicated. Three alternative Alu pairings (I–III) are predicted and experimental editing analysis indicates that all three do form in vivo. ORF, open reading frame; *, editing sites. (B) Editing analysis of AluSp+ in GPR81. Percentages of editing in human brain are indicated. The exonic sequence appears in capitals. The edited AT dinucleotide that becomes a splice donor site is underlined. These results suggest that Alu elements in human mRNAs are subject to RNA editing by ADARs because of foldback structures formed between two oppositely oriented Alus present within the same primary transcript. Editing of Alus Is Tissue Dependent and It Alters Codons and Pre-mRNA Splice Sites of Alternatively Spliced Alu Exons Exonic Alu repeat elements are predominantly located in the 5′- and 3′-UTRs of mRNAs, and as a result, most cases of Alu editing occur in noncoding regions. However, some editing events predict amino acid changes ( Table 1 ). Among the identified genes for which we performed a detailed, quantitative editing analysis several unique and recurring features emerge regarding the locations and functional implications of the editing events. The LUSTR1 cDNA codes for a G-protein-coupled seven-transmembrane receptor (also termed GPR107 or KIAA1624), with three AtoG discrepancies located within an alternatively spliced AluJo-derived exon that leads to the in-frame insertion of 29 amino acids between transmembrane regions V and VI of the protein (see Figure 1 A). The experimental analysis revealed a total of ten editing sites within this Alu element (see Figure 1 B), including two major sites that lead to amino acid changes (H/R and Q/R sites). Interestingly, editing levels at all positions were significantly different in human brain (19%–58%) compared to lung (less than 5%), suggesting a tissue-specific regulation of editing (see Figure 1 B). Analyzing the RNA editing pattern of LUSTR1 pre-mRNA revealed additional intronic editing sites, one of which represents the splice acceptor adenosine (AG to IG) in intron 15 (22% edited in brain; see Figure 1 B). Editing at this position is predicted to lead to the exclusion of the Alu exon, indicating that the alternative splicing of exon 15a might be coregulated by RNA editing of its splice junction. This is to our knowledge the first documented example where A-to-I RNA editing acts to destroy a pre-mRNA splice signal. A picture similar to LUSTR1 emerges from analysis of the gene for human inhibitor of BTKI (also termed KIAA1417; Liu et al. 2001 ; Strausberg et al. 2002 ). Again, an alternatively spliced Alu exon (located between constitutive exons 22 and 23) is affected. This time two AluSx elements are positioned in opposite orientation at the start and end of the exon (see Figure 3 ). Inclusion of the exon using the splice acceptor site provided by AluSx− leads to the premature termination of translation within this exon with all editing sites located in the 3′-UTR. Editing levels at 20 sites throughout the Alu element range from less than 5% to 31% in human brain, whereas cDNAs isolated from human lung again displayed few editing sites with low editing levels of less than 5%. A splice site is also subject to editing in BTKI, this time affecting an additional alternative splice acceptor site within AluSx−. On the pre-mRNA level this position is edited to 15%. However, in transcripts that use the weak upstream splice acceptor site (underlined with a dashed line in Figure 3 B; as in the HUGE database clone hh15303), the additional alternative splice site (underlined with a solid line in Figure 3 B) is highly edited, raising the possibility that edited BTKI pre-mRNA preferentially follows the alternative splicing pathway (data not shown). The analysis of GPR81 revealed another case of Alu exon alternative splicing and, surprisingly, a new mechanism showing how RNA editing might affect RNA processing. Within the AluSp+ element located in the 3′-UTR of GPR81 transcripts a splice donor site (AT to IT) is generated in 57% of primary transcripts by RNA editing. This is predicted to give rise to alternatively spliced mRNA products represented by GenBank entry AF385431 (see Figure 5 B). This is, to our knowledge the first reported case of potential splice donor site creation by RNA editing. It is possible that here the Alu element was inserted into the 3′-UTR exon of the GPR81 gene and has evolved into a state where it is a single mutation away from initiating the birth of a novel intron. Posttranscriptionally RNA editing provides the final base change to create the new splice site. This scenario is supported by the fact that in mice the GPR81 gene lacks introns. It is intriguing that we find cases where editing in alternatively spliced Alu exons, or within adjacent splice sites, interferes with or counteracts exon formation of an Alu repeat. It suggests that RNA editing might be more generally involved in the regulation of Alu exonization. Recently, it has been shown that more than 5% of the alternatively spliced exons in the human genome are Alu derived ( Sorek et al. 2002 ). Exonization of Alu repeats occurs via activating mutations in mostly antisense-oriented, intronic Alus generating a novel splice acceptor site ( Mitchell et al. 1991 ; Lev-Maor et al. 2003 ), and it has been speculated that exonization of transposable elements in general is a major mechanism for the generation of novel exons ( Kreahling and Graveley 2004 ). A large number of intronic Alu elements are just a single mutation away from being exonized ( Lev-Maor et al. 2003 ; Kreahling and Graveley 2004 ), and in some cases the constitutive splicing of an intronic Alu has been shown to cause a genetic disorder ( Mitchell et al. 1991 ; Knebelmann et al. 1995 ; Vervoort et al. 1998 ). In this context RNA editing may partially counteract genomic mutations that lead to the incorporation of deleterious novel exons while maintaining their potential to form exons with beneficial functions through further mutation. Furthermore, RNA editing in Alus might be involved also in the generation of novel introns as seems to be the case in GPR81. Statistically, however, the exonization of intronic Alus would be much more frequent than the intronization of exonic Alus because of the abundance of Alus in introns. A Transcriptome Wide Screen for Edited Alu Repeats The results presented above show that clusters of AtoG mismatches in cDNA/gDNA sequence comparisons represent an effective way to identify authentic editing cases with a low rate of false positives. Since all clusters of AtoG discrepancies mapped to repeat elements, we wondered how prevalent the editing of Alu or other repeat elements is in the human transcriptome. Therefore, we devised a database search procedure to identify pairs of inverted repetitive elements in human mRNAs exhibiting AtoG transitions. Initially, a limited search was carried out for closely spaced (less than 2 kb) inverted pairs of human Alu, MIR, and L1 repeat elements that overlap with exonic sequences and for which an mRNA sequence can be found in GenBank entries. This search, involving about one-third of all repeat elements in the human genome, identified 71 mRNAs with exonic repetitive-element pairs (51 Alu, six L1, six MER, and eight MIR). From those mRNAs, 27 displayed clusters of AtoG changes, all in Alu elements. Fourteen of these genes were chosen for experimental analysis, and all 14 proved to be subject to A-to-I RNA editing ( Table 2 ). Since these initial results indicated a high prevalence of editing in Alu elements, we decided to carry out a comprehensive search involving all elements present in cDNA sequences. Table 2 Edited Alu Exons from Computational Screen for Alu Element Foldback Structures a SNPs according to National Center for Biotechnology Information SNP database b Not all cDNA regions with A/G discrepancies were analyzed ND, not determined We analyzed the total of 103,723 human mRNA sequences (from the University of California, Santa Cruz [UCSC] Genome database [ Kent et al. 2002 ], July 2003 assembly) for overlaps with repetitive elements of the L1, Alu, MaLR, and MIR families. For Alus, 17,406 mRNAs (16.8%) contained a total of 31,666 complete or partial repetitive-element sequences. Comparing the cDNA sequences with their genomic counterpart revealed that the number of AtoG discrepancies within Alu repeats is more than seven times higher than the average number of the other transitions (23,204 versus 3,271 [the average for GtoA, CtoT, and TtoC transitions]). In fact, the number of AtoG mismatches is higher than all other eleven types of nucleotide discrepancies combined ( Figure 6 A). Figure 6 Mismatch Bias in Exonic Repetitive-Element Sequences (A) Plot of the nature and number of mismatches within Alu and L1 sequences present in human cDNAs. For reasons of comparison the L1 mismatch numbers have been multiplied by 2.9 so that the non-AtoG mismatch count for Alu and L1 is identical. Transition mismatches AtoG, GtoA, CtoT, and TtoC are displayed together for comparison. (B) Plotted are the total number of Alu sequences found in human cDNAs (first column) and the number of elements harboring AtoG and GtoA mismatches (second and last column). The third column indicates the high confidence set of edited elements (α = 0.000001). While the finding that non-AtoG transitions (GtoA, CtoT, and TtoC) are approximately three times more frequent than transversions is in line with results from previous studies analyzing gDNA sequences ( Lander et al. 2001 ; Venter et al. 2001 ), there is no explanation for the observed excess of AtoG mismatches relative to other base transitions. Alu sequences carry 22–23 CpG dinucleotides, which are known to show high mutation rates because of C-methylation, and as a consequence, these positions should display an elevated frequency of SNPs. Nevertheless, an elevated number of SNPs would lead to a rise in AtoG as well as GtoA mismatches when comparing a representative population of cDNAs with the corresponding genomic sequence. Thus, we concluded that the excess exclusively in the number of AtoG discrepancies in Alus over other base changes may reflect cases of bona fide A-to-I editing at the RNA level. We then devised a statistical approach to distinguish repetitive elements that show AtoG mismatches due to sequencing errors and SNPs from those that have undergone A-to-I RNA editing. The method was based on the observation above that Alu elements subject to RNA editing undergo multiple base modifications that result in a cluster of AtoG discrepancies (5–30) between cDNA and gDNA. The probability that a cluster of several AtoG discrepancies is due to sequencing errors or SNPs (in the absence of an increased number of other nucleotide discrepancies indicating low-quality sequence data) is negligible. Thus the number of clustered AtoG changes can be used to distinguish genuinely edited elements from elements with aberrant or non-editing-related base changes. For each Alu element with AtoG discrepancies, we computed the χ 2 test comparing the observed number of AtoG discrepancies with the expected number, based on the number of non-AtoG mismatches present in the same sequence. Elements with a χ 2 higher than the critical value for α = 0.00001 (corresponding to a probability of one in 100,000 that the observed AtoG transitions are due to SNPs or sequencing errors) were selected as “edited” and will be called so throughout the rest of the manuscript. Using this approach we found that out of those 17,406 mRNAs with one or more exonic Alu elements, 1,445 (8.0%) mRNAs are edited within one or more of the Alu sequences (for a full list of edited mRNAs see Table S1 ). When looking at all the 31,666 Alu elements present within these 17,406 RNAs, we find that 1,925 (6.1%) Alu elements are “edited,” while another 4,574 Alu elements (14.4%) show AtoG discrepancies but fail to pass our probability cutoff ( Figure 6 B). Thus, the total of 6,499 elements (or 20.5%) represents the upper limit of potentially edited Alus in our sample ( Figure 6 B). The total number of Alus with GtoA discrepancies in the same sequence sample is 2,002, and we considered this value to reflect base changes that are due to SNPs and sequencing errors. Assuming a similar number for random AtoG and GtoA mismatches, we can subtract this number from the total count of potentially edited Alus, obtaining 4,497 cases (14.2%) as the approximate number of actually edited elements. In order to validate our screening approach, we performed an identical analysis for GtoA, CtoT, and TtoC mismatches. Compared to the 1,925 AtoG-edited Alu elements in mRNA, we found 12 GtoA, 11 CtoT, and 11 TtoC cases of “editing.” These cases may represent false positives and thus set the error level of our screen to less than 0.6%. These results suggest that out of the 103,723 human mRNAs at least 1.4% are A-to-I edited within an exonic Alu element. Apart from Alu repeats, many more low- and high-frequency repeats exist in the human genome ( Venter et al. 2001 ) and might give rise to RNA foldback structures that result in exonic A-to-I editing. Therefore, the total number of mRNAs edited in exonic repeat sequences is probably higher than the value obtained from our analysis of Alu elements. Most Alu repeats are located in introns, and it is there where the bulk of RNA editing is expected to occur. The average number of Alu repeats per gene is 12.4 estimated for Chromosomes 21 and 22 ( Grover et al. 2003 ). This value is comparable to the 19.3 Alus/gene estimated from our data (2,003,976/103,723: total number of Alus (nonunique) in mRNA boundaries/number of mRNAs) for the whole genome. Considering that based on our analysis 14.2% of exonic Alus are edited, and assuming similar editing rates for intronic Alus, we can estimate that the probability of an average pre-mRNA to be edited is approximately 1–0.858 19.3 or 94.7% (85.0% with the 12.4 Alu/gene estimate). While this value is an approximation, assuming that all genes have similar structures, and does not take into account editing in other repeat elements, it does reflect the magnitude of repetitive-element editing. Distance, Conservation, and Tissue Localization Influence Which Pairs of Alu Elements Are Edited To gain insight into the factors that determine which Alus are subject to RNA editing, and under what circumstances, the identified set of 1,925 high-confidence cases of editing in Alu elements (contained in the 1,445 mRNAs listed in Table S1 ) was used for further computational analysis. It was assumed that the observed editing is the result of RNA foldback structures formed between intramolecular inverted Alu repeats, as we have demonstrated for all the experimentally analyzed cases. If this hypothesis is correct, then the distance between an Alu and its closest inverted pairing element should be a critical determinant for how likely it is that a given element will be targeted by the RNA editing machinery. To test this hypothesis the closest inverted Alu was identified within the same gene for all 31,666 Alu elements. A properly oriented element was found for 19,231 of those Alu elements, and a plot was made showing the percentage of edited Alus as a function of the distance between elements ( Figure 7 A). The highest level of editing (approximately 16%) was found for Alu pairs 300–400 nt apart, which corresponds to slightly more than the size of a full-length Alu repeat. Editing levels subside with increasing distance as the probability for base-pairing between the two Alu elements apparently decreases. Alu pairs with distances below 300 nt indicate partial Alu elements, and the observed decrease in editing levels is likely because of the smaller, less energetically stable foldback structures. These results suggest that the optimal configuration of an Alu-pair stem loop involves two full-length Alus forming the stem separated by a short (10–50 bp) intervening loop sequence. Interestingly, as the distance increases we ultimately arrive at a low-level plateau of approximately 1% editing without any further drop in editing levels. RNA editing in trans caused by base-pairing Alus located in separate RNA molecules might be responsible for this “background” editing. A-to-I editing in trans does occur on pre-annealed RNA duplexes in vitro ( Bass and Weintraub 1987 ; Nishikura et al. 1991 ) and could occur also in vivo if such intermolecular RNA duplexes form. In Xenopus one case of potential trans editing has been described, involving RNA duplexes formed between sense and antisense transcripts of bFGF ( Saccomanno and Bass 1999 ). Figure 7 Factors Determining Alu Targeting Probability (A) Percentage of edited elements classified in bins according to the distance separating the element and its closest inverted Alu partner. (B) Percentage of edited Alu elements clustered according to divergence from their corresponding Alu-subfamily consensus. (C) Percentage of edited Alu elements in each Alu subfamily. In (A), (B), and (C) the numbers at the bottom of the bars show the sample size in each bin. (D) Percentage of edited elements according to the tissue from which the RNA was isolated. Error bars show 95% confidence levels. The distance dependence of the extent of editing clearly suggests that the formation of Alu–Alu stem loop structures predominantly results from intramolecular Alu inverted repeats with an upper limit of approximately 1% editing that could be due to intermolecular Alu pairings. To our knowledge, these results describe for the first time the distance relationship of long-range RNA folding interactions in vivo and how their stability is influenced by distance. More important, considering the high frequency of Alu elements in primate RNA sequences and the low levels of potential intermolecular editing observed, we conclude that intermolecular duplexes between complementary RNA sequences do not form in the nucleus at a significant rate. This raises the question of how the regulation of thousands of human messages proposed by Yelin and colleagues involving antisense transcripts works ( Yelin et al. 2003 ). It might also explain why cases of editing involving endogenous sense/antisense RNA duplexes have not been reported despite evidence for extensive antisense transcription. The editing of RNAs by ADARs has been shown to be dependent on the double-stranded character of the substrates, such that editing levels and promiscuity increase with the extent of the base-paired region ( Bass 2002 ). The human Alu family is composed of several subfamilies of different genetic ages, and their consensus sequences contain diagnostic changes distinguishing one subfamily from another. The extent of base-pairing between two oppositely oriented Alu elements, and in turn the extent of A-to-I editing, depends on their sequence homology, and it is expected that highly diverged elements would form less stable foldback structures. The relationship between observed editing level in our set of 31,666 Alu repeats and the sequence divergence of each Alu repeat from the consensus of its respective subfamily is shown in Figure 7 B. A decrease in editing levels is seen with an increase in diversity, suggesting that an Alu element with lower sequence homology to most other Alu repeats has a lower probability of forming a suitable editing substrate. Unexpectedly, we also observed a drop in editing levels for Alu elements with low divergence from their subfamily consensus sequence. This trend may be caused by the distribution of Alu divergence. Within the human genome the majority of Alu elements have diverged by 10%–15% from their subfamily consensus ( Stenger et al. 2001 ). Therefore, Alu elements with lower than average divergence have a lower likelihood of encountering another element of similar divergence, resulting in low editing levels for this subset. We obtained similar results when we compared the editing levels of Alu elements with the sum of the divergence of the edited Alu and its closest inverted Alu element (data not shown). In agreement with these conclusions, we find that the most populated Alu subfamily (AluSx) and the subfamilies closely related to AluSx sequences (AluSq, Sc, and Sg) show the highest levels of editing ( Figure 7 C). The pool of mRNAs used in this study represents a heterogeneous collection of sequences from different tissues and cell types. In analyzing the editing of Alu elements as a function of tissue origin ( Figure 7 D), significant differences in editing levels were found. The highest editing activities were seen in brain tissues, in trachea and thymus. These results are in accordance with prior studies that have measured the overall activity of RNA editing enzymes in selected mammalian tissues as judged by the amount of inosine detectable in the poly(A)+ fraction of RNA ( Paul and Bass 1998 ). The two human enzymes with A-to-I RNA editing activity (ADAR1 and ADAR2) display a different but overlapping activity profile on known substrates, and their expression is highest in brain (ADAR2) and in cells of the immune system (ADAR1; Bass 2002 ). Furthermore, ADAR1 was found to be induced during inflammation leading to high activity in blood cells and thymus ( Yang et al. 2003 ). These findings are also in agreement with our experimental results, which show much higher editing levels in brain-derived RNAs than in the same mRNA isolated from lung tissue (see Figures 1 – 3 ). The pool of edited Alu elements was analyzed for other features that might influence editing levels, such as the position of the edited Alu within the mRNA (3′-UTR, 5′-UTR, and coding region) or its orientation in relation to the mRNA (sense, antisense). No significant correlation of Alu editing was detected with any of these features (data not shown). Editing of Alu Repeats Shows Sequence and Structure Preferences The availability of such a large collection of A-to-I edited sequences resulting from this analysis allowed us to examine the modification pattern of edited Alu elements for potential editing hot spots or base preferences. To this end we first aligned all 141 edited Alu sequences (greater than 260 bp) in RNAs originating from Chromosome 1 and mapped the edited sites on their consensus sequence ( Figure 8 ). Interestingly, certain adenosines are targeted in greater than 30% of the edited RNAs while other adenosines do not show any evidence of editing. The four editing hot spots all map to TA dinucleotides that are located in conserved Alu regions (greater than 80% identity), suggesting that they are base-paired in the average foldback structure. This confirms the previously proposed T/A 5′-neighbor preference for both ADARs ( Polson and Bass 1994 ). Surprisingly, most of the 22 CpGs of the consensus sequence coincide with the location of high-frequency editing events. CpGs are known to be targeted by cytosine DNA methylation, which results in a high mutation rate, turning CpGs into either CA or TG dinucleotides ( Batzer and Deininger 2002 ). Since less than 50% of the transcripts carry a CA or TG (edited in reverse complement) at these CpG consensus sites, the editing efficiency (edited adenosines/total adenosines) at these positions is comparable to that at the hot spots ( Figure 8 A, arrows). Figure 8 Sequence and Structure Preferences of Editing in Alus (A) The consensus sequence of 141 edited full-length Alu elements present within human Chromosome 1 transcripts with the number of editing events indicated for each sequence position (bars). Insertions and deletions present in fewer than five elements are not shown in the alignment for clarity. Bases conserved in more than 80% of the sequences are boxed. For the lesser conserved consensus positions the next most frequent base is listed below. Consensus CpG dinucleotides are in bold. Arrows indicate “high-efficiency” positions where more than 20% of adenosines present appear to be edited. Note the overlap of these positions with CpGs. Major features of Alu sequences, such as the A-Box and B-Box of Pol III and the Alu polyA sequence are labeled. (B) A typical Alu foldback structure and its major features as discussed in the text. Arrows indicate TA hot-spot positions. The magnifications show the two typical configurations of editing sites found in Alu pairs: mismatched A/C bulges (i) and A/U base pairs (ii). As a result of the high CpG mutation rate, frequently the Alu foldback structure of the unedited RNA is predicted to carry A–C mismatches at these positions. Editing at these sites restores the CpG repeat (CA→CI) on the RNA level and converts the A–C mismatch to an I–C base pair. Energy calculations for several predicted Alu pairs show, surprisingly, that the stability of the foldback structure is not diminished by editing but often increased because of the high frequency of I/C pair formation (data not shown). It is therefore unlikely that in the case of Alu foldback structures, RNA editing serves to resolve RNA secondary structures that interfere with the processes of splicing or translation of these RNAs, as suggested previously ( Morse et al. 2002 ). Two typical configurations of editing sites observed in Alu elements are depicted in the magnifications of Figure 8 B where either A–U pairs are turned into I–U wobble pairs in conserved regions of the sequence (ii), or A–C mismatches are converted into I–C pairs within nonconserved Alu regions (i). While the above analysis shows the qualitative features of the editing sites in Alus, determination of cis preferences was carried out by extracting 14,774 pentanucleotide sequences with the edited adenosine as the middle base and estimating the frequency of each base at positions −2, −1, 1, and 2 relative to the editing site. To correct for Alu sequence bias we performed the same analysis for a randomly chosen adenosine for each edited adenosine in our sample. We then subtracted those frequencies to obtain unbiased editing preferences ( Figure 9 ). The presence of large, unpaired poly(A)+ tails in Alus obscures our analysis for adenosines surrounding edited A's but is informative regarding other base preferences. Position −1 shows a strong preference for C and T and aversion for G in agreement with previous studies ( Bass 2002 ). Interestingly, we observe preferences for G in position +1 and for C or G at positions −2 and +2, which have not been described before. This preference pattern appears not to be linked to any Alu-specific structural feature and therefore possibly reflects the editing enzyme cis preferences. We also identify a preference for an editing site to be preceded or followed by another editing site ( Figure 9 ). This data-rich assessment of sequence preferences for edited sites might be useful in an ab initio identification of new editing sites. Taken together our results identify loose RNA duplexes carrying A–C mismatches or A/U-rich regions, as favored editing targets. The high incidence of “corrective” editing at mutated CpG consensus positions in Alus raises the possibility that posttranscriptional restoration of CpG repeats in Alu primary transcripts by RNA editing contributed to the surprising retention of CpGs in Alus during evolution ( Batzer and Deininger 2002 ). This might constitute an important consequence of A-to-I editing in view of the role of CpG islands in the regulation of gene expression. Figure 9 Cis Preferences of Editing Sites in Alus Tables (i) and (ii) show the frequency of A, G, C, T, or an A/G editing site at positions −2, −1, 1, and 2 relative to each of the 14,774 AtoG mismatch sites found within the high confidence group of Alu elements (i) and in relation to a randomly chosen adenosine from each of the those sequences for each AtoG mismatch (ii). Table (iii) shows relative editing preferences after bias removal by subtracting table (ii) from Table (i). (iv) Graphical representation of Table (iii). Potential Functional Implications of RNA Editing in Repetitive Elements Considering the available data on in vitro editing activities of ADARs on dsRNA molecules of different sequences and structures, it is not surprising that highly base-paired RNA foldback structures such as the ones induced by Alu inverted repeats are substrates for the editing enzymes. However, it is remarkable and maybe surprising that these predicted structures are edited in vivo at significant levels. This indicates that many of these structures do form in vivo and are readily accessible for ADARs in the nucleus. Alu elements are ideal for the formation of editable RNA structures because of their large numbers, size, and degree of conservation. We find no evidence for a sequence or otherwise specific interaction of the editing machinery with Alu sequences. Thus, other repetitive elements able to form similar structures should also be targets of A-to-I editing. Our data suggest, however, that editing levels in all other major repeat-element families that dominate the human genome (LINE, LTR, and other short interspersed elements) are very low compared to editing levels seen in Alu repeats (see Figure 6 A and unpublished data). The selectivity for Alus might be explained based on the distribution features of each repetitive-element family: For example full-length L1 repeats are approximately 6 kb in length, and as a consequence, most of the time they have low chance of having a base-pairing sequence in proximity. MIR repeats, although found in significant numbers, which potentially could form foldback structures, have a low average level of conservation (30%–40% divergence) and so may be inadequately double stranded to be a substrate. MaLR elements of the LTR superfamily are present in numbers such that the average distance between an inverted pair is very high (approximately 10 kb). However, our analysis suggests that all repetitive elements might become targets of RNA editing at different stages in evolution. Young repetitive elements in their expansionary phase of evolution display features that we identify as important for being editing targets. Based on these observations it will not be surprising if repeat elements that show low levels of editing in humans are major targets in other organisms. For mRNA fractions, we estimated the inosine content due to Alu editing as follows: In 103,724 mRNAs we found 23,204 AtoG mismatches, while the same sequence sample has an average for the other transitions of 3,271. Assuming an average mRNA size of 4 kb, the ratio of inosine in the sample is estimated to be one inosine every 20,814 nucleotides (103,724 × 4,000/[23,204–3271]) generated by editing in Alu sequences. This estimation for Alu editing is in the range of one inosine in 17,000 nt (brain), one in 33,000 nt (lung, heart), to one inosine in 150,000 nt (skeletal muscle) as was experimentally determined by Bass and colleagues in the polyA-fraction of rat RNAs ( Paul and Bass 1998 ). Since the rat genome lacks Alus, the total amount of inosine generated in human mRNAs may be much higher than in rats, unless a class of edited sequences in rats exists with a similar prevalence to Alus in humans. In any case, our data imply that most of the inosine detected in mRNA transcripts can be explained by the widespread A-to-I editing of repetitive elements. Repeat-element editing might therefore point toward an important housekeeping function for RNA editing. In contrast, the well-studied examples of editing that lead to single nucleotide and codon changes in mRNA might be less frequent cases of editing events. While a significant amount of editing occurs in mRNAs that contain repetitive elements in their exons, our results predict that the bulk of A-to-I editing takes place in intronic sequences missing from cDNA databases. This is suggested by the experimental results regarding the LUSTR, GPR81, p53, SIRT2, NFκB, and paraplegin genes, for which intronic data was available (see Figures 1 A, 4 , and 5 A). This extensive editing of repetitive elements in pre-mRNAs creates an enormous pool for the generation of gain-of-function mutations. The involvement of editing in creating or destroying splicing sites of alternatively spliced Alu exons, along with internal editing of those exons, suggests an intriguing new mechanism for accelerated evolution. We are now in a position to analyze the extent to which this process occurs within the human transcriptome. Such a role in “stimulating” evolution, however, is unlikely to be related to the “daily” function of A-to-I RNA editing. It has been shown that hyperedited, inosine-containing RNAs are retained in the nucleus by a protein complex containing the inosine binding protein p54 ( Zhang and Carmichael 2001 ). In view of the widespread editing of Alus this offers an intriguing mechanism to preclude aberrantly spliced mRNAs and, more generally, repetitive-element-containing RNAs from exiting the nucleus. This model, though, suggests that intronic RNA editing occurs frequently in other organisms and in other repetitive-element types as well, something that remains to be shown. A connection between A-to-I RNA editing and RNAi has recently been suggested through studies in C. elegans where inactivation of the editing machinery leads to transgene silencing ( Knight and Bass 2002 ), and subsequent inactivation of the RNAi pathway restored transgene expression ( Tonkin and Bass 2003 ). Furthermore, retrotransposon LTR sequences were shown to induce natural RNAi due to RNA duplex formation ( Sijen and Plasterk 2003 ). The RNAi machinery has been implicated in gene silencing in two independent modalities: at the RNA level through degradation of mRNAs and at the chromatin structure level through induction of methylation ( Dykxhoorn et al. 2003 ; Ekwall 2004 ). Both silencing pathways might be affected by editing of repetitive-element foldback structures. Silencing of RNAs containing such inverted repeats might be prevented through their modification by RNA editing and their subsequent nuclear retention ( Zhang and Carmichael 2001 ) or by rendering those RNAs inadequate substrates of the RNAi machinery. It is possible that the observed embryonic lethality and apoptosis in A-to-I editing-deficient mice ( Wang et al. 2000 , 2004 ; Hartner et al. 2004 ) is related to the breakdown of this control mechanism leading to the posttranscriptional silencing of essential genes. The work presented here has been based on the analysis of cellular mRNAs that contain Alu repeat elements. However, the underlying principles probably also apply to Alu RNAs generated from transcriptionally active Alu elements. Alu elements do not encode transcription termination signals ( Deininger 1989 ), and thus read-through transcription from transposition-competent Alu repeats can result in intramolecular Alu pairs, leading to the editing of a sequence that subsequently becomes retrotranscribed. Editing of primary transcripts of repetitive elements may have an important role in the control of their proliferation and a dedicated analysis of such transcripts for editing events represents an important future direction. A recent study by Levanon et al. (2004) reported a computational approach for the identification of heavily edited genes in the human transcriptome and found that editing mostly occurs in Alu repeat elements (greater than 92% of the substrates identified), giving us the opportunity to compare the two approaches and datasets. The computational strategy used by Levanon et al. (2004) differs substantially from ours both in the sequence dataset employed and in the methodology applied. The use of expressed sequence tags (ESTs; in contrast to our use of mRNA sequences) offers a much larger primary dataset for analysis; however, single-pass sequences have a higher error rate ( Liang et al. 2000 ), and EST databases are biased toward sequences near 3′-termini of mRNAs ( Liang et al. 2000 ). Levanon et al. (2004) selected candidate sequences for editing by identifying inverted repeats followed by the evaluation of AtoG mismatch rates, whereas we directly evaluated the AtoG mismatch content in repetitive elements irrespective of the presence of a nearby pairing sequence. The approach of Levanon et al. (2004) allows the discovery of edited inverted repeats that do not belong to any of the repetitive-element families (although the previously known brain substrates were missed), but it does not identify cases where a base-pairing sequence is not evident because of truncated cDNA and EST sequences and incomplete knowledge of gene boundaries. We found that for approximately one-third of the edited Alu elements a pairing Alu cannot be located within the gene boundaries as determined by known mRNAs, although in most of the cases it can be identified at the genome level. A comparison of the edited gene/mRNA datasets of the two studies shows a 34.5% overlap when gene names and symbols are compared. It should be noted, though, that editing of the same gene might reflect editing at different sites or within different Alu elements of the same gene. The two approaches are overlapping as well as complementary. Taken together, they have probably uncovered the most significant part of the heavily edited exonic sequences for which sequence data are available. From our analysis we estimate an additional approximately 4,000 edited Alu elements besides the 1,925 Alus that we have selected as a very high confidence set. Thus, it is important to note that the heavily edited sequences represent the tip of an iceberg with many more mRNAs in the human transcriptome being edited at single or a small number of positions. Materials and Methods RNA editing analysis Human brain samples were provided by the Harvard Brain Tissue Resource Center, Belmont, Massachusetts, United States; human lung cDNA was from Clontech (Palo Alto, California, United States). Total RNA isolation and reverse transcription have been described previously ( Ausubel et al. 1995 ; Maas et al. 2001 ). Gene-specific PCR was performed as described earlier ( Maas et al. 2001 ), and a list of oligonucleotide primer sequences used in this study is available on request. RNA editing analysis was done by direct sequencing of gene-specific, gel-purified RT-PCR products as described ( Maas et al. 2001 ), using an automated ABI310 (Applied Biosystems, Foster City, California, United States) capillary electrophoresis sequencer. Human gDNA used for gene-specific PCR was isolated from the same tissues according to standard protocols ( Ausubel et al. 1995 ). Computational procedures For analysis of the pool of human cDNA sequences we developed a program named Procedures for Repetitive Element Foldback Analysis (PREFA). We used the set of cDNA sequences from the UCSC database (July 2003) comprising 103,723 sequences (after removal of duplicate entries). The set of repetitive elements (for Alus 1,163,041 unique elements) and related information of the human genome (created with RepeatMasker based on the Repbase [ Jurka 2000 ] release of June 2002) was obtained from the same source. For each examined repetitive-element family we first selected the subset overlapping partially or fully with genes. For Alus the number is 2,003,976, including duplicates, or 572,107 unique sequences. From this subset we then selected those overlapping with exons (31,666). The RNA and genomic sequence for each element was extracted and compared base by base for mismatches. A small number of cases with very high non-AtoG mismatches (greater than 20/element) were discarded as misaligned or erroneous. From the repetitive elements showing at least a single AtoG change we selected those where mismatch distribution cannot be accounted for by SNPs and sequence errors using the following procedure: The overall expected ratio of AtoG discrepancies relative to the total number of mismatches was calculated from the whole sample, assuming the expected AtoG mismatches to be approximately equal to the average of the rest of the transitions: The expected probability for an AtoG mismatch at a single position in a given element was calculated from the total number of mismatches found in the element in cases where other mismatches were present (2) or from the whole sample where only AtoG mismatches were found (3): Here nAtoG and nOther is the total number of AtoG and non-AtoG mismatches found for this element: Given the probability p for an AtoG mismatch to occur at any given position, the expected values for the number of AtoG were calculated: A χ 2 test was calculated for each element and those with a χ 2 value exceeding the critical value (for α = 0.000001) were selected as edited, and these values correspond to approximately more than five AtoG changes in the absence of any other change in the approximately 300 bp of an Alu). For each element in the high-confidence set the closest inverted element was identified among the elements present in the same gene boundaries. The distance separating the pair was calculated from the location of the first base of each element, according to the genomic sequence numbering and irrespective of their orientation. The divergence of each element was derived from the corresponding entry in the UCSC annotation database (ChrN_rmsk) representing mismatches per hundred bases. Tissue of origin of the RNAs was also derived from the UCSC mRNA annotation. For RNAs described to originate from multiple tissues, the corresponding RNAs were included in the count for each of those tissues. RNAs originating from a specific subregion of a tissue, such as subareas of the brain, were counted within the subregion but not in the whole-tissue set of RNAs. Alignment of the Chromosome 1-derived Alu sequences was performed with the MegAlign program of the DNASTAR (Madison, Wisconsin, United States) package (Lasergene) using the CLUSTAL algorithm ( Jeanmougin et al. 1998 ). Further manual adjustments were necessary owing to the presence of simple repeats in Alu sequences. Analysis of the alignment and base counts surrounding the editing sites were done with PREFA. Supporting Information Table S1 Database of Computationally Identified Editing Targets The database lists the GenBank accession numbers, gene names, gene product description, chromosome location, and type of Alu element and location within the mRNA sequence, the identity of the most likely pairing Alu elements within the same gene, and the distance in base pairs (bp) between the pairing Alus. The positions of all predicted editing sites within the individual sequences can be viewed by pasting the accession number into the USCS genome browser ( Kent et al. 2002 ) at http://genome.ucsc.edu/cgi-bin/hgGateway and following the link to mRNA/Genomic alignment. We found that six cDNAs map on two chromosomes (AB095924, AK021666, AK055562, AK092837, AK094425, and BC039501); details are given for the most plausible assignment. We have observed that in the 43 cases that we experimentally analyzed, usually additional editing sites were identified when directly sequencing gene-specific PCR products. (276 KB XLS). Click here for additional data file. Accession Numbers The GenBank (( http://www.ncbi.nlm.nih.gov/Genbank ) accession numbers for the genetic sequences discussed in this paper are LUSTR (AB046844), KIAA0500 (AB007969), BTKI (AB037838), KIAA1497 (AB040930), and GPR81 (BC0067484). The Entrez Gene ( http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene ) ID numbers for ADAR1, p53, SIRT2, NFκB, and SPG7 are 103, 7157, 22933, 4790, and 6687, respectively.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526178.xml

549597

Designing Equitable Antiretroviral Allocation Strategies in Resource-Constrained Countries

Background Recently, a global commitment has been made to expand access to antiretrovirals (ARVs) in the developing world. However, in many resource-constrained countries the number of individuals infected with HIV in need of treatment will far exceed the supply of ARVs, and only a limited number of health-care facilities (HCFs) will be available for ARV distribution. Deciding how to allocate the limited supply of ARVs among HCFs will be extremely difficult. Resource allocation decisions can be made on the basis of many epidemiological, ethical, or preferential treatment priority criteria. Methods and Findings Here we use operations research techniques, and we show how to determine the optimal strategy for allocating ARVs among HCFs in order to satisfy the equitable criterion that each individual infected with HIV has an equal chance of receiving ARVs. We present a novel spatial mathematical model that includes heterogeneity in treatment accessibility. We show how to use our theoretical framework, in conjunction with an equity objective function, to determine an optimal equitable allocation strategy (OEAS) for ARVs in resource-constrained regions. Our equity objective function enables us to apply the egalitarian principle of equity with respect to access to health care. We use data from the detailed ARV rollout plan designed by the government of South Africa to determine an OEAS for the province of KwaZulu–Natal. We determine the OEAS for KwaZulu–Natal, and we then compare this OEAS with two other ARV allocation strategies: (i) allocating ARVs only to Durban (the largest urban city in KwaZulu–Natal province) and (ii) allocating ARVs equally to all available HCFs. In addition, we compare the OEAS to the current allocation plan of the South African government (which is based upon allocating ARVs to 17 HCFs). We show that our OEAS significantly improves equity in treatment accessibility in comparison with these three ARV allocation strategies. We also quantify how the size of the catchment region surrounding each HCF, and the number of HCFs utilized for ARV distribution, alters the OEAS and the probability of achieving equity in treatment accessibility. We calculate that in order to achieve the greatest degree of treatment equity for individuals with HIV in KwaZulu–Natal, the ARVs should be allocated to 54 HCFs and each HCF should serve a catchment region of 40 to 60 km. Conclusion Our OEAS would substantially improve equality in treatment accessibility in comparison with other allocation strategies. Furthermore, our OEAS is extremely different from the currently planned strategy. We suggest that our novel methodology be used to design optimal ARV allocation strategies for resource-constrained countries.

Introduction The HIV/AIDS epidemic is having a devastating impact in sub-Saharan Africa and other resource-constrained regions. Recently, the World Health Organization and other organizations have committed to expand access to antiretrovirals (ARVs) in the developing world, the United States government has pledged to provide $15 billion for AIDS in Africa and the Carribean, and drug prices have fallen [ 1 ]. However, even if these resources are provided for the global treatment of HIV, the number of individuals in need of treatment will far exceed the supply of ARVs [ 1 ]. Thus, difficult decisions will have to be made as to how to design HIV treatment strategies with these scarce resources. Resource allocation decisions can be made on the basis of many different epidemiological, ethical, or preferential treatment priority criteria. Many diverse groups have been suggested for treatment priority in resource-limited regions, including the following: only men, pregnant women, children, the sickest, the most economically productive, individuals in the military, or even individuals of the dominant ethnic group [ 2 ]. It has also been proposed that a lottery would be the only fair approach to allocating ARVs [ 3 ]. Only a limited number of ARVs will be available, and only a fixed number of health-care facilities (HCFs) can be used for ARV distribution. Thus, the resource allocation decisions that need to be made are extremely complex. Here, we use operations research to address this important resource allocation problem and to design ARV allocation strategies that are rational and equitable. The allocation decisions that we make here are based on ethical criteria, and not on epidemiological or preferential treatment priority criteria. Specifically, we determine the optimal allocation strategy that would ensure that each individual with HIV has an equal chance of receiving ARVs. We present a novel spatial mathematical model of treatment accessibility that we use in conjunction with an equity objective function to determine an optimal equitable allocation strategy (OEAS) for ARVs in a resource-constrained region. We quantify how changing the size of the catchment region surrounding each HCF, and the number of HCFs utilized for ARV distribution, alters the OEAS. Specifically, we use data from the detailed ARV rollout plan designed by the government of South Africa to determine an OEAS (based upon a variety of assumptions) for the province of KwaZulu–Natal. We also discuss how our proposed ARV allocation strategy differs from the currently proposed plan. Our current analysis is applied to the South African province of KwaZulu–Natal, although our methodology could be applied to any resource-constrained setting. KwaZulu–Natal is the largest province in South Africa with a population of approximately 9.4 million and has more people infected with HIV than any other province (approximately 21% of all cases in South Africa [ 4 ]). We use data from 51 communities (cities, towns, and villages) in the province of KwaZulu–Natal; we exclude communities with a population of less than 500 people. Data are not available on the number of individuals with HIV in each specific community, and thus we use the estimated HIV prevalence in the region (approximately 13% in urban areas and 9% in rural areas [ 4 ]) to estimate the number of infected people in each community. See Figure 1 and Table 1 for the population sizes and spatial locations of each of the 51 communities used in our analysis. For our analysis the quantity of ARVs available for distribution to the HCFs is sufficient to treat 10% of the total number of infected people, which is a realistic level during the incremental scale-up of ARV therapy over the next few years. The government of South Africa has selected 17 HCFs to participate in the ARV rollout that began in April 2004. These 17 HCFs are distributed throughout the province (see Figure 1 and Table 2 ). Some communities are close to HCFs, whilst others are a great distance from any HCF, with a range of 0–90 km ( Figure 2 A). Hence, this spatial distribution of HCFs produces large heterogeneity in accessibility to treatment. Inequality in access to health care is a common characteristic of resource-constrained regions [ 5 , 6 , 7 , 8 , 9 , 10 ]. We explicitly consider heterogeneity in treatment accessibility in our analysis of ARV allocation strategies. Figure 1 Map of South Africa Indicating the Location of the KwaZulu–Natal Province and Map of KwaZulu–Natal Black crosses indicate the location of the 17 HCFs that have been designated for ARV rollout by the South African government, and the spatial distribution of communities distinguished by the number of individuals infected with HIV (by both size and color). Durban (represented by the large red diamond) is the capital city of the province and has more individuals with HIV than any other community. Pietermaritzburg and Newcastle (represented by orange diamonds) have the next greatest numbers of individuals with HIV. Figure 2 Accessibility of Communities to HCFs (A) A histogram indicating heterogeneity in the distance from communities in KwaZulu–Natal to the closest HCF. The treatment accessibility function used in our model is a Gaussian distribution, exp(− kd 2 ), indicating that accessibility is strongly related to distance (d), and k is a dispersal length scale parameter. (B) The catchment region is shown with an effective radius of 20 km for coverage from each HCF ( k = 0.0151). (C) The catchment region is shown with an effective radius of 40 km for coverage from each HCF ( k = 0.003786). (D) The catchment region is shown with an effective radius of 60 km for coverage from each HCF ( k = 0.00168). In each case, the red dots indicate the location of the HCF, the green circles represent the locations where treatment accessibility has been reduced to 50% relative to someone located at the HCF, and the blue circles represent the locations where treatment accessibility has been reduced to 1% relative to someone located at the HCF. The locations of communities are presented as black diamonds. The large black diamonds denote large communities (with population greater than 10,000 people), and the small black diamonds denote small communities (with population less than 10,000 people). Substantially more area of the province is covered if HCFs have catchment regions of 60-km radius, relative to catchment regions of 40-km radius, and substantially less area of the province is covered if HCFs have a catchment region of only 20-km radius. However, the proportion of people with access does not differ greatly between the different catchment sizes because of the great spatial heterogeneity in the prevalence of people with HIV. Table 1 Communities in KwaZulu–Natal with Populations of at Least 500 People The location and population of each community (city, town, or village) is indicated. Sources: [ 33 ] and the World Gazetteer ( http://www.world-gazetteer.com ) Table 2 HCFs Proposed for Use in the Rollout of ARVs in KwaZulu–Natal Source: KwaZulu–Natal Department of Health ( http://www.kznhealth.gov.za/default.htm ) We have developed a novel spatial mathematical model of treatment accessibility that we use to determine an OEAS for ARVs in a resource-constrained region. To the best of our knowledge, this is the first analysis to address how to deal with the extremely difficult problem of allocating a scarce supply of ARVs in order to design a rational and equitable allocation strategy. We model the “spatial diffusion of treatment” to the locations of disease, rather than modeling the “spatial diffusion of disease,” which is the conventional approach [ 11 , 12 , 13 , 14 , 15 , 16 ]. Our spatial model includes HCFs and the HIV-infected communities surrounding these HCFs; we refer to the region around each HCF as the catchment region. Thus, the radius of the catchment region specifies the approximate maximum distance that we assume infected people would be willing (or able) to travel for treatment. Each HCF can serve many communities, and some communities can access multiple HCFs; our model sums the number of people with HIV in each HCF's catchment region who could potentially travel to the HCF to receive ARVs (we define this number as the “effective demand” on that specific HCF). Thus, the “effective demand” on each HCF is a direct function of the number of individuals with HIV in the catchment region, weighted by their distance from the HCF. By including a weighting function we explicitly model heterogeneity in accessibility to treatment based on distance from the HCF. Here, the distance from a HCF becomes the main determinant influencing whether or not an individual with HIV has access to treatment. We developed an equity objective function to assess how the limited supply of ARVs should be allocated to each HCF to ensure that an equal proportion of infected people in each community receive treatment. To apply our theoretical framework to KwaZulu–Natal we model the specific location of the 17 HCFs and the 51 communities of 500 or more individuals (see Figure 1 ); for these conditions we determine an OEAS. We compared our OEAS with two other allocation strategies: (i) allocating ARVs only to Durban, the major urban area (i.e., concentrating ARVs where there is the best health-care infrastructure) and (ii) allocating ARVs equally to all 17 HCFs. We conduct our analysis assuming three different radii of catchment regions: 20 km, 40 km, and 60 km. We then extend this analysis and recalculate the OEAS assuming that more than 17 HCFs are available to distribute ARVs. This analysis is useful because there is a second potential pool of 27 ARV-implementation HCFs in the South African operational plan for ARV rollout [ 17 ]. We analyze this case, in which 27 HCFs are utilized in the ARV rollout, and we also analyze how optimal ARV allocation would change if all 54 hospitals in KwaZulu–Natal were operational for the rollout of ARVs. Methods Calculating Demand and Treatment Access We assume that the number of people with HIV who will travel to a specific HCF is directly proportional to the number of individuals with HIV in that particular community, but that the probability of an individual traveling to receive ARVs (i.e., the treatment accessibility) decreases with distance from the HCF. We define d i,j as the distance from community i to HCF j, f ( d i,j ) as a weighting function that determines the treatment accessibility to a HCF based upon distance d i,j , and I i as the number of people with HIV in community i . The distance, d ij , between community i and HCF j is based on the longitude (lon) and latitude (lat) of each location and is determined by where R is the radius of the earth, taken to be 6,371 km, and the angles are in radian measure. We calculate the “effective demand” of community i on HCF j to be the number of people with HIV in community i that will travel to HCF j for ARV regimes, namely, f ( d i,j ) I i . Thus, demand on HCFs for ARVs is reduced by the treatment accessibility function. Our model is conceptually similar to the “gravity” models that have been used to predict retail travel [ 18 ], plan land use [ 19 ], and determine accessibility of primary care [ 20 ]. However, this is to our knowledge the first time this approach has been used to calculate ARV allocations. We use a Gaussian to model treatment accessibility, f ( d ) = exp(− kd 2 ), where k is a dispersal length scale parameter determining the radius of the catchment region. The size of the actual catchment regions is unknown, but based upon distances from communities to HCFs in KwaZulu–Natal (see Figure 2 A) we assume that individuals are likely to travel a maximum distance of approximately 40 km to a HCF ( k = 0.003786). We vary the catchment region by considering a 20-km radius ( k = 0.0151) and a 60-km radius ( k = 0.00168). The different catchment regions that we simulate (with radii of 20 km, 40 km, and 60 km) for each HCF are illustrated in Figure 2 B– 2 D. The number of people with HIV throughout the province that have access to HCFs is approximately 86% of the total number of people with HIV for the case of a 20-km catchment region, 89% for a 40-km catchment region, and 93% for a 60-km catchment region. Modeling the Distribution of Treatment To determine how many ARVs should be allocated to each HCF, we first calculate how a given supply of ARVs will be distributed from each HCF to the surrounding communities in the catchment region. We calculate the “effective demand” on HCF j, D j , to be which sums the “effective demand” of all communities on HCF j (where there are m communities). Then, we model the distribution of ARVs from a HCF to each community within the catchment region as the proportion of the “effective demand” on HCF j that is contributed by the respective community. Accordingly, ARVs will be distributed from HCF, j, to each community as the ratio Therefore, the number of people treated in community i by the drug supply allocated to HCF j is where S j is the number of regimes allocated to HCF j . Hence, the total number of people with HIV treated in community i,T i , summing over all n HCFs is The Equity Objective Function We establish an equity objective function to determine the optimal equitable allocation of ARVs to each HCF so that all individuals with HIV have an equal chance of receiving treatment. To obtain the same fraction of treated individuals in each community, given that there are A ARV regimes for a total of individuals with HIV, the resulting objective function to minimize (based on least squares) becomes Our goal is to minimize E, by solving for the number of ARVs to be allocated to each HCF ( S 1 , S 2 ,…, S n ), whilst enforcing the following three constraints: (i) ensure that the total number of ARVs available is equal to the sum of the supply allocated to all HCFs, (ii) ensure that only a positive number of ARVs are allocated to each HCF ( S j ≥ 0, j = 1… n ); and (iii) ensure that the number of people treated in each community is not greater than the number of people with HIV in the community ( T i ≤ I i , i = 1… m ). We note that if a different objective is required, then all of our preceding analysis still holds and only the functional form of the objective function needs to be altered. To solve the problem, and determine the OEAS, we used successive linear programming operations research techniques [ 21 ]. Results The OEAS of ARVs in KwaZulu–Natal that we determined is complex (see Figure 3 A and 3 B). According to our OEAS, the majority of ARVs should be allocated to HCFs in Durban, and the remaining ARVs should be allocated to the other HCFs throughout the province (with two non-Durban HCFs receiving 5%–15% of the total ARVs and the remaining non-Durban HCFs each receiving less than 5% of the total ARVs available). We note that our OEAS does not produce perfect equality; however, our optimal strategy significantly improves equality in obtaining treatment over the two other allocation strategies that we analyzed for comparison: (i) ARVs allocated only to one HCF (in the largest city, Durban) (see Figure 3 D and 3 E), and (ii) equal quantities of ARVs allocated to each HCF throughout the province (see Figure 3 G and 3 H). For comparison of allocation strategies (in Figure 3 ) we used an effective catchment radius of 40 km ( k = 0.003786). The proportion of infected individuals that are treated at each location is displayed graphically in Figure 3 for our OEAS ( Figure 3 C) and the two comparison allocation strategies ( Figure 3 F and 3 I). The best achievable outcome, given the limited treatment resources available, is that 10% of people with HIV are treated in each community throughout the province, yielding the map shown in Figure 3 C, 3 F, and 3I, but with dark blue/magenta over the entire province. Whilst our OEAS does not fully achieve this, it is considerably better than both of the comparison ARV allocation strategies. Furthermore, the equity objective function evaluates to E = 0.27 for our OEAS, compared with (i) E = 0.50 and (ii) E = 133.88 for the comparison allocation strategies. There is large diversity in the fraction of individuals with HIV treated per community when equal quantities of ARVs are given to each HCF, evidenced by an inter-quartile range of 0.025%–41.746% compared with inter-quartile ranges of 0%–0% and 0.011%–9.982% for the first comparison strategy and our OEAS, respectively. Therefore, equal access is not obtained if equal quantities of ARVs are allocated to each HCF. Obviously, allocating to only one HCF (the first comparison strategy) could also be considered unequal because although the inter-quartile range is minimal, effectively only one community (Durban) receives ARVs. Our OEAS, while not perfect, achieves the best equality possible given the accessibility constraints and limited ARV supply. Figure 3 Pie Charts of the Three Strategies for Allocating ARVs to HCFs The three strategies considered are as follows: allocation of ARVs according to the results of minimizing our objective function (first row) allocation of ARVs only to one HCF in Durban (second row), allocation of ARVs equally to each of the 17 HCFs (third row). The proportion of ARVs allocated by these strategies to the 17 different HCFs is indicated in (A), (D), and (G), with each HCF represented by a different color. The spatial allocation of ARVs is shown in (B), (E), and (H), respectively. The respective percentage of infected people that are treated throughout the KwaZulu–Natal province is simulated in (C), (F), and (I). Here, the x–y plane represents spatial location, and the shaded color at a location refers to the proportion of individuals with HIV that are treated at the specified location. The plots were obtained by generating an interpolating surface where the z -ordinate, colored by magnitude, represents the proportion of treated individuals, and then orientating the view of the surface normal to the x–y plane. We performed surface data interpolation using the method of translates [ 32 ]. The catchment region for HCFs is a factor of large uncertainty. We considered three catchment region sizes: radii of 20 km, 40 km, and 60 km. We also simulated two additional cases with increased numbers and locations of HCFs (27 HCFs as suggested in South Africa's official ARV rollout operational plan [ 17 ]; and all 54 hospitals in KwaZulu–Natal). In Figure 4 we present box plots of the percentage of infected people that obtain treatment per community for the three sets of HCFs and the three catchment region sizes we simulate. For each specified condition we calculate the OEAS. It is apparent that equality in access to ARVs is improved substantially if the radius of each catchment region is increased and/or the number of HCFs is increased ( Figure 4 ). Our results show that the number of HCFs utilized is of greater importance than the size of the catchment region. If 54 HCFs are used, then even a (small) catchment radius of 20 km results in the ideal median proportion of 10% of people with HIV in each community receiving ARVs. In the case of 27 HCFs, 88% of all people with HIV have access to HCFs for a 20-km catchment region, 91% for a 40-km catchment region, and 96% for a 60-km catchment region. In the case of 54 HCFs, 90% of all people with HIV in the province have access to HCFs for a 20-km catchment region, 94% for a 40-km catchment region, and 99% for a 60-km catchment region. Therefore, increasing the number of HCFs available for an ARV rollout is effective in significantly increasing equality in treatment accessibility as shown in Figure 4 . Furthermore, if catchment regions actually have a radius of 60 km, or can be increased to this size through improvements in transportation, this would enable access to HCFs for almost all people in the province, as shown in Figure 4 . The actual HCF allocations determined by our model and optimization for the cases of 17, 27, and 54 HCFs (and for all catchment sizes we consider) are presented as pie charts in Figure 5 . It is clear from our analysis that the equality criterion, such that each individual with HIV in KwaZulu–Natal has an equal chance of receiving ARVs, can best be satisfied by utilizing all 54 HCFs for ARV distribution and ensuring that each HCF serves a catchment region of 40 to 60 km. Figure 4 Percentage of People with HIV That Obtain Treatment per Community for Various Approaches Box plots of the percentage of infected people that obtain treatment per community for the three different sets of HCFs simulated in our analysis for ARV rollout, namely, using the 17 HCFs likely to be used, the 27 HCFs suggested by the South African government as potential implementation points, and all of the 54 hospitals in the KwaZulu–Natal province. These cases are represented for each of the three catchment region sizes we considered (with radii of 20 km, 40 km, or 60 km) and referenced against the ideal fraction treated (dotted blue line) under perfect conditions of egalitarian distribution, given the limited ARV supply. The red crosses indicate the median percentage of people with HIV that obtain treatment per community. Figure 5 Actual Allocation of ARVs to HCFs These pie charts show ARV allocation to HCFs according to our model and optimization for the cases of 17 , 27 , and 54. The allocation is shown for each of the catchment region sizes considered: 20-km radius, 40-km radius, and 60-km radius. Discussion We have established an elegant and simple theoretical framework for determining an equitable and rational allocation of ARVs to HCFs in resource-constrained countries. To the best of our knowledge, this is the first analysis to address this very difficult problem. We determined that increasing the size of the catchment region of each HCF can improve access to HCFs considerably for rural populations. We suggest that studies be performed to collect data on the distance that individuals with HIV are willing and able to travel for treatment. This will facilitate discussions of this important issue, which must be considered in the making of policy decisions. A database consisting of such information has been proposed for South Africa [ 22 ]. In an effort to provide equal access to communities with relatively little access to ARV therapy, the concept of a mobile clinic that would travel between communities to take health-care workers and resources to the location of the demand is a new initiative in Nigeria (S. Agwale, personal communication) that could also be considered in other regions. We calculated the optimal allocation of ARVs to available HCFs so that all infected individuals will have as close as possible to an equal chance of obtaining treatment. We have shown that increasing the number of HCFs involved in ARV distribution can improve equality of access to ARVs substantially. The current plan in KwaZulu–Natal is to use only 17 HCFs. However, our results clearly show that in order to achieve an optimal equitable allocation strategy, all existing infrastructure (i.e., all 54 HCFs) should be used. The strategy that we are advising may be fairly easy to accomplish at the policy level because the health-care infrastructure (specifically these HCFs) already exists, although consideration must be made for issues such as the training and transportation that is necessary, which may be costly. In contrast, increasing the size of catchment regions may be very difficult. Obviously, increasing both the number of HCFs and the size of the catchment region each services would substantially increase equality of access to health care in KwaZulu–Natal. Future modeling studies could extend our work by not making the simplifying assumption that all patients have similar ease of travel over the same distance and by including weighting functions on distance impedance for different communities (based on the quality of the road infrastructure, for example, and the availability of transportation) (D. P. Wilson, J. O. Kahn, S. M. Blower, unpublished data). Here, we have shown how to calculate optimal ARV allocation strategies based upon the principle of equity. Future research is necessary to compare ARV allocation strategies based upon the principle of efficiency (i.e., allocating ARVs to maximize epidemic reduction) in order to determine whether utilizing different principles for optimization would result in similar (or different) allocation strategies. The World Health Organization and the Joint United Nations Programme on HIV/AIDS have identified three core principles that should underlie the effort to fairly distribute ARVs, namely: urgency, equity, and sustainability [ 23 ]. They state that policy decisions for the fair distribution of ARVs should be based upon the following ethical principles: (i) the principle that like cases should be treated alike, (ii) the utilitarian principles of maximizing overall societal benefits, (iii) the egalitarian principles of equity (distributing resources, such as health care, equally among different groups), and (iv) the Maximin principle (which prioritizes individuals that are the least advantaged) [ 24 ]. Here, we investigated the level of decision-making associated with allocating ARVs to HCFs, and we have applied the egalitarian principle of equity with respect to access to health care. We suggest that allocating ARVs to HCFs to achieve equality in accessibility could be carried out, and then individual-level ethical considerations could be thought out at the next level of deliberation. Future research is necessary to identify alternative (and more detailed) ethical ARV allocation strategies. Although we have focused on one equitable strategy, there are many other ARV allocation strategies that are ethical. Uneven access to HIV treatment has the very real potential to fracture social and political structures and could lead to intrastate and/or interstate conflict [ 2 ]. Government decisions on ARV allocation have potentially socially destabilizing ramifications because essentially the decisions determine who lives and who dies. Resource allocation decisions will have to be made at a number of levels: it must be decided what proportion of the available ARVs should be allocated to each province; then it must be decided how many ARVs should be allocated to each HCF within each region; and finally, particular groups of individuals may be chosen to have treatment priority. Treatment priority decisions for individuals could be based on many different criteria, including disease progression (CD4 cell counts and viral load), socioeconomic status, ethnicity, and who is thought to have the greatest risk of transmitting infections (for example, pregnant women with HIV or female sex workers). Although it could be argued that behavioral core groups should be targeted to receive ARVs because this may have the greatest epidemiological impact, such an allocation strategy would be neither feasible nor practical to implement. For example, sex workers are an obvious behavioral core group, but many women would likely claim to be sex workers if they knew that ARVs were only available to sex workers. Additionally, the ethics of targeting such groups in favor of other societal groups must be questioned. It could also be argued that, to maximize the preventative effect of ARV therapy, ARVs should be concentrated in virological core groups (i.e., people with the highest viral load) [ 25 , 26 ]; this novel approach of targeting the virological core group has recently been proposed for controlling HSV-2 epidemics [ 27 ]. Identifying individuals in the virological core group would be far easier than identifying individuals in the behavioral core group. These individuals are likely to be the sickest and those with evidence of disease-related symptoms. Treatment allocation strategies could also be designed based on reducing the future epidemic impact and disregarding treatment equality amongst currently infected people. Such strategies place different social value on currently infected people in comparison with future infected people; such strategies therefore may not be ethical even though they may be epidemiologically sound (also, it is important to note that any epidemic predictions have large uncertainty ranges [ 28 , 29 ]). Our model has been applied to the South African province of KwaZulu–Natal, but it can be applied by government health officials in any resource-constrained country. In many of the countries worst affected by the HIV pandemic, scarcity of resources will mean that not everyone that could potentially benefit from ARVs will be able to access them. Many of the decisions that must be made to develop an effective response to the HIV/AIDS epidemic are inevitably underpinned by ethical considerations. Leadership in most resource-constrained regions cannot avoid these decisions. Whilst there has been considerable attention given to South Africa, many other countries worldwide either have plans in place (e.g., Brazil, Thailand, and Botswana) or are in the process of developing national programs for ARV distribution through the public health system (e.g., Mozambique, Malawi, and Kenya) [ 1 ]. Legitimate authorities in each nation must come to their own consensus on the priorities and objectives of an ARV rollout, which is not a trivial matter [ 1 , 30 ]. Our objective function and model can be used to calculate allocation strategies that provide equity in access (compensating for geographical isolation), but if authorities in a given nation prioritize a different goal for ARV rollout, then an objective function to optimize can be formulated to reflect the specific national policy goal. Our model can be used by policy makers to determine an optimal scientifically based allocation strategy, based upon the specific objective function. As the ARV rollout commences in KwaZulu–Natal, difficult decisions will have to be made as to how to allocate scarce resources. We have shown that it is possible to obtain a mathematical solution to an equity problem. We suggest that our novel approach could be used to determine optimal equitable allocation strategies for many other resource-constrained countries that are just beginning to receive ARVs [ 31 ]. Patient Summary Background Antiretroviral drugs can change the lives of patients with HIV/AIDS. Their high price, however, means that many poor countries do not have enough of these drugs to treat all the people who need them. The decision of who will get treatment is very difficult, and different ways to come up with ethical solutions to the problem have been proposed. Why Was This Study Done? One of the approaches is to try to make sure that every infected person has the same chance to get antiretroviral drugs. David Wilson and Sally Blower, the authors of this study, wanted to find a scientific strategy to achieve this goal of equal access. What Did the Researchers Do? They used mathematical models to calculate how to distribute available drugs among hospitals and doctor's offices so that each patient in a particular area had an equal chance to get treated. What Did They Find? When they used their approach on a real example, the South African province of KwaZulu–Natal, they found that making some changes to the current plans for drug distribution would lead to more equal access among all of the individuals with HIV in the province. Instead of only 17 out of the 54 health care facilities in KwaZulu–Natal distributing the drugs (which is the current plan of the South African government), Wilson and Blower calculate that it would be fairer if all 54 facilities distributed the medicines. What Does This Mean? Mathematical models like the one used here are always based on assumptions and simplifications. As a consequence, they are never perfect matches for a real-life situation, but they can help to guide complicated decisions. This article suggests that the approach Wilson and Blower developed could help to determine strategies for equitable allocation of limited HIV treatment resources. What Next? The authors hope that the tools they developed will be used by policy makers in resource-poor countries to guide their strategies. They are keen to work with these policy makers to adapt and optimize the method to local settings and priorities. More Information Online Report by the World Health Organization and the Joint United Nations Programme on HIV/AIDS on ethics and equitable access to HIV/AIDS treatment: http://www.who.int/hiv/pub/advocacy/en/ethicsmeetingreport_e.pdf Ruth Macklin's report on ethics and equity in access to HIV treatment: http://www.who.int/ethics/en/background-macklin.pdf The Pro-Poor Health Policy Team's report on priority in HIV/AIDS treatment: http://www.who.int/ethics/en/background-pro-poor3.pdf News article from the World Health Organization Bulletin on the South African HIV/AIDS treatment program: http://www.who.int/bulletin/volumes/82/1/en/news.pdf

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549597.xml

548676

Comparative performance of the Mbita trap, CDC light trap and the human landing catch in the sampling of Anopheles arabiensis, An. funestus and culicine species in a rice irrigation in western Kenya

Background Mosquitoes sampling is an important component in malaria control. However, most of the methods used have several shortcomings and hence there is a need to develop and calibrate new methods. The Mbita trap for capturing host-seeking mosquitoes was recently developed and successfully tested in Kenya. However, the Mbita trap is less effective at catching outdoor-biting Anopheles funestus and Anopheles arabiensis in Madagascar and, thus, there is need to further evaluate this trap in diverse epidemiological settings. This study reports a field evaluation of the Mbita trap in a rice irrigation scheme in Kenya Methods The mosquito sampling efficiency of the Mbita trap was compared to that of the CDC light trap and the human landing catch in western Kenya. Data was analysed by Bayesian regression of linear and non-linear models. Results The Mbita trap caught about 17%, 60%, and 20% of the number of An. arabiensis , An. funestus , and culicine species caught in the human landing collections respectively. There was consistency in sampling proportionality between the Mbita trap and the human landing catch for both An. arabiensis and the culicine species. For An. funestus , the Mbita trap portrayed some density-dependent sampling efficiency that suggested lowered sampling efficiency of human landing catch at low densities. The CDC light trap caught about 60%, 120%, and 552% of the number of An. arabiensis , An. funestus , and culicine species caught in the human landing collections respectively. There was consistency in the sampling proportionality between the CDC light trap and the human landing catch for both An. arabiensis and An. funestus , whereas for the culicines, there was no simple relationship between the two methods. Conclusions The Mbita trap is less sensitive than either the human landing catch or the CDC light trap. However, for a given investment of time and money, it is likely to catch more mosquitoes over a longer (and hence more representative) period. This trap can therefore be recommended for use by community members for passive mosquito surveillance. Nonetheless, there is still a need to develop new sampling methods for some epidemiological settings. The human landing catch should be maintained as the standard reference method for use in calibrating new methods for sampling the human biting population of mosquitoes.

Background Mosquito sampling is a prerequisite to most vector population studies [ 1 ]. The entomological parameter being studied and the behaviour of the mosquito species being sampled determine the choice of a sampling method [ 2 ]. However, most of the available mosquito sampling methods may not allow for such rational choices to be made, as there are major limitations associated with their use [ 3 ]. Therefore, new tools for sampling mosquito vector populations must be continuously developed. Nonetheless, even these new sampling tools must be calibrated against the existing ones in different vectorial systems [ 4 ] if they are to be adopted for conventional use. A new trap, the Mbita trap has been developed [ 5 ] and separately evaluated in quite different vectorial systems in Western Kenya and Madagascar [ 6 , 7 ] with varying degrees of success. The Mbita trap was originally developed in semi-field systems with an Anopheles gambiae colony originating from southern Tanzania [ 5 ] and proved a sensitive and representative way to sample An. gambiae , Anopheles arabiensis and Anopheles funestus in Western Kenya [ 6 , 7 ]. In sharp contrast, the Mbita trap proved highly insensitive for catching An. funestus and An. arabiensis in rice-growing communities in the highlands of Madagascar [ 6 , 7 ]. In this study, the performance of the Mbita trap compared to the CDC light trap hung adjacent to a human-occupied bednet and the human landing catches in the sampling of An. arabiensis An. funestus and culicines species of mosquitoes in a rice-growing community in western Kenya with relatively high mosquito densities is reported. Methods Description of the study area These studies were carried out in a village adjacent to the Ahero rice irrigation scheme in western Kenya. Populations of An. arabiensis and An. funestus [ 8 ] as well as culicine species [ 1 ] are predominant in this area. The characteristics of the mosquito population and malaria vectorial system in this area have been described in detail elsewhere [ 1 , 9 ]. Sampling In Ahero, three houses were selected upon receiving consent from the household heads. Occupants were given a non-impregnated bed net per sleeping space and trained in their correct use. With informed consent, three young men who had earlier been trained in mosquito sampling [ 6 ] were recruited to act as bait in the three alternative mosquito collection methods. On each experimental night, one of the three subjects slept in the Mbita trap (BNT), another slept in a bed net with a CDC light trap suspended beside it (CDC) and the third conducted a human landing catch (HLC) [ 3 ]. Both the Mbita trap and the CDC light trap-bed net system were set on mattresses placed on mats laid on the floor and not on beds. In all the experiments, a standard miniature CDC light trap (Model 512; John W. Hock Company, Gainesville, Florida, USA) with an incandescent light bulb was used. The trap was hung beside the bed net on the foot side of the sleeping person with its shield touching the side of the net and its inlet about 25 cm above the sleeping person [ 10 ]. Each of the three sampling methods was allocated to one of the three houses on a given night in a 3 × 3 randomised Latin square experimental design replicated 3 times. The human baits did not move around the sites so that the effects of a particular site and the attractiveness of the human bait associated with it were combined for simplified statistical analysis. Sampling was carried out from 20.00 hrs to 06.00 hrs between October and November 2002. Ethical considerations Informed consent was obtained from all the participants. Thick and thin blood smears were regularly taken from the participants to examine for the presence of malaria parasites and, when found positive, they were treated with pyrimethamine-sulfadoxine (Fansidar ® ). A follow-up was made to ensure that any parasitaemia was fully cleared. If parasitaemia did not clear, the participants were referred to hospital for further treatment with second line drugs. The Kenya Medical Research Institute (KEMRI) through the KEMRI/National Ethical Review Committee granted ethical approval (KEMRI/7/3/1) for this study. Mosquito processing Mosquitoes were taken to the laboratory and killed by suffocation with chloroform vapour. They were counted and identified morphologically using taxonomic keys [ 11 , 12 ] and then desiccated over anhydrous copper sulphite and kept at room temperature until further processed. Abdomens of An. gambiae sensu lato were analysed by PCR for sibling species identification [ 13 ]. Statistical methods The simple expedient of adding one to each mosquito count in order to cater for zero counts can be misleading [ 14 ]. Therefore, Winbugs version ® 1.4 was used to fit regression-based models to the data. The conceptual basis of this Bayesian regression has been described in detail elsewhere [ 15 ]. The following models were fitted to the data: Scenario A: A linear model for sampling proportionality E ( y i ) = α t β c E ( x i ) Where: E ( y i ) is the expected number of mosquitoes caught using the method being tested; E ( x i is the expected number of mosquitoes caught using the human landing method (assuming the same mosquito collector as bait); α t is a multiplication factor corresponding to trapping method t in relation to the reference trapping method which was human landing catch; and β c is a multiplication factor corresponding to human bait c compared to the reference catcher, assigned number 1, whose value is set to 1. Scenario B: A non-linear model for sampling proportionality E ( y i ) = α t β c ( E ( X i )) yt All the terms for model B are identical to model A except that it includes y t which is the exponent corresponding to trapping method t . A value of y t different from 1 indicates a lack of proportionality between the methods. Both models assumed Poisson errors in the numbers of mosquitoes caught by any of the three methods. Results Overall, the Mbita trap, the human landing collection and the CDC light trap-bednet method caught 135, 576, and 474 An. arabiensis and 309, 427, and 470 An. funestus respectively. The corresponding figures for culicines mosquitoes (mainly Culex species) were 32, 121 and 578. Also, 30 male mosquitoes were caught, 29 of them by the CDC light trap and one in the landing collections. The parameter estimates from our models (Table 1 , Figure 1 ) indicate the Mbita trap caught about 17%, 60%, and 20% of the number of An. arabiensis , An. funestus , and culicine species caught in the human landing collections respectively. There was consistency in the sampling proportionality between the Mbita trap and the human landing catch for both An. arabiensis and the culicine species whereas for An. funestus , the Mbita trap portrayed some density-dependent sampling efficiency. More specifically, the Mbita trap appears more sensitive than human landing catch at low mosquito densities. The CDC light trap, on the other hand, caught about 60%, 120%, and 552% of the number of An. arabiensis , An. funestus , and culicine species caught in the human landing collections respectively. There was consistency in the sampling proportionality between the CDC light trap and the human landing catch for both An. arabiensis and An. funestus , whereas for the culicines, there was no simple relationship between the CDC light trap catches and the landing catches (Table 1 , Fig. 1 ). From PCR identification, all the successfully amplified specimens of An. gambiae s.l. were found to be An. arabiensis . Table 1 Point estimates and 95% confidence intervals for model parameters. An. arabiensis An. funestus Culicines Model A α t : BNT versus HLC 0.17 (0.14, 0.21) 0.73 (0.62, 0.85) 0.20 (0.13, 0.29) α t : CDC versus HLC 0.56 (0.49, 0.66) 1.19 (1.03, 1.37) 4.84 (3.81, 6.21) β c : Person 2 vs person 1 0.38 (0.32, 0.45) 0.89 (0.77, 1.02) 0.54 (0.41, 0.72) β c : Person 3 vs person 1 0.54 (0.47, 0.62) 0.77 (0.66, 0.90) 0.59 (0.42, 0.80) Model B α t : BNT versus HLC 0.80 (0.52, 1.12) 0.39 (0.30, 0.50) 0.71 (0.00, 2.02) α t : CDC versus HLC 1.04 (0.79, 1.42) 0.84 (0.66, 1.06) 5.52 (3.08, 7.46) Figure 1 Numbers of female mosquitoes caught by the three sampling methods in 9 nights in western Kenya. Regression lines (unbroken) depict the fitted simple proportionality model (Model A ) and the non-proportional (broken lines), density-dependent sampling efficiency model (Model B ). Discussion The results obtained from this study indicate a three-fold decrease in efficiency for both the Mbita trap and the CDC light trap when used to sample An. arabiensis compared to their reported performance for nearby An. gambiae s.l. population that comprised of roughly equal numbers of An gambiae sensu stricto and An. arabiensis [ 6 ]. However, the consistency in the proportionality of their catches relative to the human landing collections was maintained. Several factors might explain this observation. First, this area is dominated by An. arabiensis , a mosquito species that is usually largely zoophagic but endophilic [ 16 ]. Therefore, protecting all the people in a bednet, as was the case in the houses where the Mbita trap and the CDC light trap were used, might have prompted the indoor resting An. arabiensis to seek alternative hosts outdoors. For the case of the human landing collection, the human bait was more readily available for such indoor resting mosquitoes. Second, unlike in Lwanda [ 6 ], where no cattle were present in any of the homesteads that were sampled, large numbers of cattle were present in all the homesteads sampled at Ahero. Therefore, there was an alternative source of blood meal to this more flexible species, which can utilize both domestic [ 16 ] and wild bovids [ 17 ]. The availability of cattle could possibly account for the reported poor performance of the Mbita trap in sampling An. arabiensis in the highlands of Madagascar [ 7 ]. Other studies in Ahero have reported similar CDC light trap to human landing catch ratios [ 1 ] for An. arabiensis as this study found but with no correlation between the two methods. The performance of the Mbita trap relative to the human catch for An. funestus in Ahero was similar to that reported for Lwanda [ 6 , 7 ] but showing density dependent sampling efficiency as density increased. Specifically, it appears that the Mbita trap may be more sensitive at low densities (Figure 1 ). It was considered whether this could be caused by the lowered attentiveness of individuals conducting tedious human landing catches when few mosquitoes are present. However, this was not found to be the case as the sampling efficiency of the CDC light trap, relative to the human landing catches showed no density dependence. The efficiency of the CDC light trap for An. funestus was about 2.5-fold that in Lwanda. The relatively high densities of this species in Ahero compared to Lwanda, might, at least partly, account for these observations. At Lwanda, where the densities of An. funestus were low, no density-dependent sampling efficiency was noted for the Mbita trap while some density-dependent sampling efficiency was noted for the CDC light trap [ 6 ] suggesting that the Mbita trap is more sensitive in low densities while the CDC light trap is better at higher densities of this species. Many studies have evaluated the performance of the CDC light trap relative to the human landing catch but it is very difficult to compare the results due to the different methodologies and sampling procedures applied. In this study, the three methods were used concurrently in different houses on the same night while in other studies [ 2 , 18 , 19 ] the methods were used in the same houses but on different nights. Small differences in sleeping arrangements, availability of alternative hosts, temperatures, humidity, and wind speed and direction between the different days might introduce some sampling bias in this case. Furthermore, the procedures used for conducting human landing catch also vary appreciably: some studies have used one human per house to perform landing catches [ 20 ] while others [ 2 , 18 ] have used two catchers in the same house. There is, therefore, a need to standardize the operational conditions and sampling procedures used if valid comparisons between various studies in are to be made Conclusions Although the Mbita trap is less sensitive than either the human landing collections or the CDC light trap, for a given investment of time and money [ 5 ], it is likely to catch more mosquitoes over a longer period, larger number of sampling sites or both. Adult mosquito densities are highly aggregated in space and time, resulting over 80% of transmission occurring in 20% of places and time [ 21 ], and the importance of catching them across large numbers of sampling points and frequent intervals to obtain representative samples of the vector population has recently been emphasized [ 22 ]. The Mbita trap may therefore be very useful for enabling community members in collecting large numbers of samples that are representative of the overall vector population at a less cost [ 5 ] than a smaller number of light traps/human catchers. Used in this way, rather than as a direct replacement for the CDC light trap-bednet method, this trap will surely find a place in community-based malaria vector surveillance. It might be important to note that some community members in Rusinga, an island adjacent to ICIPE-Mbita point where the trap was developed, have adopted this trap for passive mosquito surveillance with some encouraging results. However, this trap might not work in all epidemiological settings [ 7 ] and therefore more mosquito behavioural studies should be carried out in order to gain more insight to guide further development of mosquito sampling and control tools. The human landing catch should be maintained as the standard reference method for use in calibrating new methods for sampling the human biting population of mosquitoes. Authors' contributions EMM designed the trap, designed the study and drafted the manuscript. GOM & DOO supervised the fieldwork and recorded the data. LWI & PNN were involved in the study design and drafting the manuscript, TAS carried out the data analysis and helped in the interpretation of results. GFK guided the experimental design, data analysis and drafting of manuscript. BGJK conceived the initial idea of developing the trap and solicited for funds used in the trap development and trials. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548676.xml

517399

Slowly developing depression of N-methyl-D-aspartate receptor mediated responses in young rat hippocampi

Background Activation of N-methyl-D-aspartate (NMDA) type glutamate receptors is essential in triggering various forms of synaptic plasticity. A critical issue is to what extent such plasticity involves persistent changes of glutamate receptor subtypes and many prior studies have suggested a main role for alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptors in mediating the effect. Our previous work in hippocampal slices revealed that, under pharmacological unblocking of NMDA receptors, both AMPA and NMDA receptor mediated responses undergo a slowly developing depression. In the present study we have further adressed this phenomenon, focusing on the contribution via NMDA receptors. Pharmacologically isolated NMDA receptor mediated excitatory postsynaptic potentials (EPSPs) were recorded for two independent synaptic pathways in CA1 area using perfusion with low Mg 2+ (0.1 mM) to unblock NMDA receptors. Results Following unblocking of NMDA receptors, there was a gradual decline of NMDA receptor mediated EPSPs for 2–3 hours towards a stable level of ca. 60–70 % of the maximal size. If such an experimental session was repeated twice in the same pathway with a period of NMDA receptor blockade in between, the depression attained in the first session was still evident in the second one and no further decay occurred. The persistency of the depression was also validated by comparison between pathways. It was found that the responses of a control pathway, unstimulated in the first session of receptor unblocking, behaved as novel responses when tested in association with the depressed pathway under the second session. In similar experiments, but with AP5 present during the first session, there was no subsequent difference between NMDA EPSPs. Conclusions Our findings show that merely evoking NMDA receptor mediated responses results in a depression which is input specific, induced via NMDA receptor activation, and is maintained for several hours through periods of receptor blockade. The similarity to key features of long-term depression and long-term potentiation suggests a possible relation to these phenomena. Additionally, a short term potentiation and decay (<5 min) were observed during sudden start of NMDA receptor activation supporting the idea that NMDA receptor mediated responses are highly plastic.

Background Hippocampal synapses display a variety of activity dependent changes that may represent basic elements of memory. Of foremost interest are long-term potentiation (LTP) and depression (LTD), especially forms that depend on N-methyl-D-aspartate (NMDA) receptor activation and therefore can attain "associative" properties [ 1 - 3 ]. The selective induction of LTP versus LTD has been attributed to differing amounts of Ca 2+ ions entering via postsynaptic NMDA receptor channels [ 4 ]. Depending on type of stimulation, enzymes with different sensitivities to Ca 2+ may be engaged and change the balance between kinase and phosphatase activities, leading to either phosphorylation or dephosphorylation of postsynaptic target proteins, such as ionotropic receptors [ 2 ]. It has been shown that afferent stimulation by frequencies in the range 0.5 to 5 Hz reliably produces LTD whereas higher frequencies, 50–100 Hz, lead to LTP [ 5 ]. Several studies suggest that temporal factors are also important, implying that LTD requires a longer time to be induced than LTP [ 6 ]. We have previously demonstrated that under conditions of facilitated activation of NMDA receptors by low extracellular Mg 2+ synaptic plasticity can be induced by frequencies as low as 0.1–0.2 Hz when applied for prolonged periods of time [ 7 ]. Following an initial phase of transient potentiation there was a substantial depression that developed gradually during several hours and that remained stable after termination of NMDA receptor activation. Although the relation to "standard LTD" was not fully clarified, such slowly developing depression in low Mg 2+ solution may provide a useful model for studying certain forms of NMDA receptor dependent depression. In the present study, we will further develop the concept of gradually decaying responses. One critical issue regarding LTP, LTD as well as other forms of glutamatergic synaptic plasticity, is the relative contribution of different glutamate receptor subtypes in creating the synaptic modification. Knowledge about this matter may be helpful in elucidating the underlying modification. While a selective change of alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptors has been cherished [ 8 - 10 ], especially in the case of LTP, several studies also observed NMDA receptor mediated changes in both LTP and LTD [ 11 - 14 ]. Previous work on LTD in our lab described an equal change of AMPA and NMDA responses [ 15 ]. However, it was reported by others that the relative contributions of AMPA and NMDA responses during LTD depend on experimental conditions, an equal change being one possible outcome [ 12 ]. In our recent examination of a slowly developing depression using composite AMPA-NMDA excitatory postsynaptic potentials (EPSPs) [ 7 ], the two responses declined in close parallel, indicating a common factor. Such an equal change is compatible with both a coordinated change of receptors and a presynaptic one via a decrease of glutamate release. However, in view of other studies reporting a coupling between responses via AMPA and NMDA receptors [ 16 , 17 ], one may ask whether our observation of declining NMDA responses could be secondary to the change of AMPA. In the present study, isolated NMDA receptor mediated EPSPs were shown to decline progressively during prolonged low frequency activation (0.1 Hz). Moreover, following sudden start of stimulation there was an initial, transient potentiation. Our findings also resolved some questions regarding input specificity and durability of the slow decay, which were previously addressed only for AMPA EPSPs. Results Isolated NMDA EPSPs show a progressive decay AMPA EPSPs were initially recorded in low Mg 2+ solution in the presence of NMDA receptor antagonist AP5 to allow for pathway equalization (see Methods) without evoking NMDA EPSPs. Synaptic transmission was then entirely blocked by adding AMPA receptor antagonist CNQX, followed by unblocking of NMDA receptors by wash out of AP5. During this time, only one pathway was stimulated, keeping the other one silent for later use. As illustrated in Fig. 1A (upper part), an NMDA receptor mediated EPSP appeared within 10 min and reached maximum about 30 min after switching to AP5-free solution. During the following recording period of nearly 2 h, the NMDA EPSP decayed substantially, on average down to 58 ± 6 % of peak value (n = 8). Control experiments showed that isolated AMPA EPSPs recorded in low Mg 2+ remained stable for several hours (changing to 96 ± 5 % of baseline after 2 h, n = 5, not illustrated). Reinduction in a naive pathway To exclude the possibility that the observed decay of NMDA responses were due to deterioration of slices, implying a general decrease of essential physiological processes, the experiment was repeated for the other pathway, i.e. the one that had not previously expressed NMDA responses (continuation in the same set of 8 slices). As seen in Fig. 1A (lower part), a similar result was obtained as above. The NMDA EPSP peaked at 98 ± 7 % and declined to 63 ± 8 % relative to the peak in the first experimental session. As illustrated in Fig. 1B , the curve obtained for a naive pathway during the second session of NMDA unblocking was similar to the one obtained for the pathway activated during the first session, the two curves overlapping closely for the entire recording period. Comparison between pathways Specificity and persistency The same experiment was used to address the question of persistency as well as input specificity of the slowly developing depression of NMDA EPSPs. As can be seen in Fig. 1A , the pathway that was active during the first experimental session was retested during the second one together with the pathway receiving novel NMDA receptor activation. This allowed for a comparison between the pathways (for convenience, the peak in the first session is still used as reference for the values in the following). During the second session of NMDA receptor unblocking, the previously treated pathway displayed a substantially smaller peak than the naive one (48 ± 3 % vs. 98 ± 7 % p < 0.05, n = 8), and the two curves were still different by the end of the session (44 ± 3 % vs. 63 ± 8 %; difference 19 ± 4 %, p < 0.05). Since the latter time point was located at 3 h after the end of the first session it is evident that the depression of NMDA EPSPs lasted for at least 3 h. It is noteworthy that the previously depressed pathway showed no significant decay during the second session, passing from 48 to 44 %, i.e. a relative change by 92 % (p > 0.05), as if it had already been saturated. For a comparison, the naive pathway changed by a factor 63/98 = 64 % (p < 0.05) (see curves in Fig. 1C ). A graphic summary of all "peak" and "end-of-session" values is given in Fig. 3B . It can also be noted that NMDA EPSPs recorded for contiguous time intervals up to 4 h reached a saturation level after 2–3 h (n = 3, not illustrated). Instantaneous versus persistent depression The above results suggest that the progressive decay observed in a single pathway as an instant event actually represents a long-term, pathway-specific change that can be assessed long later by comparison across pathways. To further examine the relation between the instantaneously recorded depression and the one measured about 90 min later, the relation between the two was plotted as illustrated in Fig. 3C . The two variables were found to be positively correlated (r = 0.71, p < 0.05, n = 8), implying that depression to a lower level in one pathway led to smaller responses in that pathway at later times compared to another pathway. The regression line, passing below rather than through the point of no depression (100 %, 100 %), indicates that even a slight instantaneous decay may be coupled to a noticeable change in the long term. Possibly, the declining trend was partially masked by recovery from AP5 leading to an underestimation of it. Effect of NMDA receptor blockade on subsequent NMDA EPSP decay To pinpoint the induction mechanism, in terms of a pre- versus postsynaptic location, experiments were carried out in a similar way as above except for keeping AP5 in the solution during the first session (see Fig. 2A ). Hence, the stimulus pattern included a 3 h long interval with no stimulation in one of the pathways. The other pathway was stimulated during that time, and most likely releasing glutamate, but no postsynaptic response was expressed due to blockade of NMDA receptors. It can be argued that successful blockade of depression would predict a postsynaptic mechanism whereas a failure to do so would predict a presynaptic one. Fig. 2B shows that the depression was actually blocked, demonstrating the importance of NMDA receptor activation in the induction process. As illustrated, the two curves obtained during unblocking of NMDA receptors in the second session were quite similar. The continuously stimulated pathway, being depressed in the standard case, peaked at a level of 105 ± 5 % relative to the control pathway (n = 5; see also graphic summary of values in Fig. 3B ). Phase trajectories as indicators of waveform change Depending on the type of synaptic modification, EPSP waveforms may change in different ways, and previous work in our lab has demonstrated that NMDA EPSPs are more prone than AMPA EPSPs to show these changes [ 18 ]. Taking advantage of such waveform analysis, which might shed light on the underlying mechanism, we examined phase plots based on measures of the initial part and the later part of the NMDA EPSP (see Methods) on the X- and Y-axis, respectively. The curve in Fig. 3D is based on a total of 20 experiments (pathways), including some with only a single session. As shown, the phase trajectory displayed a loop indicating a difference between the effect of AP5 and the gradual decay of NMDA EPSPs. On the average the encircled area was 15 ± 2 % (p < 0.05, n = 20; scaling peak × peak as 100 %, clockwise being positive). While our data show that the time window matters for measuring NMDA EPSPs, all the above results were qualitatively similar regardless of which window was used. It can be noted, however, that the depression of NMDA EPSPs by the end of a recording session was less pronounced using early measurements than late ones (responses attaining 72 ± 3 % vs. 61 ± 3 % of the peak value, p < 0.05, n = 20). In 12 out of the 20 experiments, the fiber volley was well separated from the stimulus artifact allowing it to be properly measured. No significant change was detected, the end-of session value amounting 103 ± 2 % of the value at the EPSP peak (p > 0.05, not illustrated). Short-term effects induced by sudden onset stimulation In the above, NMDA receptor activation occurred gradually while the antagonist AP5 was washed out. This is in line with the experimental protocol used in our previous work on composite EPSPs containing both AMPA and NMDA components [ 7 ]. However, a natural question is whether sudden, novel activation of NMDA receptors is equivalent in producing the results observed here. We therefore pursued experiments in which stimulation was silenced until washout of AP5 was complete. One pathway, receiving such sudden stimulation, was compared to a control pathway subjected to gradual NMDA receptor activation during a single recording session. Fig. 4A shows an essential difference in behavior between the pathways, the sudden start of activation leading to substantially larger responses for about 5 min. Fig. 4B reveals additional complexity, the initial responses showing actual growth of responses for about a minute before they started to decay, implying an early potentiation process. The total range of responses was substantial, from a peak above 200 % to about 70 % by the end of the recording session (relative to the peak of the control pathway), i.e. about 3 times. In order to determine whether the transient potentiation had any obvious relation to the slow depression of NMDA EPSPs, the relation between the two was examined. Thus, the degree of initial potentiation was calculated by comparing the pathways just after stimulation was started and the depression was determined, as before, by comparing the end-of-session value with the peak value (see legend of Fig. 4 for further details). The two variables, illustrated by the XY-plot in Fig. 4C , were found to have no significant correlation (r = 0.35, p > 0.05). Discussion Our study revealed a progressive decline of pharmacologically isolated NMDA EPSPs, as observed for several hours in response to low rate (0.1 Hz) activation of afferents. The decline was found to be a form of long-term synaptic depression with an induction linked to NMDA receptor activation and with an expression that was maintained through periods without such activation. Several of its basic characteristics were similar to those of conventional LTP and LTD, suggesting a possible relation to these phenomena. Synapse specificity and NMDA-dependent induction Decaying responses is a potential side effect in long-term, electric recording in vitro due to declining viability of biological tissue or other experimental imperfections. Such unspecific "run down" can not account for the present findings since the gradual depression of responses could be repeated in the same slice, using a previously undepressed pathway. On the other hand, if the experiment was repeated twice in the same pathway, the second occasion revealed a diminished NMDA EPSP that showed little further decay. Together, these results show that the depression is input specific and long lasting and that it can saturate. Moreover, the lack of associated changes of the fiber volley speaks against a failure of axon conductance [ 19 ], favoring a synaptic localization of the process. While both pre- and postsynaptic expression mechanisms appear feasible, certain mechanisms of induction can be excluded. For instance, a decrease in probability of glutamate release due to a direct depletion of the vesicle pool is unlikely since AMPA EPSPs could be evoked for several hours without significant decay (see also [ 7 , 20 ]). Even so, a use-dependent reduction of vesicle content may affect NMDA responses selectively under certain conditions by restricting "glutamate spillover" [ 21 ]. The most critical data with respect to the induction mechanism is that a period of conditioning stimulation, normally leading to reduction of NMDA EPSPs in the same pathway later on, was ineffective if delivered during blockade of NMDA receptors. This implies that the induction of the depression requires activation of NMDA receptors, most likely postsynaptically. Other observations of decaying NMDA EPSPs The input specificity and NMDA dependent induction of the current depression conform with basic properties of conventional LTP and LTD [ 22 ]. The depression might then be a case of LTD, although induced by an alternative protocol. In fact, LTD was shown to be associated with changes involving both AMPA and NMDA receptors, although the linkage between the two contributions is controversial [ 12 , 15 ]. Moreover, both of the cited studies demonstrated LTD of isolated NMDA EPSPs induced by 1–2 Hz stimulation. In contrast, experiments in cultures, inducing LTD by field stimulation at a higher frequency (5 Hz), reported only AMPA receptor mediated changes [ 10 ]. Direct interaction tests may further clarify the relation between the present depression and LTD. Gradually decaying, NMDA receptor mediated responses were observed previously in our lab during recording of composite AMPA-NMDA EPSPs for several hours [ 7 , 23 ]. Attempts to relate the decay to LTD demonstrated a weak reduction of subsequent LTD of AMPA responses suggesting at least some elements in common [ 7 ]. In view of studies reporting forms of AMPA-NMDA coupling [ 16 , 17 ], it is arguable that the studies demonstrating a decay of both components could have been influenced by the use of composite responses. In one of our studies [ 7 ], the observed depression of the AMPA component of composite EPSPs was verified by additional comparison between isolated AMPA EPSPs obtained under blockade of NMDA receptors. A similar verification was lacking for the depression of the NMDA response. By recording of isolated NMDA EPSPs, the present study ascertains that NMDA receptor mediated responses undergo a use-dependent depression, which is manifested in the absence of AMPA receptor activation. However, the decay was less pronounced than that reported previously for the NMDA component of composite EPSPs (average reduction to 60 % of peak as compared to 40 % in the previous study [ 7 ]). While we observed that isolated NMDA EPSPs decay "spontaneously", most prior studies employing such EPSPs did not report a decay. It might be that limitations of recording time concealed the effect and cell dialysis during whole cell recording could also be a limiting factor. Actually, a recent study, recording "novel" responses under whole cell conditions, reported on decaying AMPA EPSPs but constant NMDA EPSPs [ 24 ]. The possibility of AMPA receptor LTD under the present conditions could not be excluded as the blockade of the receptors may just conceal the effect. Further studies may help to reveal this matter. Persistency and saturability Standard LTP/LTD experiments compare relatively stable periods of recording before and after induction of the synaptic modification. This was not possible in the present case, since merely test stimulation evoked the decay. Therefore, comparisons were generally made between synaptic pathways subjected to different stimulus paradigms. The induction of depression in a single pathway during an initial 2 h period caused a subsequent difference between NMDA responses of the two pathways throughout a subsequent test period. The degree of initial decay was closely related to the later difference between pathways, suggesting that once depression occurred it could be maintained through periods of receptor blockade until testing was performed. Our data suggest a duration of the depression of more than 3 h after the initial induction period. This is in the range commonly referred to as "late", which is believed to involve special biochemistry such as gene expression and protein synthesis [ 25 , 26 ]. Whether, the presently studied depression involves such changes remains to be determined. The gradual depression of NMDA EPSPs was found to saturate after 2–3 h as evidenced by both single and double session experiments. This is in line with several other forms of NMDA-dependent plasticity, including LTP, LTD and chemically induced variants, which are shown to be saturable [ 27 - 29 ]. Whether, the saturation observed here is a "true one" at the level of expression is not known. Alternatively it could be a phenomenon at the induction level, related to weaker induction due to the diminished NMDA response. Possible expression mechanisms Previous work on conventionally induced LTD revealed an essential role for protein phosphatases in mediating the synaptic modification [ 30 , 31 ]. Consistent with the idea that changes of AMPA receptors mediate NMDA-dependent synaptic plasticity [ 32 , 33 ] it was demonstrated that certain sites of the GluR1 subunit were targeted in LTP/de-potentiation and other ones in LTD/de-depression [ 2 , 34 ]. Less is known about mechanisms underlying NMDA receptor changes in LTP/LTD as well as in the current depression. A previous study in our lab recording composite EPSPs reported that LTD of the NMDA component was blocked by a phosphatase inhibitor in a similar manner as "standard LTD" [ 15 ]. Hence, one can envisage that NMDA receptors would be controlled via dephosphorylation in a similar manner as inferred for AMPA receptors. NMDA receptors also have a number of other regulatory sites, allowing for modulation by glycine, polyamines, calcium, and redox agents [ 35 ] and they have shown to be mobile as well [ 36 - 38 ], in keeping with the idea of mobile AMPA receptors in LTP/LTD [ 39 , 40 ]. Regardless of details, additional factors are needed to stabilize the synaptic modification in the long term, perhaps via synthesis of new proteins as previously demonstrated for LTP and LTD lasting longer than about 3 h [ 25 , 28 , 41 ]. Changes in synaptic morphology and altered subunit composition of receptors are examples of protein synthesis dependent mechanisms that have been implied in late forms of plasticity [ 32 , 42 ]. Although a postsynaptic modification appears to be the primary choice, a presynaptic one that is initiated postsynaptically is also conceivable. In previous attempts to distinguish between pre- and postsynaptic mechanisms, LTD was compared with depression caused by various pharmacological agents with respect to the ability to influence the waveform of EPSPs [ 18 ]. While LTD in that study was found to affect isolated NMDA EPSPs in a uniform manner, i.e. no waveform change, the present data appeared to be less clear-cut. Nevertheless, the relation between early and late EPSP measurements differed for the initial AP5 washout period and the following period of actively induced depression, indicating a change in EPSP waveform. The depression therefore appeared to be distinct from a postsynaptic modification via modulation of channel gating. However, a clear test of the pre-post issue still remains. Unfortunately, the MK-801 test of release probability [ 43 ] does not appear useful when dealing with decaying responses as in the present case. Short-term changes and their possible mechanisms While the main line of experiments employed a smooth start of NMDA receptor activation following the gradual washout of AP5, another set of experiments made use of sudden activation by awaiting full washout until stimulation was started. Compared to smooth activation, there was an additional, transient potentiation that largely decayed within 20–30 stimuli. This is in accord with a previous study in hippocampal slices showing that stopping stimulation of composite AMPA-NMDA EPSPs for 10–60 min (and one case of isolated NMDA EPSP for 10 min) resulted in a transient potentiation when stimulation was resumed [ 23 ]. Several other studies describe decaying NMDA responses in relation to inactivation or desensitization of receptors [ 44 , 45 ]. Accordingly, synaptically evoked NMDA responses in cell cultures were found to inactivate (i.e. decay) within a few minutes in much the same manner as observed here [ 45 ], a process shown to be triggered by postsynaptic influx of Ca 2+ via the NMDA channels. Similar mechanisms of receptor desensitization/inactivation might be responsible in the present case in forming the transient phase after starting stimulation. Some details remain unexplained by this simple model, such as the biphasic character of the transient phase in terms of initial growth and subsequent decay. One can speculate that a minor LTP, or short-term potentiation, might be induced by the sudden activation of NMDA receptors and so would contribute to the initial growth, although the underlying cause is not addressed in this kind of explanation. Conclusions The above results emphasize that NMDA receptor mediated responses are highly plastic and that mere test stimulation can induce a short-term potentiation as well as a slowly developing depression that persists for several hours. The depression was input specific and saturable, and its induction required NMDA but not AMPA receptor activation in conformity with conventionally induced LTP and LTD, suggesting a relation to these phenomena. While a low Mg 2+ solution was used in our case to unblock NMDA receptors, similar unblocking may occur naturally in response to depolarization. Several important issues are still not settled. Is the saturation of the NMDA EPSP depression an absolute matter or can it be overcome, leading to further down regulation and possibly silencing of synapses? Conversely, is it possible to reverse, i.e. de-depress, the change by LTP or similar processes, allowing for bidirectional control? Further research is needed to resolve these questions. Methods Experiments were performed on 12 to 18 day old Sprague-Dawley rats. The animals were decapitated after isoflurane (Forene) anesthesia in accordance with the guidelines of the Swedish Council for Laboratory Animals. All animal procedures were approved by the Local Ethics Committee at Göteborg University. The brain was removed and placed in an ice-cold artificial cerebrospinal fluid solution containing (in mM) NaCl 119, KCl 2.5, CaCl 2 2, MgCl 2 2, NaHCO 3 26, NaH 2 PO 4 1 and glucose 10, oxygenated by 95% O 2 , 5% CO 2 . The hippocampus was dissected out and transverse 400 μm thick slices were prepared by a vibratome or tissue chopper. The slices were initially kept in the same solution at room temperature for at least 60 min. As required, slices were then transferred to one or several "submerged type" recording chambers. During the experiment, slices were perfused at 30°C by a solution similar to that above except that the concentration of Mg 2+ was 0.1 mM. The usage of low Mg 2+ allowed for expression of NMDA receptor mediated responses. Stimulation was delivered as 0.1 ms negative constant current pulses via monopolar tungsten electrodes. For each slice, two stimulating electrodes were placed in the apical dendritic layer of CA1 pyramidal cells on either side of the recording electrode to provide for stimulation of two separate sets of afferents. Field EPSPs were recorded by using a glass micropipette filled with 3 M NaCl (4–10 MΩ resistance). The basal test stimulus frequency was 0.1 Hz with stimuli delivered alternately to the two electrodes, successive stimuli being separated by 5 s. To test the effect of stimulus interruption, one of the two electrodes was given no stimulation during a certain time, the other one remaining stimulated at 0.1 Hz. Recording commenced by monitoring isolated AMPA EPSPs in the presence of AP5 (50 μM) to block NMDA responses. A low concentration of CNQX (1 μM) was used to partially suppress the AMPA responses. In this way, somewhat larger stimulus strengths could be applied, suitable for evoking isolated NMDA EPSPs in the later part of the experiment. During the time of AMPA EPSP recording, the stimulus strengths were adjusted for each slice to equalize the synaptic inputs of the two pathways. This was essential for later comparison of NMDA EPSP across pathways. After obtaining a baseline of equal AMPA responses, the concentration of CNQX was raised to 10 μM which entirely blocked synaptic responces. The remaining non-synaptic response, consisting of stimulus artifact and presynaptic volley, was used to define "true zero". To study NMDA receptor mediated responses, CNQX (10 μM) was maintained in the solution while AP5 was washed out for one or several 2 h periods, referred to as sessions in the following. In between the sessions as well as afterwards, synaptic transmission was again blocked by applying AP5 (50 μM), framing in the sessions by periods of recording non-synaptic responses. Under the sessions, various tests were made depending on the purpose of investigation. Usually one input remained silent during the first session and stimulation was not resumed until after synaptic transmission was reblocked. In another kind of experiment, the initially silent pathway was reactivated in the early part of the first session after NMDA receptors were unblocked, providing a means for sudden start of NMDA receptor activation. Signals were amplified, filtered and transferred to a PC clone computer for on-line and off-line analysis by specially designed electronic equipment (based on an Eagle Instruments multifunction board) and own developed computer software. AMPA EPSPs were measured using an early time window (first 1.5 ms after the fiber volley) while NMDA EPSPs were measured using both an early (first 5 ms after volley) and a late (35–45 ms after artifact) time window. The late measurement was used in presenting most of the results, allowing easy comparison with previous work in our lab that estimated the NMDA component of composite EPSPs via a late measurement [ 7 ]. Similar albeit not identical results were obtained with early and late measurements (see illustration in Fig. 3D ). Measurements were calculated by integrating the curve along the specified time window after substraction of the prestimulus baseline. All values were corrected by substracting the corresponding measurements of the non-synaptic potential obtained after total blockage of the EPSPs (except when measuring the fiber volley). The final data were quantified as relative values compared to a reference level defining 100 %. While the initial baseline formed a natural reference for AMPA responses, the choice was less obvious for NMDA responses, leading us to use the highest level of responses for one of the pathways in one of the experimental sessions (selected to make sense). Results are expressed as mean ± S.E.M. Statistical comparisons were made using Student's t-test. Drugs were obtained from Tocris Cookson, UK; prefabricated stimulating electrodes were obtained from World Precision Instruments, FL USA, type TM33B. Abbreviations AMPA, alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid; AP5, D(-)-2-amino-5-phosphonopentanoic acid; CA, cornu ammonis; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; EPSP, excitatory postsynaptic potential; LTD, long-term depression; LTP, long-term potentiation; NMDA, N-methyl-D-aspartate; Authors' contributions MD planned and carried out most of the experiments including data analysis, and compiled the manuscript. RL carried out experiments, participated in the planning process and helped in shaping the final manuscript. HPX carried out the initial experiments establishing the effect of NMDA EPSP depression. BJ was responsible for logistics planning and participated in experiments. HW conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC517399.xml

514492

New Route to Longer Life

Ever since the early Greeks recast humans as the center of the universe and remade God in their own image, Western philosophers and poets have grappled with the limits of human mortality. Philosophers found relief from Keats's “unwilling sleep” by dividing human existence into body and soul and asserting that the true essence of humanity lies in the immortal soul, not in the body. Ironically, as this decidedly nonscientific subject has lost favor with modern-day philosophers, it has captured the imagination of scientists. But, for now at least, the interest is in prolonging life rather than escaping mortality. Over the past twenty years, mounting evidence from a wide range of organisms indicates that a longer life awaits those who eat less. In yeast, calories can be restricted directly, by limiting yeast's glucose supply, or indirectly, by inhibiting yeast's ability to metabolize glucose. Either way, many studies have suggested that the increased longevity associated with calorie restriction is linked to increased activity of a gene called SIR2 . Now, Brian Kennedy and colleagues show that calorie restriction and SIR2 promote longevity through distinct genetic pathways—and that aging in yeast and higher organisms may be more similar than previously thought. One of the causes of aging in yeast is the accumulation of coiled bits of DNA, called extrachromosomal ribosomal DNA circles (ERCs), in the nucleus of a mother cell (which divides to create two identical daughter cells). An overabundance of these rDNA circles wreaks havoc on a cell and eventually kills it. Genetic mutations that reduce their levels are linked to increased life span. Mutations that disrupt the FOB1 gene, for example, dramatically reduce ERC levels and increase the reproductive life span of cells by 30%–40%. In contrast, mutations that disrupt SIR2 increase ERC levels and cut life span in half, while increasing SIR2 activity increases life span by 30%–40%. In previous experiments, several groups have identified a link between calorie restriction, SIR2 , and the accumulation of ERCs. The idea is that calorie restriction somehow activates the protein encoded by SIR2 , which in turn decreases ERC accumulation. Now, Kennedy's team has found that the combination of calorie restriction and FOB1 mutation increases life span more than either approach does alone. This finding was unexpected because previous studies showed that combining increased SIR2 activity with FOB1 deletion mutations did not extend life span. If calorie restriction extends life through SIR2 , then combining either caloric restriction or SIR2 overexpression with FOB1 mutations should produce the same result. This contradiction raised the possibility that calorie restriction operates through another mechanism, independent of SIR2 . In support of this view, caloric restriction enhances life span to a greater extent in FOB1 mutants lacking SIR2 than in FOB1 mutants with an intact SIR2 gene. This and other genetic experiments indicate that calorie restriction does not always work through SIR2 . That suggests, the authors explain, that calorie restriction functions either by regulating ERC levels or by some still unknown molecular pathway. They conclude that the enhanced longevity seen in calorie-restricted FOB1 mutants is not related to ERCs, because these yeast strains already have low ERC levels. Since calorie restriction is the only demonstrated approach to increasing life span in a diverse range of organisms, including mammals, and since there's no evidence that ERCs affect the aging of any organism besides yeast, these results bode well for understanding how calorie restriction works in higher organisms. And the finding that calorie restriction and SIR2 operate through genetically distinct pathways in yeast, the authors conclude, suggests that certain aspects of both pathways might have been conserved through evolution. Working out the details of these pathways in yeast is the first step toward understanding which, if any, of these components might enhance longevity in humans. Of course, as any student of Greek mythology knows, longevity without eternal youth comes with a price.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514492.xml

517831

How to Make a Mother in Five Easy Steps

Assembling a complex structure like an automobile requires the tight coordination of hundreds of independent entities—parts must be shipped and arrive on time, workers with the right skills must be in the right place on the assembly line, four (not three, not five) wheels must be bolted into place just so. Overseeing the entire operation is a cadre of managers, whose ability to monitor and respond to changing conditions keeps the entire process moving forward on time and in step. A pair of Bacillus subtilis sporangia , consisting of a large mother cell (green) and a forespore (red) A living cell is orders of magnitude more complex, and yet has no omniscient manager at the helm. So how does a cell make anything happen on time, and equally important, keep everything from happening all at once? These central questions in developmental biology now have the outlines of an answer in one species, Bacillus subtilis . In this issue, Richard Losick and colleagues show that spore formation in this bacterium is ultimately governed by the temporal interactions of five genes, which together coordinate the activity of almost 400 others. When conditions are right, B. subtilis divides to form two different cell types: one is a resistant spore, and the other is a mother cell, which engulfs the spore and surrounds it with a protective coat. Building on the large literature addressing the genetic events underlying mother-cell development, Losick et al. performed a variety of experiments to determine exactly which genes turned on and off when, and which genes controlled which others. Over the five-hour process of mother-cell development, they determined that 383 individual genes were activated, representing 9% of the bacterium's 4,106 genes. The instigator of the entire process is a protein called sigmaE (σ E ). Sigma factors, such as the sigma-E protein, bind to RNA polymerase, and in so doing, increase its affinity for, and therefore its ability to activate, other genes. Thus, sigma factors preferentially activate a specific set of genes. Sigma-E turns on 262 genes (which together compose its “regulon”), kick-starting a variety of processes in the early development of the mother cell. Importantly, two of its targets, SpoIIID and GerR, are genes for DNA-binding proteins, which themselves modulate the expression of genes in the middle phase of development. Part of SpoIIID's portfolio is turning off transcription of a portion of the sigma-E regulon, and amplifying transcription of another portion. This type of control circuit, in which A leads to B, and then both A and B influence C, is called a feed-forward loop. Among the joint targets for sigma-E and SpoIIID is another sigma factor, sigma-K (σ K ). By generating the DNA-binding protein GerE, this factor begins a second feed-forward loop, and together, sigma-K and GerE activate the final set of genes needed for mother-cell development. In outline, the system looks like this: sigma-E → SpoIIID/GerR → sigma-K → GerE. The consequence of all this activity is a series of transcriptional pulses, timed to supply proteins just as they are needed, and then turn off their production when the need passes. For instance, to form the multilayered coat around the spore, sigma-E turns on genes that form the bottom layer, or substratum; these are turned off by SpoIIID. Genes for outer layers, also turned on by sigma-E, are not turned off by SpoIIID, but instead by GerE. Sigma-K turns on genes which form the polysaccharide surface of the coat, which is needed later on. The elucidation of this complex pattern of gene expression doesn't by any means answer every question about B. subtilis development, let alone development in more complex organisms. There is much still to be learned about how genes lower down in the hierarchy—the “middle managers”—do their jobs, and how the system is fine-tuned by environmental conditions. And while the general scheme of feed-forward loops and hierarchical control is likely to apply to multicellular, eukaryotic organisms, the details are certain to be different, and much more complex.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC517831.xml

539266

TGF β1 and PDGF AA override Collagen type I inhibition of proliferation in human liver connective tissue cells

Background A marked expansion of the connective tissue population and an abnormal deposition of extracellular matrix proteins are hallmarks of chronic and acute injuries to liver tissue. Liver connective tissue cells, also called stellate cells, derived from fibrotic liver have been thoroughly characterized and correspond phenotypically to myofibroblasts. They are thought to derive from fat-storing Ito cells in the perisinusoidal space and acquire a contractile phenotype when activated by tissue injury. In the last few years it has become evident that several peptide growth factors such as PDGF AA and TGF-β are involved in the development of fibrosis by modulating myofibroblast proliferation and collagen secretion. The fact that during the development of chronic fibrosis there is concomitant deposition of collagen, a known inhibitory factor, and sustained cell proliferation, raises the possibility that stellate cells from chronic liver fibrosis patients fail to respond to normal physiologic controls. Methods In this study we address whether cells from fibrotic liver patients respond to normal controls of proliferation. We compared cell proliferation of primary human liver connective tissue cells (LCTC) from patients with liver fibrosis and skin fibroblasts (SF) in the presence of collagens type I and IV; TGF-β, PDGF AA and combinations of collagen type I and TGF-β or PDGF AA. Results Our results indicate that despite displaying normal contact and collagen-induced inhibition of proliferation LCTC respond more vigorously to lower concentrations of PDGF AA. In addition, we show that collagen type I synergizes with growth factors to promote mitogenesis of LCTC but not SF. Conclusions The synergistic interaction of growth factors and extracellular matrix proteins may underlie the development of chronic liver fibrosis.

Background In normal liver, connective tissue cells are rare and mostly restricted to the periportal and pericentrovenular spaces and to the Glisson's capsule. A minor population of connective tissue cells is present inside the hepatic lobule, in the perisinusoidal space. They are known as hepatic stellate cells (HSC) and are considered to be one of the major contributors for fibrogenesis in liver [ 1 ]. Chronic and acute injuries to liver tissue induce a marked expansion of the connective tissue cells and concomitantly an abnormal deposition of extracellular matrix proteins [ 2 ]. Extensive studies of experimental models and humans have shown that these cells are of the myofibroblast lineage, characterized by the expression of smooth muscle α-actin [ 3 - 5 ]. In fact, there is now increasing evidence that are several populations of myofibroblasts in the diseased liver in addition to those derived from HSC [ 6 , 7 ]. The origin of these cells is still debated. In experimental models, it is considered that an acute liver injury is followed by activation and increase of stellate cells, while chronic injuries elicit activation of similar cells that can be either of lobular or periportal origin [ 7 ]. In humans, extensive primary periportal fibrosis such as schistosomal fibrosis was related to expansion of myofibroblasts originating from the portal vein wall [ 8 ], while in cirrhosis resident periportal connective tissue cells, as well as lobular stellate cells and pericentrovenular cells were reported to convert into myofibroblasts [ 3 , 9 , 10 ]. Independently of their origin, connective tissue cells of human fibrotic and cirrhotic liver tissues were shown to have homogeneous morphologic characteristics, as well as a specific growth patterns [ 11 - 15 ]. Recent studies on the mechanisms controlling connective tissue cell expansion and abnormal collagen deposition during liver fibrosis have shown that several peptide growth factors are involved in the development of fibrosis. Evidence from in vivo and in vitro studies indicated that TGF-β has a relevant role in the development of fibrosis, being a potent regulator of proliferation and of extracellular matrix protein synthesis [ 16 - 18 ]. TGF-β is highly expressed in fibrotic regions and overexpression of TGF-β in animal models results in fibrosis [ 18 , 19 ]. Conversely, TGF-β antagonists can prevent experimentally induced fibrosis [ 20 - 22 ]. PDGF AA was shown to have a pivotal role in controls of normal and pathologic proliferation of two myofibroblastic cell lineages: mesangial cells in kidney and alveolar smooth muscle cells in lungs [ 23 , 24 ]. On the other hand, it has also been demonstrated in vitro that the extracellular matrix can in its turn modulate cellular responses to peptide growth factors [ 25 ]. In vitro , fibroblast proliferation can be regulated by the presence of collagen type I. It has been suggested that collagen induces a quiescent state by decreasing the responsiveness to different growth factors [ 26 , 27 ]. An increase in cell density is also known to inhibit fibroblast proliferation in vitro . During the development of chronic fibrosis there is concomitant deposition of collagen and sustained cell proliferation, raising the possibility that liver connective tissue cells (LCTC) from chronic liver fibrosis patients fail to respond to normal physiologic controls. In this study we addressed the question of whether cells from fibrotic liver patients respond to normal physiologic controls of proliferation. We compared the behavior of primary human LCTC from patients with liver fibrosis with skin fibroblasts (SF). We compared cell proliferation in the presence of collagen types I and IV, of TGF-β and PDGF AA, and in combinations of collagen type I and TGF-β or PDGF AA. Our results indicate that LCTC respond more vigorously to lower concentrations of peptide growth factors than SF. Interestingly, collagen type I synergizes with the growth factors tested in promoting mitogenesis of LCTC but not of SF. We believe that this mechanism may underlie the development of chronic liver fibrosis. Methods Cell lines Liver tissue fragments and skin biopsies were obtained through collaboration with the University Hospital, Department of Surgery, Federal University of Rio de Janeiro (O.M. Vieira, M.D.). An informed consent was obtained from every patient, and the study was conducted in agreement with the ethical guidelines of the Federal University of Rio de Janeiro. Diagnoses were given by the pathological anatomy service of the hospital. All surgical liver biopsies were obtained for diagnostic purposes during spleno-renal or porto-caval anastomosis. Skin fibroblasts were obtained from the mid-abdominal incision performed for the surgery. We studied six primary normal SF lines and ten primary LCTC lines derived from patients with schistosomal fibrosis or alcoholic cirrhosis. Biopsies were collected and brought to the laboratory in chilled Dulbecco's modified Eagle's medium (DMEM) (Sigma, St. Lois, MO) with 10% fresh human serum and 60 μg/ml of gentamicin (Schering, Rio de Janeiro, Brazil). They were cut into pieces of approximately 1 mm 3 , washed in fresh medium and plated, 6 or 9 pieces per 25 cm 2 flask. They were maintained in DMEM supplemented with 4 g/liter HEPES and 10% fresh human serum, at 37°C in a humid incubator with 5% CO 2 , 95% air. All the cultures used in this study were finite cell lines derived from the primary cultures described above, and were discarded before the 10th passage. LCTC have been described in detail previously, and compared to skin fibroblasts and vascular smooth muscle cells [ 12 , 13 , 15 ]. They were characterized by electron microscopy, immunofluorescence, morphology and proliferation in culture and ability to contract a collagen matrix. These established primary cell cultures are homogeneous in terms of production of smooth muscle α-actin, fibronectin, collagen I and III, and elements of the basement membrane such as collagen IV and laminin [ 12 , 13 , 15 ]. Hence, they can be phenotypically characterized as myofibroblasts and different from fibroblasts or smooth muscle cells. Serum and growth factors Fresh human citrated plasma was obtained from the hemotherapy service of the Hospital dos Servidores do Estado (Rio de Janeiro, Brazil). Plasma was coagulated with calcium and sera from several patients were pooled. Human TGF-β1 and human recombinant PDGF-AA were purchased from Sigma Chemical Company, St Louis, MO. TGF-β1 was activated by incubation in bovine serum albumin 1 mg/ml and 4 mM HCl. Preparation of collagen coated dishes One hundred μl of solutions with varying concentrations of collagen type I (rat tail tendon collagen, prepared in our laboratory as previously described [ 13 ]) or collagen type IV (Sigma) were dispensed onto a 12-well or a 96-well plate and dried under a laminar flow for 24 hr at room temperature. This procedure allowed us to control for the exact amount of collagen added to each plate. The plates used were freshly prepared and were washed three times in serum free medium. Proliferation Cell proliferation was assayed by cell counting and [H 3 ]-thymidine incorporation. These methods correlated well with autoradiography results in previous studies [ 15 ]. For the experiments to assay contact inhibition of proliferation, approximately 1 × 10 3 cells/cm 2 were plated. Each day, during a period of 8 days, cells were trypsinized and counted in a hemocytometer and medium was changed every two days. To assay the proliferative response induced by growth factors and extracellular matrix proteins cells were trypsinized and plated at a concentration of 5 × 10 3 and allowed to adhere for 2 hr in DMEM with 10% serum. Then, cells were starved in serum-free medium for 18 hr to synchronize the cell population. When testing for growth factors effects, serum-free medium was replaced by medium containing 1% serum, supplemented with the growth factors and 0.5 μCi/ml [H 3 ]-thymidine. When assessing the effect of extracellular matrix proteins, serum-free medium was replaced by medium containing 10% serum. Cells were collected 24 or 32 hr after the stimulus and lysed in 10 N NaOH. The trichloroacetic acid precipitable material was then spotted on a filter and counted on a scintillation counter. Kinetics of thymidine incorporation Cells were plated at a concentration of 2.5 × 10 3 per well (0.5 cm 2 ) in 96 well plates in DMEM with 10% serum. Cells were allowed to adhere and spread for two hr. Medium was then replaced with serum-free medium for 24 hr. Subsequently, cells were incubated in different media containing 1% serum and growth factors. [H 3 ]-thymidine (1 μCi/ml) was added at 6, 24, 32 and 40 hr and incubated for 2 hr. Cells were harvested and processed as described above. Adhesion and recovery To determine cell adhesion on different substrata plates were prepared with 0.3 mg/ml collagen type I or IV. Cells were plated on each substratum and the supernatant was removed at 0, 10, 20, 40 and 60 min and the cells remaining in suspension counted in a hemocytometer. To determine cell recovery, cells plated after 2 hr were trypsinized and counted. Statistical analysis Difference between the means of various subgroups was assessed using the Mann-Whitney U-test. Results Chronic liver fibrosis is characterized by abnormal proliferation of connective tissue cells. Therefore, we hypothesized that connective tissue cells from fibrotic lesions have lost normal proliferation controls. By using a series of primary connective tissue cell lines from skin and from fibrotic livers, we investigated several parameters of cell growth in vitro in order to identify potential mechanisms to explain the excessive proliferation of LCTC during fibrosis. We observed no differences among the cell lines obtained from patients with schistosomal fibrosis or alcoholic cirrhosis, and all the studied primary LCTC lines were included in a single experimental series. We established and characterized 16 primary human cell lines. The experiments performed were done at or before the tenth passage and every experiment was performed with at least two cell lines from each type (LCTC or SF) isolated from different individuals. Contact inhibition of proliferation To assess whether connective tissue cells from fibrotic livers (LCTC) and normal skin fibroblasts (SF) responded to contact inhibition of proliferation, we measured cell proliferation during eight days. Both LCTC and SF showed a comparable pattern of growth and reached a plateau at a similar cell density (Figure 1A,1B ). Reproducibly, LCTC declined in number after reaching a maximal density (Figure 1B , day 8). To confirm these results, we measured [H 3 ]-thymidine incorporation in various cell densities (Figure 1C ). Both LCTC and SF showed maximum [H 3 ]-thymidine incorporation at the same cell density (1 × 10 4 /cm 2 ) and a marked decline at higher densities. We also examined the response of both cell types to increasing concentrations of fresh human serum. Both SF and LCTC reached maximum proliferation, as measured by [H 3 ]-thymidine incorporation at 20% serum (not shown). In conclusion, primary SF and LCTC were similarly responsive to normal contact inhibition of proliferation and serum concentration, although SF seemed to respond more efficiently to contact inhibition. Figure 1 Contact inhibition of proliferation. A. Proliferation of normal human skin fibroblasts (SF) in vitro. B. Proliferation of human liver connective tissue cells (LCTC). Cells were plated at 1 × 10 3 cells/cm 2 . Each day, during a period of 8 days, cells were trypsinized and counted in a hemocytometer ( n = 3). Medium containing 10% serum was replaced every two days. C. [H 3 ]-Thymidine incorporation in SF and LCTC at increasing cell density ( n = 4). The effects of extracellular matrix Despite the abnormal deposition of extracellular matrix in chronic fibrotic livers, LCTC continuously proliferate suggesting that these cells are not responsive to normal inhibition of proliferation by collagen I. To determine the influence of the extracellular matrix on LCTC, we analyzed DNA synthesis in the presence of collagen type I and type IV, two major components of extracellular matrix in fibrotic livers. Collagen type I inhibited DNA synthesis in a dose-dependent manner, as measured by [H 3 ]-thymidine incorporation in LCTC and SF (Figure 2A,2B ). In SF, [H 3 ]-thymidine incorporation at the highest collagen concentration was approximately 30% of the control, suggesting that this cell type is more susceptible to collagen-mediated inhibition of proliferation. However, the level of the inhibition was variable among the studied patients, and a less marked reduction, (~50%) was observed in SF derived from some of the studied cases (not shown). We did not observe any change in cell morphology dependent upon the substrate. Figure 2 Collagen-mediated inhibition of proliferation. A. Cell proliferation of LCTC plated on increasing collagen concentration as measured by [H 3 ]-Thymidine incorporation ( n = 4). Note inhibition of incorporation in a dose-dependent manner. B. Cell proliferation of SF plated on increasing collagen concentration as measured by [H 3 ]-Thymidine incorporation ( n = 4). Note inhibition of incorporation in a dose-dependent manner. (*) denotes significant difference from control; p ≤ 0.05 . C. Left panel. Kinetics of DNA synthesis of LCTC comparing growth on plastic (empty circles) and on collagen (filled circles) ( n = 4). Right panel, top. Adhesion of cells plated on plastic (filled circles), collagen IV (empty triangles) and collagen I (empty squares) ( n = 3). Right panel, bottom. Recovery of cells from plastic or collagen I ( n = 4). The question of whether the difference observed between collagen-coated and plastic plates resulted from an alteration in kinetics of cell cycle progression or to an actual block in DNA synthesis was addressed in the following experiment. After serum-mediated activation of quiescent cells, the cultures were pulsed with [H 3 ]-thymidine for 2 hr, at 6, 20, 26, 32 and 40 hr. [H 3 ]-thymidine incorporation reached its peak at 32 hours after activation in control experiments (plated on plastic), and cells plated on collagen-coated plates did not show any delayed peak, demonstrating that the effect of collagen was not delaying the progression of cell cycle but blocking DNA synthesis (Figure 2C ). These results indicate that LCTC are responsive to normal inhibition of proliferation mediated by collagen I. Next, we asked whether collagen type IV could also influence the proliferation of the studied cells. Even at high concentrations (90 μg/cm 2 ), collagen type IV did not induce any significant change on [H 3 ]-thymidine incorporation in two LCTC cell lines, neither in two SF cell lines (not shown). Again, no difference in morphology was noted. To rule out the possibility that the results were being masked by differential adhesion to collagen or differential cell recovery when measuring the [H 3 ]-thymidine incorporation, we monitored cell adhesion on plastic, on collagen type I and on collagen type IV. Although by 10 min there were a significantly less cells attached on collagen IV, by 40 min the same number of cells had attached to the plates in all substrates (Figure 2C , right panel). These results rule out the possibility that lower [H 3 ]-thymidine incorporation found in cells plated on collagen I was due to differential adhesion, since our experiments allowed 2 hr for attachment. We also counted cell numbers on the samples recovered for scintillation counting and found it to be comparable (Figure 2C , right panel), thus ruling out the possibility that the difference was due to differential recovery. Taken together these results indicate that LCTC are responsive to normal inhibition of proliferation mediated by collagen I in a manner similar to that of normal SF. Effects of PDGF AA and TGF-β on proliferation To investigate the effects of PDGF AA we starved cells for 18 hr and stimulated them with increasing concentrations of PDGF AA. As shown in Figure 3A , LCTC were more sensitive to lower concentrations of the growth factor (5, 10 and 20 ng/ml), reaching a peak at 10 ng/ml. Although reaching the same level of stimulation at 40 ng/ml, SF were less sensitive to PDGF AA at lower concentrations. PDGF AA did not affect timing of cell cycle progression, since cells in the presence and in the absence of PDGF AA displayed a peak of [H 3 ]-thymidine incorporation at 26 hr after release in medium containing serum and factors (results not shown). In conclusion, LCTC appear to be more sensitive to lower concentrations of PDGF AA. Figure 3 Mitogenic effects of PDGF AA and TGF-β on SF and LCTC. A. Effect of increasing PDGF AA concentrations on growth of LCTC (black bars) and SF (gray bars) ( n = 4). B. Effect of increasing concentrations of TGF-β on growth of LCTC (black bars) and SF (gray bars) ( n = 4). (*) denotes significant difference between SF and LCTC; p ≤ 0.05 . C. Kinetics of DNA synthesis in LCTC as measured by [H 3 ]-thymidine incorporation in the presence (filled circles) and absence (empty circles) of TGF-β ( n = 4). Note second delayed peak of incorporation at 40 hr (black arrow). D. Kinetics of DNA synthesis in SF as measured by [H 3 ]-thymidine incorporation in the presence (filled circles) and absence (empty circles) of TGF-β ( n = 4). Note the absence of the second delayed peak of incorporation at 40 hr (black arrow). To test the responsiveness of LCTC and SF to TGF-β, cells were starved overnight and stimulated with increasing concentrations of acid-activated TGF-β. In Figure 3B we note that at lower concentrations, stimulation of LCTC and SF is comparable. However, higher concentrations of TGF-β (0.1 ng/ml) seemed to be less effective in promoting mitogenesis of SF. Interestingly, at the highest concentration (1 ng/ml) TGF-β was inhibitory to SF but still highly stimulatory for LCTC. When we analyzed the kinetics of thymidine incorporation during TGF-β treatment (Figure 3C,3D ) we observed a second delayed peak of DNA synthesis at 40 hr (Figure 3C and 3D , arrow). In smooth muscle cells this second delayed peak has been shown to be due to a PDGF AA autocrine loop [ 28 ]. While in LCTC the second peak was 85% of the first peak, in SF it was slightly under 50%. These results indicate that TGF-β can induce a prolonged and strong induction of DNA synthesis in LCTC even at concentrations that are inhibitory for SF. In this set of experiments we were able to compare matched LCTC and SF from the same patient. Our results indicate that the observed differences between LCTC and SF are not due to individual variation or genetic background. Synergy of extracellular matrix and growth factors Next, we assessed the interplay between growth factors and the extracellular matrix on the stimulation of proliferation. We compared proliferation of cells cultured on collagen film using the highest inhibitory concentration (42 μg/cm 2 ; Figure 2 ) and cultured on plastic. Initially, we treated cells with 10 ng/ml of PDGF AA in the presence and absence of collagen after 24 hr (Figure 4A ). Both LCTC and SF were efficiently inhibited by collagen, in accordance with the experiments in Figure 2A and 2B , performed with different cell lines. In the absence of collagen, PDGF AA stimulated proliferation comparable to that obtained in Figure 3A and 3B . These conditions provide an internal control to distinguish effects that are cell type-specific from individual variations. Surprisingly, PDGF AA was able to override the collagen-mediated inhibition only in LCTC and not in SF. Figure 4 Interaction between peptide growth factors and extracellular matrix. A. Effects of PDGF AA on the growth of LCTC (black bars) and SF (gray bars) plated on collagen or plastic ( n = 4). B. Effects of TGF-β on the growth of LCTC (black bars) and SF (gray bars) plated on collagen or plastic ( n = 3). (*) denotes significant difference between SF and LCTC; p ≤ 0.05 . Next, cells were stimulated with TGF-β (Figure 4B ) and proliferation measured. Since there was a possibility that the autocrine loop could induce a late response, we assessed [H 3 ]-thymidine incorporation at 40 hr. (Figure 4B ). Collagen inhibited these cell lines as in previous experiments with LCTC and SF cell lines. TGF-β alone had little effect on proliferation. However, similarly to the experiment with PDGF AA, TGF-β was not only able to override the collagen-mediated inhibition but showed a synergistic effect only in LCTC. SF were not able to escape inhibition mediated by collagen. Discussion Injuries to liver tissue usually involve transient or long-term development of fibrosis. In acute injuries, fibrotic tissue is frequently reabsorbed and normal tissue architecture is restored. In cases when the primary agent or the secondary pathogenic mechanisms are persistent, the fibrotic reaction can be perpetuated causing a severe impairment of organ function [ 1 ]. It is therefore important to identify the factors that are involved in this perpetuation in order to devise more efficient preventive measures and therapies. However, there is a dearth of knowledge about the behavior of the human LCTC isolated from fibrotic livers. In fact, most studies done to date used LCTC cells from experimental models such as rats and mice and the few studies dealing with myofibroblasts from human liver derived the cells from normal tissue. In order to evaluate the hypothesis that LCTC from fibrotic livers had lost control of proliferation we investigated their growth in tissue culture. Contact inhibition of proliferation is a characteristic of normal cells and is lost in neoplastic cell lines. Our results showed that LCTC respond normally to contact inhibition of proliferation. Collagen type I, which is the predominant extracellular matrix protein deposited during hepatic fibrogenesis [ 2 ], has been shown to be a potent inhibitor of mesenchymal cell proliferation [ 26 , 29 , 30 ]. Since the perpetuated proliferation of LCTC is a hallmark of chronic liver diseases, it was conceivable that these cells had lost normal collagen-mediated inhibition. However, our results indicate that LCTC derived from fibrotic livers are responsive to collagen I to an extent comparable to normal skin fibroblasts. In addition, collagen IV a major component of basement membranes did not affect the growth of either cell type studied. While recent evidence from rat vascular smooth muscle cells indicates that PDGF AA promotes only protein synthesis without activation of DNA synthesis [ 31 ], we demonstrate that PDGF AA is active as a mitogen for human LCTC. These results are in accordance with previous data using connective tissue cells derived from normal liver [ 32 ]. Interestingly, the same authors have found that in fibrotic livers there is marked increase in the expression of both PDGFα receptor and its ligand PDGF AA [ 33 ]. Taken together, these results strengthen the notion that PDGF-AA is an important mediator of connective tissue expansion during liver fibrosis. Although generally agreed to contribute to the fibrogenic process by upregulating expression of genes encoding extracellular matrix proteins, a role for TGF-β in promoting mitogenesis has emerged in the last few years. TGF-β displayed a bi-phasic curve with low concentrations stimulating and higher doses inhibiting the proliferation of SF. On the other hand, TGF-β was a potent mitogenic stimulus for LCTC even at high concentrations. Interestingly, it has been shown that TGF-β also stimulates an additional delayed growth response mediated by an autocrine loop of PDGF AA [ 28 ], which we observed only in LCTC. One surprising outcome of this study was that not only the peptide growth factors tested were able to override the collagen-mediated inhibition, but that they were also able to synergize with the extracellular matrix. Since the collagen we used was a crude preparation from rat tail tendon we cannot rule out the possibility that the observed effect may depend on other associated proteins that are known to be accessory factors in promoting more efficient ligand-binding to the receptor. These results reinforce the notion that connective tissue cells can be therapeutically controlled using a strategy to inhibit the action of peptide growth factors. It is well known that continuous presence of the injury agent determines the evolution of the disease, favoring its progression and perpetuation (establishment of a permanent scar tissue). On the other hand, it is largely unclear which are the other physiologic determinants for disease evolution. Genetic background is known to influence the development of hypertrophic cheloid skin scars as well as the outcome of schistosomal liver fibrosis [ 34 ]. In experiments in which we were able to use matched LCTC and SF from the same individual our results suggest that tissue origin was consistently more important than the genetic background, since LCTC behaved differently from the same-patient's SF but similarly to LCTC from different individuals. Conclusions The results obtained in this study using human primary liver connective tissue cells suggest a plausible scenario for the development of liver fibrosis. Stellate quiescent cells are triggered to proliferate by high molecular weight serum factors [ 15 ] and may be subjected to a positive autocrine and paracrine feedback loop of growth factor stimulation. Whereas under normal conditions connective tissue cells are kept under strict check by the extracellular matrix, PDGF AA and TGF-β are capable of not only overriding the inhibitory effect caused by collagen type I on LCTC but also synergize to provide a stronger stimulus for proliferation. This interplay of extracellular signals may underlie the development of irreversible liver fibrosis. List of abbreviations HEPES: N-[2-Hydroxyethyl] piperazine-N'-[2-ethanesulfonic acid]; HSC: hepatic stellate cells; LCTC: liver connective tissue cells; SF: skin fibroblasts; PDGF: platelet-derived growth factor; TGF-β: transforming growth factor-β. Competing interests The author(s) declare that they have no competing interests. Authors' contributions AG established primary cell line cultures, carried out the cell biological studies, participated in designing and interpreting the experiments. MC provided technical support for the cell cultures and participated in the writing of the manuscript. RB participated in the design of the study and in the interpretation of the results. AM established primary cultures, conceived of the study, and participated in its design and coordination. 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/PMC539266.xml

549554

Nine-year comparison of presentation and management of acute coronary syndromes in Ireland: a national cross-sectional survey

Background Shorter time to treatment is associated with lower mortality in acute coronary syndromes (ACS). A previous (1994) survey showed substantial delays for acute myocardial infarction (AMI) in Ireland. The present study compared current practice with 1994 and surveyed acute coronary syndromes as a more complete contemporary evaluation of critical cardiac care than assessing AMI alone. Methods Following ethics committee approval, all centres (N = 39) admitting acute cardiac patients to intensive/coronary care unit provided information on 1365 episodes. A cross-sectional survey design was employed. Results Since 1994, median hospital arrival to thrombolysis time was reduced by 41% (76 to 45 minutes). Thrombolysis was delivered more often in the emergency department in 2003 (48% vs 2%). Thrombolysis when delivered in the emergency department was achieved faster than thrombolysis delivered in intensive/coronary care (35 mins v 60 mins; z = 5.62, p < .0001). Suspected AMI patients who did not subsequently receive thrombolysis took longer to present to hospital (5 h vs 2 h 34 mins; z = 7.33, p < .0001) and had longer transfer times to the intensive/coronary care unit following arrival (2 h 17 mins vs 1 h 10 mins; z = 8.92, p < .0001). Fewer confirmed AMI cases received thrombolysis in 2003 (43% vs 58%). There was an increase in confirmed cases of AMI from 1994 (70% to 87%). Conclusions Substantial improvements in time to thrombolysis have occurred since 1994, probably relating to treatment provision in emergency departments. Patient delay pre-hospital is still the principal impediment to effective treatment of ACS. A recent change of definition of AMI may have precluded an exact comparison between 1994 and 2003 data.

Background Ireland has one of the highest mortality rates from cardiovascular disease in the European Union [ 1 ]. Acute coronary syndrome (ACS) is a major portion of cardiovascular diseases. ACS includes unstable angina and both persistent-ST-segment elevation and non-ST segment elevation acute myocardial infarction (AMI) [ 2 - 4 ]. Thrombus formation is the primary reason for myocardial infarction [ 5 ]. This usually occurs after a complex interaction between coronary atherosclerosis, plaque rupture and platelet activation. Thrombolysis is an appropriate treatment for thrombus formation in ST-elevation AMI and when delivered in a 'timely' manner, preferably within 6 hours but including up to 12 hours after symptom onset [ 6 , 7 ], can significantly reduce morbidity and mortality from AMI. Each hour of time saved can lead to a decrease of about 1.6 deaths per 1000 patients treated [ 7 ]. Therefore, shortening of time to treatment for AMI patients is an important life-saving goal for health services [ 6 ]. International guidelines, for example those from the European Society of Cardiology [ 8 , 9 ] and the British Heart Foundation [ 10 ], have proposed a 'call-to-needle' time of 90 minutes for thrombolysis administration. The National Service Framework (NSF) in the United Kingdom (UK) further reduced the recommended time in 2000 with a proposed 'call-to-needle' time of 60 minutes [ 11 ]. Furthermore, eligible patients should be thrombolysed within 30 minutes of arrival at hospital [ 9 ]. In Ireland, the 1994 national census found a median time to treatment of 4 hours 30 minutes [ 12 ]. This compares unfavourably to more recently recorded median treatment times in other countries, e.g. 2 hours 45 minutes in the UK [ 13 ] or 2 hours 54 minutes in Switzerland [ 14 ]. Approximately 50% of AMI deaths in the community occur within two hours from the onset of symptoms [ 15 ]. Early management of AMI patients with thrombolysis significantly reduces morbidity and mortality [ 16 ]. Time to treatment can be shortened by thrombolysing patients in the emergency department prior to transfer to intensive/coronary care. This strategy has lead to significant reductions in delay [ 17 ]. Since 1994, there has been no examination of time to treatment and the extent to which international guidelines for treatment of AMI are being achieved in Ireland. This study assessed the presentation and management of a national cohort of suspected ACS patients admitted to intensive/coronary care units (I/CCUs) in all 39 Republic of Ireland hospitals providing such care. We decided to extend the range of patients assessed from those with suspected AMI (as was the case in previous surveys [ 12 , 18 ]) to all suspected ACS patients, to profile the pool of possible patients presenting from which testing determines eligibility for reperfusion therapy. Differing proportions of reperfusion-eligible patients over time or centre, alongside the absolute number of patients presenting, may influence the speed of management of eligible patients. Also, some treatments for other ACS (e.g. unstable angina) can be similar to treatment for AMI, depending on the severity of the event [ 19 ] (e.g. angioplasty treatment for unstable angina). Changes in the definition of AMI in recent years have lead to an increase in the proportion of diagnosed AMI patients and a decrease in the proportion of patients diagnosed as having unstable angina [ 20 - 22 ]. which may preclude exact comparison to previous findings. Where possible we compare data to results from 1994. Methods Sample All Irish centres admitting suspected ACS patients to I/CCU (N = 39) agreed to participate following relevant ethics approval [ 23 ]. Data collection was conducted from January to October 2003. Four hospitals had not recruited 25 suspected AMI patients by the study cut-off date. Suspected acute coronary syndrome (ACS) patients admitted to I/CCU were recruited. Staff were provided with the consensus definition of ACS as agreed in 2000 by the Joint European Society of Cardiology/American College of Cardiology Committee [ 4 ]. This definition uses enzyme (troponin) change as a marker of myocardial necrosis. The survey was of suspected ACS and the main focus was on how patients with suspected ACS are treated in the early phase of their hospital admission. Therefore the admission diagnosis was used to categorise patients (e.g. if a patient was admitted to I/CCU with a diagnosis of 'chest pain – query AMI', they were listed as suspected AMI for the purposes of this study; if patients were admitted with suspected ACS or suspected unstable angina, they were categorised as 'other ACS'). Data on successive admissions were audited anonymously from hospital charts. Participating hospitals recruited all consecutive suspected ACS patients, until 25 suspected cases of AMI had been admitted to I/CCU. Data on a total of 1365 episodes were collected (935 suspected AMI and 430 suspected other ACS admitted contemporaneously). Data collected assessed demographic details, clinical history, risk factors, presentation and management profile. Eligible patients were also approached to participate in a follow-up survey (results to be reported elsewhere). Analysis Analysis was conducted on the data using STATA/SE 8.0. Mann-Whitney U tests were used to test for significance between treatment times, χ 2 was used for categorical variables, and t-tests were used for continuous variables. Results from 1994 are reported, but significance tests were not conducted between 1994 and 2003 data (as raw data from 1994 was unavailable). Total time to treatment was defined as follows: symptom onset to reperfusion (thrombolysis or direct infarct angioplasty). Inpatient and other hospital transfer times were not included in the analysis for 1994 or 2003 samples. Results Baseline characteristics The sample consisted of 1365 episodes, 935 suspected AMI and 430 suspected other ACS patients. The gender breakdown has been described elsewhere (manuscript submitted for publication). The overall mean age was 64 years (std dev = 13; median = 65; range = 20–100 yrs). Admission characteristics are shown in table 1 . Table 1 Comparative demographic and admission profile in 1994 and 2003 Demographic 1994 (N = 950 suspected AMI) 2003 Suspected AMI (N = 935) Suspected Other ACS (N = 430) Combined (N = 1365) Mean age (years) (mean) (std dev) - 66 (13) 61*** (14) 64 (13) Men 64 (12) 64 (13) 60*** (13) 63 (13) Women 69 (11) 71 (13) 64*** (14) 69 (13) Referral source (%) Primary care physician 69 53 55 53 Self 24 39 33 38 Other 7 8 12 9 Admission mode to hospital (%) Ambulance 46 48 38** 45 Car (passenger) 42 42 43 42 Car (driver) 8 6 12*** 8 Other 4 4 7 5 Distance from hospital at symptom onset Median (range) miles 9 (0–165) 9 (0–80) 8 (0–150) 8 (0–150) Previous CHD history (%) AMI 24 16 31*** 21 Unstable angina 14 14 28*** 19 Coronary artery bypass graft 4 5 13*** 7 Percutaneous coronary intervention 2 6 18*** 10 (*p < .05, **p < .01, ***p < .001) The demographic profile of patients admitted for suspected AMI in 2003 appears similar to those admitted in 1994. Other ACS patients differed from suspected AMI patients in 2003 in the following aspects: they were younger, less likely to be admitted by ambulance, more likely to drive themselves to hospital, and had a higher prevalence of previous ACS and coronary interventions. Thrombolysis Both location of, and speed of administration of thrombolysis have changed considerably in a positive direction since 1994 (Figure 1 ). In 1994, 38% of suspected AMI and 58% of confirmed AMI patients received thrombolysis, which occurred in I/CCU (96%), emergency department (2%) or other location (2%). In 2003, 41% of suspected AMIs and 44% of confirmed AMIs were thrombolysed in I/CCU (48%), emergency department (48%) or other location (4%). A further 4% of suspected AMIs received direct infarct angioplasty. Figure 1 Thrombolysis administration locations and treatment times in 1994 and 2003 Emergency department delivered thrombolysis occurred in a significantly shorter time (median 35 minutes) than I/CCU administered thrombolysis (median 60 minutes) in 2003 (z = 5.62, p < .0001). In 2003, 29% of those thrombolysed were treated within 90 minutes of calling for professional help. This rose to 42% and 62% within 2 and 3 hours respectively. On hospital arrival, 35% of patients were thrombolysed within 30 minutes, rising to 60% and 74% within 60 and 90 minutes respectively. Thirty-six per cent of hospitals thrombolysed 80% or more of patients in the emergency department while 56% of hospitals thrombolysed over 50% of patients in the emergency department. Time to treatment Suspected AMI patients waited a similar length of time to get to hospital from onset of symptoms in 2003 as 1994 (Table 2 ). Symptom onset to hospital arrival time was significantly higher for patients admitted with suspected other ACS in 2003 (3 h 35 mins vs 4 h 39 mins, z = 2.99, p < .01). Table 2 Median overall time to treatment for all patients in 1994 and 2003 Treatment times 1994 Suspected AMI (n = 950) 2003 Thrombolysed AMIs (n = 382) Suspected AMI – non-thrombolysed (n = 553) All suspected AMI (n = 935) Suspected other ACS (n = 430) Total (n = 1365) Symptom onset to hospital 3 h 30 mins 2 h 34 mins 5 h 00 mins*** 3 h 35 mins 4 h 39 mins** 3 h 56 mins Call-to-thrombolysis Unavailable 2 h 20 mins - 2 h 20 mins - - Hospital arrival to I/CCU 55 mins 1 h 10 mins 2 h 17 mins*** 1 h 40 mins 2 h 43 mins*** 1 h 55 mins Hospital arrival to thrombolysis 76 mins 45 mins - 45 mins - - I/CCU admission to thrombolysis 25 mins 20 mins - 20 mins - - (*p < .05, **p < .01, ***p < .001) Patients waited longer to be admitted to I/CCU in 2003, but received thrombolysis more quickly (45 mins v 76 mins) after hospital arrival. This represents a 41% decrease in time-to-thrombolysis since 1994. In 2003, total time to treatment for suspected AMI patients who received reperfusion was 4 hours 00 mins. Patients with suspected AMI who were subsequently thrombolysed (thrombolysed AMIs) presented to hospital (2 h 34 mins vs 5 h, z = 7.33, p < 0.0001) and had a faster I/CCU transfer time (1 h 10 mins vs 2 h 17 mins, z = 8.92, p < 0.0001) than suspected AMI patients who were not thrombolysed (non-thrombolysed AMIs). Suspected other ACS patients also waited significantly longer than non-thrombolysed AMIs for hospital transfer to I/CCU (2 h 43 mins vs 2 h 17 mins; z = 2.128, p < 0.05). In 1994, treatment times for patients referred by primary care physicians were significantly longer than those who self-referred (symptom onset to hospital arrival: primary care physician-referred 4 h 15 mins, self-referred 2 h 05 mins, p < 0.001). In 2003, primary care physician-referred suspected AMI patients also had a significantly longer pre-hospital delay (5 h vs 2 h 28 mins, z = 7.9, p < 0.001). For suspected AMI patients, a previous experience of AMI made no difference to hospital presentation time (3 h vs 3 h 45 mins, z = 1.2, p > 0.05). There were no gender differences in pre-hospital delay time (3 h 53 mins for men vs 3 h 14 mins for women, z = 0.63, p > 0.05), or hospital arrival to thrombolysis time (45 mins for men vs 50 mins for women, z = 1.43, p > 0.05) for suspected AMI patients in 2003. Discharge For suspected AMI patients, there was an increase in those patients diagnosed with myocardial infarction of 21% from 1994 to 2003, and a 10% reduction in the diagnosis of unstable angina (Table 3 ). Table 3 I/CCU discharge diagnoses and hospital mortality (%) Discharge diagnosis 1994 Census (suspected AMI) % CCU 2003 Survey Thrombolysed AMIs (n = 382) Suspected AMI – non-thrombolysed (n = 553) Suspected AMI Suspected Other ACS Total Myocardial infarction 70 97 86*** 91 19*** 68 Unstable angina 14 <1 6*** 4 47*** 17 Other cardiac 9 5 11** 9 32*** 16 Non cardiac 7 1 5** 3 17*** 8 Mortality 11 9 10 10 1*** 7 Patients may have more than one diagnosis (*p < .05, **p < .01, ***p < .001) In 2003, thrombolysed AMIs were more likely to be discharged as having myocardial infarction than non-thrombolysed AMIs, but were less likely to receive a discharge diagnosis of unstable angina, other cardiac or non-cardiac diagnoses. All suspected AMI patients in 2003 were more likely to be discharged as having had myocardial infarction and were more likely to die in hospital than suspected other ACS patients, but were less likely to have discharge diagnoses of unstable angina, other cardiac or non cardiac. Discussion The present survey outlines the current presentation and management of ACS in Ireland. This study built on the previous research conducted in 1994, but also expanded its findings beyond suspected AMI patients to all suspected ACS patients. Pre-hospital patient delay remains stable and substantial, while a considerable reduction (41%) in time to thrombolysis from hospital admission has occurred since 1994. This can probably be attributed in large part to the relocation of thrombolytic administration to the emergency department, thereby reducing the 'door-to-needle' times in 2003. Similarly, treatment of AMI patients in emergency departments prior to transfer to I/CCU may account for longer I/CCU transfer times in 2003. Significant progress has been made in the treatment of AMI patients who receive thrombolysis, which has yielded faster 'door-to-needle' times. Suspected AMI patients who received thrombolysis in the 2003 sample were treated more quickly than 1994 (median 45 mins v 76 mins). The transfer of thrombolytic administration from the I/CCU (96% in 1994) to the emergency department (48% in 2003, with 48% administered in I/CCU) is probably the main reason for the decreased time to treatment for thrombolysed patients. The present survey found, for instance, that 56% of hospitals thrombolysed over half of their patients in the emergency department. It is not clear, however, that an additional shift of thrombolysis to the emergency department in the remaining hospitals would also result in a further reduction of 'door-to-needle' time. This is because some hospitals already adopt a 'fast-track' policy, where chest pain patients are admitted directly to CCU, bypassing emergency department assessment. Adopting a strategy of emergency department thrombolysis may have little or no effect in these cases. Nonetheless, only 35% of patients received thrombolysis within 30 mins of hospital arrival, which compares unfavourably to other surveys (e.g. in England and Wales in 2003, the MINAP study found that over 80% of patients were thrombolysed within 30 mins of hospital arrival [ 24 ]). These comparisons must be interpreted with some caution however, since MINAP does not count cases which have a component of extra delay due to clinical reasons (i.e. patients presenting with contraindications to thrombolysis), and reports on all eligible STEMI cases treated within 30 mins of hospital arrival. Our analysis included all patients who eventually received thrombolysis, even those patients who developed ST-elevation some time after hospital arrival. While our results may not be completely comparable to other similar surveys, the overall message that thrombolysis is unsatisfactory both in absolute and comparative terms is clear. Substantial pre-hospital and in-hospital delays were seen for non-thrombolysed AMIs. Suspected AMI patients who did not receive thrombolysis waited significantly longer for transfer to I/CCU than those who received thrombolysis. Indeed, this group of AMIs had comparable times to patients with suspected other ACS. These patients presented to hospital more slowly than thrombolysed AMIs. This may indicate the less severe symptoms of unstable angina and non-ST-elevation myocardial infarction. Also, recent changes in AMI definition have lead to increases in proportions of diagnosed AMI patients and decreases in proportions of patients diagnosed with unstable angina [ 20 - 22 ]. These changes probably preclude exact comparison with 1994 findings. Adding credence to this hypothesis was the large change in discharge diagnoses. Considering suspected AMIs, the proportion of patients diagnosed with myocardial infarction increased by 21% from 1994 to 2003, with a corresponding 10% reduction in unstable angina diagnoses. The reduction in 'non cardiac' discharges for suspected AMI patients (7% in 1994 to 3% in 2003), may be partially attributed to definition change [ 21 ]. Finally, the 14% decrease in thrombolysed confirmed AMIs may also be attributed to the definition change, with a larger portion of AMI patients being ineligible for thrombolysis in 2003. Suspected AMI patients who did not receive thrombolysis took significantly longer from hospital arrival to arrive in I/CCU than did patients who received thrombolysis. Suspected other ACS patients presented to hospital in time frames similar to suspected AMI patients who did not receive thrombolysis, but were not admitted to I/CCU as quickly. The development of chest pain assessment units and other similar units (e.g. medical admission units), which involve the initial screening of chest pain patients to determine whether the pain is cardiac in origin prior to transfer to I/CCU, may have skewed the data in the current survey. A typical scenario may be that a patient currently arrives at hospital, without ST-elevation when assessed by ECG, and is assessed for some time in the chest pain assessment unit to either 'rule-in' or 'rule-out' ACS. In the past, such patients may have been admitted directly to I/CCU. In addition, these findings may reflect the triaging of patients in the emergency department, where patients labelled as suspected AMI were treated more quickly than those labelled as suspected other ACS. Patients can be triaged into groups reflecting the necessity for immediate treatment. Non-ST-elevation myocardial infarction and unstable angina patients may be detained for observation in the emergency department. It might also be that pressure for I/CCU bed places is more quickly resolved for the more acute ST-elevation AMI patients. However, further prospective observational research on this aspect of care is required. Patient delay prior to hospital arrival is still the biggest impediment to improving treatment times for AMI and other ACS. Patients who were referred by primary care physicians had a significantly longer time to treatment than those who self-referred in both 2003 and 1994. Clearly, in the present system, although substantial improvements have been made since the mid-1990s, a 'call-to-treatment' standard of 90 minutes is not currently being met. The present survey found that 29% of patients were thrombolysed within 90 minutes of calling for professional help. Future research and resources need to focus on a reduction in pre-hospital delay factors by the services and professional groups concerned. A number of psychological studies have provided detailed examination of patient contributions to delayed help-seeking for AMI [ 25 - 27 ]. These highlight possibilities for intervention to reduce symptom onset to help-seeking times but caution against a simplistic public education campaign approach. Since general public advertising/education campaigns have little efficacy [ 25 , 26 , 28 ] and even a previous experience of AMI has no effect on pre-hospital call for assistance times, sophisticated strategies are needed to address this problem. As regards reducing health services delay, one method is to authorise health professionals in the pre-hospital setting to administer thrombolysis. Ambulance personnel and primary care physicians are the two most obvious choices. The administration of thrombolysis by ambulance personnel has been shown to reduce time to treatment [ 29 ]. A recently completed study on thrombolysis administered by primary care physicians showed the practicalities of this approach in an Irish setting [ 30 ]. In the hospital setting, involving nurses in decision-making processes for thrombolysis has been shown to be effective in reducing treatment times [ 31 , 32 ]. The provision of more mobile coronary care units may also reduce delay times in rural areas [ 33 ]. The present study has highlighted the need for a national prospective registry of ACS. The value of such registries has been shown in other countries. For example, registries increase the use of appropriate reperfusion therapy, but to ensure this practice continues they need to be ongoing [ 34 ]. In the UK, MINAP [ 35 ] has contributed to an increase in the numbers of patients thrombolysed within recommended timescales. For example, following the publication of the NSF guidelines on CHD in 2000 [ 11 ], the proportion of patients receiving thrombolysis within 30 minutes of hospital arrival more than doubled (79% in 2002, compared with 38% pre-2000) [ 36 ]. Registries can also underscore the extent of adherence to international guidelines. The current Irish data highlight progress in some areas over a nine-year period but indicate that improvements need to occur in other aspects. A national registry can provide the continuous feedback needed to keep a focus on areas for improvement. The monitoring of guideline adherence should even be considered a part of optimal practice procedures [ 37 ]. Conclusions Significant progress in some aspects (e.g. time to thrombolysis) of care for suspected ACS has occurred, but there is still scope for improvement. A necessary goal is to increase the proportions of patients seen within the recommended time. Other aspects of ACS care also now require attention. Patient delay pre-hospital should remain the main focus of future research. The implementation of a national registry would allow resources to be focused on these aspects and facilitate routine health care monitoring. Competing interests The author(s) declare that they have no competing interests. Authors' contributions FD carried out the survey, analysed the data and drafted the manuscript. HM, DD & ES conceived of the study and participated in its design and coordination. RC participated in the study design and coordination, and provided statistical input on data analysis and interpretation. All authors provided critical input on manuscript drafts, and approved the final version. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549554.xml

516044

Loss of KCNJ10 protein expression abolishes endocochlear potential and causes deafness in Pendred syndrome mouse model

Background Pendred syndrome, a common autosomal-recessive disorder characterized by congenital deafness and goiter, is caused by mutations of SLC26A4, which codes for pendrin. We investigated the relationship between pendrin and deafness using mice that have ( Slc26a4 +/+ ) or lack a complete Slc26a4 gene ( Slc26a4 -/- ). Methods Expression of pendrin and other proteins was determined by confocal immunocytochemistry. Expression of mRNA was determined by quantitative RT-PCR. The endocochlear potential and the endolymphatic K + concentration were measured with double-barreled microelectrodes. Currents generated by the stria marginal cells were recorded with a vibrating probe. Tissue masses were evaluated by morphometric distance measurements and pigmentation was quantified by densitometry. Results Pendrin was found in the cochlea in apical membranes of spiral prominence cells and spindle-shaped cells of stria vascularis, in outer sulcus and root cells. Endolymph volume in Slc26a4 -/- mice was increased and tissue masses in areas normally occupied by type I and II fibrocytes were reduced. Slc26a4 -/- mice lacked the endocochlear potential, which is generated across the basal cell barrier by the K + channel KCNJ10 localized in intermediate cells. Stria vascularis was hyperpigmented, suggesting unalleviated free radical damage. The basal cell barrier appeared intact; intermediate cells and KCNJ10 mRNA were present but KCNJ10 protein was absent. Endolymphatic K + concentrations were normal and membrane proteins necessary for K + secretion were present, including the K + channel KCNQ1 and KCNE1, Na + /2Cl - /K + cotransporter SLC12A2 and the gap junction GJB2. Conclusions These observations demonstrate that pendrin dysfunction leads to a loss of KCNJ10 protein expression and a loss of the endocochlear potential, which may be the direct cause of deafness in Pendred syndrome.

Background Pendred syndrome is a relatively common autosomal-recessive disorder characterized by deafness and goiter [ 1 ]. The syndrome is caused by mutations of the PDS gene SLC26A4, which codes for the protein pendrin [ 2 ]. Deafness is congenital and generally profound although sometimes late in onset and provoked by light head injury. Vestibular dysfunction is uncommon. Goiter is variable and generally develops around puberty [ 3 ]. The cause of goiter appears to be an impairment of iodide fixation in the follicular lumen due to a reduced rate of iodide transport across the apical membrane of thyroid gland epithelial cells [ 4 ]. A positive perchlorate discharge test and an enlarged vestibular aqueduct appear to be the most reliable clinical signs of Pendred syndrome [ 5 ]. Pendrin is an anion exchanger that can transport Cl - , I - , HCO 3 - and formate [ 6 , 7 ]. Expression has been found in the inner ear and thyroid gland consistent with the clinical signs of deafness and goiter [ 2 , 3 , 8 ]. In addition, pendrin expression has been found in the kidney [ 9 ], mammary gland [ 10 ], uterus [ 11 ], testes [ 12 ] and placenta [ 13 ]. No expression was found in fetal or adult brain, consistent with a peripheral cause of deafness [ 2 , 11 ]. Expression of pendrin mRNA in the inner ear has been found in several places including the cochlea, the vestibular labyrinth and the endolymphatic sac [ 8 ]. The precise location of pendrin protein expression, however, has not yet been determined. The variability of deafness in Pendred syndrome and the observation that deafness is sometimes late in onset suggest that pendrin dysfunction may not be the direct cause of deafness. It is conceivable that pendrin dysfunction favors changes in the expression levels of proteins that are critical for the maintenance of the hearing function. Detailed studies aimed at identifying the direct cause of deafness in Pendred syndrome have recently become possible due to the generation of a pendrin-specific polyclonal antibody [ 9 ] and the development of Slc26a4 -/- mice, which bear a targeted disruption of the mouse Slc26a4 gene [ 14 ]. The aim in the present study was to determine the location of pendrin and the cause of deafness in Slc26a4 -/- mice. Methods The endocochlear potential and the endolymphatic and perilymphatic K + concentrations were measured in young adult mice (1–4 month of age) that either have ( Slc26a4 +/+ ) or lack ( Slc26a4 -/- ) a functional gene for pendrin [ 14 ]. The mouse strain 129Sv/Ev (Taconic, Germantown, NY) was chosen as the source of Slc26a4 +/+ controls, since Slc26a4 -/- mice were propagated in the this strain and generated using a stem cell line derived from 129Sv/Ev. Slc26a4 +/+ and Slc26a4 -/- were agouti. They did not differ in coat color. Differences in pigmentation were verified using coisogenic age-matched Slc26a4 +/+ and Slc26a4 -/- that were bred in parallel. Expression of key proteins involved in the generation of the endocochlear potential and the transport of K + was studied using confocal immunocytochemistry and quantitative RT-PCR. K + secretion and the generation of the endocochlear potential were measured using electrophysiological techniques. All experiments were approved by the Institutional Animal Care and Use Committee of Kansas State University. Confocal immunocytochemistry Animals were deeply anesthetized with sodium pentobarbital (100 mg/kg i.p.) and transcardially perfused with a Cl - -free phosphate-buffered Na-gluconate solution containing 4% paraformaldehyde. Temporal bones were removed and cochleae fixed by perilymphatic perfusion, decalcified in EDTA, processed through a sucrose gradient and infiltrated with polyethylene glycol. Mid-modiolar cryosections (12 μm, CM3050S, Leica, Nussloch, Germany) were blocked in PBS with 0.2% Triton-X (PBS-TX) and 10% bovine serum albumin. Slides were incubated overnight at 4°C with primary antibodies in PBS-TX with 1–3% BSA [rabbit anti-pendrin, 1:500 (h766–780); rat anti-ZO-1, 1:100 (Chemicon, Temecula, CA); goat anti-KCNQ1, 1:200 (C20, Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-KCNE1, 1:200 (Alomone, Jerusalem, Israel); rabbit anti-KCNJ10, 1:300 (Alomone); rabbit anti-SLC12A2, 1:100 (Chemicon); and rabbit anti-connexin 26, 1:100 (Zymed, San Francisco, CA)]. Slides were washed in PBS-TX and incubated for 1 h at 25°C with appropriate secondary antibodies at a 1:1000 dilution in PBS-TX with 1–3% BSA [donkey anti-rabbit Alexa 488, chicken anti-rat Alexa 594, and chicken anti-goat Alexa 594 (Molecular Probes, Eugene, OR)]. Actin filaments were visualized by staining with Alexa 488 conjugated phalloidin (Molecular Probes). After incubation, slides were washed with PBS-TX, mounted with FluorSave (Calbiochem, La Jolla, CA), and viewed by confocal microscopy (LSM 5 Pascal or LSM 510 Meta, Carl Zeiss, Jena or Göttingen, Germany). Laser scanning brightfield images were collected to document structural preservation, for morphometric analysis and for analysis of pigmentation. Quantitative RT-PCR Animals were deeply anesthetized with sodium pentobarbital (100 mg/kg i.p.). Kidneys and brain were removed and pulverized in liquid N 2 . Temporal bones were removed and two preparations of inner ear tissues were obtained by microdissection: (1) stria vascularis without spiral ligament, and (2) spiral ganglia including the surrounding bone and the organ of Corti. The dissection medium was changed twice during the microdissection and isolated tissues were washed three times to minimize cross-contamination. The Slc26a4 -/- genotype was verified by the observation of large otoconia in the utricular macula [ 14 ]. Total RNA was isolated and residual DNA contamination was removed by enzymatic digestion (RNeasy micro, Qiagen, Valencia, CA). Quality and quantity of 18S rRNA obtained from kidneys and brain were determined (Nano Assay, 2100 Bioanalyzer, Agilent, Palo Alto, CA). Further, the quality of RNA preparations obtained from stria vascularis and spiral ganglia was assessed (Pico Assay, 2100 Bioanalyzer). Real time RT-PCR was performed on RNA obtained from individual animals (OneStep RT-PCR, Qiagen; Smart Cycler, Cepheid, Sunnyvale, CA) in the presence of 0.2× SYBR green I (Molecular Probes). Transcripts of 18S rRNA and mRNA for the K + channels KCNJ10, KCNQ1 and KCNQ4 were amplified using gene-specific primers (Table 1 ). RT was performed for 30 min at 50°C and 15 min at 95°C. PCR consisted of 50 cycles of 1 min at 60°C, 1 min at 72°C, 7s heating to hot-measurement temperature, 13s hot-measurement at 3–5°C below product melting temperature (Table 1 ) and 1 min at 94°C. Hot-measurements were performed to eliminate the detection of primer-dimers that had melting temperatures between 72 and 75°C [ 15 ]. PCR was followed by a melt (60 to 95°C). The generation of a single product of appropriate size was routinely checked by the presence of a single melt peak and by agarose gel-electrophoresis. Product identity was confirmed by sequencing. Table 1 Primers Template Primers Product Length Melting Temperature Hot Measurement Fidelity n 18S gag gtt cga aga cga tca ga ( sense ) 316 bp 83.2 ± 02°C 78°C 1.89 ± 0.02 17 tcg ctc cac caa cta aga ac (antisense) KCNJ10 tgg tgt ggt gtg gta tct gg ( sense ) 411 bp 83.2 ± 02°C 78°C 1.86 ± 0.02 15 tga agc agt ttg cct gtc ac ( antisense ) KCNQ1 ttt gtt cat ccc cat ctc ag ( sense ) 239 bp 82.5 ± 02°C 79°C 1.85 ± 0.02 17 gtt gct ggg tag gaa gag ( antisense ) KCNQ4 ccc gga aac cct tct gtg tc ( sense ) 245 bp 83.2 ± 02°C 80°C 1.87 ± 0.01 17 aaa gat gag cac cag gaa cc ( antisense ) Template molecules (T) were quantified according to T = P / (F ^C t ) where P is the number of product molecules, F is the fidelity of the reaction and C t is the cycle at which the number of product molecules reaches a chosen threshold [ 15 ]. Fidelity (F) was obtained from the slope (S) of the log-linear phase of the growth curve via a best-fit fifth-order polynomial: F = 7.39 + 3.80 × S + 1.05 × S 2 + 0.15 × S 3 + 11.38 × 10 -3 * S 4 + 3.39 × 10 -4 × S 5 . The number of product molecules at threshold (P Ct ) was determined by amplification of known amounts of 18S rRNA according to P Ct = T × F ^C t . Quantifications of 18S rRNA were used to compare tissue amounts. Genomic contamination of inner ear samples was assessed to be <0.02% by omission of the RT step. In vitro electrophysiology Animals were deeply anesthetized with sodium pentobarbital (100 mg/kg i.p.). Stria vascularis without spiral ligament was obtained by microdissection. Currents generated by the stria marginal epithelium were recorded [ 16 ]. A Pt-Ir wire microelectrode with a Pt-black tip was positioned 20–30 μm from the apical surface of the epithelium and vibrated at 200–400 Hz by piezo-electric bimorph elements (Applicable Electronics, Forest Dale, MA; ASET version 1.05, Science Wares, East Falmouth, MA). A Pt-black electrode (26-gauge) served as reference in the bath chamber. The signals from the phase-sensitive detectors were digitized (16 bit) at a rate of 0.5 Hz. The output was expressed as current density at the electrode. In situ electrophysiology Animals were anesthetized with inactin (thiobutabarbital sodium salt, 140 mg/kg ip). The endocochlear potential and the endolymphatic [K + ] were measured with double-barreled microelectrodes [ 17 ]. Measurements were made in the basal turn by a round-window approach through the basilar membrane and in the apical turn after thinning the bone over the stria vascularis and picking a small hole (~30 μm). K + -selective electrodes were calibrated in solutions of constant cation (K + and Na + ) concentration of 150 mM. The minor selectivity of the K + electrodes for Na + produced a nonlinearity in the calibration curve, which was closely fitted by the Nicolski equation using nonlinear curve-fitting software (OriginLab, Northampton, MA): V = V i + S × log ([K + ] + A × [Na + ]), where V i is an offset term, S is slope, A is selectivity, and [Na + ] is Na + concentration. Calibrations were made immediately after withdrawal of the electrodes from the cochlea. Plasma K + concentrations were obtained using a blood analyzer (Stat Profile M, Nova Biomedical, Waltham, MA). Data are presented as mean ± sem; n denotes the number of experiments. Differences were considered significant when p < 0.05. Results and Discussion In situ hybridization in the cochlea has suggested that pendrin mRNA is expressed in cells that reside immediately beneath the spiral prominence on the lateral wall of the external sulcus [ 8 ]. To determine the location of pendrin protein expression, we performed confocal immunocytochemistry on cryosections prepared from temporal bones of normal ( Slc26a4 +/+ ) mice using an established antibody [ 3 ]. Staining was absent when the primary antibody was pre-absorbed with the antigenic peptide ( data not shown ). Strong expression of pendrin was observed not only in outer sulcus epithelial cells, as predicted from in situ hybridization data, but also in root cells, in apical membranes of spiral prominence surface epithelial cells and in apical membranes of spindle-shaped cells that are part of stria vascularis (Fig. 1a,1b,1c ). The presence of pendrin in spindle-shaped cells suggests that these cells secrete HCO 3 - into endolymph. Pendrin-mediated HCO 3 - transport has previously been shown in the kidney [ 9 ]. HCO 3 - in the cochlea may be generated from CO 2 by carbonic anhydrase located in strial intermediate cells [ 18 , 19 ]. CO 2 may be supplied by the metabolically highly active stria marginal cells. It is conceivable that pendrin dysfunction interrupts HCO 3 - secretion and leads to an accumulation of HCO 3 - in stria vascularis. Preliminary data (Wangemann et al., unpublished ) support a role for pendrin in HCO 3 - secretion into endolymph. Figure 1 Protein localization of pendrin, KCNQ1, ZO-1 and F actin in cochlea and vestibular labyrinth of Slc26a4 +/+ and Slc26a4 -/- mice. a: Overview of cochlea; bar = 100 μm. b-f: Detail of cochlear lateral wall; bar = 10 μm. g: Detail of utricle; bar = 10 μm. h-i: Detail of ampulla; bar = 10 μm. RM, Reissner's membrane; SC, spindle-shaped cells, SMC, strial marginal cells; SV, stria vascularis; SL, spiral ligament; LIM, spiral limbus. BC, basal cells; SP, spiral prominence epithelial cells; RC, root cells; OS, outer sulcus epithelial cells; VHC, vestibular hair cells; VTC, vestibular transitional cells; VDC, vestibular dark cells; arrows, basal cells at the top and bottom of stria vascularis form tight junctions with surface epithelial cells. The extent of pendrin expression in spiral prominence cells and stria vascularis was determined by labeling KCNQ1, a K + channel that is expressed in strial marginal cells, and ZO-1, a tight junction protein that labels basal cells and thereby delineates the boundaries of stria vascularis. Dual labeling experiments demonstrated that pendrin is expressed in spindle-shaped cells, which are surface epithelial cells in stria vascularis adjacent to strial marginal cells near the borders of both the spiral prominence and Reissner's membrane (Fig. 1c ). In situ hybridization in the vestibular labyrinth suggested that pendrin mRNA is expressed in non-sensory cells [ 8 ]. Using confocal immunocytochemistry on cryosections, we observed strong expression of the pendrin protein in the apical membrane of vestibular transitional cells in the utricle and ampullae (Fig. 1g,1h,1i ). Dual labeling with KCNQ1 demonstrated that pendrin expression was clearly limited to vestibular transitional cells and did not extend to other non-sensory cells such as vestibular dark cells, which were clearly identified by the expression of KCNQ1 and KCNE1 in their apical membranes. The onset of pendrin expression during development of the mouse inner ear has been determined to be embryonic day (ED) 13 [ 14 ]. Morphologically detectable differences in the inner ears of Slc26a4 +/+ and Slc26a4 -/- mice become evident as early as ED 15, when Slc26a4 -/- mice start to develop an enlarged endolymphatic space that persists into adulthood [ 14 ]. Interestingly, sensory hair cells in the cochlea appear normal until postnatal day (PD) 7 but show clear evidence of degeneration by PD 15 [ 14 ]. These observations suggest that the cochlear environment supports the survival of sensory hair cells in spite of the enlargement of the endolymphatic duct. A normal endolymphatic K + concentration, which is critical for hair cell survival [ 20 ], is established at PD 3 [ 21 ] and may persist in Slc26a4 -/- at least until PD 7. The time period between PD 7 and 15 is the time when the endocochlear potential develops at the onset of hearing [ 22 ]. We hypothesized that a lack of a normal endocochlear potential or an alteration of the endolymphatic K + concentration could account for deafness in Slc26a4 -/- mice. Measurements revealed that the endocochlear potential was absent but that the endolymphatic K + concentration was normal in adult Slc26a4 -/- (Fig. 2a ). No significant differences between Slc26a4 +/+ and Slc26a4 -/- mice were found in perilymphatic (Fig. 2a ) or plasma K + concentrations ( Slc26a4 +/+ , 4.9 ± 0.3 mM, n = 6; Slc26a4 -/- , 5.1 ± 0.3 mM, n = 6). These observations suggest that a primary event leading to deafness in Slc26a4 -/- mice, and potentially in patients suffering from Pendred syndrome, is the loss of the endocochlear potential. Degeneration of hair cells is probably a response to the loss of the endocochlear potential. Figure 2 Potential, K + concentrations and pigmentation of stria vascularis in Slc26a4 +/+ and Slc26a4 -/- mice. a: Endocochlear potential and K + concentrations in endolymph and perilymph at the apex (A) and base (B) of the cochlea. Numbers adjacent to symbols denote number of measurements. b-d: Pigmentation of stria vascularis in Slc26a4 +/+ and Slc26a4 -/- mice. b: View of stria vascularis through the bony capsule of the cochlea. OW, oval window, RW, round window; arrows, stria vascularis. c-d: Whole-mounts of stria vascularis isolated from age-matched mice. c: Laser-scanning images, bar = 10 μm, d: Quantification of pigmentation based on optical density. The endocochlear potential is generated by stria vascularis in the lateral wall of the cochlea [ 17 , 23 ]. The potential is generated across the basal cell barrier of stria vascularis by the K + channel KCNJ10 located in intermediate cells [ 24 ], which are connected to basal cells by a high density of gap junctions [ 25 ]. Marginal cells of stria vascularis, which form the barrier toward endolymph, transport K + from the intrastrial space into endolymph and keep the K + concentration low adjacent to the KCNJ10 K + channel [ 26 ]. To determine the cause of the loss of the endocochlear potential in Slc26a4 -/- mice, we first determined whether intermediate cells are present in stria vascularis, since a loss of intermediate cells is known to lead to a loss of the endocochlear potential [ 27 , 28 ]. Intermediate cells of stria vascularis were visualized by their pigmentation. Pigmentation was present in stria vascularis of Slc26a4 -/- mice (Fig. 2b ), which suggests that intermediate cells are present. Interestingly, pigmentation of stria vascularis was much stronger in Slc26a4 -/- than in Slc26a4 +/+ mice. To determine in greater detail the cause of the loss of the endocochlear potential in Slc26a4 -/- mice, we isolated total RNA from stria vascularis and spiral ganglia of young (1–4 month) Slc26a4 +/+ and both young and old (~12 month) Slc26a4 -/- mice, assessed amounts of isolated tissues by quantification of 18S rRNA, and quantified the expression of KCNJ10 mRNA. Quantities of stria vascularis isolated from these different mice were similar, since no significant differences were found in the numbers of 18S rRNA molecules (log(rRNA) = 9.46 ± 0.08, n = 17, pooled data ). In contrast, quantities of spiral ganglia isolated from young and old Slc26a4 -/- mice (log(rRNA) = 9.04 ± 0.18, n = 4 and 9.29 ± 0.15, n = 6) were significantly smaller than in Slc26a4 +/+ mice (log(rRNA) = 9.48 ± 0.19, n = 7), consistent with morphometric data ( see below ). Expression of KCNJ10 mRNA was normal in stria vascularis and spiral ganglia of young Slc26a4 -/- mice but significantly reduced in old Slc26a4 -/- mice (Fig. 3 ). Quantifications of KCNQ1 and KCNQ4 mRNA were used to assess possible cross-contamination between the stria vascularis and spiral ganglia preparations based on the assumptions that (1) KCNQ1 is expressed in cells of the stria vascularis but not the spiral ganglia preparation and (2) KCNQ4 is expressed in cells of the spiral ganglia but not the stria vascularis preparation. The number of KCNQ1 mRNA molecules in stria vascularis was 24-fold greater than in spiral ganglia, and the number of KCNQ4 mRNA molecules in spiral ganglia was 5-fold greater than in stria vascularis. These observations validate our measurements of KCNJ10 and KCNQ1 by demonstrating that the microdissected preparations of stria vascularis and spiral ganglia were 98% and 78% pure, respectively. Figure 3 Quantification of KCNJ10 and KCNQ1 mRNA expression in stria vascularis and spiral ganglia of Slc26a4 +/+ and Slc26a4 -/- mice. a: Electropherogram of total RNA isolated from stria vascularis microdissected from one mouse. The amount of total RNA was obtained from the total integral ( shaded ) and the amount of 18S rRNA was obtained from the integral of the 18S peak. Sharp peaks representing 18S and 28S rRNA demonstrate the high quality of RNA. Insert: Genotype of Slc26a4 -/- mice was verified by the observation of one or few very large rhomboedric otoconia in the utricular macula ( arrow ). A, crista ampullaris; U, utricular macula. Scale bar: 100 μm. b: Example of real-time RT-PCR data used for quantification of 18S, KCNJ10, KCNQ1 and KCNQ4. Known quantities of 18S rRNA were used to calibrate the threshold. SV, stria vascularis; SG, spiral ganglia. c: Quantification of KCNJ10 and KCNQ1 mRNA in young Slc26a4 +/+ and young and old Slc26a4 -/- mice. The presence of KCNJ10 mRNA in stria vascularis of Slc26a4 -/- mice supports the finding that intermediate cells are present. Expression of the KCNJ10 protein was assessed in young Slc26a4 -/- mice by confocal immunocytochemistry. Interestingly, expression of the KCNJ10 protein was absent in stria vascularis but normal in spiral ganglia (Fig. 4 ). The absence of the KCNJ10 K + channel in stria vascularis is sufficient to explain the loss of the endocochlear potential [ 17 ]. Figure 4 Protein localization of KCNJ10 in the cochlea of Slc26a4 +/+ and Slc26a4 -/- mice. a: Overview of cochlea; bar = 100 μm. Compare to Fig. 1a to note the enlarged scala media and the distended Reissner's membrane. b-c: Detail of lateral wall and spiral ganglia ( insert ); main bar: 10 μm, insert: 5 μm. Expression of KCNJ10 in Slc26a4 -/- mice was absent in stria vascularis but unchanged in spiral ganglion cells. RM, Reissner's membrane, SV, stria vascularis; SP, spiral prominence; SL, spiral ligament; LIM, spiral limbus; SG, spiral ganglion. Histological evaluation of cryosections revealed an enlargement of scala media with a large bulging of Reissner's membrane and an apparent degeneration of the organ of Corti, as described earlier [ 14 ]. Tissue height of stria vascularis was normal (Fig. 5 ), consistent with the similar numbers of 18S rRNA molecules in isolated preparations ( see above ). The absence of a change in tissue height is consistent with the presence of intermediate cells [ 28 ]. Further, we observed an apparent loss of tissue masses in areas that are normally occupied primarily by type I and II fibrocytes. Spiral prominence in Slc26a4 -/- mice was less prominent, spiral ligament thinner and spiral limbus flatter (Fig. 5 ). The observation that tissue masses were lost in the spiral limbus region is consistent with the finding of a reduced number of 18S rRNA molecules in the spiral ganglia preparation ( see above ). In addition, we observed an apparent degeneration of stria vascularis, including an increase in pigmentation and an irregular pattern of the tight junctions of marginal cells (Fig. 2c , 6 , 7 ). Tight junctions were visualized by F-actin expression. Marginal cells appeared to form a continuous layer. Figure 5 Morphometric analysis of cochlear tissue masses in Slc26a4 +/+ and Slc26a4 -/- mice. a: locations of measurements. Thickness of stria vascularis (SV) was obtained as average of three distance measurements perpendicular to the surface of marginal cells. Thickness of spiral prominence (SP) was measured perpendicular to a tangential line ( dashed ) that connects the surface of the outer sulcus (OS) with the basal layer of stria vascularis. Thickness of spiral ligament (SL) was measured perpendicular to the tangential line as distance between the surface of spiral prominence and the interface between spiral ligament and bone (B). Thickness of spiral limbus (LIM) was obtained perpendicular to the surface of the bone as a tangential line that touches the inner sulcus (IS) and reaches from the surface of the spiral limbus to the interface between spiral limbus connective tissue and bone. b: Summary. Data from 7–8 animals contributed to each column. Figure 6 Analysis of marginal and basal cell barriers by in Slc26a4 -/- mice. Tight junctions were visualized by F actin. Whole-mounts of stria vascularis were viewed either from the basal cell side ( a-f ) or from the marginal cell side ( g-l ). Bright field images verify that the same area was viewed from either side ( b and h ). Colored bright field images were mixed with images of F actin staining to indicate the position of pigment granules ( d , j , f and l ). Focus was varied to either visualize the marginal cell barrier (SMC, c-d and i-j ) or the basal cell barrier (BC, e-f and k-l ). Both the marginal cell ( e-f ) and the basal cell barrier ( i-f ) appeared to be intact. It was critical for this finding that pigmentation did not block the path of the laser. Blockage of the laser by pigmentation produces the untrue impression of a discontinuous marginal cell barrier ( c-d ) or basal cell barrier ( k-l ). Comparison of images is aided by marking a significant area with a star . Bars = 10 μm. Figure 7 Analysis of marginal and basal cell barriers in Slc26a4 +/+ and Slc26a4 -/- mice. Tight junctions were visualized by F actin in whole-mounts of stria vascularis. Bright field images were taken to evaluate pigmentation ( a , d and g ). Note the intact marginal cell ( b ) and basal cell ( c ) barriers in Slc26a4 +/+ mice. Minimal pigmentation of Slc26a4 +/+ mice did not compromise F actin localization. Whole-mounts of stria vascularis from Slc26a4 -/- mice were viewed either from the basal cell side ( e-f ) or from the marginal cell side ( h-i ). Bright field images verify that the same area was viewed from either side ( d and g ). Focus was varied to either visualize the marginal cell barrier (SMC, b , e and h ) or the basal cell barrier (BC, c , f and i ). Both the basal cell ( f ) and the marginal cell ( h ) barriers appeared to be intact. Blockage of the laser by pigmentation produces the untrue impression of 'holes' in the marginal cell barrier ( e ) or basal cell barrier ( i ). Comparison of images is aided by marking three significant areas with colored stars . Bars = 10 μm. Pendrin-expressing surface epithelial cells in the spiral prominence region are located in an area where basal cells, which are interconnected by tight junctions, form additional tight junctions with surface epithelial cells [ 29 ]. A discontinuity of this complex junction in Slc26a4 -/- mice would explain the absence of the endocochlear potential. To evaluate the presence of this connection, we determined by confocal immunocytochemistry the expression of ZO-1 and of F-actin, which associate with tight junction complexes. ZO-1 and F-actin expression revealed a continuous layer of basal cells, including a junction of basal cells with surface epithelial cells in Slc26a4 -/- mice, as observed in normal mice (Fig. 1c,1d,1e,1f , 6 , 7 ). These observations make it unlikely that the primary cause of the loss of the endocochlear potential is a compromise in the basal cell barrier. The observation that endolymphatic and perilymphatic K + concentrations were normal in Slc26a4 -/- mice suggests that stria vascularis was able to secrete K + in spite of the apparent signs of degeneration. The rate of K + secretion necessary to maintain endolymphatic K + concentration, however, may be less than necessary in Slc26a4 +/+ mice given the reduced numbers of sensory hair cells, which provide a major pathway for K + exit from endolymph [ 30 , 31 ]. In order to substantiate the view that stria vascularis in Slc26a4 -/- secretes K + , we measured the magnitude of the bumetanide-sensitive current exiting stria vascularis across the apical membrane of strial marginal cells and determined the expression of the proteins KCNQ1, KCNE1 and SLC12A2, which are essential for K + secretion [ 20 , 32 , 33 ], and of GJB2 (Cx26), which is thought to contribute to K + cycling [ 23 ]. K + secretion is known to be sensitive to 10 -5 M bumetanide, a loop-diuretic that inhibits the Na + /2Cl - /K + cotransporter SLC12A2 [ 26 ]. Bumetanide-sensitive currents were found in both Slc26a4 +/+ and Slc26a4 -/- mice although the magnitude of the current was significantly smaller in Slc26a4 -/- mice (22 ± 6 μA/cm 2 , n = 4 versus 14 ± 2 μA/cm 2 , n = 4). KCNQ1 and KCNE1, subunits of the secretory K + channel, were co-localized in the apical membrane of strial marginal cells; the Na + /2Cl - /K + cotransporter SLC12A2, which is located in the basolateral membrane of strial marginal cells, was found in stria vascularis; and the gap junction protein GJB2 was found in spiral ligament of Slc26a4 +/+ and Slc26a4 -/- mice (Fig. 8 ). Co-localization of KCNQ1 and KCNE1 proteins was also observed in vestibular dark cells (Fig. 1i ). Expression of KCNQ1 protein in stria vascularis of Slc26a4 -/- mice is consistent with the finding of KCNQ1 mRNA expression (Fig. 3 ). These observations make it unlikely that the primary cause of the loss of the endocochlear potential is a compromise of K + secretion by strial marginal cells or a compromise of gap junction mediated K + cycling. Figure 8 Protein localization of KCNQ1, KCNE1, SLC12A2 and GJB2 in the cochlear lateral wall of Slc26a4 +/+ and Slc26a4 -/- mice. a-f: bars: 10 μm. SMC, strial marginal cells; SV, stria vascularis; BC, basal cells; SL, spiral ligament. Conclusions We described locations of pendrin expression in the cochlea and vestibular labyrinth and detected normal endolymphatic K + concentrations in spite of an enlargement of this fluid compartment. We found that Slc26a4 -/- mice lack the endocochlear potential because they do not express KCNJ10 protein. Intermediate cells protect stria vascularis by converting CO 2 to HCO 3 - and by detoxifying free radicals (Fig. 9 ). CO 2 and free radicals are generated by the large numbers of mitochondria in the metabolically highly active strial marginal cells. Intermediate cells employ carbonic anhydrase to convert CO 2 to HCO 3 - [ 18 , 19 ] and catalase to detoxify free radicals. To protect themselves from free radical damage, intermediate cells generate glutathione and melanin pigment [ 34 - 37 ]. It is conceivable that loss of pendrin, which may secrete HCO 3 - into endolymph, results in an accumulation of HCO 3 - and an alkalinization of the intrastrial spaces. This extracellular alkalinization may enhance free radical stress, since it may inhibit the uptake of cysteine and thereby limit production of the protective glutathione [ 38 ]. Support for the hypothesis of enhanced free radical stress comes from the observed hyperpigmentation in mice lacking pendrin. Strial hyperpigmentation has also been observed in other conditions that are associated with free radical stress, such as acoustic trauma [ 37 ]. Alterations in the cytosolic pH in conjunction with free radical stress may lead to the loss of KCNJ10 protein expression in strial intermediate cells. Function and expression of other K + channels has been shown to be controlled by the cytosolic pH and free radicals, which encode the metabolic state of the cell [ 39 ]. Suppression of the KCNJ10 K + channel in strial intermediate cells, which is essential for the generation of the endocochlear potential, is probably the direct cause of deafness in Slc26a4 -/- mice and patients suffering from Pendred syndrome. Figure 9 Model for the loss of KCNJ10 in the absence of pendrin expression in stria vascularis. Cys, cysteine, Glu, glutamate, Gly, glycine, CA, carbonic anhydrase, GST, glutathione-S-transferase, GSH, glutathione. Competing interests None declared. Authors' contributions PW designed and coordinated the immunocytochemical, morphometrical and molecular biological experiments. EMI and BA carried out confocal immunocytochemistry on cryosections. SJ and SVJ carried out confocal microscopy on whole-mounts of stria vascularis. SVJ and RJM carried out quantitative RT-PCR. DCM designed and coordinated electrophysiological experiments. TW carried out measurements of the endocochlear potential and the endolymphatic K + concentration. JHL carried out current measurements on strial marginal cells. SMW, LAE, IER and EDG provided the mice and the pendrin-specific antibody. PW and DCM conceived the study. PW wrote 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/PMC516044.xml

539272

Critical appraisal skills training for health care professionals: a randomized controlled trial [ISRCTN46272378]

Introduction Critical appraisal skills are believed to play a central role in an evidence-based approach to health practice. The aim of this study was to evaluate the effectiveness and costs of a critical appraisal skills educational intervention aimed at health care professionals. Methods This prospective controlled trial randomized 145 self-selected general practitioners, hospital physicians, professions allied to medicine, and healthcare managers/administrators from the South West of England to a half-day critical appraisal skills training workshop (based on the model of problem-based small group learning) or waiting list control. The following outcomes were assessed at 6-months follow up: knowledge of the principles necessary for appraising evidence; attitudes towards the use of evidence about healthcare; evidence seeking behaviour; perceived confidence in appraising evidence; and ability to critically appraise a systematic review article. Results At follow up overall knowledge score [mean difference: 2.6 (95% CI: 0.6 to 4.6)] and ability to appraise the results of a systematic review [mean difference: 1.2 (95% CI: 0.01 to 2.4)] were higher in the critical skills training group compared to control. No statistical significant differences in overall attitude towards evidence, evidence seeking behaviour, perceived confidence, and other areas of critical appraisal skills ability (methodology or generalizability) were observed between groups. Taking into account the workshop provision costs and costs of participants time and expenses of participants, the average cost of providing the critical appraisal workshops was approximately £250 per person. Conclusions The findings of this study challenge the policy of funding 'one-off' educational interventions aimed at enhancing the evidence-based practice of health care professionals. Future evaluations of evidence-based practice interventions need to take in account this trial's negative findings and methodological difficulties.

Introduction For clinicians to make sense of scientific evidence and follow an evidence-based approach to their practice it has been stated they should be able to: (1) turn problems of their clinical practice into focused questions; (2) comprehensively search for literature to address these questions; (3) critically appraise this literature for its usefulness and scientific validity; and, (4) apply the results of this appraisal to their practice [ 1 ]. McColl and colleagues undertook one of the few studies of the prevalence of critical appraisal skills (CAS). In a sample of family practitioners, it was reported that only about one third claimed they "understood and could explain to others" terms which are intimately associated with an ability to critically appraise research [ 2 ]. A number of approaches have been developed to help clinicians enhance their CAS, including the publication of a number of critical appraisal checklists and the introduction of CAS teaching into undergraduate and postgraduate education in UK and abroad [ 3 , 4 ]. In UK and abroad, the Critical Appraisal Skills Programme (CASP) has become one of the most widely been disseminated forms of CAS training [ 5 ]. Four systematic reviews have been published that explore the effectiveness of CAS training [ 6 - 9 ]. These reviews observed marked heterogeneity in the nature of education intervention across individual studies, particularly in terms of duration (which varied across studies from 1 hour or less to 10 hours or more). However, these reviews consistently reported that CAS training results in small improvements in participants' knowledge of methodological and statistical issues in clinical research and enhances their attitudes towards the use of medical literature in clinical decision making. Nevertheless these findings need to be interpreted with considerable caution as most of the studies had poor internal validity. Only one randomized controlled trial was identified [ 10 ] and, in general, studies failed to blind outcome assessment. A focus on classroom-based interventions delivered to either medical students or medical residents, also limits the generalisability of the current evidence base. The aim of this study was to undertake a randomized controlled trial to assess the effectiveness and cost of CAS training in a range of practising healthcare professionals using a range of validated outcomes. Given its wide dissemination, the CASP model of CAS was evaluated in this trial. Methods Study design The study was a prospective randomized controlled trial. Study outcomes were not assessed at baseline to avoid a pre-test effect. The possibility of a pre-assessment leading to a higher post assessment score due to an item-practice effect is well recognised in the educational evaluative literature [ 11 ]. However, trial participants' characteristics (i.e. gender, age, attitude towards the use of evidence about healthcare research, and details of previous training in research, epidemiology, or statistics) were collected by questionnaire prior to randomization and used as covariates to reduce variation from individual differences. Ethical approval for the study was obtained from all of the local district ethics committees from which the participants were drawn. Selection of subjects & setting Over a three-month period, 1,305 practitioners, working within the South and West Regional Health Authority in England, were sent an invitation to participate in one of a number of CAS workshops being run across the region. Invitations were sent to the health authority offices and all general practices in the geographical area. The letters of invitation included an explanation that agreement to take part in the workshops would include a formal evaluation. Applying to attend, which involved completion of a questionnaire with baseline questions, was taken as consent to enter the study. On receipt of a completed questionnaire, participants were randomized to either intervention or control. The intervention group were given a date to attend a CAS workshop and the control participants assigned to a waiting list to attend a workshop. The only exclusion criterion for entry into the study was attendance at a previous CAS workshop. Sample size determination The target sample size was 200, 100 in each group, which was chosen to allow the study to detect a 'moderate' effect size difference of 0.4 standard deviation units (in any outcome) at 80% power and a 5% significance level (2-tailed) [ 12 ]. Randomization and blinding An independent researcher used computer generated codes to allocate applicants randomly to intervention (attend a critical appraisal workshop) or control group ('waiting list'), stratified by occupation: manager/administrator; medically qualified practising physician; nurse/profession allied to medicine and 'other' professions. The researchers who scored study outcomes were blinded to the allocation of participants at all times. Intervention group The teaching programme used in this study was based on the Critical Appraisal Skills Programme (CASP). The half-day workshop centres upon facilitating the process by which research evidence is systematically examined to assess study validity, the results and relevance to a particular clinical scenario. Participants practise these skills, during the workshop, by critically appraising a systematic review article and then receive follow up materials following the workshop (see Appendix 1 for details of intervention). Development of outcomes Given the absence of suitable validated outcomes measures, the outcomes were developed for use in trial. A questionnaire was developed and validated (reliability and internal consistency) to assess the following outcomes – knowledge of the principles necessary for appraising evidence; attitudes towards the use of evidence about healthcare; evidence seeking behaviour; perceived confidence in appraising evidence; and, knowledge of the principles necessary for appraising evidence; attitudes towards the use of evidence about healthcare; evidence seeking behaviour; perceived confidence in appraising evidence. A copy of the outcome questionnaire can be found in Appendix 2 (see Additional file 1 ). Full details of the validation process can be found elsewhere [ 13 ]. The questionnaire included 18 multiple-choice knowledge questions, 7 attitude statements and 6 confidence statements. Possible response categories to the knowledge questions were 'true', 'false' or 'don't know'. Correct, incorrect and don't know responses were awarded scores of 1, -1 and 0 respectively. Knowledge scores across question were summed giving a possible range of scores from -18 to +18. Attitude statements were scored on a five-point Likert scale. A 'strongly agree' to a positive attitude statement or 'strongly disagree' to a negative attitude statement was given a score of 5. Conversely, a 'strongly disagree' with a positive attitude statement and 'strongly agree' with a negative attitude statement was give a score of 1. Attitude scores were summed giving a possible range of scores from 7 to 35. The 6 statements of confidence in critical appraisal skills statements were scored using a 1 to 5 Likert scale and summed. A minimum overall score of 5 indicated 'little or no confidence' while a maximum total score of 30 indicated 'complete confidence'. Critical appraisal ability was assessed through the appraisal of a systematic review article. Participants' critiques were independently assessed by two of the authors (BR & PE) using a 5-point visual analogue scale, a high score indicating a superior level of appraisal skill. A framework for scoring the reviews was developed and agreement assessed; a random sample of 20 appraisals (10 control and 10 intervention) was assessed using this framework. Intra-class correlation coefficients were calculated for each of the three aspects of critical appraisal skills assessed: 'methodology' (0.86), 'results' (0.84) and 'relevance/generalisability' (0.70), indicating satisfactory inter-assessor agreement. Assessment of outcomes Six months after the CAS workshop, the intervention group were asked to complete the outcome questionnaire and undertake the critique of a systematic review article (different to article used in the workshop). Five to six months after randomisation, and about one month prior to attending the workshop, controls were asked to complete the same outcomes. Thus, outcomes were obtained from both groups at about the same time after randomisation. Statistical analysis Primary analysis of the difference between CAS training and control groups was performed on an intention-to-treat basis, adjusting for baseline characteristics. Given that not all participants in the intervention group attended a CASP workshop, a secondary explanatory analysis was also conducted, i.e. according to whether participants received the intervention or not (see Figure 1 ). For continuous outcomes, multiple linear regression modeling was used to adjust for potential confounding arising from baseline differences in prognostic variables between groups. Regression model goodness of fit was checked by examining model residuals. Ordinal outcomes were compared by Mann-Whitney U tests, and binary outcomes were compared by Chi-squared analyses. Percentages and time variables were analysed as continuous variables. All analyses were carried out using STATA. All statistical tests used a level of significance of 0.05 and two-sided hypothesis testing. 95% confidence intervals (95% CI's) were calculated for differences between the two groups. No adjustment for multiple comparisons was made. However, all analyses were planned a priori and reported in full. Costs were analysed using recognized methods [ 14 ]. Figure 1 Flow diagram summarising participant recruitment and receipt of outcomes Cost analysis A detailed analysis of the costs of setting up and delivering the program of CAS workshops was undertaken. This cost analysis was carried out from the perspective of the NHS. Based on information about the resources and associated costs of providing the workshops, the following items were considered – costs of inviting and processing applications to attend a workshop, time of workshop organizers in the Regional R&D Office, hire of workshop venue and catering, time and expenses of workshop tutors associated with preparing and delivering the workshops, time and expenses (including locum cover) of workshop participants associated with attending the workshops. Published health and social care costs [ 15 ], local costs (e.g. NHS trust costs) and Whitley Council pay scale were used to estimate the value of staff time. Results Subject enrolment Despite intensive efforts, the trial failed to recruit the target number of individuals. A revised power calculation estimated that, at 5% significance and 80% power, the 145 participants actually recruited would enable the trial to detect a difference of 0.47 standard deviation units (~20% larger than the originally powered difference). 72 were randomized to the control group and 73 to the intervention group. A total of 61 (85%) and 44 (60%) questionnaires and 43 (60%) and 21 (29%) appraisals were returned by the control and CAS training participants respectively (see Figure 1 ). The two groups were well balanced for baseline demographic characteristics (see Table 1 ). Table 1 Distribution of baseline characteristics of health care practitioners randomized to two groups. Values are numbers (percentages) unless otherwise stated. Characteristics CAS training N = 73 Control N = 72 Sex, male 48 (65.8) 46 (63.9) Age (years) <30 2 (2.7) 4 (5.5) 30–39 20 (27.4) 20 (27.8) 40–49 37 (50.6) 32 (44.4) 50–59 12 (16.4) 13 (18.0) 60 + 2 (2.7) 3 (4.2) Access to medical library 71 (97.3) 68 (97.1) Prior experience of searching literature 47 (64.4) 45 (64.3) Received formal education* in research methods 31 (42.5) 33 (47.1) Received formal education* in epidemiology 24 (32.9) 22 (31.4) Received formal education* in statistics 36 (49.3) 39 (55.7) Prior involvement in research 50 (68.5) 41 (58.6) *: postgraduate education Study outcomes 1. Knowledge of the principles necessary for appraising evidence Participants were asked to answer six knowledge questions, each of which had three parts. The frequency of correct answers to 4 of the 6 questions was higher in the CAS training group than the control. Total knowledge score was significantly higher for the CAS training group than controls [ITT mean difference: 2.6 (95% CI: 0.6 to 4.6); explanatory analysis mean difference 3.1 (95% CI: 1.1 to 5.2)] (see Table 2 ). A difference in total knowledge score of 2.0 and 3.0 corresponds to difference of 0.2 to 0.3 standard deviation units respectively i.e. below the cut off of 0.4 standard deviations units corresponding to a 'moderate' effect size [ 12 ]. Table 2 CAS training and control groups total score for knowledge of the principles necessary for appraising evidence, attitude towards the use of evidence, perceived confidence and appraisal skill. CAS training Mean (SD) Control Mean (SD) Intention to treat analysis Mean difference+ (95% CI) Explanatory analysis Mean difference+ (95% CI) Knowledge [range -18 to 18] 9.7 (5.3) 8.0 (5.1) 2.6 (0.6 to 4.6)* 3.2 (1.1 to 5.2)* Attitude [range 7 to 35] 25.0 (3.8) 24.8 (4.0) 0.04 (-1.5 to 1.6) -0.04 (-1.7 to 1.6) Confidence [range 6 to 30] 15.0 (5.3) 13.8 (5.1) 1.4 (-0.5 to 3.3) 1.13 (-0.8 to 3.1) Appraisal skill [all range 1 to 5] Methodology 2.4 (2.5) 2.0 (2.1) 0.6 (-0.8 to 1.9) 0.6 (-0.9 to 2.1) Results 2.6 (2.8) 1.7 (1.8) 1.2 (0.01 to 2.4)* 1.1 (-0.2 to 2.4) Relevance/Generalisability 2.7 (2.2) 2.4 (1.7) 0.3 (-0.8 to 1.4) 0.6 (-0.6 to 1.8) + Adjusted for sex, age, attendance at previous educational activity, access to medical library, prior experience of searching literature, formal education in research methods and/or epidemiology and or statistics, prior involvement in research * Statistically significant at P ≤ 0.05 2. Attitudes towards the use of evidence about healthcare With the exception of a more positive response to one attitude statement (' systematic reviews play a key role in informing evidence-based decisions '), in the CAS training group compared to control there were no other significant differences between groups in attitude statements. There was no evidence of difference in overall attitude score between groups (see Table 2 ). 3. Perceived confidence in appraising a published paper There was no evidence of a statistically significant difference between groups in total confidence score (see Table 2 ). 4. Ability to appraise a systematic review There was some evidence of the ability of participants in the CAS training group to appraise 'results' of the systematic review article [ITT mean difference: 1.2 (95% CI: 0.01 to 2.4)]. However, the difference was not significant when assessed using explanatory analysis. No difference between groups was observed in the ability to appraise 'methodology' or 'relevance/generalisability' of evidence (see Table 2 ). 5. Reading and evidence seeking behaviour A comparison of various aspects of evidence seeking behaviour is detailed in Tables 3 and 4 . The participants in the CAS training group self reported to: (1) read more articles, both for keeping up-to-date and for solving healthcare problems; (2) spend less time reading professional literature for keeping up-to-date, but spend more time reading professional literature for solving healthcare problems; (3) read 'thoroughly' a higher proportion of articles; and (4) use of the Cochrane library more frequently and, (5) read research reports, textbooks and other resources less frequently for solving healthcare problems. However, with the exception of (4), none of these differences were statistically significant in comparison to control Table 3 CAS training and control groups reported number of articles read, and number of hours spent reading. CAS training Mean (SD) Control Mean (SD) Intention to treat Mean difference+ (95% CI) Explanatory analysis Mean difference+ (95% CI) No. articles looked at or read thoroughly each week for keeping up-to-date 5.7 (6.4) 5.1 (4.3) 0.9 (-0.6 to 1.2) 0.5 (-0.7 to 1.3) No. hours spent reading professional literature each week for keeping up-to-date 2.2 (1.9) 2.5 (3.9) 0.9 (-0.6 to 1.2) 0.9 (-0.6 to 1.3) No. articles looked at or read thoroughly each week to solve a health care problem 1.1 (0.8) 0.9 (0.8) 1.5 (-0.8 to 2.7) 1.4 (-0.8 to 2.7) No. hours spent reading professional literature to solve a health care problem 0.9 (0.7) 0.9 (0.6) -0.02 (-0.4 to 0.3) -0.1 (-0.5 to 0.2) Proportion of articles read thoroughly 21.9 (23.6) 19.2 (19.9) 1.3 (-0.8 to 2.0) 2.6 (-0.7 to 1.8) Proportion of articles skim read 37.0 (20.8) 42.3 (24.9) -5.7 (-15.4 to 4.1) -8.2 (-18.1 to 1.6) Proportion of articles for which only abstracts read 49.7 (23.4) 40.8 (26.7) 7.9 (-3.3 to 19.1) 12.0 (1.0 to 23.0)* + Adjusted for sex, age, attendance at previous educational activity, access to medical library, prior experience of searching literature, formal education in research methods and/or epidemiology and or statistics, prior involvement in research * Statistically significant at P ≤ 0.05 Table 4 CAS training and control groups use of the resources for solving a health care problem CAS training Median (LQ, UQ) Control Median (LQ, UQ) Median Difference (p-value) † Review articles 2.0 (1.0, 3.0) 2.0 (1.25, 2.0) 0 (0.66) Research reports 1.0 (1.0, 2.0) 2.0 (1.0, 2.0) 1.0 (0.97) Secondary journals 2.0 (1.0, 3.0) 2.0 (1.0, 2.0) 0 (0.22) Textbooks 2.00 (2.0, 3.0) 3.0 (1.0, 3.0) 1.0 (0.77) Worldwide Web 1.0 (0, 2.0) 1.0 (0, 2.0) 0 (0.98) Guidelines 2.0 (2.0, 3.0) 2.0 (2.0, 3.0) 0 (0.64) Cochrane Library 1.0 (0, 2.0) 0 (0, 1.0) 1.0 (0.05) Colleagues 3.0 (2.75, 3.0) 3.0 (2.0, 3.0) 0 (0.55) Other resources 2.0 (0, 3.0) 3.0 (1.5, 3.75) 1.0 (0.41) † Mann Whitney test; Likert Scale: '0': 'never'; '1': 'rarely'; '2': 'occasionally'; '3': 'often' & '4': 'very often'; UQ: upper quartile; LQ: lower quartile Costs The mean cost to the NHS of conducting the CAS workshops was £250 per person (see Table 5 ). The majority of this cost (approximately £140) resulted from salary costs associated with the time of the participants attending the workshop. The remaining costs of the workshops were associated with the administration (approximately £25 per person), venue hire (approximately £42 per person), and tutors' time and travel (approximately £49 per person). There was some variation in the cost (from approximately £240 – £340 per person) across the 7 workshops, due to the attendance level, i.e. workshops with the most participants tended to have the lower cost. Table 5 Summary of costs of CAS training Workshop I II III IV V VI VII Total Total per Head No. Attendees 16 19 9 14 18 5 7 88 Administration* £322 £329 £313 £318 £316 £295 £300 £2,193 £25 Venue £1028 £670 £285 £916 £475 £215 £74 £3,663 £42 Participants' costs† £2,074 £2,825 £1,567 £1,878 £2,179 £793 £992 £12,310 £140 Tutors' costs £709 £719 £890 £712 £555 £636 £73 £4,294 £49 Total £4,132 £4,542 £3,057 £3,824 £3,525 £1,940 £1,439 £22,460. £255 Total per head £258 £239 £340 £273 £196 £388 £206 £255 * Costs of initial invitations, copyright permission for use of paper, invitations to participants, pre workshop & post workshop pack production, postage costs, preparation for workshop, and travel bookings. † Costs included participants' time at workshop with the exception of GPs, in these cases the cost of locum cover was applied. Discussion The results of this prospective randomized controlled trial demonstrates that a half-day CAS workshop can elicit small improvements in healthcare professionals' knowledge of the principles and theory of evidence-based practice and some improvement in aspects of their critical appraisal skills ability. Nevertheless, we found little evidence of any improvement, as a result of CAS training, in the other study outcomes, i.e. participants' attitude towards evidence or their evidence seeking behaviour. Taking into account the set up costs and of time and locum expenses of participants, the mean cost of conducting these CAS training workshops was about £250 per person. The lack of substantive improvements in knowledge, skills and attitudes outcome observed in this trial are consistent with previous studies of CAS training [ 6 - 9 ]. Potential limitations of this study The number of participants recruited was less than that intended, not all participants provided outcomes and the trial was about 20 percent under the desired power. Nevertheless this study remains the largest randomized controlled trial to date and some statistically significant differences were observed. The educational context in which this randomized trial was undertaken imposed certain constraints on its conduction and execution. As a result, poor recruitment, loss to follow up and poor uptake of the CAS training experienced by this trial may have threatened both its internal validity and generalisability. However, efforts were made in the analysis of the findings of this trial to overcome these limitations. The return of outcomes in this trial could not be mandatory. Despite considerable efforts by the project team (reminders and personal telephone calls from the trial principle investigator to participants), we failed to obtain a substantial proportion of outcomes in the trial participants – 60% and 85% of the knowledge, attitude and behaviour outcomes were obtained for CAS training and control groups respectively, and even less for the critique of the published systematic review. It is plausible that respondents may have differed in some way to non-respondents, such as in their level of motivation, and may therefore responded more positively to this educational intervention. However, this was not supported by the poor outcome response rate. Moreover there was no evidence of a difference in the baseline characteristics of participants who returned their outcomes, and those who did not. A differential response rate across the two study groups possibly reflects a greater reluctance in those individuals who had undertaken the educational intervention to return their outcomes (i.e. 'more to lose') compared to those in the control group. If true, the direction, in terms of over- or underestimating the impact of the intervention, is uncertain. An interview-administered assessment, rather than a mail based one, may have enhanced outcome response rate. Of the 73 participants allocated to receive CAS training only 52 (71%) actually attended. The reasons for this were unclear, and were not formally addressed within this study. In addition to conventional intention-to-treat analyses, secondary explanatory analyses, i.e. based upon the participants who actually did attend the workshop, were undertaken. That there were no differences between groups for most outcomes, irrespective of whether an intention-to-treat or explanatory analysis, was used (see Tables 2 and 3 ) suggests that the poor intervention uptake was not important source of bias. Implications of findings With the drive to evidence-based practice in recent years, considerable efforts have been made in providing CAS training as part of healthcare professionals' undergraduate and postgraduate activities in many countries. The findings of this study, the largest randomized controlled trial to date, provide only limited support for such training. However, it is important to put this finding in the appropriate educational context. The half-day CASP workshop evaluated in this trial has been widely disseminated and its duration and format is consistent with many previous CAS interventions [ 9 ]. Nevertheless it is probably unrealistic to expect that the half-day workshop evaluated in this trial would in itself result in changes in professional behaviour. This is supported by a large body of evidence and theory on changing professional practice [ 17 ]. Therefore it is important to see, and assess, CAS training, not in isolation, but as one part of education approach towards evidence-based practice or as a part of the undergraduate and postgraduate curriculum. It is also important to reassess the objective of CAS training. With increasing availability of carefully appraised evidence such as secondary journals (e.g. Evidence Based Medicine ) and on-line critically appraised topics ('CATs'), the most important role of CAS training may be simply be to sensitise participants to the availability of high quality evidence. Further debate is therefore needed about refocusing critical appraisals skills training towards finding such evidence and the role of healthcare librarians and the new initiatives such as the National Electronic Library for Health. A number of commentators have criticised previous evaluations of CAS training for not using experimental designs [ 6 - 9 ]. However, the experience of this study has demonstrated some of the difficulties in implementing an evaluation of 'real life' educational intervention using such an experimental design. The difficulty of employing randomized controlled trials in the evaluation of educational interventions has been highlighted by others [ 18 ]. Future evaluations of CAS and other educational interventions aimed at promoting evidence-based practice need to take into account both these perspectives. Conclusions This prospective randomized controlled found small improvements in self-selected healthcare professionals' knowledge and understanding of the medical literature and appraisal skills with critical appraisal skills training. No improvement was observed in attitudes towards the use evidence and evidence-seeking behaviour. The findings of this study challenge the policy of funding in isolation 'one-off' educational interventions aimed at enhancing the evidence-based practice of health care professionals. Future evaluations of evidence-based practice interventions need to take in account both this trials' negative findings and methodological difficulties. List of abbreviations CAS – critical appraisal skills CASP – Critical Appraisals Skills Programme 95% CI – 95 percent confidence interval ITT – intention to treat NHS – National Health Service R&D – research and development Competing interests The author(s) declare that they have no competing interests. Author's contributions RT, BR and PE conceived, designed and secured funding for the trial. RST drafted the paper. RJT collected the study outcomes and undertook the data analysis. RST is a guarantor for the study. Funding NHS R&D Executive: Evaluating methods to practice the implementation of R&D [project no. IMP 12-9] Appendix 1. Objectives, syllabus, and delivery methods of critical appraisal skills workshop for health care decision makers Workshop objectives (taken from workshop materials) • To critically appraised a published review article. • To understand the terms systematic review and meta-analysis. • To be able to explain why critical appraisal skills are important for provision of health care. • To have greater confidence in your ability to make sense of the research evidence. Workshop format 3 hours attendance (also advised to undertake at least 1 hour preparation reading the article to be appraised in the workshop and address a written 'clinical scenario') • Introductory talk: overview of the importance of evidence based health care practice, the theoretical basis of the appraisal of a systematic review, and orientation to the JAMA appraisal guideline (~60 mins). • Small group work: appraisal of a published systematic review (~60 mins). • Plenary session: feedback from the small group, general discussion of the relevance of the appraisal to clinical scenario and ballot of opinions on the clinical scenario. (~60 mins) All workshops were run by 3 to 4 individuals each of whom had a formal training in health services research methods and were experienced in delivering CASP workshops. Workshop materials One to two weeks prior to the workshop, a pre-workshop pack was sent to participants. • Workshop objectives. • Orientation guide. • Clinical scenario and questions • Systematic review paper. • Glossary. One to two weeks post workshop, a post workshop pack was sent to participants: • Introductory talk slides. • Systematic review checklist. • JAMA guidelines for systematic review [ 15 ]. Educational rationale The workshop is based on the Critical Appraisal Skills Programme (CASP) developed by Oxford Regional Health Authority and developed from the educational methods of McMaster University in Canada [ 5 ]. The 'McMaster model' key features include, self-directed learning, small group teaching methods and the importance of grounding education within the clinical decision making process. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 MS-Word format, contains study outcome questionnaire Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539272.xml

526193

The impact of employee level and work stress on mental health and GP service use: an analysis of a sample of Australian government employees

Background This study sought to identify the extent to which employee level and work stressors were associated with mental health problems experienced by Australian government employees, and with their use of primary care services. Methods 806 government employees aged between 40 and 44 years were surveyed as part of an epidemiological study conducted in Australia. Data collected from participants included sociodemographic attributes, physical health, psychological measures and work stressors relating to job control, job demands, job security and skills discretion at work. For 88% of these participants, information on visits made to general practitioners (GPs) for the six months before and after their survey interview was obtained from health insurance records. Results When work stress and personal factors were taken into account, men at more junior levels reported better mental health, more positive affect and used fewer GP services. Women at middle-management levels obtained less GP care than their more senior counterparts. Both men and women who reported higher levels of work stress were found to have poorer mental health and well-being. The impact of such stressors on GP service use, however, differed for men and women. Conclusion Measures of work stress and not employee level affect the mental health and well-being of government employees. For governments with responsibility for funding health care services, reducing work stress experienced by their own employees offers potential benefits by improving the health of their workforce and reducing outlays for such services.

Background In 1999, the World Health Organization reported that workers continued to suffer high levels of work-related injuries and deaths [ 1 ]. It also flagged, however, the increase in mental health problems reported by workers in industrialized countries as a result of their experiencing psychological stress and excessive job demands in the workplace [ 1 ]. The health consequences of such psychosocial aspects of the work environment have been examined in a range of settings across different countries. Much of this research has drawn on the model developed and refined by Karasek who proposed that work-related mental strain and associated psychiatric disorder result from combinations of, and interactions between, four different employment factors: heavy job demands, limited input to decision making processes, lack of skill discretion within the job and poor work-based social support [ 2 , 3 ]. Such factors, in particular those concerning decision making, skill discretion and social support have been found to be most problematic for those in lower grades of employment and to be less prevalent among employees in higher ranking positions [ 4 - 6 ]. The applicability of this model for the government sector is well supported by cross-sectional and longitudinal studies drawing on the Whitehall II study of a large cohort of 10,308 London-based government employees. Again, such studies have found that those in lower grades report that they have less job control, less variety in their work, and less job satisfaction [ 7 ]. Those reporting higher levels of such work stress have also been found to have greater risk of cardiovascular health problems [ 8 ] and poorer psychiatric health [ 5 , 9 , 10 ]. There has been little research undertaken on the health impact of job level and work stressors for government employees in Australia. An earlier study that explored the relationships between work stressors and blood pressure in Australian government employees, found chronic perceived work stress to be associated with blood pressure change [ 11 ]. The impact of job level and work stress on Australian government employees' mental health has not been previously explored. We have been able to explore these issues using data collected from 806 government employees who participated in the PATH Through Life Project, a large community-based study being conducted by the Centre for Mental Health Research in Canberra, Australian Capital Territory (ACT). Survey participants provided information on sociodemographic measures, mental and physical health, employment level and work-related stress. For 88% of these participants, independently collected information on their use of general practitioner services was also available. These data have allowed us to examine the impact of employment level and work-related stress on Australian government employees' mental and physical health, their psychological well-being, and also their use of general practitioner care. We hypothesised that those working in lower level government positions would report higher levels of work-related stress, that they would be found to have more mental and physical health problems and that they would use higher numbers of primary medical services. Methods Subjects The PATH Through Life Project is a longitudinal study of individuals living in the community with participants being drawn from three age groups: 20–24, 40–44 and 60–64 years. Those in the age group of interest for this study were aged from 40 to 44 years on 1 January 2000 and drawn from the Australian Electoral Rolls for Canberra in the Australian Capital Territory and adjacent town of Queanbeyan in New South Wales. Enrolment on these rolls is compulsory for all Australians aged 18 and over. Potential participants were drawn from a 10-year age range, the minimum range then released for research purposes by the Australian Electoral Commission. The number of potential participants found, and in the required age range, was 3919, of whom 2530 participated in the survey, giving a response rate of 64.4%. Canberra is the national capital of Australia and many Australian Government entities are based in the ACT, including both houses of parliament, and the 16 major agencies that currently comprise the Australian Public Service. In this study, 806 respondents aged between 40 and 44 reported that they worked in office-based government administrative positions, developing and implementing government policy. As well as providing information on labour force status and the type of position held, respondents who worked in government positions were specifically asked to provide details on the level of the position they occupied. Five mutually exclusive employee categories were formed by grouping together those whose levels of employment were broadly comparable as follows. Employees who occupied positions at Australian Public Service (APS) Levels One to Four were grouped together as Junior employees and those at both the APS Levels 5 and 6 were classified as belonging to the Intermediate category of employees. Employees in the next two classifications of the APS (Executive Levels 1 and 2) were allocated to separate groups, Senior 1 and Senior 2. While those in Executive Level 1 positions develop policy and implement government programs, those at the Executive 2 level are primarily managers with direct responsibility for managing a number of employees from APS Level 1 to Executive Level 1 [ 12 ]. Finally, all respondents in the Senior Executive Service of the government were allocated to one category, Executive. The number of participants in each of these five categories of employees is given in Table 1 , together with a short description of the positions included in each of those categories. Table 1 Descriptions of government employee categories Position level in Australian Public Service (APS) Description of positions covered by these levels Employee category Number, % of participants % female Mean years of education APS Level 1 APS Level 2 APS Level 3 APS Level 4 Work is always supervised; can include: drafting correspondence, organising travel, filing, other routine clerical work. Junior 123 15.26 74.80 13.46 APS Level 5 APS Level 6 Work includes: supervising junior staff, liaising with external bodies, supporting project managers, drafting complex correspondence and policy papers, undertaking research. Inter-mediate 215 26.67 51.63 14.87 APS Level 7 Work includes: managing government programs and contracts, supervising staff, preparing high level policy advice, developing and implementing government policy Senior 1 220 27.30 37.73 15.69 APS Level 8 Work includes: managing a Section of staff, providing policy, financial, or administrative advice to government, representing department at external meetings. Senior 2 193 23.95 33.16 16.02 Senior Executive Service Responsible for: overall management of large numbers of staff; achieving government objectives through development and implementation of innovative and financially sound policy. Executive 55 6.82 30.91 16.58 Total 806 100.0 Mean measure 45.53 15.27 Measures Survey participants completed a questionnaire that included socio-demographic characteristics and measures of physical and mental health, and well-being. Participants in the workforce were asked 22 questions relating to their work situation. These questions matched those used in the Whitehall II study [ 7 , 13 ]. Nine questions related to job control, and reflected the amount of authority the worker has over decision-making [ 9 ]. Four questions concerned the manageability of job demands; the extent to which the worker is faced with difficult time and workload pressures and conflicting demands. Finally six questions addressed skill discretion and related to the variety of tasks to be done and the breadth of skills needed to undertake those tasks. For each of these 19 questions, respondents could answer: often, sometimes, rarely or never. Responses were given values of 1 to 4 with the highest score allocated to the less stressful work circumstances: those in which the individual had more job control, more manageable job demands, and higher levels of skill discretion. Participants were also asked the number of hours they usually worked per week and their assessment of their employment security and future employment opportunities. Answers to the last two questions used four point Likert-type scales and, again, were coded to give higher scores to those who reported that they had a more secure position or could obtain another job relatively easily. The mean of these two scores was used as an overall measure of job security. Socio-demographic measures used in these analyses included sex, age, years of education, level of household responsibilities, and experience of any of six life events during the previous six months. Since each of these factors has the potential to modify an individual's mental health independently of their work stress, we adjusted for these in our final analyses. Scores for level of household responsibility were drawn from participants' responses to questions concerning the extent to which, in their household, they were responsible for four areas: household tasks, childcare, financial management, and providing money. Comparable scores for participants who did not have children in their households were then derived by calculating the mean of their measures for household tasks, financial management and providing money, and adding this to their total household responsibility score. Scores for these measures could range from zero to 16 with higher sores representing more household responsibilities. Health measures obtained from participants and used in these analyses included: scores on Goldberg's depression and anxiety scales [ 14 ] and state measures of positive and negative affect using the Positive and Negative Affect Scales (PANAS) [ 15 ]. Measures of self-rated health, mental and physical health were taken from respondents' answers to the Medical Outcomes Study 12-item Short-Form Health Survey (SF-12) [ 16 ]. The first of these is measured by a single question in which participants rate their health as excellent, very good, good, fair or poor with higher scores indicating poorer self-rated health. Records on participants' visits to general practitioners (GPs) were also obtained. In Australia, the costs of most health care visits made to medical practitioners by Australians with citizenship or residency status are subsidised, either partly or totally, through the Australian Government funded universal health insurance scheme, Medicare. Information on the number of such visits is collected by the Health Insurance Commission. These data are used for administrative purposes and identify general practitioner and specialist services, but not the health problems addressed during each visit. While these records cover most visits made to general practitioners, they will not include a small number of services, paid for by patients but not claimed against Medicare. All participants were asked if they would consent to the researchers being provided information on the number of visits they made to general practitioners for specific periods before and after their interview. 709 (88.0%) of the 806 participants consented to this request and information on the number of GP visits they made during the six months preceding and the six months following their PATH interview was obtained from the Health Insurance Commission. Statistical analyses Analyses of variance (ANOVAs) were first undertaken to examine the extent to which sociodemographic measures and work stress measures changed with level of employment. Similar analyses, conducted separately for men and women, then compared mean mental health measures across the five employee levels. Finally, regression analyses were used to examine the contribution of employee category and work attributes in explaining participants' health and health service measures, whilst controlling for the following possible modifying factors: participant's age, years of education, level of household responsibilities and life events experienced in the past six months. For these analyses, categorical variables identifying each of the five government employee categories were created and the first four of these included in the regression equations, taking the most senior category, Executive, as reference group. After initial testing indicated that two dependent variables, negative affect and use of GP services, were not normally distributed and more closely fitted the negative binomial and Poisson distributions respectively, analyses of these two measures used negative binomial and Poisson regressions respectively. Strength of associations between dependent variables and predictor variables were measured using R 2 for linear regressions and Incidence Rate Ratios when the Poisson or negative binomial regression model was used. Incidence Rate Ratios (IRRs) are interpreted in a similar manner to odds ratios and represent the expected change in the dependent variable in response to one unit change in the predictor variable. The contribution of employee level and work stress measures in explaining variation in health measures was also obtained. This contribution was measured using change in R 2 for linear regressions and the change in the Chi-square estimate of the fit of model for the negative binomial and the Poisson regression analyses. A final analysis examined the impact of employee category and work stressors on use of GP services, taking into account demographic, lifestyle and health measures. Analyses were undertaken using SPSS 11.5 and STATA 7 [ 17 ]. Results Across the five categories of government employees, there was no significant difference in education level or in numbers of life events experienced in the past six months. Employees working at higher levels, however, reported lower levels of household responsibility and had more opportunities to develop and use different skills in their work, reported more job control and felt more secure in their current jobs. As expected, those in more senior positions also had less manageable job demands and worked longer hours. Analyses were then performed, separately for men and women, to examine differences in measures of mental health, well-being and GP service use across the five government employee categories. Level of physical health, as measured by the SF-12, was also examined for comparative purposes. The only measure to differ significantly across employee categories was level of positive affect in the past four weeks (Table 3 ). For both men and women, those in higher categories recorded higher scores on this measure. Table 3 Mean health scores by government employee category Government employee category: Junior Intermediate Senior I Senior 2 Executive P Men Self-rated health 2.48 (2.17–2.80) 2.31 (2.14–2.48) 2.26 (2.11–2.42) 2.25 (2.10–2.40) 2.08 (1.82–2.34) 0.42 SF-12 mental health 51.91 (49.60–54.22) 50.92 (49.17– 52.66) 49.35 (47.64– 51.07) 50.52 (49.13– 51.91) 49.71 (47.50– 51.92) 0.49 SF-12 physical health 52.91 (50.91–54.92) 51.60 (50.25–52.95) 52.56 (51.36–53.76) 52.83 (51.81–53.86) 54.65 (52.47-56.83) 0.18 Goldberg Depression Score 2.50 (1.61–3.39) 2.09 (1.64–2.53) 2.17 (1.75–2.59) 2.15 (1.80–2.50) 1.45 (1.01–1.88) 0.35 Goldberg Anxiety Score 3.30 (2.27–4.33) 3.12 (2.58–3.65) 3.18 (2.69–3.66) 3.65 (3.22–4.08) 3.13 (3.22–4.08) 0.54 Positive affect 31.53 (28.89–34.18) 30.25 (28.91–31.59) 31.65 (30.46–32.84) 32.87 (31.70–34.04) 34.29 (32.74–35.84) 0.01 Negative affect 16.93 (14.98–18.89) 16.25 (14.96–17.54) 16.31 (15.34–17.29) 16.31 (15.32–17.30) 15.45 (14.00–16.90) 0.88 GP services used 2.86 (1.54–4.17) 3.24 (2.39–4.08) 3.12 (2.45–3.79) 2.14 (1.97–2.85) 2.30 (1.37–3.23) 0.33 Women Self-rated health 2.36 (2.17–2.55) 2.24 (2.09–2.39) 2.13 (1.94–2.32) 2.14 (1.90–2.38) 2.35 (1.81–2.90) 0.42 SF-12 mental health 47.98 (45.52–50.30) 48.12 (45.95–50.30) 48.40 (46.19–50.61) 51.11 (49.18–53.03) 52.60 (49.48–55.71) 0.91 SF-12 physical health 51.68 (50.12–53.24) 52.55 (51.10–53.99) 52.29 (50.58–54.01) 51.11 (49.18–53.03) 52.60 (49.48–55.74) 0.76 Goldberg Depression Score 2.76 (2.20–3.32) 2.58 (2.13–3.02) 2.35 (1.84–2.86) 2.02 (1.46–2.58) 2.24 (1.12–3.35) 0.39 Goldberg Anxiety Score 3.96 (3.37–4.54) 3.78 (3.28–4.29) 4.01 (3.39–4.63) 3.38 (2.70–4.05) 3.41 (2.21–4.62) 0.62 Positive affect 29.59 (28.14–31.03) 31.78 (30.63– 32.94) 32.05 (30.72– 33.37) 32.30 (30.68–33.91) 32.59 (30.30–34.87) 0.03 Negative affect 17.89 (16.12–19.66) 17.31 (16.05–18.57) 17.70 (16.16–19.24) 16.64 (14.76–18.52) 17.59 (14.70–20.48) 0.87 GP services used 4.21 (3.36–5.06) 5.11 (4.03–6.19) 4.47 (3.33–5.62) 3.15 (2.35–3.96) 4.73 (2.33–7.13) 0.16 We next used regression analyses to examine the impact of employee category and work stress on participants' measures of health and well-being and on their use of GP services. In preliminary testing, we found that the two measures – job control and skill discretion – both contributed significantly and independently to mental health measures, hence these measures were not combined but included separately in the analyses. The results of these analyses for men are in Table 4 and for women in Table 5 . After controlling for socio-demographic and work stress measures, men in the lowest levels of employment reported significantly better mental health as measured by the SF-12 Mental Health score, higher levels of positive affect and used fewer GP services. Other measures of mental health, including self-rated health, and symptoms of anxiety and depression, did not vary significantly with employee level. Men with more manageable job demands reported better mental health, fewer depressive and anxiety symptoms and less negative affect. For men, there was a consistent association between less work stress and better health. Those with more job security or higher levels of skill discretion reported significantly better self-rated health, mental health, fewer depressive and anxiety symptoms, more positive and less negative affect and also used fewer GP services. However, those who worked fewer hours per week made more visits to GPs. Table 4 Associations between health measures, and government employee category and work stressors – men Health measure Predictor variables: Self-rated health SF-12 Mental health SF-12 Physical health Goldberg Depression Goldberg Anxiety Positive Affect Negative Affect GP services obtained Beta c Beta c Beta c Beta c Beta c Beta c Incident Rate Ratio a Incident Rate Ratio a Government employee category Junior -0.033 0.161* -0.014 0.029 -0.038 0.137* 1.011 0.492*** Intermediate -0.002 0.115 -0.183 0.096 0.021 -0.034 1.029 0.742 Senior 1 0.017 0.027 -0.151 0.141 0.040 -0.032 1.049 0.864 Senior 2 0.036 0.089 -0.120 0.112 0.078 -0.008 1.032 0.797 Work stress measures Job control -0.060 0.031 0.019 -0.022 -0.016 0.082 0.966 0.808* Manageable job demands 0.010 0.150** 0.037 -0.145** -0.245*** 0.019 0.879*** 1.018 Usual hours per week 0.066 -0.047 -0.013 0.103 0.050 0.009 1.002 0.982*** Job security -0.180*** 0.193*** 0.007 -0.171** -0.154** 0.125** 0.942* 0.868** Skill discretion -0.171** 0.302*** 0.024 -0.247*** -0.237*** 0.398*** 0.792*** 0.766** Δ R 2 attributable to level & work stress 0.085*** 0.176*** 0.016 0.134*** 0.152*** 0.228*** 75.77 b *** 91.20 b *** All analyses controlling for age, education, household responsibilities, and life events in past 6 months a Incidence Rate Ratio from negative binomial or Poisson regression b Chi-square estimate of the improvement in fit of regression model due to employee level and work stress factors. c Standardised Beta * p < 0.05; ** p < 0.01; *** p < 0.001 Table 5 Associations between health measures, and government employee category and work stressors – women Health measure Predictor variables: Self-rated health SF-12 Mental health SF-12 Physical health Goldberg Depression Goldberg Anxiety Positive Affect Negative Affect GP services obtained Beta c Beta c Beta c Beta c Beta c Beta c Incident Rate Ratio a Incident Rate Ratio a Government employee category Junior -0.178 0.044 -0.015 -0.041 0.057 -0.030 0.970 0.807 Intermediate -0.138 -0.022 0.000 0.044 0.113 0.013 1.014 1.015 Senior 1 -0.156 0.005 0.011 0.008 0.104 0.009 1.022 0.928 Senior 2 -0.073 0.066 -0.119 -0.043 0.003 0.008 0.914 0.737* Work stress measures Job control -0.049 0.120* 0.015 -0.099 -0.098 0.027 0.826*** 0.864* Manageable job demands -0.123* 0.173** 0.141* -0.245*** -0.265*** 0.146* 0.852*** 1.213** Usual hours per week 0.004 -0.086 0.058 0.043 -0.023 -0.010 1.002 1.001* Job security -0.054 0.100 -0.076 -0.101 -0.095 0.159** 0.925 1.018 Skill discretion -0.157* 0.139*** 0.061 -0.310*** -0.111 0.273*** 0.967** 0.994 Δ R 2 attributable to level & work stress 0.037* 0.083 0.033 0.150*** 0.093*** 0.121*** 68.60 b *** 27.52 b ** All analyses controlling for age, education, household responsibilities, and life events in past 6 months a Incidence Rate Ratio from negative binomial or Poisson regression; b Chi-square estimate of improvement in fit of the regression model due to employee level and work stress factors. c Standardised Beta * p < 0.05; ** p < 0.01; *** p < 0.001 For women, we found no effect of employee level on their measures of mental or physical health. However, employment level was associated with GP service use with those in middle management positions being less likely to have obtained GP care, compared with those at the executive level. Women's levels of mental health and well-being were better when they worked in a job that offered higher levels of skill discretion. Manageability of job demands impacted on all health measures considered while those who worked longer hours were more likely to have obtained GP services. Finally, we explored the contributions of employment level and work stress factors in explaining participants' use of GP services, after including in the model SF-12 measures of mental and physical health that could also contribute to explaining such service use. Again, men and women were considered separately. Overall, controlling for mental and physical health in addition to sociodemographic measures had little impact on our findings. Men in lower employee categories continued to use significantly fewer GP services compared with their counterparts in executive positions while those with less job control and less job security again obtained more care. After controlling for mental and physical health, women in middle management levels again used fewer GP services compared with those at the executive level. Similarly, women with more manageable job demands and those working longer hours continued to be higher users of GP care. Discussion This paper has reported on the associations between categories of employee levels, work stressors and mental health measures of 806 government employees aged between 40 and 44 who participated in the PATH Through Life Project being conducted in Canberra, Australia's national capital. Impact of employee level on work stressors, health and GP service use Our first hypothesis, that government employees working at lower levels would report higher levels of work-related stress, is supported in the main by our research. Overall, participants at more senior levels reported that they had more control over aspects of their work, greater opportunities to do interesting work using a range of skills, more job security but also that they were subject to higher job demands. These results closely match those reported by Marmot and colleagues in their 1991 study of over 10,000 British civil servants aged between 35 and 55 [ 7 ]. We had hypothesized that those in lower level positions would have poorer physical and mental health compared with more senior staff. None of our results supports this hypothesis. Although both male and female participants in the higher grades reported better well-being as measured by higher positive affect, we found no mental health benefit associated with having a more senior position. Furthermore, for men, we found this result to be reversed when the analyses controlled for work stress factors. Men in lower level positions reported higher levels of positive affect and better mental health, as measured by the SF-12. Women's mental health and negative affect were not affected by their employee level both when this factor was considered alone and when work stress factors were taken into account. We were unable to replicate the finding by Marmot and colleagues that those in higher positions had significantly better physical health compared with their subordinates. This difference in our results might be explained by our having a smaller sample. However, it also indicates that while those working at lower levels are more likely to experience some types of work stressors, working at those lower levels, per se , is not automatically associated with poorer mental or physical health for Australian government employees. While these findings are unusual, they do align with some previous research undertaken in the UK showing that those working in lower grades had better mental health [ 18 ]. One possible interpretation of our results is that men and women of this age group who have continued to work at lower employee levels may be pursuing satisfying goals in other areas of their lives, for example, family, outside business, recreational pursuits. Of course, confirmation of such an interpretation would require more detailed information from participants concerning their working arrangements and life priorities. We found government employee level affected men's use of GP services but the direction of this effect was the reverse of that hypothesised. After taking into account education, other personal measures and levels of work stress, men at lower levels of the public service used fewer GP services than their superiors. This aligns with the previous finding that men at lower level positions also had better mental health [ 18 ] and suggests that further research is needed on the health drawbacks for men of their rising through levels of government employment. We also found that women in middle-management levels were less likely to have obtained GP services compared with their more senior counterparts. The impact of work stress on health and GP service use Work stress factors experienced at all levels of employment played a more significant role in affecting participants' health and well-being. For male and female participants, those whose job demands were more manageable had significantly better mental health. This measure of work stress was particularly important for women, for whom manageability of demands was significantly associated with all health measures. Correspondingly, level of skill discretion had a greater impact on men's mental health. This last finding concurs with a recent Australian study of government employees which reported that, for men, having interesting work was an important reason for their staying in the government sector whereas women saw their relative job security as the advantage of this form of employment [ 19 ]. Since previous research has linked job insecurity and poorer mental health [ 20 , 21 ], we expected similar findings in this current study. Women with less job security had less positive and more negative affect. Men with less job security, on the other hand, scored significantly worse on all six mental health measures. This finding, with that of the Australian study reported above, suggests that, while women place value on security, this and other work attributes may play a less important role in affecting their overall mental health, compared with their male counterparts. Previous studies have also reported sex differences concerning the mental health consequences of job insecurity [ 22 ]. Job security has also previously been found to affect the physical health of men more than women [ 23 , 24 ]. However, the one measure of physical health used in this study was not directly affected by job insecurity for men or women. A number of work stress factors contributed significantly to explaining men's and women's use of GP care. Women who used more services had less job control while men were more likely to visit a GP if they reported less skill discretion, job control or felt more insecure about their job. Research reported in the 1980's also found that factory workers with higher job insecurity used more health services [ 25 ]. This finding in our study of government employees suggests that job insecurity has an important effect on men's, but not women's, self-assessment of their levels of health and well-being. Women who worked longer hours used slightly more GP services and also had more manageable job demands. Men who worked fewer hours, however, obtained more GP care. These findings indicate that time spent at work is unlikely to have deterred women from addressing their health care needs. For men, on the other hand, working fewer hours per week might have given them needed, additional opportunity to set aside time for a GP visit. In Australia, many GP surgeries are open during standard working hours and offer only limited access outside of these times [ 26 ]. Limitations This study is limited by having access only to cross-sectional data, since the direction of relationships between psychological stress and work stressors cannot be confirmed. It may be that those with fewer symptoms of depression or anxiety, for example, also have more positive views about the current or potential benefits of their positions. A number of issues raised here may become clearer when data from the next wave of the PATH project are collected. Conclusions These findings have implications for governments in their role as major employers. Large organisations in the public and private sectors inevitably develop hierarchical systems of employment as efficient mechanisms for allocating work functions and responsibilities. This study suggests that employees' health is less affected by their position within this structural hierarchy and is more associated with various work stressors that can be experienced by individuals across all levels of employment. For a large employer, reducing the impact of work stress on its workforce may be beneficial, not only for individual employees, but also for the productivity of the organisation as a whole. These findings also have implications specific to the Australia setting where GP services are provided and subsidised through the universal health insurance system, Medicare. This component of health care is the financial responsibility of the Australian Government, which is also the employer of the majority of participants in this current study. Initiatives aimed at reducing work stress experienced by government employees, and correspondingly, the numbers of GPs services obtained by that workforce, might prove to be a judicious use of Australian Government resources. Such potential benefits may well apply to other governments that have responsibility for funding health care services. Competing interests The authors declare that they have no competing interests. Abbreviations ACT Australian Capital Territory APS Australian Public Service GP General practitioner IRR Incidence rate ratio PANAS Positive and Negative Affect Scale SF-12 12-item Short Form Health Survey SPSS Statistical Package for Social Scientists Authors' contributions RP contributed to the conception of this study, and was responsible for all analyses of data, interpretation of results and writing of successive drafts of the paper. AFJ was leader in designing and running the PATH Through Life Project, contributed to interpretation of findings and edited successive drafts of the paper. HC was co-leader in designing and running the PATH Through Life Project, contributed to interpretation of findings and edited successive drafts of the paper. DB, LS and RD contributed to the conception of the study and editing successive drafts of the paper. All authors read and approved the final version of the manuscript. Table 2 Mean sociodemographic measures, work stress and working hours by government employee categories Government employee categories Junior Intermediate Senior 1 Senior 2 Executive P Years of education 13.46 14.87 15.69 16.03 16.58 0.37 Number of life events in past 6 months 1.02 0.95 0.92 0.89 0.73 0.59 Level of household responsibility 9.90 9.80 9.66 9.57 8.42 0.01 Job control score* 2.91 3.12 3.25 3.29 3.35 <0.01 Manageable job demands score* 2.33 2.28 2.03 1.81 1.63 <0.01 Skill discretion score* 2.76 3.13 3.35 3.42 3.54 <0.01 Job security* 2.64 2.76 2.78 2.89 2.91 0.02 Number of hours worked per week 36.78 39.20 43.20 48.06 53.73 <0.01 * Individuals' responses were scored from 1 to 4. Higher scores were given to work circumstances in which the individual had more job control, more manageable job demands, higher skill discretion and greater job security. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526193.xml

526187

Integrating linkage and radiation hybrid mapping data for bovine chromosome 15

Background Bovine chromosome (BTA) 15 contains a quantitative trait loci (QTL) for meat tenderness, as well as several breaks in synteny with human chromosome (HSA) 11. Both linkage and radiation hybrid (RH) maps of BTA 15 are available, but the linkage map lacks gene-specific markers needed to identify genes underlying the QTL, and the gene-rich RH map lacks associations with marker genotypes needed to define the QTL. Integrating the maps will provide information to further explore the QTL as well as refine the comparative map between BTA 15 and HSA 11. A recently developed approach to integrating linkage and RH maps uses both linkage and RH data to resolve a consensus marker order, rather than aligning independently constructed maps. Automated map construction procedures employing this maximum-likelihood approach were developed to integrate BTA RH and linkage data, and establish comparative positions of BTA 15 markers with HSA 11 homologs. Results The integrated BTA 15 map represents 145 markers; 42 shared by both data sets, 36 unique to the linkage data and 67 unique to RH data. Sequence alignment yielded comparative positions for 77 bovine markers with homologs on HSA 11. The map covers approximately 32% of HSA 11 sequence in five segments of conserved synteny, another 15% of HSA 11 is shared with BTA 29. Bovine and human order are consistent in portions of the syntenic segments, but some rearrangement is apparent. Comparative positions of gene markers near the meat tenderness QTL indicate the region includes separate segments of HSA 11. The two microsatellite markers flanking the QTL peak are between defined syntenic segments. Conclusions Combining data to construct an integrated map not only consolidates information from different sources onto a single map, but information contributed from each data set increases the accuracy of the map. Comparison of bovine maps with well annotated human sequence can provide useful information about genes near mapped bovine markers, but bovine gene order may be different than human. Procedures to connect genetic and physical mapping data, build integrated maps for livestock species, and connect those maps to more fully annotated sequence can be automated, facilitating the maintenance of up-to-date maps, and providing a valuable tool to further explore genetic variation in livestock.

Background Genome maps for livestock species are necessary to identify genes affecting economically important production traits. Linkage maps, based primarily on highly polymorphic, anonymous microsatellite markers, have been important for identifying chromosomal regions influencing economically important traits in cattle [ 1 - 3 ]. Because a lack of recombination between closely linked markers limits resolution, and because cattle linkage maps [ 4 , 5 ] contain few genes, linkage maps are of limited value for ordering closely linked markers and identifying genes underlying quantitative trait loci (QTL). The radiation hybrid (RH) approach allows mapping monomorphic markers for genes and can provide a higher resolution for ordering close markers [ 6 , 7 ], but high breakage frequency RH data are less reliable than linkage data for ordering widely separated groups of markers [ 8 ]. Integrating linkage and RH data into a single map will refine marker order to facilitate genomic sequencing and will also increase the efficiency of identifying genes associated with QTL. Integrated analysis of both linkage data and RH data allows each source of information to complement the other, providing coarse to intermediate scale maps of the bovine genome, populated with gene markers to facilitate discovery of positional candidate genes for QTL. These integrated maps will lack the fine scale of complete genome sequence, but represent a resource useful for gene identification through comparative mapping approaches, using more complete genome sequence and annotation from other organisms. Similarity between segments of bovine DNA and genomic sequence from other species may supplement integrated data to predict the location of unmapped genes in the bovine genome [ 9 ]. A comprehensive integrated map, containing all identified genes and markers, will simplify database queries and reduce ambiguity inherent in mining information from other mammals. An integrated map can also provide a framework for assembling bovine genomic sequence as data becomes available. A well-ordered map of sequence-tagged-sites (STS) was essential for assembling the human sequence [ 10 ]. The National Institutes of Health (NIH) identified the bovine genome as high priority for sequencing [ 11 ], and sequencing is underway. One pivotal criterion to classifying the bovine genome as ready for sequencing was the availability of well-maintained genetic and physical maps; integrating these maps will provide additional support for sequence assembly. Integration of linkage and RH maps has been reported for a number of species [ 11 - 14 ] and individual bovine chromosomes [ 15 - 17 ]. The general approach to integrated mapping has been to score several markers from linkage maps on an RH panel, then align the independent maps via common markers. Nadkarni [ 18 ] and White et al. [ 19 ] described procedures to synthesize information from multiple independent maps onto a single merged map. These approaches do not directly use data contributing to each map, but merge results of independent analyses. A fundamentally different approach is to merge independent data sets with common markers, so each data set contributes to constructing a single integrated map. Agarwala et al. [ 20 ] developed procedures for integrating RH maps where markers common to independent RH panels contributed to the solution of a comprehensive RH map. Schiex et al. [ 8 ], developed procedures and released CarthaGene software [ 21 ] to merge and solve integrated maps representing multiple linkage and RH data sets. A large volume of data are being generated in cattle and other livestock species that is not rapidly reflected in current map representations. The result is a lack of truly up-to-date maps of any livestock species, as the maps may lag by months or years in their representation of existing data. It is not feasible to devote significant human resources to constantly maintain and update these maps, so it is critical that automated procedures be developed to free human map curators from many of the time-consuming, error-prone tasks experienced in the mapping process. Existing map construction software is automated to the extent that the likelihoods of many alternative marker orders can be evaluated with a single command, but the entire process of gathering and formatting raw data, constructing maps, examining results and publishing on the internet, or elsewhere, requires human intervention at several stages. Automated procedures will streamline the process in order to focus human effort on the critical stages of verifying raw data and examining the resulting maps. Bovine chromosome (BTA) 15 provides an interesting example to study the integration of linkage and RH data, and comparison of the bovine to the human genome. A QTL for meat tenderness has been reported on bovine chromosome 15 [ 22 , 23 ]. Comparative mapping indicates that alternating segments of human chromosome (HSA) 11 are conserved on BTA 15 and BTA 29 [ 15 , 23 , 24 ]. We combined the available linkage and RH data to further examine BTA 15. An integrated linkage and RH map was constructed using CarthaGene software (version 0.99 [ 21 ]), and the comparative positions of DNA sequences shared by segments of HSA 11 and the integrated BTA 15 map were established. We also assessed the potential for automating integrated mapping procedures, anticipating a need to extend integration to the entire bovine genome in order to provide up-to-date maps. Results and discussion The low resolution of the bovine linkage map is indicated by multiple markers sharing the same map position, even when they may be separated by a substantial physical distance. Inclusion of RH data provides additional evidence by which markers that are inseparable only with linkage data can be ordered. The BTA 15 linkage map (Figure 1A ; Additional file 1 ) shows 78 markers placed in 54 distinct positions, with ten positions representing a pair of markers and seven representing three markers. Marker separation on the higher resolution RH map is greater (Figure 1B ; Additional file 1 ), with 109 markers mapped to 105 distinct positions. Projected onto a common scale, the integrated map represents 145 markers in 118 different positions (Figure 1C ; Additional file 1 ). Eighteen positions contain two markers, at three positions three markers are represented, and one position is occupied by four markers. Figure 1 Linear representations of bovine chromosome 15 (BTA15) linkage (A), radiation hybrid (RH; B) and integrated linkage/RH maps (C). Named markers are common to both linkage and RH data sets. Tick marks without a marker name represent markers unique to an individual data set. The linkage map was solved with CRIMAP, and the RH map solved using Carthagene diploid RH data. The integrated linkage/RH map was ordered using CarthaGene with backcross linkage data merged by order with RH data. Integrated RH and linkage maps Markers common to both the linkage and RH data sets provide a basis for integrating the data and constructing maps representing both types of data. Primer sequences associated with the RH and linkage markers indicated 42 common markers in the two data sets, with 36 markers unique to the linkage data and 67 unique to RH, for a total of 145 markers represented on the integrated linkage-RH map. Four sets of markers with different primer sequences matching the same bovine sequence were identified. In two instances (MB064 and HBBMS matching Genbank accession AC130787; T608B5 and SP608B5 matching Genbank accession NM_001752), markers in the set were placed adjacent to each other by the map building routine. In the two other cases (FSHB, FSHBMS, and CSPS101 matching accession Genbank M83753; NCAM1MS and MB085 matching Genbank accession X16451), markers in the set were separated by several markers after initial map construction. In both cases, the map could be reordered so markers in each set were placed next to each other without decreasing likelihood of the map. The final integrated order includes these manual adjustments, so that in all cases of different markers matching the same sequence, the markers are adjacent on the map. Comparison of the integrated map to independently solved linkage (Figure 2A ) and RH (Figure 2B ) maps indicates relatively good agreement between the maps. Product-moment correlations between independent (CRIMAP linkage map; CarthaGene diploid RH map) and integrated (CarthaGene backcross linkage data merged by order with diploid RH data) map positions were greater than 0.99 for both the linkage and RH maps. The final integrated map did suggest some rearrangement of both the linkage and RH maps. Solved using CRIMAP, the integrated map order of linkage markers was somewhat more likely than the order of the independent linkage map (lod score of 2.4 favors integrated order). This result suggests that the most likely order identified by the integrated mapping process had not been evaluated while using CRIMAP to construct the linkage map. Because of differences in speed, CarthaGene can feasibly evaluate many more orders than CRIMAP; even without integration with RH data, CarthaGene might be utilized to identify errors in marker order and refine linkage maps. Figure 2 Comparison of independent bovine chromosome 15 (BTA15) linkage (A) and radiation hybird (RH; B) maps with the integrated BTA15 map. The independent linkage map was solved with CRIMAP, and the independent RH map solved using Carthagene diploid RH data. The integrated linkage/RH map was ordered using CarthaGene with backcross linkage data merged by order with RH data. Tick marks along each axis represent positions of markers on the respective linear map. Symbols indicate the intersection of the maps. Symbols forming a straight line indicate agreement between the maps, while deviations from a straight line indicate inconsistencies between the maps. Syntenic group segments are indicated by shading on the RH map (B). Comparison of the integrated map to the RH map shows the markers remained in the five blocks identified by Gautier et al. [ 24 ], and the order of those blocks is the same for both maps (Figure 2B ). Some markers were reordered within blocks of the RH map. As with the linkage map, the integrated map order was more likely than the original independent map order (lod score of 3.4 favors integrated order; both likelihoods solved using CarthaGene with a diploid RH model). Comparative bovine and human map Comparative map positions for 77 markers mapped to BTA15 were established using primersearch [ 26 ] to identify bovine DNA sequence associated with each marker, and subsequent BLASTN against HSA11 contig sequences. Positions of the bovine-human matches were between 4.16 Mbp and 135.59 Mbp on the HSA11 draft sequence (Build 31). Percentage identities of the matches ranged from 83% (475/570 bases) to 100% (1941/1941 bases), with a mean of 93% (449/475 bases). The syntenic group segments (S1, S2, S3, S4 and S4') identified by Gautier et al. [ 24 ] were retrieved in the comparison of the integrated BTA15 map with HSA11 (Figure 3 ). The integrated BTA15 map covers approximately 32% of HSA11. There are eight gaps in coverage containing between 4.2 and 25.6 Mbp of HSA11 sequence. Boundaries of the syntenic segments encompass 36% of the loci on HSA11 (Table 1 ), not counting the 76 loci within large internal gaps in S1 (7.8 Mbp) and S4 (8.9 Mbp). Some of these gaps in HSA11 coverage are syntenic with BTA29 [ 15 , 22 , 23 ]. Our current BTA29 linkage map places at least one marker in each of the previously identified segments shared by HSA11 and BTA29, accounting for another 15% of HSA11 sequence. Accounting for segments shared with BTA29 leaves 7 gaps containing from 4.9 to 16.1 Mbp of HSA11 sequence that has not been shown to be homologous to mapped regions of bovine chromosomes 15 and 29, although two of the gaps are located within syntenic segments S1 and S4. Figure 3 Comparison of the integrated bovine chromosome 15 (BTA15) map with human chromosome 11 (HSA11) DNA sequence (Build 31). Tick marks along the HSA11 axis indicate positions of HSA11 sequence homologous to bovine sequence mapped to either BTA15 or BTA29. Tick marks along the BTA15 axis indicates positions of markers on BTA15. Shading marks regions shared by HSA11 and BTA29. Boxes indicate syntenic group segments. Table 1 Loci and gene ontology (GO) annotation of human chromosome 11 (HSA11). HSA11 position (Mb) Number of Loci Loci with GO Term Unique GO Terms Segment a Start End S1 104.0 124.4 200 70 176 S1 gap 104.0 111.8 44 15 54 S2 3.9 18.3 228 52 135 S3 73.3 78.5 80 27 68 S4 30.8 47.6 101 31 86 S4 gap 35.9 44.8 32 6 19 S4' 58.6 60.9 63 12 31 All syntenic regions 672 192 344 internal gaps removed 596 171 310 Entire chromosome 0.0 1640 433 578 >QTL 16.3 20.3 53 16 46 <QTL 122.4 126.4 63 9 25 a Syntenic group segments S1, S2, S3, S4, S4' identified by Gautier et al. (2002). Gaps are relatively long segments within a syntenic group that do not contain sequence common to HSA11 and bovine chromosome 15 (BTA15). Segments designated >QTL and <QTL are 4 Mb segments of HSA11 centered around syntenic markers defining boundaries of S1 (<QTL) and S2 (>QTL), flanking the BTA15 meat tenderness QTL identified by Keele et al. (1998). All syntenic regions represents the union of S1, S2, S3, S4 and S4'. Entire chromosome includes all loci with a position established on HSA11 sequence. Markers more recent [ 23 , 24 ] than the original description of the meat tenderness QTL [ 22 ] have resulted in some rearrangment of the BTA15 map, so position of the QTL must be shifted to current positions of markers defining the QTL region. The syntenic segment S1 contains several markers that were within the 95% confidence interval surrounding the QTL, but the two markers most closely flanking the QTL peak, HEL1 and BMS1782, could not be matched to HSA11 sequence and are between defined boundaries of syntenic group segments S1 and S2. Because this QTL region includes a break in bovine-human synteny, the ends of both syntenic segments should be examined to identify positional candidate loci influencing the tenderness QTL. Human loci, in two 4 Mbp segments surrounding the boundaries of S1 and S2 that flank the QTL peak, were identified and associated with gene ontology (GO; [ 27 ]) terms to further describe genes near the QTL. These two segments contain 116 loci (Table 1 ); 25 of these loci have GO annotation [ 28 ]) with terms representing various biological processes, cellular components and molecular functions (Figure 4 ). The GO annotation of loci in both syntenic segments near the QTL may guide further marker development to fine-map the QTL by associations between new markers and tenderness. Adding new markers to this region will also refine boundaries of S1 and S2, and position of the breakpoint between these two segments. Figure 4 Gene ontology classification of loci on human chromosome 11 (HSA11) in regions near a quantative trait loci (QTL) for meat tenderness. Bovine markers flanking the QTL peak are between defined syntenic regions, so loci in two 4 Mbp regions of HSA 11 (16.3 to 20.3 Mbp; 122.4 to 126.4 Mbp) surrounding markers that define syntenic regions were identified and classified by gene ontology annotation. Order is well conserved within syntenic group segments S1, S3, S4' and portions of S2 and S4. The most notable rearrangements within segments are an inversion of several markers in the center of S2, and inconsistent ordering within a subset of S4. The internal rearrangements within syntenic groups found here, pig-human rearrangements [ 29 ], and mouse-human rearrangements [ 30 ] suggest that precise ordering requires reliable data from the species of interest. Comparative information can be used to predict gene location in regions where within-species mapping data are not available [ 9 ] or the available data are ambiguous, and may guide marker development and fine-mapping efforts in specific regions [ 23 , 24 ]. Marker orders based on comparative data, however, should be used with caution. For each systenic segment of BTA 15, marker orders predicted from human order were less likely than the order identified from bovine data (Table 2 ). Table 2 Comparison of integrated bovine chromosome 15 (BTA15) map where marker order is based on bovine data with alternative maps where segments are reordered according to order of human chromosome 11. log 10 likelihood b Total LOD Map Linkage RH Total Bovine data -790.2 -890.4 -1680.6 Syntenic segments a reordered S1 -790.3 -896.5 -1686.8 -6.2 S2 -859.6 -916.4 -1776.0 -95.4 S3 -790.2 -900.0 -1690.2 -9.6 S4 -922.0 -924.6 -1846.6 -166.1 S4' -790.2 -892.1 -1682.3 -2.3 a syntenic group segments described by Gautier et al. (2002). b likelihoods computed with backcross linkage data merged by order with diploid RH model data Challenges for building high-resolution integrated maps and leveraging data from various sources, both within and across species, will be to determine regions where additional data may be informative and placing appropriate emphasis on the different sources of information at different levels of resolution. Linkage maps can provide the scaffold for ordering an entire chromosome, so linkage data may receive the greatest emphasis for initially determining a coarse order. Increased emphasis should be given to higher resolution RH and other physical mapping data to resolve order where placement of linkage markers is uncertain, and markers are too close to provide definitive order. Comparative sequence and mapping information from other species should be most useful to position markers within regions where physical data have insufficient resolution and within-species sequence data are not available. Using appropriate weights to combine genetic and physical mapping data, within-species sequence and comparative sequence data should allow the different data sources to complement each other, resulting in consensus maps supported by the combined sources of information. Automation Genome maps of livestock species need to represent current information in order to maximize utility of the maps. Positions of putative QTL may become misleading if QTL positions are not updated to reflect subtle rearrangements resulting from new mapping data. Genes associated with phenotypic variation will be more readily identified if available information to link mapping data to genes and their function is maintained. Continually updating the maps to depict relevant existing information will be facilitated by automation, but a number of issues must be addressed for implementation of automated procedures to be fruitful. Access to dynamic sources of mapping data must be maintained, so that new information can be incorporated into the maps soon after it is generated. Information to connect data from various sources must be available to expedite integration. Map computation strategies deserve some attention, to minimize the delay between acquiring new data and appearance of those data in subsequent maps. Procedures developed to integrate BTA15 linkage and RH data can be applied to available data for the entire bovine genome. The integration effort will be more valuable, however, if sources of data for the integrated map are periodically updated. Success of a comprehensive integration effort will also depend on information available to establish connections between the data sets. One alternative is to resolve marker nomenclature, perhaps by developing and maintaining a database of marker names and synonyms. A more straightforward, and easily automated, approach is to use primer sequences as universal identifiers to establish connections between mapping data sets. Database curation efforts to associate mapping records (animal genotypes and RH vectors) with primer pairs may be more worthwhile than attempts to resolve all possible names for a given marker. Primer sequence can also be used to establish connections to sequence databases. Sequence similarity searches should reveal connections to STS sequences associated with markers; the process will also identify connections to other sequences, including more completely annotated and assembled sequence. Sequences identified in this process can be used to establish connections with human and other well annotated, assembled genomic sequence for comparative mapping. Similar associations between mapping and sequence data may be established using marker and locus names, provided that marker nomenclature can be resolved Sequence-based connections between mapping data sets, integrated maps, and genomic sequence may be more reliable and are more amenable to automation than attempts to connect sources using names and other information. Connections between maps and annotated sequence can accelerate positional candidate gene discovery if the sequence annotation includes functional information. Harhay and Keele [ 31 ] used GO and GO-annotated human sequence to link livestock EST with function; mapping the EST can extend their procedures to relate map position to function. Connecting map positions to GO terms requires synchronizing several information sources, including livestock maps, human sequence annotated GO terms, and GO databases. Placement of new markers on integrated maps must keep pace with new marker development, if integrated maps are to remain current with available mapping and sequence data, The basic concept of map construction, finding the most likely marker order out of all possible orders, is conceptually simple but computationally demanding, because the number of possible orders increases factorially with the number of markers. Evaluating all possible marker orders is not feasible when mapping data represents more than twenty or thirty markers on a chromosome. Cost and time constraints limit map construction to strategies that evaluate a sufficient number of possible orders to ensure that a reasonably good order is identified. As bovine sequence data becomes available, methods to exploit that resource to refine both the integrated maps and sequence assemblies must be implemented. Advent of whole-genome sequence assemblies has not diminished the value of maps in sequenced species. Discrepancies between human maps and sequence assemblies have been noted [ 32 , 33 ], although concordance between a SNP linkage map and sequence assemblies increased in later assemblies [ 33 ]. A comprehensive linkage-RH map has been used to validate mouse sequence assemblies, revealing cases of significant inversions and translocations in sequence, as well as confirming sequence order in other regions where the sequence order disagrees with previous mouse RH maps [ 34 ]. An integrated linkage-RH map of the rat suggests some errors in the draft sequence, but more importantly, provides a mechanism to anchor QTL on the genomic sequence [ 35 ]. The strategies employed must be sufficiently flexible to allow manual manipulation of the resulting maps. Some evidence, such as STS markers sharing the same sequence, and ordering information from other species, is not easily represented in linkage and RH mapping data. In some cases of markers sharing the same sequence, markers can be forced to share the same position, or data from multiple markers can be combined to create a single haplotype representing multiple markers. Marker orders suggested by maps of other species may be compared with likely orders identified from within-species data. Incorporating information not directly represented in mapping data can require manually evaluating additional orders, and making some judgement about which results are most acceptable. In exploratory analyses merging BTA15 linkage and RH data, simulated annealing and taboo search algorithms in CarthaGene were explored as methods of initially ordering the integrated map, before refinement with the polish and flips routines. Resulting maps were similar to the map presented, but required more than 24 hours to compute. The map presented was initially ordered by placing each marker against a pair of markers common to both data sets, and was constructed in less than four hours. Another approach involved initially placing markers against the set of all markers common to both the linkage and RH data sets, in the linkage map order. While map construction was somewhat faster using this approach, the resulting map was less likely than the map initiated from a pair of markers and showed greater disagreement with the linkage map. Parallelization of the mapping algorithms can substantially increase the speed of map construction. Likelihoods of a number of alternative orders must be computed at several steps during the map building process. If these calculations are distributed across multiple processors, time required to compute all likelihoods and arrive at a final order will be reduced because computations are performed simultaneously. Increased parallelization should also increase the feasibility of implementing more thorough algorithms that examine a larger number of possible orders, therefore increasing the probability of identifying more likely maps. Conclusions Linkage and radiation hybrid maps are powerful tools to facilitate discovery of genomic regions and ultimately genes influencing livestock production traits. Combining linkage and RH data can provide more accurate, consolidated maps representing more information, especially if the maps are connected to well annotated genomic sequence. Automating map construction and comparative mapping procedures will expedite construction of whole-genome integrated maps and maintaining a comprehensive resource as new data becomes available. Success of automated procedures to connect data from various sources and construct integrated maps depends on information available to establish connections between data; sequence-based approaches to connect data are preferrable. Methods Data sets for integrated map construction Linkage data for 78 markers in the BTA15 linkage group were obtained from the U.S. Meat Animal Research Center (MARC) reference population (224 animals; [ 4 ]). Radiation hybrid data for 109 markers were obtained from the ComRad project radiation hybrid panel (94 cell lines; [ 7 , 24 ]). These data include two newly developed microsatellite markers genotyped in the MARC families (Table 3 ), and seventeen previously unpublished markers with RH data(Table 4 ). Table 3 Description of previously unpublished linkage markers placed on the integrated bovine chromosome 15 map. Marker Name Forward Primer Accession number Reverse Primer DIK2411 CTAACGCCCCTGAGACAGAC AB112806 GTGGCGTTAGTTGGTCCTTC DIK2374 CCTGTTTGGGACACTCTCCT AB112803 GAATCTCTTCAATGCCGAATG Table 4 Description of previously unpublished radiation hybrid markers placed on the integrated bovine chromosome 15 map. Locus Symbol Gene Name Forward Primer Accession Number Reverse Primer C11ORF15 chromosome 11 open reading frame 15 GCATCCTAGAACAGACTGGCT AW657178 GGAGGCAACCGGAACTCCAGT DKK3 dickkopf (Xenopus laevis) homolog 3 CGAAGACCATTATCAGCCACA AW336328 CTCTGGATGCATACATGAAGGA EIF4G2 eukaryotic translation initiation factor 4 gamma, 2 AGCTTGAGGCCTGCTCAGTCT AV602677 GTCCCAAAGGTGGCGTTTGA FLJ11790 protocadherin 16 dachsous-like (Drosophila) CCCAGCTTCTCACCTTCACTA AW428073 GATATGGAGCTCGGTGTCGTCT INPPL1 inositol polyphosphate phosphatase-like 1 CAGCTCAACTTGGAGCGGGAA BF705795 GAACCCCGCTCATAGCGGTAA KIAA0750 hypothetical protein KIAA0750 GTGGGAAGCTGGCTATTGCA AW652984 GAAGATGAAAGCCACACCGCT MRPL17 mitochondrial ribosomal protein L17 CACCTGTTGCAGAACTTGCTT BE899833 CCCAGCTTCCCGTAGTCAATA PARVA parvin, alpha GCCGTATCCCTCAACTCCTTT BE477207 CTCAAGAGTCCCTGTTGAAGA PSMA1 proteasome (prosome, macropain) subunit, alpha type, 1 GAATATGCAATGGAAGCTGTC AV602233 GCTGCAAGTTCTGACTGTGCT RANBP7 RAN binding protein 7 GGGTGAAGAGATGAGGAAGAT BF45355 CTGATACTCATCAACAGGGTT RNF21 tripartite motif-containing 34 GAAGAGAAACTCCTACTCTTCT AW447003 CTCCTGAGATCGTTCACAAAGA ST5 suppression of tumorigenicity 5 CGCTGCTCTGGTCTATCACTT BF604586 ATTGCCAGCCCCTGGCAGGAA STIM1 stromal interaction molecule 1 GCCCTCCAGGCTAGCCGAAAT BE756550 CACTGCCACCCCCATCCTGTT TAF2H TAF10 RNA polymerase II, TATA box binding protein (TBP)-associated factor, 30 kDa TGGTGTCCAGCACGCCTCTA AW315164 GTAGTAACCAGTCACTGCATCA UVRAG UV radiation resistance associated gene GTACATTTTCAGCTGAGCACC BE590188 CGCGGTACACTCCTTTCTCAA WEE1 wee1+ (S. pombe) homolog GATGGATGCGTTTATGCCATA AV598317 CGAACTACATGAGAATGTTGC ZFP26 C3HC4-like zinc finger protein CTGCTAAAGTGGCTTCTGGC BF04414 GGTACAGACCACTCGTACAA All bovine sequence information stored in GenBank was identified using the taxonomy ID field of the sequence file annotation and obtained from NCBI. Provisional sequence data consisting of tentative consensus clustering of bovine EST data was obtained from the Bos taurus gene index (BTGI; [ 36 ]) assembled by The Institute for Genonomics Research (TIGR, [ 37 ]). Other sources of sequence were the NCBI nt database (NT; [ 38 ]), and human chromosome 11 draft sequence contigs (Build 31;[ 38 ]). Data integration Connections between data sets are necessary for integrated analyses of those data to be meaningful. Because some marker names were ambiguous, connections between markers in the linkage and RH data were established using primer sequence. Markers with identical primers were considered to be the same, regardless of marker name. Primer sequence was also used to establish connections with human sequence. Primer pairs were matched against bovine sequence from GenBank, NT and BTGI databases using the EMBOSS [ 26 ] primersearch tool. The longest matching sequence having one or fewer mismatches and an amplimer less than 600 bp was selected for homology search against HSA11 contigs. The selected sequences were examined for gaps, and where gaps occurred, only the ungapped pieces matching a primer pair was used in the homology search. Connections between the individual sequences matching bovine markers and human sequence were then determined via BLASTN [ 39 ] with an expectation value of e -20 , and default values for other parameters. Connections between human position and functional GO annotation were extracted from the downloadable LocusLink database [ 38 ]. Procedures using the GO database [ 40 ] and perl API [ 41 ] were developed to classify specific GO terms into general categories described by higher level terms. Integrated map construction Observations for RH data are binary (0/1), indicating absence or presence of a particular marker in a cell line, where each cell line represents a relatively short segment of DNA on a chromosome. Physically close markers are more likely to be observed on the same cell line than distant markers. Linkage data includes pedigree information and marker genotypes, where individual genotypes represent alleles inherited from each parent. Alleles for physically close markers on a single chromosome are more likely to be inherited from the same grandparent; the likelihood of marker alleles with different grandparental origin appearing on the same chromosome increases with distance between markers. These chromosomes can be represented in a binary, RH-like format that can be merged with RH data using CarthaGene. Analagous to RH data representing presence or absence of a marker in a cell line, binary representation of linkage data indicates presence or absence of a maternal allele on an individual chromosome. The chrompic option of CRIMAP [ 42 ] was used to construct these individual chromosomes, using the most likely order identified by an automated linkage mapping routine. No distinction was made between definite phase-known maternal and paternal inheritance, and statistically predicted inheritance when phase could not be determined. An interface to the CarthaGene shared library was developed using perl and the perl Inline modules [ 43 ] to automate map construction (see Additional file 2 ). This interface includes procedures to initially place markers on a map and refine map order, as well as a number of utility routines. A map construction script using this interface was also developed (see Additional file 3 ). The script to order markers on the integrated map starts by merging the binary backcross representation of linkage data with the haploid model RH data, assuming common marker order ( dsmergor ). Two markers shared by the linkage and RH data sets are identified, and all other markers are inserted, one at a time, into the most likely position using the CarthaGene buildfw procedure. Once all markers from both data sets are placed, the marker order is refined iteratively, cycling through polish and flips routines until likelihood does not improve. The polish procedure individually tests each marker in all alternative positions, and flips evaluates permutations of all sets of six adjacent markers. After convergence using the map construction script, further evaluation of alternative orders was carried out with the backcross linkage data merged with a diploid model of the RH data, again assuming common marker order. Marker orders consistent with available sequence information were evaluated. Where primer paris for different markers matched the same bovine sequence, but the markers were separated by one or more other markers by the map construction routine, likelihoods of orders with the matching markers placed adjacent to each other were determined. Likelihoods of marker orders consistent with human sequence within each syntenic segment were also computed. The sequence-based orders were used in the final integrated map if they did not decrease likelihood of the map. Log-likelihoods of the final integrated map order were computed with the RH and linkage data sets for comparison to the independent maps, using CarthaGene for the RH map and CRIMAP for the linkage map. The final integrated marker order was projected onto a common relative scale representing all markers. This was accomplished by merging the linkage data with RH data, modeled as backcross, using dsmergen . Marker order was set to the final integrated order, map distances computed, then scaled to range from zero to 100. Computation All computation was performed on a 10-node Linux cluster, each node configured with 2 AMD 1900+ CPUs and 3 Gb RAM. When practical, computation was parallelized using perl scripts and open source Grid Engine software [ 44 ] to distribute tasks to each node in the cluster. Steps that were parallelized included matching primers to sequence, and the Blast searches to align bovine with human sequence. Authors' contributions WMS developed perl scripts for automated analyses, conducted analyses to match markers from linkage and RH data sets, constructed the linkage and integrated maps, and drafted the manuscript. MG and AE developed RH markers and constructed the RH map. AE and JWK conceived the research, and contributed to planning analyses and evaluating results. RTS and TPLS assisted with evaluation of the integrated map. GPH conducted BLAST analyses and associated bovine sequence with human GO annotation. NI, AT, HT developed linkage markers. YS and GLB coordinated linkage map data collection. GLB curates the MARC linkage data and linkage maps. Supplementary Material Additional File 1 linkage, RH and integrated maps of BTA 15 Click here for file Additional File 2 perl interface to CarthaGene, requires perl Inline::Tcl module Click here for file Additional File 3 perl script to construct integrated maps using CarthaGene, requires carthaPerl.pl interface to CarthaGene Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526187.xml

387274

A Pacific Culture among Wild Baboons: Its Emergence and Transmission

Reports exist of transmission of culture in nonhuman primates. We examine this in a troop of savanna baboons studied since 1978. During the mid-1980s, half of the males died from tuberculosis; because of circumstances of the outbreak, it was more aggressive males who died, leaving a cohort of atypically unaggressive survivors. A decade later, these behavioral patterns persisted. Males leave their natal troops at adolescence; by the mid-1990s, no males remained who had resided in the troop a decade before. Thus, critically, the troop's unique culture was being adopted by new males joining the troop. We describe (a) features of this culture in the behavior of males, including high rates of grooming and affiliation with females and a “relaxed” dominance hierarchy; (b) physiological measures suggesting less stress among low-ranking males; (c) models explaining transmission of this culture; and (d) data testing these models, centered around treatment of transfer males by resident females.

Introduction A goal of primatology is to understand the enormous variability in primate social behavior. Early investigators examined interspecies differences, e.g., that pair-bonding is more common among arboreal than terrestrial primates ( Crook and Gartlan 1966 ). Attention has also focused on geographical differences in behavior within species ( Whiten et al. 1999 ). Often, such differences reflect environmental factors (e.g., a correlation between quantities of rainfall and foraging time) or, in theory, could reflect genetic drift. However, increasing evidence suggests that group-specific traits can also represent “traditions” or “cultures” (the latter term will be used, commensurate with the near consensus among primatologists that the term can be appropriately applied to nonhuman primates). As traditionally applied to humans, such “culture” can be defined as behaviors shared by a population, but not necessarily other species members, that are independent of genetics or ecological factors and that persist past their originators ( Kroeber and Kluckhohn 1966 ; Cavalli-Sforza 2000 ; de Waal 2000 ; de Waal 2001 ). Thus defined, transmission of culture occurs in apes ( McGrew 1998 ; Whiten et al. 1999 ; van Schaik et al. 2003 ), monkeys ( Kawai 1965 ; Cambefort 1981 ; Perry et al. 2003 ), cetaceans ( Noad et al. 2000 ; Rendell and Whitehead 2001 ), and fish and birds ( Laland and Reader 1999 ; Laland and Hoppitt 2003 ). As particularly striking examples, chimpanzees (Pan troglodytes) across Africa demonstrate variability in 39 behaviors related to tool use, grooming, and courtship ( Whiten et al. 1999 ), and the excavation of near-millenium-old chimpanzee tools has been reported ( Mercader et al. 2002 ). Nearly all such cases of nonhuman culture involve either technology (for example, the use of hammers for nut cracking by chimpanzees), food acquisition, or communication. In this paper, we document the emergence of a unique culture in a troop of olive baboons (Papio anubis) related to the overall structure and social atmosphere of the troop. We also document physiological correlates of this troop atmosphere, the transmission of relevant behaviors past their originators, and possible mechanisms of transmission. Results/Discussion Circumstances Leading to the Emergence of a Unique Culture In the early 1980s, Forest Troop slept in trees 1 km from a tourist lodge. During that period, an open garbage pit was greatly expanded at the lodge. This attracted an adjacent baboon troop, Garbage Dump Troop, which slept near the pit and foraged almost exclusively there. By 1982, many Forest Troop males went to the garbage pit at dawn for food. While such refuse eaters did not differ in age distribution (data not shown) or average dominance rank from non–refuse eaters, they were more aggressive ( Table 1 ); such aggressiveness could be viewed as a prerequisite in order to compete with Garbage Dump males for access to refuse. Refuse eaters were also involved in more dominance interactions within Forest Troop than were non–refuse eaters (note that frequency of dominance interactions is independent of outcome, and thus of rank). Table 1 Characteristics of Forest Troop Males As a Function of Whether They Competed for Refuse with the Garbage Dump Troop Statistical comparisons by unpaired t-test, n = 7 and n = 8 for refuse eaters and remaining males, respectively. Dominance rank based on approach–avoidance criteria ( Altmann 1974 ). Data concerning refuse eaters were derived solely from their time in the troop, rather than including time spent with the Garbage Dump Troop. Rate of male–male aggression consisted of aggression with any other adult or subadult male in the troop. Rate of aggression directed at females included all adult and subadult females. Rates of behaviors are per 100 h of focal observation, except for grooming, which is per 10 h. Data are mean ± standard error of the mean (SEM) In 1983, an outbreak of bovine tuberculosis occurred, originating from infected meat in the dump. From 1983 to 1986, most Garbage Dump animals died, as did all refuse-eating Forest Troop males (46% of adult males); no other Forest Troop animals died ( Tarara et al. 1985 ; Sapolsky and Else 1987 ). These deaths greatly altered Forest Troop composition, such that there were fewer adult males and more adult females; this more than doubled the female:male ratio ( Table 2 ). By 1986, troop behavior had changed markedly, because only less aggressive males had survived. Table 2 Troop Composition Before and After the Tuberculosis Outbreak Data from annual troop censuses. Census numbers include both “subadult” animals (undergoing the emergence of secondary sexual characteristics) and “fully adult” (fully emerged secondary sexual characteristics) Because of these events, observations of the troop were stopped, and only censusing was done until 1993. Research was begun on Talek Troop, approximately 50 km away. In 1993, informal observation of Forest Troop indicated that the behavioral features seen by 1986 had persisted. Critically, by 1993, no adult males remained from 1983–1986; all current adult males had joined the troop following 1986. Thus, the distinctive behaviors that emerged during the mid-1980s because of the selective deaths were being carried out by the next cohort of adult males that had transferred into the troop. Focal sampling on Forest Troop recommenced in 1993, in order to document this phenomenon. Data from Forest Troop 1993–1996 (henceforth, F93–96) were compared with two other data sets that served as controls: observations from 1993–1998 on the Talek Troop (henceforth T93–98), and observations of Forest Troop itself prior to the deaths (1979, 1980, 1982; henceforth F79–82). These two control data sets did not differ significantly from each other and were combined, henceforth T93–98/F79–82. Atypical Features of the Behavior of Forest Troop Males Male–male dominance interactions Males of F93–96 and T93–98/F79–82 had similar rates of approach–avoidance dominance interactions (data not shown). Moreover, dominance stability did not differ, as measured by the percentage of approach–avoidance interactions which represented a reversal of the direction of dominance within a dyad of males of adjacent rank (16% ± 5% and 20% ± 5% for F93–96 and T93–98/F79–82, respectively, n.s.). There was also no difference in the average tenure length of the highest-ranking male (approximately a year). Despite those similarities, dominance behavior in F93–96 differed from the two control cases in ways that, arguably, made for less stress for low-ranking males. A first example concerns approach–avoidance dominance interactions between males more than two ranks apart in the hierarchy. The overwhelming majority of such interactions were won by the higher-ranking individual. Because a male is rarely seriously threatened by an individual more than two ranks lower in the hierarchy, interactions between individuals that far apart typically represent harassment of or displacement of the subordinate by the higher-ranking male, rather than true competition. In T93–98 and F79–82, approximately 80% of approach–avoidance interactions were between males more than two ranks apart in the hierarchy. In contrast, a significantly smaller percentage of approach–avoidance interactions were soin F93–96 ( Figure 1 A). Instead, a disproportionate percentage of F93–96 dominance interactions occurred among males of adjacent ranks (with, as noted, no difference in dominance stability)( Figure 1 B). Moreover, high-ranking males in F93–96 were more “tolerant” of very low-ranking males, as there was a disproportionately high number of reversals with males more than two steps lower in the hierarchy ( Figure 1 C). Thus, in F93–96, with a typical level of dominance stability, approach–avoidance dominance interactions were concentrated among closely ranking animals, with low-ranking males being more tolerated and less subject to harassment and/or displacement by high-ranking males. Figure 1 Quality of Male–Male Dominance Interactions (A) Percentage of male approach–avoidance dominance interactions occurring between males more than two ranks apart. (B) Percentage of male approach–avoidance interactions occurring between males of adjacent ranks. (C) Percentage of approach–avoidance interactions representing a reversal of the direction of dominance within a dyad by a male more than two steps lower ranking. Mean ± SEM, ** and *** indicate p < 0.02 and p < 0.01, respectively, by t-test, treating each male/year as a data point. Data were derived from a total of ten different males in F93–96, 31 different males in T93–98, and 19 different males in F79–82. Potentially, the result in (B) could have arisen from different numbers of males in F93–96 versus the other two troops (a smaller group size does not change the number of adjacent animals available to any given subject, but decreases the number of nonadjacent animals available). However, the same results were found if the numbers of males in the three troops were artificially made equal by excluding excess males from either the top or the bottom of the hierarchy (data not shown). Aggression Patterns of aggression also differed between F93–96 and T93–98/F79–82 in a way that suggested a less stressful environment for subordinates in F93–96. The troops had similar overall rates of aggressive interactions ( Table 3 ). However, aggression in F93–96 was more likely than in the control troops to occur between closely ranked animals (i.e., within two rank steps), rather than to reflect high-ranking males directing aggression at extremely low-ranking ones; the latter type of interaction is particularly stressful for a subordinate, because of its typical unpredictability. Moreover, F93–96 males were less likely than T93–98/F79–82 males to direct aggression at females. Table 3 Patterns of Aggression in Forest Troop 1993–1996 versus Talek Troop 1993–1998 and Forest Troop 1979–1982 ** and *** indicate p < 0.025 and p < 0.01 by unpaired t-test, respectively. Observed/expected ratios were derived by comparing observed frequencies of behavior with the frequencies expected with even distribution of aggressive interactions across all dyads; a ratio of 1.0 indicates the behavior occurring at the expected rate. Data from T93–98 and F79–82 did not differ significantly, and thus were pooled. Data were derived from a total of ten different males in F93–96, 31 different males in T93–98, and 19 different males in F79–82 We examined the data for reconciliative behavior (i.e., affiliative behaviors between pairs following aggressive interactions [ de Waal and van Roosmalen 1979 ]) in F93–96 and T93–98/F79–82. However, we saw no male–male reconciliation in any troop, in agreement with prior reports ( Cheney et al. 1995 ). Affiliative behaviors Quantitative data on affiliative behaviors were not available for F79–82. However, F93–96 males socially groomed more often than did control T93–98 males ( Figure 2 A) this difference was due to more grooming between males and females. F93–96 males were also in close proximity to other animals more often than were T93–98 males ( Figure 2 B). While males did not differ between troops in the average number of adult male neighbors (i.e., within 3 m), F93–96 males were more likely than T93–98 males to have adult females, infants, adolescents, and juveniles as neighbors. Figure 2 Quality of Affiliative Behaviors (A) Amount of grooming involving adult males in Forest Troop 1993–1996 and Talek Troop 1993–1996. The first pair of columns represents mean time adult males spent grooming adult females; the second pair, mean time adult males were groomed by adult females. (B) Comparison of average number of neighbors (i.e., within 3 m) of adult males in the two troops. Mean ± SEM. *, **, and *** indicate p < 0.05, 0.02, and 0.01, respectively, by unpaired t-test. Data were derived from a total of ten different males and 17 different females in F93–96, 31 different males and 21 different females in T93–98, and 19 different males and 23 different females in F79–82. Sexual behavior Sexual behavior did not differ between F93–96 and T93–98/F79–82. The percentages of nonpregnant, nonlactating females in estrus per day did not differ (27% ± 7% and 30% ± 4%, respectively, n.s.). Moreover, the relationship between male rank and reproductive success did not differ ( R 2 of correlation between rank and reproductive success: 0.25 ± 0.25 and 0.54 ± 0.10, respectively, n.s.). Physiological Correlates of Behavioral Features of Forest Troop Thus, F93–96 males had high rates of affiliative behaviors, and low-ranking males were subject to low rates of aggressive attack and subordination by high-ranking males. In a stable hierarchy, low-ranking baboon males show physiological indications of being stressed, including elevated basal levels of glucocorticoids (the adrenal hormones secreted in response to stress), hypertension, and decreased levels of high density lipoprotein cholesterol, growth factors, and circulating lymphocytes ( Sapolsky 1993 ; Sapolsky and Share 1994 ; Sapolsky and Spencer 1997 ). We tested whether subordinate males in F93–96 were spared the stress-related physiology of subordination seen in other troops. This was the case ( Figure 3 A). In F79–82, i.e., prior to the tuberculosis outbreak, subordination was associated with elevated basal levels of glucocorticoids, as in other species in which subordination entails extensive stressors and low rates of coping outlets ( Sapolsky 2001 ). While glucocorticoids aid in surviving an acute physical stressor, chronic overexposure increases the risk of glucose intolerance, hypertension, ulcers, and reproductive and immune suppression ( Sapolsky et al. 2000 ). In contrast to this picture in F79–82, in which subordination was associated with a physiology suggesting a chronic state of stress, subordinate F93–96 males did not have elevated basal glucocorticoid levels (levels were unavailable for T93–98). Figure 3 Stress-Related Physiological Profiles (A) Basal glucocorticoid levels (μg/100 ml). Males were split into higher- and lower-ranking 50%, by approach–avoidance criteria. The primate glucocorticoid, cortisol, was measured by radioimmunoassay. (B) Number of anxiety-related behaviors observed 10–20 min after β-carboline-3-carboxylic acid administration (M-156, Research Biochemicals International, Natick, Massachusetts, United States), after subtracting the number observed 10–20 min after vehicle administration (dextrin in 1 ml saline); 0.5 g of the drug in 1ml saline was delivered intramuscularly by dart syringe (Pneu-Dart, Inc., Williamsport, Pennsylvania, United States) fired from a blowgun at 5 m. Mean ± SEM. * and *** indicate p < 0.05 and p < 0.01, respectively, by unpaired t-test. Data were derived from a total of ten different males in F93–96, 31 different males in T93–98, and 18 different males in F79–82. Subordinate F93–96 males were spared another stress-related physiological marker. Experimental anxiety was induced by darting males, intramuscularly, with β-carboline-3-carboxylic acid, a benzodiazepine receptor antagonist which induces behavioral and physiological indices of anxiety in primates (benzodiazepine receptors bind tranquilizers such as valium and librium and mediate their anxiolytic effects)( Ninan et al. 1982 ). Males were darted on days when they had not had a fight, injury or mating. As a control, they were darted on separate days with vehicle alone (order of dartings randomized). Males were then monitored by an observer unaware of treatment. β-carboline-3-carboxylic acid had no effect on behavior in high-ranking males in T93–98 or F93–96 ( Figure 3 B).The drug increased anxiety-related behaviors in low-ranking males in T93–98 but not in F93–96 (the recorded anxiety-related behaviors were self-scratching, rhythmic head shaking, assuming a vigilant stance, repeated wiping of nose, and jaw grinding in a solitary male [ Ninan et al. 1982 ; Aureli and van Schaik 1991 ; Castles et al. 1999 ]). Thus, in the more typical F79–82 and T93–98 troops, subordination had distinctive stress-related physiological correlates. In contrast, F93–96 males lacked these rank-related differences. Potential Mechanisms Underlying Transmission of This Culture A decade after the deaths of the more aggressive males in the troop, Forest Troop preserved a distinct social milieu accompanied by distinct physiological correlates. Critically, as noted, no adult males in F93–96 had been troop members at the end of the tuberculosis outbreak. Instead, these males had subsequently transferred in as adolescents, adopting the local social style. A number of investigators have emphasized how a tolerant and gregarious social setting facilitates social transmission (e.g., van Schaik et al. 1999 ), exactly the conditions in F93–96. The present case of social transmission is reminiscent of some prior cases. For example, juvenile rhesus monkeys (Macaca mulatta) housed with stumptail macaques (M. artoides) assume the latter's more conciliatory style ( de Waal and Johanowicz 1993 ). Moreover, anubis baboons (Papio anubis) and hamadryas baboons (P. hamadryas) differ in social structure, and females of either species experimentally transferred into a group of the other adopt the novel social structure within hours ( Kummer 1971 ). Several models have been hypothesized to explain transmission of cultures ( Whiten et al. 1999 ; de Waal 2001 ; Galef 1990 ). For clarity, it is useful to first consider their application to an established example of transmission of a “technology” before then applying them to the transmission of the social milieu of F93–96. An example of the former is the nut cracking with stone hammers by West African chimpanzees ( Boesch and Boesch 1983 ; Boesch 2003 ), a trait transmitted transgenerationally. In “instructional models” of chimpanzee tool use, young are actively taught hammer use. In the case of F93–96, instructional models would involve new transfer males being subject to socially rewarding interactions (e.g., grooming) or aversive ones (e.g., supplantation or attack) contingent upon their assimilating the troop tradition. In such models, a key question is who “instructs.” Much as with the term “culture” being used with respect to animal behavior, the use of the term “instruction” has also generated some controversy, with some preferring the concept of “active behavioral modification” by others bringing about the change. As a striking example of that, when young male cowbirds learn to produce their local song, they initially produce an undifferentiated repertoire of songs, and females react to the production of appropriate dialect with copulation solicitation displays, thus providing positive reinforcement and shaping those behaviors ( Smith et al. 2000 ). In “observational models” applied to chimpanzee tool use, young learn nut cracking by observing and copying adults. As applied to F93–96, transfer males would model behavior upon that of resident males. In “facilitation models” of the chimpanzee example, proximity to adults and their hammers increases the likelihood of the young experimenting with hammers and deriving the skill themselves. As applied to the baboons, male F93–96 behaviors would be an implicit default state where, in the absence of the more typical rates of male aggression (either male–male or male–female), females broadly tend to become more affiliative, and in the context of more affiliative female behavior, transfer males broadly tend to become less aggressive. As perhaps a way of stating the same, the default state may emerge because of the atmosphere of a troop with a high female:male ratio (with less need for male competition for access to estrus females). Finally, a “self-selection model” may apply to the baboons, in which particular kinds of males were more prone to transfer into such a troop (note that the fact that males transferred in from an array of surrounding troops rules out the possibility of an additional model, in which the culture was continued by genetic means). We assessed these models by analyzing cases where adolescent males transferred on known dates and were observed for at least 2–6 mo afterward. Thus, we searched for behavioral patterns involving new transfer males that might differ between F93–96 (five such transfers) and T93–98/F79–82 (12 transfers). Many interactions involving new transfer males did not differ ( Table 4 ). Transfer males in F93–96, T93–98, and F79–82 all attacked and supplanted females from feeding or resting sites at equal rates. Moreover, despite the different dominance structure among resident F93–96 males, resident males in F93–96, T93–98, and F79–82 all treated new transfer males similarly. There were similar latencies until transfer males were first lunged at by residents, and transfer males were involved in dominance and aggressive interactions at similar rates in all three troops (note that because there were half as many resident males in F93–96 as in T93–98 or F79–82, the rate of such interactions within any given resident/transfer male dyad would differ). We examined instances where resident males acted aggressively towards transfer males, determining whether such behaviors were more prevalent during the 20 min after aggressive behavior by the transfer male than at other, randomly selected times ( de Waal and Yoshihara 1983 ; de Waal and Johanowicz 1993 ). We found no evidence for such contingent behavior (data not shown). Table 4 Behaviors of Newly Transferred Males Subjects consisted of the five transfer males in F93–96 and the 12 in T93–98/F79–82 We then examined affiliative interactions between females and new transfer males, and found striking differences between F93–96 and T93–98/F79–82, in that F93–96 females treated new transfer males in the same affiliative manner that they treated resident males. F93–96 transfer males had a shorter latency until first being groomed by or presented to by a female than did T93–98/F79–82 transfer males ( Figure 4 A). (The differences between F93–96 and T93–98 did not arise from a single F93–96 female accounting for the much shorter latencies until presentation and grooming: three different females accounted for the first interactions with the five F93–96 transfer males). Moreover, F93–96 transfers sat in closer proximity to and had more grooming bouts with females than did T93–98/F79–82 transfers ( Figure 4 B). While estrous females are more likely than nonestrous females to interact with transfer males ( Smuts 1999 ), the percentage of females in estrus did not differ among the troops (see above). In addition, F93–96 females did not seem to treat transfer males in a contingent manner ( de Waal and Yoshihara 1983 ; de Waal and Johanowicz 1993 ). To test for this, we first examined instances where resident females were affiliative towards transfer males, determining whether this was more likely during the 20 min following an affiliative behavior on the part of the transfer male than at other, randomly selected times. Second, we determined whether females were less likely to be affiliative during the 20 min following an aggressive behavior on the part of a transfer male. We found no evidence for either pattern (data not shown). Figure 4 Quality of Interactions between Resident Females and Transfer Males (A) Latency, in days, until a newly transferred male is first groomed by a female (left) or presented to by a female (right). (B) Average number of adult female neighbors per scan (i.e., within 3 m; left) and average number of grooming bouts with females per 100 h of observation (right) for transfer males. Mean ± SEM. * and *** indicate p < 0.05 and p < 0.01, respectively, by unpaired t-test. Latency until first presented to by a female approached significance ( p < 0.08). Data were derived from a total of ten different males and 17 different females in F93–96, and 31 different males and 21 different females in T93–98. These data allow some insight as to the mechanisms of social transmission in F93–96 (without remotely allowing an analysis fine-grained enough to see whether these mechanisms were equally relevant to the transmission of all the components of the F93–96 culture, namely the low rates of male aggression, the high rates of female affilitation, and the relaxed dominance structure). The lack of contingency in the treatment of transfer males by residents argues against instruction; commensurate with this, there is relatively little evidence for “instruction” in nonhuman primate cultural transmission ( de Waal 2001 ; for an exception, see Boesch 1991 ). The similar rates of displacement behaviors by transfer males onto females in all three troops argue against self-selection (i.e., the possibility that F93–96 transfer males already behaved differently than transfer males elsewhere). This is not surprising. While adolescent male baboons may transfer repeatedly before choosing a troop ( Pusey and Packer 1986 ), as well as later in life ( Sapolsky 1996 ), we have seen little evidence among these animals of the systematic sampling of different troops required by a self-selection model. The data instead support either observational or facilitative/default models. Insofar as resident males in all troops interacted with transfer males similarly, transmission in F93–96 could have involved observation only if such observations were of how resident males interacted with females or each other. Some, but not all, studies support observational models of social transmission in other primates ( Visalberghi and Fragaszy 1990 ; Whiten 1998 ; Boesch 2003 ; Whiten et al. 2003 ); there are few data at present from baboons concerning this issue. As shown, F93–96 transfer males were had high rates of affilitative interactions with females. The preponderance of females in F93–96 is a plausible explanation for their unconditional (or, at least, less conditional) increase in tolerance of and affiliation with males (including transfer males), insofar as males in the troop had less numeric means to be aggressive to females. (Note that this skewed sex ratio continues in this troop to the present, for unknown reasons.) Thus, affilative data support a facilitative/default model only if it involves preferential sensitivity to the quality of interactions with females. This analysis raises the possibility that there is no social transmission, but that the F93–96 pattern is merely the emergent outcome of the 2:1 female:male ratio. To test this, we analyzed the five available studies of baboon troops with adult female:male ratios of 2 or more which contained quantitative data comparable to the present data ( Seyfarth 1976 , 1978 ; Strum 1982 ; Bercovitch 1985 ; Noe 1994 ). The key question was whether those prior data more closely resembled those of F93–96 or the control troops. Previous data more closely resembled, and did not differ significantly from, data from the control troops for the percentage of time males groomed females (based on Seyfarth 1978 ), the percentage of time females groomed males ( Seyfarth 1978 ), the rate of intersexual aggression ( Seyfarth 1976 , 1978 ), the structure of male–male dominance ( Noe 1995 ), or the structure of of male–male aggression ( Strum 1982 ; Bercovitch 1985 ). In contrast, no quantitative measures more closely resembled F93–96. This strongly suggests that the F93–96 pattern is unique and is being uniquely maintained, rather than being the social structure that automatically emerges whenever a female-skewed female:male ratio occurs. Thus, insofar as a facilitative/default model is operating in this troop, it cannot be a relative paucity of males which “activates” a default state; instead, it would likely be the paucity of aggressive males. The unconditional (or less conditional) nature of the default model is puzzling, in that it requires that females be relatively affiliative to recent transfer males who, nonetheless, are initially aggressive to them. This seems counter to the long-standing emphasis in primatology on individual relations (i.e., females are unlikely to be unable to distinguish between relatively unaggressive resident males and relatively aggressive newly transferred males). Precedent for this unexpected implication comes from the social epidemiology literature concerning “social capital,” in which health and life expectancy increase in a community as a function of communitywide attributes that transcend the level of the individual or individual social networks ( Kawachi et al 1997 ). In summary, we have observed circumstances that produced a distinctive set of behaviors and physiological correlates in a troop of wild baboons. Moreover, these behaviors were taken on by new troop members; while obviously not conclusive, the data suggest that this most likely occurs through observational or facilitative/default models. Finally, somewhat uniquely in nonhuman primate studies, these findings concern the intergenerational transfer of social, rather than material culture. These findings raise some issues. There appear to be adverse health consequences of the stress-related physiological profile of subordination in typical baboon troops ( Sapolsky 1993 ; Sapolsky and Share 1994 ; Sapolsky and Spencer 1997 ). The distinctive rank-related patterns of physiology in F93–96 suggest that subordinate males in that troop may be spared those pathologies. Another issue concerns the consequences of the culture of F93–96 remaining stable over some time. A hallmark of human culture is that it is cumulative (i.e., innovations are built upon each other), and there is only scant evidence, at best, for the same in nonhuman primates ( Boesch 2003 ). It would thus be interesting to see if additional features of the F93–96 social tradition emerge with time. A converse issue concerns circumstances that might destroy the F93–96 culture. The culture might be destroyed if numerous males transfer into the troop simultaneously, or if a male transfers in who, rather than assuming the F93–96 culture, instead takes advantage of it. Game theory suggests that F93–96 would be vulnerable to such “cheating.” Another issue concerns the fate of natal males from F93–96 when they transfer elsewhere. Reciprocal altruism models ( Axelrod and Hamilton, 1981 ) suggest that if one F93–96 male transfers elsewhere and continues his natal behavioral style, he will be at a competitive disadvantage. However, should two F93–96 males simultaneously join another troop and maintain F93–96–typical interactions between them, they might be at a competitive advantage. This might represent a means to transmit this social style between troops. Finally, these findings raise the issue of their applicability to understanding human social behavior and its transmission. Human history is filled with examples of the selective loss of demographic subsets of societies (e.g., the relative paucity of adult men following the American Civil War or the relative paucity of girls in contemporary China due to male-biased reproductive technology practices and female-biased infanticide). The present data suggest that demographic skews may have long-term, even multigenerational consequences, including significant changes in the quality of life in a social group. Materials and Methods Subjects were a troop, Forest Troop, of olive baboons (Papio anubis) living in the Masai Mara Reserve of Kenya. Olive baboons live in multimale troops of 30–150 animals, with polygamy and considerable male–male aggression. Males change troops at puberty and, as adults, achieve ranks in somewhat fluid dominance hierarchies. In contrast, females remain in their natal troop, inheriting a rank one below that of their mother. Subjects were observed each summer from 1978–1986, and continuously since 1993. An additional troop, Talek Troop, was observed continuously since 1984. Behavioral data were collected as 20-min focal samples ( Altmann 1974 ). During years of only summer observation (Forest Troop, 1978–1986), 45 samples were collected per subject per season; otherwise, an average of three samples per subject per week were collected throughout the year. Sampling was distributed throughout the day in the same fashion for each individual. During samples, social behavior, feeding, and grooming were recorded. Rankings were derived from approach–avoidance interactions, which included avoidances, supplants, and presentations, in the absence of aggression. Escalated aggression included open-mouthed lunges, chases, and bites. Nearest neighbor scans were done before and after each sample. Reproductive success was indirectly estimated from frequencies of matings and consortships (maintenance of exclusive mating with and proximity to an estrous female for at least one sample). The value of any given consortship or mating was adjusted by the probability of a fertile mating occurring that day ( Hendrickx and Kraemer 1969 ). Endocrine data were collected under circumstances allowing for measures of basal steroid hormone levels ( Sapolsky and Share 1997 ). Subjects were darted unaware with anesthetic from a blowgun syringe between 7 A.M. and 10 A.M., and only on days on which they were not sick, injured, in a consortship, or had not recently had a fight. Blood samples were collected within 3 min of anesthetization.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC387274.xml

509414

The Case of the Noisy Neurons

People are unpredictable. One night you may crave Italian food, but another only Thai will do. One day you might finish a crossword puzzle in record time, and the next not a single clue prompts an answer. Such behavioral variation has been found in laboratory studies, too: a person's ability to find a faint image on a screen varies widely from one viewing to the next. Similarly, when an animal repeatedly receives the same stimulus—for example, a faint image—a neuron in a region of the animal's visual brain might be very active upon one presentation and relatively quiet the next. Across the cerebral cortex—the brain region that integrates the senses and controls voluntary movement—neurons are notorious for their unpredictable behavior. The neurons themselves don't create this noise; when directly stimulated with an electrode multiple times, neurons will give the same response every time. Most neurons, however, receive signals from a host of other neurons. These various signals combine to form a seemingly noisy electrical input, which shows up as fluctuations in the recipient neuron's membrane potential—a difference in electrical charge between the inside and outside of the cell's membrane. Neuron function is intimately tied to the membrane potential, which is usually maintained within a narrow range, called the resting potential. But incoming signals can push the resting potential higher or lower. If the membrane potential rises above a certain threshold, the neuron fires, sending an electrical signal down its length. In this way, the brain relays and processes information. Since the 1960s, neuroscientists trying to account for the cortex's variable responses have pointed to noisy inputs from other parts of the brain as the prime suspect. In this issue of PLoS Biology , Matteo Carandini addresses this longstanding mystery of neuron variability and comes up with a different answer. Carandini simultaneously measured the membrane potentials and firing patterns of individual neurons in the cat visual cortex. He found, surprisingly, that the membrane potentials varied much less than the firing patterns, ruling out noisy inputs as the cause of neurons' noisy outputs. Instead, the neurons amplified noise in the signals they received. Noise and threshold make neurons unpredictable Carandini then used a simple model of neuron behavior to explain why this would occur. He started with a tried-and-true approximation of neuron behavior, called the rectification model: a neuron doesn't fire until its membrane potential rises above a threshold, but once it crosses this threshold, its firing rate is correlated with the strength of incoming signals. Then he added the assumption that the neurons receive signals with some randomness. Given these minimal assumptions, Carandini showed that neurons fed a noisy signal will tend to amplify the noise in the signal. Importantly, his model reproduced a well-known phenomenon: as cortical neurons' average firing rate goes up, their firing rate also becomes more variable—that is, they get noisier. Carandini's model also predicted something new: as the firing rate continues to increase, the firing rate should become more consistent and less noisy—which he calls saturation of variability. Carandini's measurements in cats showed neurons actually behave this way, a key validation of his model. It's not clear whether this amplification of variability is something that helps or hampers the brain. Despite being a nuisance to neuroscientists, such fluctuations could be crucial to how the brain functions, Carandini speculates. Without some variability in their cortex, animals would act like cameras or other simple machines that respond the same way each time to a stimulus. It's advantageous for behavior, and hence brains, to be adaptable. But amplifying noise in a signal seems to run counter to relaying and processing the information in the signal. Carandini suggests that what appears as noise in the experiments are signals from other parts of the cortex—that is, noise is in the eye of the beholder. Now that the source of the variability is clear, neuroscientists can study whether it serves a function in the brain.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509414.xml

549540

Mass spectrometrical analysis of recombinant human growth hormone (Genotropin®) reveals amino acid substitutions in 2% of the expressed protein

Background The structural integrity of recombinant proteins is of critical importance to their application as clinical treatments. Recombinant growth hormone preparations have been examined by several methodologies. In this study recombinant human growth hormone (rhGH; Genotropin ® ), expressed in E. coli K12 , was structurally analyzed by two-dimensional gel electrophoresis and MALDI-TOF-TOF, LC-MS and LC-MS/ MS sequencing of the resolved peptides. Results Electrospray LC-MS analysis revealed one major protein with an average molecular mass of 22126.8 Da and some additional minor components. Electrospray LC-MS/MS evaluation of the enzymatically digested Genotropin ® sample resulted in the identification of amino acid substitutions at the residues M 14 , M 125 , and M 170 ; di-methylation of K 70 (or exchange to arginine); deamidation of N 149 , and N 152 , and oxidation of M 140 , M 125 and M 170 . Peak area comparison of the modified and parental peptides indicates that these changes were present in ~2% of the recombinant preparation. Conclusion Modifications of the recombinant human growth hormone may lead to structural or conformational changes, modification of antigenicity and development of antibody formation in treated subjects. Amino acid exchanges may be caused by differences between human and E. coli codon usage and/or unknown copy editing mechanisms. While deamidation and oxidation can be assigned to processing events, the mechanism for possible di-methylation of K 70 remains unclear.

Background The structural integrity of recombinant products generated by prokaryotic and eukaryotic organisms is a major concern. Modifications such as amino acid sequence substitution/mutations of recombinant proteins may lead to pharmacological inactivation, autoimmune phenomena [ 1 - 3 ] and adverse effects [ 4 , 5 ]. Human growth hormone (hGH) replacement is a frequent therapeutic intervention [ 6 , 7 ]. Genetic changes in human growth hormone have been linked to biological inactivity and disease: Lewis et al (2004) reported that a growth hormone variant I 179 _M 179 showed decreased ability to activate the extracellular signal-regulated kinase pathway and Binder et al. (2002) described hGH deficiency due to mutations of the coding regions of the growth hormone-1 gene [ 8 , 9 ]. Zhu et al. (2002) reported a case of hGH R 183 _H 183 . This single mutation causes autosomal dominant growth hormone deficiency type II by prolonged retention time of R 183 _H 183 aggregates into secretory granules [ 10 ]. However, although such changes can be detrimental, non functional sequence alteration induced by poor editing of recombinant proteins may act as a marker of growth hormone abuse in situations such as athlete doping. We therefore were highly interested in the homogeneity and structure of rhGH preparations. Genotropin ® is expressed by E. coli, strain K12 . It consists of a single polypeptide chain containing 191 amino acids and two disulfide bonds (C 53 -C 165 ; C 182 -C 189 ) [ 11 ] with a molecular mass of 22 124 Da – representing the most abundant growth hormone form in humans [ 12 ]. In humans two major hGH splicing variants have been described, a 22 kDa protein and a 20 kDa protein, that bind different sites at the growth hormone receptor and serve different biological activities [ 13 , 14 ]. The genetic origin of hGH is the hGH-N gene, located on the long arm of chromosome 17, in a 66-kbp cluster region closely related to four other genes: hGH-V, hCS-A, hCS-B and hCS-L. The hGH-N gene is expressed in both, pituitary and several nonpituitary sites [ 12 ], all other gene products are produced by placental syncytio-trophoblasts. A series of posttranslational modifications of hGH have been described and range from N-glycosylation, acetylation, deamidation, oxidation at M 14 and M 125 to polymerisation [ 12 , 15 - 18 ]. As mentioned above, Genotropin ® is expressed by E. coli. Since the fidelity of hGH translation in E. coli cannot rely on copy editing [ 19 , 20 ], nor on correct codon usage [ 21 - 23 ], there is a large potential for sequence errors. That's why investigations of structural/sequential integrity, including amino acid exchanges/mutations, and post translational modifications of rhGH Genotropin ® is of particular interest to for modern medicine and pharmacotherapy. The aim of the present study was to investigate the homogeneity of a commercial available rhGH, Genotropin ® . This was achieved using two dimensional gel electrophoresis (2-DE), matrix-assisted laser desorption/ionisation mass spectrometry (MALDI-MS) followed by tandem mass spectrometry (MALDI-MS/MS) and liquid chromatography mass spectrometry (LC-MS) followed by tandem mass spectrometry (LC-MS/MS). These modern analytical tools provide definitive structural analysis independent of antibody availability and specificity. Results Two dimensional gel electrophoresis Two dimensional gel electrophoresis (2-DE) of 1 mg Genotropin ® showed a multiple spot pattern with masses between 20 000 and 35 000 Da and pIs from 4.5 to 7.0. Several two dimensional (2D) gels with sample amounts of 0.5, 1, 2, 5, 10, 20, 50, 100, 200 and 500 g of Genotropin ® were performed. Decreasing protein load showed reduction of spot size and number and finally, limitation to two spots of 22000 Da with pI of 5.3 and 5.4. Neither MALDI-TOF-TOF nor LC-MS analysis of picked gel spots indicated any modifications or isoforms in an amount, that would explain differences between the two spots. Electrospray LC-MS measurements of the Genotropin ® sample Electrospray liquid chromatography – mass spectrometry (LC-MS) measurements of the intact Genotropin ® have shown that the main product was a molecule with an average molecular weight (MW) of 22126.8 Da. The manufacturers had determined the average MW of Genotropin ® to be 22124 Da. The mass difference of approximately 3 Da may originate from the deconvulation of some broader, lower intensity peaks. Several minor components could be also detected. (Figure 1 ) The mass differences between the main product (Nr. 1) and components Nr. 2–5 respectively indicate the oxidation of several amino acid residues. Components Nr.6 and Nr.7 show a mass discrepancy of approximately +268 Da and -19 Da respectively. According to the ratios of peak areas (Table 1 ), the sample consists to 84.4% of the unmodified main component; the oxidation products are present to 13.7% of the whole sample and the ratio of other minor components, which may represent additional modifications or amino acid substitutions, is approximately 2%. Figure 1 Reconstructed electrospray LC/MS spectrum of the Genotropin ® sample. The spectrum was recorded in positive ionization modus; 1 pM of the protein was injected. Detected average molecular masses: Nr.1.:22126.8 Da; Nr.2: 22143.5 Da; Nr.3: 22158.7 Da; Nr.4: 22174.4 Da; Nr.5: 22240.7 Da; Nr.6: 22395.9 Da; Nr7.: 22107.6 Da Table 1 Relative peak areas of the components detected by electrospray LC/MS measurement Molecular mass (Da) Area (cps) Area (%) 22126.9 4909.8 84.4 22143.5 479.7 8.2 22158.7 121.5 2.1 22174.4 120.8 2.1 22107.6 76.9 1.3 22240.7 69.9 1.2 22395.9 40.4 0.7 Average molecular masses of the components detected in the LC/MS spectrum of 1 pM Genotropin ® sample. The peak areas were calculated from the reconstructed spectrum (Figure 1) recorded in positive ionization. Electrospray LC-MS/MS measurements following the tryptic digestion of the Genotropin ® sample Electrospray tandem liquid chromatography – mass spectrometry (LC-MS/MS) measurements of the samples prepared from one dimensional SDS-PAGE indicated mass differences at several peptides. Doubly or triply charged ions were chosen for all MS/MS experiments due to their better fragmentation pattern. Table 2 shows the sequences of the modified peptides and possible explanations for the mass discrepancies. Table 2 Results of electrospray LC-MS/MS and MALDI-TOF MS/MS measurements Sequence M calculated (Da) M observed (Da) delta M (Da) Modification EETQQ K SNLELLR 1586.83 1614.74 27.91 K 70 : di-methylation or → R * FDT N SH N DDALLK 1488.68 1489.60 0.92 N 149 / N 152 deamidation R LEDGSPR 928.47 900.46 -28.01 R 127 → Q or K * LFDNA M LR 978.50 994.40 15.90 M 14 : oxidation D M DKVETFLR 1252.61 1268.40 15.79 M 170 : oxidation DLEEGIQTL M GR 1360.61 1376.66 16.05 M 125 : oxidation LFDNA M LR 978.50 960.40 -18.10 M 14 → I DLEEGIQTL M GR 1360.61 1342.60 -18.01 M 125 → I D M DKVETFLR 1252.61 1234.60 -18.01 M 170 → I L FDNAMLR 978.50 1035.40 56.90 Carbamidomethyl – N terminus • Nearly isobaric mass differences Sequences and mass differences of modified peptides detected in the Genotropin sample. Column 1: sequence of modified peptides following tryptic digestion of Genotropin ® ; column 2: calculated monoisotopic masses of the unmodified tryptic peptides; column 3: observed monoisotopic masses; column 4: differences of calculated and observed masses; column 5: possible explanations of the mass differences. A mass difference of +28 Da was detected at the position K 70 (Figure 2 ), could be explained by the di-methylation of this residue, or by the exchange of this lysine to an arginine. These modifications result in a mass difference of 28.03 Da and 28.01 Da respectively. The accuracy of the mass spectrometric detection was not high enough to differentiate between these possibilities. Figure 2 shows the fragment spectrum of the peptide EETQQKSNLELLR. Intensive y ions verify that all residues have unchanged masses except of K 70 , which makes the localization of the mass discrepancy on that lysine residue unambiguous. Figure 2 Electrospray LC-MS/MS spectrum of the modified peptide EETQQK*SNLELLR. Positive ionization product ion spectrum of the tryptic peptide m/z 808.4 generated by a linear ion trap mass spectrometer. Intensive y fragment ions verify that the K 170 residue shows a mass increment of 28 Da compared to its theoretical mass. Deamidation of the amino acids N 149 and N 152 was also detected. The molecular mass of the peptide RLEDGSPR was decreased with 28 Da. The mass difference could be localized to the N terminus of the peptide and might indicate the substitution of R 127 with a lysine or glutamine. The mass difference of these residues is only 0.04 Da and the accuracy of the mass spectrometric detection was not high enough to differentiate between these amino acids. Residues M 14 , M 125 and M 170 were observed partly oxidized and in some cases the non oxidized residue showed a mass discrepancy of -18 Da (Table 2 ). This phenomenon is illustrated by Figure 3 , which shows a product ion spectrum of the modified peptide LFDNAMLR. Fragment ions from the y series verify the mass reduction at the M 14 residue. This mass difference can be explained by the replacement of these methionines with isoleucines, which can be originate from the substitution of the last base in the genetic codon of methionine (M:ATG; I:ATT/C/A). According to the ratios of the peak areas of the peptides containing the unmodified and possibly substituted methionines, these changes were present at < 2% of the whole protein amount. A mass increase of 57 Da was detected at the peptide LFDNAMLR. It could be localized at the N terminus of the peptide and it is supposed to be an artefact of the alkylation step during sample preparation. All modifications were partial; in each case peptides with both modified and unmodified residues were present. LC-MS/MS spectra for all modified peptides are available as supplementary material. Figure 3 Electrospray LC-MS/MS spectrum of the peptide LFDNAM*LR. Positive ionization product ion spectrum of the tryptic peptide m/z 482.3 generated by a linear ion trap mass spectrometer. Ions of the y series verify the mass discrepancy of -18 Da at the M 14 residue compared to its theoretical value. MALDI analysis of Genotropin ® Approximately 96 spots were excised from a 2D gel with a sample load of 1 mg Gentotropin ® and identified by MALDI-TOF on the basis of peptide mass matching [ 24 ] following in gel digestion with trypsin. Those samples which were analysed by peptide mass fingerprinting from MALDI-TOF were additionally analysed using LIFT-TOF/TOF MS/MS from the same target. A maximum of three precursor ions per sample were chosen for MS/MS analysis. Genotropin ® was unambiguously identified by MS and MS/MS Data (Figure 4 and 5 ), with a maximum of 24 matching peptides, representing a sequence- coverage of 86% to human growth hormone sequence present in database (Figure 4 , Table 3 ). All picked and analysed spots showed similar peptide mass fingerprints. Only the oxidation status of M varied, represented by a mass difference (ΔM) of 16 Da. Oxidation at M 14 was demonstrated in 59,52% of analysed spots, 80,91% of M 125 and 54,87% of M 170 showed oxidation too (Table 3 ). Neither changes in amino acid sequence, nor post translational modifications like phosphorylation or deamidation could be detected by this method. Figure 4 MS spectrum of Genotropin showing oxidation of M 14 , M 125 and M 170 . (a) MS spectrum of Genotropin ® generated by an Ultraflex™ TOF/TOF (Bruker Daltonics) operated in the reflector mode for MALDI-TOF peptide mass fingerprint (PMF). Enlarged sections (b-d) from PMF of Genotropin ® showing oxidised/non-oxidised status (ÄM+16Da) of Methionine. (b) first peak: LFDNAMLR (979.529 Da)/ second peak: LFDNAMLR + oxidation of Methionine (995.519 Da); (c) first peak: DMDKVETFLR (1253.587 Da) / second peak: DMDKVETFLR + oxidation of Methionine (1269.572); (d) first peak: SVFANSLVYGASDSNVYDLLKDLEE-GIQTLMGR (3605.040) / second peak: SVFANSLVYGASDSNVYDLLKDLEE-GIQTLMGR + oxidation of Methionine (3621.064) Figure 5 MS/MS spectra of Genotropin ® . LIFT-TOF/TOF (MS/MS) (a, b) spectra of Genotropin ® generated by an Ultraflex™ TOF/TOF (Bruker Daltonics) operated in LIFT mode for MALDI-TOF/TOF fully automated using the FlexControl™ software. The parent ions (m/z 1205.56 and m/z 2342.12) were selected for further analysis by MS/MS and the amino acid sequences NYGLLYCFR (5a) and LHQLAFDTYQEFEEAYIPK (5b) were unambiguously assigned to human growth hormone. Table 3 Sequence coverage and peptide masses of Genotropin ® Sequence Coverage: 86% Matched peptides shown in Bold 1 FPTIPLSRLF DNAMLRAHRL HQLAFDTYQE FEEAYIPKEQ KYSFLQNPQT 51 SLCFSESIPT PSNREETQQK SNLELLRISL LLIQSWLEPV QFLRSVFANS 101 LVYGASDSNV YDLLKDLEEG IQTLMGR LED GSPRTGQIFK QTYSKFDTNS 151 HNDDALLKNY GLLYCFRKDM DKVETFLR IV QCRSVEGSCG F Start – End Observed Mr(expt) Mr(calc) Delta Miss Sequence 1–8 930.61 929.60 929.53 0.07 0 FPTIPLSR 1–16 1906.97 1905.96 1906.01 - 0.05 1 FPTIPLSRLFDNAMLR Oxidation (M) 9–16 979.51 978.50 978.50 0.01 0 LFDNAMLR 17–38 2706.34 2705.33 2705.32 0.00 1 AHRLHQLAFDTYQEFEEAYIPK 20–38 2342.12 2341.11 2341.13 - 0.02 0 LHQLAFDTYQEFEEAYIPK 20–41 2727.35 2726.34 2726.32 0.02 1 LHQLAFDTYQEFEEAYIPKEQK 39–64 3058.51 3057.50 3057.45 0.05 1 EQKYSFLQNPQTSLCFSESIPTPSNR 42–64 2673.26 2672.26 2672.25 0.00 0 YSFLQNPQTSLCFSESIPTPSNR 42–70 3416.79 3415.78 3415.60 0.18 1 YSFLQNPQTSLCFSESIPTPSNREETQQK 65–77 1587.80 1586.79 1586.83 - 0.03 1 EETQQKSNLELLR 78–94 2055.19 2054.18 2054.19 - 0.01 0 ISLLLIQSWLEPVQFLR 95–115 2262.09 2261.08 2261.12 - 0.04 0 SVFANSLVYGASDSNVYDLLK 95–127 3621.02 3620.02 3619.77 0.25 1 SVFANSLVYGASDSNVYDLLKDLEEGIQTLMGR Oxidation (M) 116–127 1377.62 1376.62 1376.66 - 0.05 0 DLEEGIQTLMGR Oxidation (M) 141–158 2097.01 2096.00 2095.98 0.02 1 QTYSKFDTNSHNDDALLK 146–158 1489.66 1488.65 1488.68 - 0.03 0 FDTNSHNDDALLK 146–167 2676.26 2675.25 2675.24 0.01 1 FDTNSHNDDALLKNYGLLYCFR 159–167 1205.56 1204.55 1204.57 - 0.02 0 NYGLLYCFR 159–168 1333.64 1332.63 1332.66 - 0.03 1 NYGLLYCFRK 169–178 1253.58 1252.57 1252.61 - 0.04 1 DMDKVETFLR 169–178 1269.57 1268.56 1268.61 - 0.04 1 DMDKVETFLR Oxidation (M) Sequence coverage of 86% and 21 matching peptides of Genotropin ® to human growth hormone could be detected using MALDI MS/ MS-MS data. Discussion The dominant protein in the Genotropin ® preparation has an average molecular mass of 22127 Da. Electrospray LC-MS/MS evaluation of the trypsinized recombinant human growth hormone (rhGH) resulted in the identification of amino acid substitutions at residues M 14 , M 125 and M 170 . Di-methylation of K 70 or exchange to arginine, deamidation of N 149 and N 152 , and oxidation of M 14 , M 125 and M 170 were also observed. These sequence alteration account for 2% of the recombinant protein. Amino acid exchanges of a rhGH has been described before: Gellerfors et al. (1990) describe exchanges rhGH Q 65 _V 65 and rhGH Q 66 _K 66 [ 25 ]. Since the product was not identified we cannot compare our results. Binding of recombinant human growth hormone to the GH receptor may be modified by the five amino acid exchanges observed in the present study. Pal et al. (2003) calculated binding energy differences between modified human growth hormone (hGHv; M 14 _W 14 ) and wild type human growth hormone (hGHwt; M 14 ) with the result that the hGHv had more binding affinity to its receptor than hGHwt [ 26 , 27 ]. Cunningham et al. showed that M 14 influenced binding, even if it is not a "hot spot" for linkage to its receptor. Furthermore, the amino acid exchanges detected may very well lead to antigenic differences and thus form the molecular basis for eliciting immune responses. The underlying cause of amino acid exchanges may be codon usage and/or absence of copy editing in E. coli: The M_I exchanges may be due to miscast of the third nucleoside of the cognate anticodon at the so-called Wobble-position, i.e. switch cytosine to guanine/adenosine, a phenomenon described by Crick as "Wobble- hypothesis [ 28 ]. Crick (1966) postulated a certain amount of wobble at the third base position of the codon allowing more than one possible codon-anticodon- base pairing. Methionine (M)_I exchange of rhGH Genotropin ® may have been generated as the base pair G-G / G-A was replacing G-C. Arginine (R)_K/Q and R_G exchanges of rhGH Genotropin ® may be due to difficulties in translation of the rare codon AGG. Kane et al. (1995) predicted translational problems with an abundant mRNA species containing an excess of rare tRNA codons that may arise after the initiation of transcription of a cloned heterologous gene in the E. coli host [ 21 ]. Recent studies suggest clusters of AGG/AGA codons can reduce both quantity and quality of the synthesized protein [ 22 , 29 ]. Translational modification normally does not include amino acid exchanges but rather frameshift mutations/deletions [ 21 , 29 - 31 ]. In summary, we found two different pathways for amino acid exchanges in Genotropin ® : translation errors due to usage of (1) the rare codon AGG in E. coli and (2) incorrect codon usage consisted with Crick's "Wobble-hypothesis". Oxidative modification of a recombinant human growth hormone has been described by Karlsson et al. (1999) who demonstrated M 14 and M 125 oxidation as detected by LC-MS [ 32 ]. No other group have reported oxidation of M 170 as in our study. Indeed Teh et al. (1987) oxidised natural hGH extracted from pituitary glands and detected M 14 and M 125 oxidation by reversed phase chromatography [ 33 ]. Gellerfors et al. (1990) oxidised rhGH with hydrogen peroxide but again failed to show oxidation of M 170 , as detected by reversed phase chromatography [ 25 ]. It is not known whether oxidation of methionines in recombinant human growth hormone leads to functional impairment but conformational changes are unlikely as proposed by circular dichroism and 1H-NMR studies [ 33 ]. It is worth mentioning that oxidised methionines are not localised at the receptor binding site. Post translational modification such as N-acetylation, N-glycosylation, deamidation and oxidation have been reported for rhGH, hGH and bovine growth hormone (bGH) [ 15 - 17 , 34 ]. Dimethylation of K 70 in rhGH and hGH have not previously been reported. Whether transmethylation occurred during processing or is a post translational event during rhGH production in E. coli is unknown. Nevertheless, Martal et al. (1985) demonstrated reduction of biological activity of hGH and bGH by methylation and ethylation of its residues K 41 , K 70 , and K 115 [ 35 ]. Therefore, dimethylation of K 70 in Genotropin ® could have biological relevance, probably reducing its pharmacotherapeutic activity. Deamidation of N 149 and N 152 may be due to technical processing, probably by heat treatment or lyophilisation and has already been reported by Gellerfors et al (1990) and Karlsson et al. (1999) [ 25 , 32 ]. Though these appear to have no function significance [ 26 , 27 , 36 ]. Modifications of the recombinant human growth hormone, as shown in this study, may effect functionality and safety depending on the prevalence of such forms in the preparation. As already mentioned above, impaired binding to the receptor, conformational changes leading to impaired function, amino acid exchanges as mutations may well lead to immune phenomena or even disease [ 1 - 3 , 37 ]. In addition, such modifications may act as markers of these proteins in situations like rhGH doping. Conclusions Using one- and two-dimensional gel electrophoresis, electrospray LC-MS, LC- MS/MS and MALDI-TOF-TOF mass spectrometry we detected a series of modifications of the recombinant human growth hormone (Genotropin ® ) including amino acid exchanges, oxidation, di-methylation and deamidation. This analytical battery is a reliable, specific and sensitive analytical tool for this purpose. Methods Sample preparation Genotropin © MiniQuick 1,0 mg (Pharmacia & Upjohn; Stockholm, Sweden) was suspended in 0,5 ml of sample buffer consisting of 8 M urea (Merck, Darmstadt, Germany), 4% CHAPS (3- [(3-cholamidopropyl) dimethylammonio]-1-propane-sulfonate) (Sigma, St. Louis, MO, USA), 10 mM 1,4-dithioerythritol (Merck, Germany) and 0,5% carrier ampholytes "Resolyte" 3,5–10 (BDH Laboratory Supplies, Electran ® , England). The suspension was transferred into Ultrafree-4 centrifugal filter units (Millipore, Bedford, MA), for desalting and concentrating proteins. Protein content of the supernatant was quantified by the Bradford protein assay system [ 38 ]. The standard curve was generated using bovine serum albumin and absorbance was measured at 595 nm. One-dimensional SDS-polyacrylamide gel electrophoresis One dimensional SDS-polyacrylamide gel was performed as described by Laemmli [ 39 ]. Samples of 0.5, 1, 2, 5, 10, 30, 50 and 100 μg were loaded on the gel. For determination of molecular weight 10 μl of precision plus protein standards, all blue (Bio Rad, California, USA), were applied on the gels. Two-dimensional gel electrophoresis (2-DE) 2 DE was performed essentially as reported [ 40 ]. Samples of 1 mg protein were applied on immobilized pH 3–10 nonlinear gradient strips in sample cups at their basic and acidic ends. Focusing was started at 200 V and the voltage was gradually increased to 8000 V at 4 V/min and then kept constant for a further 3 h (approximately 150,000 Vh totally). After the first dimension, strips (18 cm) were equilibrated for 15 min in the buffer containing 6 M urea, 20% glycerol, 2% SDS, 2% DTT and then for 15 min in the same buffer containing 2.5% iodoacetamide instead of DDT. After equilibration, strips were loaded on 9–16% gradient sodium dodecylsulfate polyacrylamide gels for second-dimensional separation. The gels (180 × 200 × 1.5 mm) were run at 40 mA per gel. Immediately after the second dimension run, gels were fixed for 12 h in 50% methanol, containing 10% acetic acid, the gels were stained with Colloidal Coomassie Blue (Novex, San Diego, CA) for 12 h on a rocking shaker. Molecular masses were determined by running standard protein markers (Biorad Laboratories, Hercules, CA) covering the range 10–250 kDa. pI values were used as given by the supplier of the immobilized pH gradient strips (Amersham Bioscience, Uppsala, Sweden). Excess of dye was washed out from the gels with distilled water and the gels were scanned with ImageScanner (Amersham Bioscience). Electronic images of the gels were recorded using Adobe Photoshop and Microsoft Power Point Softwares. Matrix-assisted laser desorption ionisation mass spectrometry Spots were excised with a spot picker (PROTEINEER sp™, Bruker Daltonics, Germany), placed into 96-well microtiter plates and in-gel digestion and sample preparation for MALDI analysis were performed by an automated procedure (PROTEINEER dp™, Bruker Daltonics) [ 41 , 42 ]. Briefly, spots were excised and washed with 10 mM ammonium bicarbonate and 50% acetonitrile in 10 mM ammonium bicarbonate. After washing, gel plugs were shrunk by addition of acetonitrile and dried by blowing out the liquid through the pierced well bottom. The dried gel pieces were reswollen with 40 ng/μl trypsin (Promega, U.S.A.) in enzyme buffer (consisting of 5 mM Octyl β-D-glucopyranoside (OGP) and 10 mM ammonium bicarbonate) and incubated for 4 hrs at 30°C. Peptide extraction was performed with 10 μl of 1% TFA in 5 mM OGP. Extracted peptides were directly applied onto a target (AnchorChip™, Bruker Daltonics) that was load with α-cyano-4-hydroxy-cinnamic acid (Bruker Daltonics) matrix thinlayer. The mass spectrometer used in this work was an Ultraflex™ TOF/TOF (Bruker Daltonics) operated in the reflector mode for MALDI-TOF peptide mass fingerprint (PMF) or LIFT mode for MALDI-TOF/TOF fully automated using the FlexControl™ software. An accelerating voltage of 25 kV was used for PMF. Calibration of the instrument was performed externally with [M+H] + ions of angiotensin I, angiotensin II, substance P, bombesin, and adrenocorticotropic hormones (clip 1–17 and clip 18–39). Each spectrum was produced by accumulating data from 200 consecutive laser shots. Those samples which were analysed by PMF from MALDI-TOF were additionally analysed using LIFT-TOF/TOF MS/MS from the same target. A maximum of three precursor ions per sample were chosen for MS/MS analysis. In the TOF1 stage, all ions were accelerated to 8 kV under conditions promoting metastable fragmentation. After selection of jointly migrating parent and fragment ions in a timed ion gate, ions were lifted by 19 kV to high potential energy in the LIFT cell. After further acceleration of the fragment ions in the second ion source, their masses could be simultaneously analysed in the reflector with high sensitivity. PMF and LIFT spectra were interpreted with the Mascot software (Matrix Science Ltd, London, UK). Database searches, through Mascot, using combined PMF and MS/MS datasets were performed via BioTools 2.2 software (Bruker). A mass tolerance of 100 ppm and 2 missing cleavage sites for PMF and MS/MS tolerance of 0.5 Da and 1 missing cleavage sites for MS/MS search were allowed and oxidation of methionine residues was considered. The probability score calculated by the software was used as criterion for correct identification. The algorithm used for determining the probability of a false positive match with a given mass spectrum is described elsewhere [ 43 ]. Nano-electrospray LC-MS and LC-MS/MS analysis Genotropin ® MiniQuick 0.6 mg (Pharmacia & Upjohn; Stockholm, Sweden) was suspended in the solution provided in the two-chamber cartridge and diluted with 1% formic acid (Merck; Darmstadt, Germany) in water (Maxima, Elga; High Wycombe, UK) to 1 pM/μl. 1 μl of this solution was used for the nano-electrospray LC-MS investigation. The HPLC used was an UltiMate™ system (Dionex Corporation; Sunnyvale, CA, USA) equipped with a PepMap C18 purification column (300 μm × 5 mm) and a 75 μm × 150 mm analytical column of the same material. 0.1% TFA (Pierce Biotechnology Inc.; Rockford, IL, USA) was used on the Switchos module for the binding of the peptides and a linear gradient of acetonitrile (Chromasolv ® , Sigma-Aldrich; Seelze, Germany) and 0.1% formic acid in water was used for the elution. The gradient was (A = 5% acetonitrile / 0.1% formic acid in water; B = 80% acetonitrile / 0.1% formic acid in water) 0% B for 12 min, 80% B in 30 min, 100 % B in 3 min, 100% B for 10 min, 0% B in 2 min, 0% B for 23 min. The flow rate was 240 nl/min. The LC-system was coupled on-line to a QSTAR Pulsar hybrid mass spectrometer (Applied Biosystems; Foster City, CA, USA). The nanospray source of Proxeon (Odense, Denmark) was used with the distal coated silica nanospray capillaries of New Objective (Woburn, MA, USA). The electrospray voltage was set to 1800 V. Spectra were acquired over the mass range of m/z 600–1600. The accumulation time was 1 sec. Protein spectra were deconvoluted by Analyst ® (Applied Biosystems; Foster City, CA, USA). LC-MS/MS analyses were carried out also with the UltiMate™ system interfaced to the QSTAR Pulsar or to an LTQ (Thermo; San Jose, CA, USA) linear ion trap mass spectrometer. The gradient was (A = 5% acetonitrile / 0.1% formic acid in water B = 80% acetonitrile / 0.1% formic acid in water) 0% B for 12 min, 60% B in 88 min, 100 % B in 5 min, 100% B for 10 min, 0% B in 5 min, 0% B for 20 min. Peptide spectra were recorded over the mass range of m/z 450–1300, MS/MS spectra were recorded in information dependent data acquisition over the mass range of m/z 50–1600. One peptide spectrum was recorded followed by two MS/MS spectra on the QSTAR Pulsar instrument; the accumulation time was 1 sec for peptide spectra and 2 sec for MS/MS spectra. The collision energy was set automatically according to the mass and charge state of the peptides chosen for fragmentation. One full spectrum was recorded followed by 3 MS/MS spectra on the LTQ instrument, automatic gain control was applied and the collision energy was set to the arbitrary value of 35. Doubly or triply charged ions were selected for product ion spectra. MS/MS spectra were interpreted by Mascot (Matrix Science Ltd, London, UK). Competing interests The author(s) declare that they have no competing interests. Authors' contributions Felizardo, Maureen carried out MALDI-TOF-TOF analysis. Raja, Karlin participated in sequence alignments and in the design of the study. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549540.xml

548689

Evaluating the evidence for models of life course socioeconomic factors and cardiovascular outcomes: a systematic review

Background A relatively consistent body of research supports an inverse graded relationship between socioeconomic status (SES) and cardiovascular disease (CVD). More recently, researchers have proposed various life course SES hypotheses, which posit that the combination, accumulation, and/or interactions of different environments and experiences throughout life can affect adult risk of CVD. Different life course designs have been utilized to examine the impact of SES throughout the life course. This systematic review describes the four most common life course hypotheses, categorizes the studies that have examined the associations between life course SES and CVD according to their life course design, discusses the strengths and weaknesses of the different designs, and summarizes the studies' findings. Methods This research reviewed 49 observational studies in the biomedical literature that included socioeconomic measures at a time other than adulthood as independent variables, and assessed subclinical CHD, incident CVD morbidity and/or mortality, and/or the prevalence of traditional CVD risk factors as their outcomes. Studies were categorized into four groups based upon life course design and analytic approach. The study authors' conclusions and statistical tests were considered in summarizing study results. Results Study results suggest that low SES throughout the life course modestly impacts CVD risk factors and CVD risk. Specifically, studies reviewed provided moderate support for the role of low early-life SES and elevated levels of CVD risk factors and CVD morbidity and mortality, little support for a unique influence of social mobility on CVD, and consistent support for the detrimental impact of the accumulation of negative SES experiences/conditions across the life course on CVD risk. Conclusions While the basic life course SES study designs have various methodologic and conceptual limitations, they provide an important approach from which to examine the influence of social factors on CVD development. Some limitations may be addressed through the analysis of study cohorts followed from childhood, the evaluation of CVD risk factors in early and middle adulthood, and the use of multiple SES measures and multiple life course analysis approaches in each life course study.

Background The adult behavioral lifestyle theory of cardiovascular disease (CVD) describes an adult's lifestyle choices and levels of physiologic risk factors as the primary predictors of CVD risk [ 1 - 3 ]. This approach is supported by a relatively consistent literature demonstrating an inverse, graded relationship between SES and CVD [ 4 - 12 ]. There is also a growing literature that focuses upon the influence of SES at different points in life on adult CVD risk. A force behind the interest in the impact of early-life conditions on adult health is the fetal origins (Barker) hypothesis. This hypothesis, which has been met with considerable support and criticism, posits that poor nutrition during fetal and early infant development ("critical periods") can increase risks for adult disease. While the dominance of the adult lifestyle model of chronic disease development may not be threatened by evidence for the fetal origins hypothesis, a reconsideration of the primary importance of adult behaviors and risk factors is underway. A life course approach to chronic disease proposes that the combination, accumulation, and/or interaction of the social environments and biological insults experienced throughout the life course impact current and future events, environments, and health conditions and thus ultimately impact adult health [ 13 , 14 ]. Various interrelated theories have been put forward, [ 14 - 23 ] and many study designs have been utilized to examine the impact of life course SES [ 13 , 17 , 19 , 23 - 30 ]. This review describes the major groups of conceptual life course SES models, and then categorizes and summarizes studies that examine the associations between life course SES and CVD risk. Studies are grouped by the basic life course design utilized; summaries include methodologic critiques and descriptions of certain key studies. Evidence supporting each conceptual life course model is considered and future research directions are discussed. Life course SES conceptual models The different extant hypotheses on the influence of life course SES on CVD can be grouped into four broad conceptual models: the latent effects, pathway, social mobility, and cumulative life course models. Most studies reviewed tested the influence of life course SES on CVD outcomes via the operation of one or more of these models. The latent effects model The "latent effects" life course conceptual model hypothesizes that adverse early life experiences increase the risk of CVD in later life, independent of intervening SES, lifestyle, or traditional CVD risk factors. Power & Hertzman (1997) describe a latent effects model wherein certain early life events may have strong independent effects on adult health [ 14 ]. Kuh and Ben-Schlomo (1997) propose the related concept of biological chains of risk, wherein prenatal and early life socioeconomic factors affect biologic resources and directly influence adult health [ 13 ]. A recent formulation suggests the operation of biological or developmental influences during early "sensitive periods" which permanently impact the organism [ 28 ]. The pathway model In this life course model, early life events and environments influence later life experiences, opportunities, and health risk factors. Hertzman et al. (2001) propose a developmental process linking early-life psychosocial environments with adult health risk via pathway effects, wherein early experiences place an individual onto a certain "life trajectory," eventually impacting adult health [ 28 ]. Similarly, Kuh et al. (1997) use the concept of "social chains of risk," whereby early events influence later life experiences, thus impacting adult disease risk [ 20 ]. Blane (1999) describes an "ongoing social process" wherein a continuity of social circumstances are linked and may create a "chain of disadvantage [ 31 ]." While a pathway life course model is intuitively appealing, its operation is difficult to test empirically. Life course studies typically collect information on participants at two or three time points, which does not permit the continuous, lifelong operation of pathway effects to be observed. The social mobility model Social mobility theories all hypothesize that SES mobility across the life course impacts adult health, although the different proposed theories posit different health effects. Forsdahl (1978) hypothesized that deprivation in early life followed by later affluence combine to produce elevated CHD mortality risk, partly via elevation of adult cholesterol levels [ 32 ]. Others proposed that natural "health selection" occurs, wherein less healthy individuals tend to have downward social mobility and healthier individuals tend to be upwardly mobile [ 30 , 33 ]. In contrast, the "health constraint" hypothesis contends that socially mobile individuals possess health characteristics of both the SES group they leave and the one they join, so that social mobility minimizes the health differences between SES groups [ 15 , 34 ]. The cumulative model A cumulative SES life course model hypothesizes that psychosocial and physiological experiences and environments during early and later life accumulate to influence adult disease risk. Davey-Smith et al. (2002) suggest that if factors operating at different life stages are combined, large differences in CVD risk will be observed [ 35 ]. Kuh et al. (1997) describe an individual's biological resources accumulated over the life course as their 'health capital' [ 20 ], which describes and influences current and future health. The "accumulation of risk" model described by Ben-Schlomo and Kuh (2002) proposes that the impacts of different life course events accumulate but do not interact [ 26 ]. Methods We conducted a MEDLINE search using keywords and MeSH terms related to SES, life course, early life, longitudinal studies, CVD and CVD risk factors. References identified and books on life course research were also examined for study citations. Inclusion criteria were: (1) publication date between January 1966 and July 2003; (2) SES or related measures as independent variables; (3) outcomes of subclinical CHD, CVD morbidity and/or mortality, or traditional CVD risk factors. Behavioral risk factors were limited to: adult diet, physical activity, alcohol consumption, and smoking. Physiological outcomes were: measures of obesity, blood pressure, intima-media thickness, fibrinogen, insulin resistance, dyslipidemia, and lung function. Ecologic studies, studies primarily evaluating the fetal origins hypothesis, and studies only examining the impact of birth weight, height, upper leg length, patterns of growth, household crowding, geographic mobility or numbers of siblings were excluded. Forty-nine studies met the inclusion criteria. Quantitative summarization of the study findings is problematic due to varied study populations, SES measures and CVD outcomes evaluated, time points evaluated, and designs utilized. Moreover, a tally approach to summarizing across studies based on statistical significance can be misleading due to the different strengths, weaknesses, power and biases of each study (p. 671) [ 36 ]. Summaries therefore consist primarily of qualitative information on the types and directions of associations observed and the findings reported by authors. Although we avoid focusing on statistical significance, in most studies the authors' decision to declare a positive finding was tied to a null-hypothesis significance test using a p-value of < 0.05. Categorization of life course studies by life course study design Studies were categorized by their life course study design(s) and study question(s) into four groups: early SES → outcome, early SES → risk factor, social trajectory, and cumulative SES studies. Several studies appear in multiple categories, as they employed more than one of these designs. Table 1 describes the research questions and hypotheses typically considered by each design. Table 1 Life course study designs: hypotheses tested and typical study questions posed Life course study design Typical chronological analysis set-up used in study design Typical study question Early SES → Outcome Early-life SES variable(s) used to predict CVD (e.g., CHD, stroke). Adjusted for later-life events, behaviors, risk factor levels to determine "direct" effect of early life SES. Is there a significant independent effect of early-life (childhood) SES on the adult risk of CVD after adjusting for later-life SES and risk factors? Early SES → Risk Factor Early-life SES variable(s) used to predict adult CVD risk factors levels. How does early-life SES affect later-life levels of behavioral and physiological CVD risk factors? Social Trajectory Inter- or intragenerational movement from one SES level to another (i.e., Low SES to High SES) used to predict adult CVD risk factors levels or CVD outcomes. How does social mobility from one point to another during the life course affect the risk of CVD? Cumulative SES A summary variable indicating number of negative SES events/environments over the life course used to predict adult CVD risk factors levels or CVD outcomes. How does the accumulated number of negative SES-related exposures across the life course influence the risk of CVD? CHD = Coronary heart disease. Results Early SES → outcome life course studies: overview These studies examined the effect of childhood and/or adolescent SES on risk of adult CVD (e.g., incident myocardial infarction (MI), CVD death, incident/fatal stroke) [ 37 - 56 ] and typically tested the latent effects hypothesis (i.e., examining the "direct" effect of early-life SES on adult CVD risk). They commonly measured the independent effect of childhood SES by statistically adjusting for later-life SES and CVD risk factors in regression models. Early SES → outcome studies: summary of results Table 2 lists these 20 studies according to CVD outcome utilized (see Additional file 1 for greater detail on these studies). All 19 studies conducting unadjusted or age-adjusted analyses reported a point estimate consistent with an inverse association between early-life SES and risk of one or more of the adult cardiovascular outcomes. Authors of 14 studies concluded that their findings supported some or all of their study hypotheses [ 39 , 41 - 45 ], 47 [ 49 - 55 ]. Fourteen of 16 studies adjusting for adult SES and/or CVD risk factors reported some indication of an inverse association; however, in only five did the authors suggest that their results support the hypothesis of an inverse association between early life SES and CVD. Table 2 Summary of SES – CVD life course studies using an early SES → outcome design 1 st Author, year & reference number Study name Early life SES measures Variables adjusted for other than age Acute/ Survived MI, CVD Burr 1980 [49] South Wales hospital cohort Father's occup (RG, 3 groups), father unemployed (> 1 year), family size Current SES Notkola 1985 [40] East-West Study 5-level index: father's occup & farm size Current SES, CVD RF's Coggon 1990 [37] Stoke-on-Trent & Newcastle study Father's occup (RG, 5 groups), height, sibling death Current SES, smoking Hasle 1990 [48] Danish worker's union study 8 variables (yes/no) on parent's occup, health, household, residence, edu, illness None Kaplan 1990 [43] Kuopio Study Factor analysis of edu, occup, farm, farm size, perceived wealth CVD RF's Lundberg 1993 [51] Swedish population cohort study 4 yes/no variables: economic hardship, large family, broken family, dissension Current SES, gender Gliksman 1995 [39] Nurses' Health Study Father's occup at 16 years Current SES, CVD RF's Lamont 2000 [41] Newcastle 1,000 Families Cohort Birth: father's occup; 5 & 10 years: parent's occup, housing, # of adverse life events Current SES, CVD RF's Marmot 2001 [46] Whitehall II Study Childhood: father's occup (RG, 4 groups), age at leaving edu Labor force entry: Occup Current & child SES Wamala 2001 [45] Stockholm Study Early-life SES disadvantage index (0–3) of 3 variables: large family, born last, low edu Current SES, CVD RF's Stroke Gliksman 1995 [39] Nurses' Health Study Father's occup (4 groups) at 16 years Current SES, CVD RF's Coggon 1990 [37] Stoke-on-Trent & Newcastle study Father's occup (RG, 5 groups), height, sibling death Current SES, smoking Davey Smith 1998 [44] Collaborative Study Father's occup (RG, 4 groups), also mnl vs. non-mnl Current SES, CVD RF's Frankel 1999 [54] Boyd Orr Cohort Father's occup (RG, 5 groups) Townsend area deprivation score Dedman 2001 [56] Boyd Orr Cohort Persons/room, tap water (yes/no), toilet type, ventilation, cleanliness (3 levels each) Childhood SES, area deprivation score CHD Mortality Notkola 1985 [40] East-West Study 5-level index using father's occup & farm size Current SES, CVD RF's Lynch 1994 [38] Kuopio Study SES index (3 groups), by parents' edu, occup, farm, perceived wealth Current SES Vagero 1994 [50] Uppsala Birth Cohort Study Occup of head of household (mnl, non-mnl, unemployed) Current SES Gliksman 1995 [39] Nurses' Health Study Father's occup (4 groups) at 16 years Current SES, CVD RF's Davey Smith 1998 [44] Collaborative Study Father's occup (RG, 4 groups), also divided into mnl vs. non-mnl Current SES, CVD RF's, area deprivation Hart 1998 [42] Collaborative Study Early SES: father's occup; Labor force entry: Occup; At screening: occup None Frankel 1999 [54] Boyd Orr Cohort Father's occup (RG, 5 groups) Townsend area deprivation score Davey Smith 2001 [47] Glasgow Alumni Cohort Father's social class, (RG, 5 groups) CVD RF's Dedman 2001 [56] Boyd Orr Cohort Persons/room, tap water (yes/no), toilet type, ventilation, cleanliness (3 levels each) Childhood SES, Townsend area deprivation score Davey Smith 2002 [52] Collaborative Study Father's occup (mnl/non-mnl) Current SES, CVD RF's Claussen 2003 [53] Oslo Mortality Study Index of housing conditions items Current SES Osler 2003 [55] Project Metropolit Father's social class (3 groups) by occup Birth weight, IQ at age 12 CHD = Coronary heart disease; Edu = Education; IHD = Ischemic heart disease; MI = Myocardial infarction; Mnl = Manual occupational class; Non-mnl = Non-manual occupational class; Occup = Occupation; RF = Risk factor; RG = Registrar General's social class categories. For studies with CHD and stroke outcomes limited to early-life SES (i.e., not controlling for adult SES or other risk factors) almost all included a point estimate consistent with an inverse association. However, in adjusted analyses, authors of only one study with a CHD-related outcome and fewer than half the studies of CVD mortality reported inverse adjusted associations. Thus, the existence of a "direct effect" of early-life SES after adjusting for adult risk factors was not strongly supported. Early SES → outcome studies: methodologic issues To prevent statistical confounding by adult conditions, most studies employed statistical models that adjusted the association between childhood SES and adult CVD risk for adult SES and behavioral/ physiological CVD risk factors. This may be an over-adjustment, as certain risk factors (e.g., BMI, smoking) are part of the pathways through which low childhood SES may influence CVD risk. Little information is usually available on how participants' early-life SES may have influenced levels of adult CVD risk factors. Additionally, if unknown variables influence both the distribution of a CVD risk factor adjusted for and the distribution of the outcome measure, then adjustment will result in an incorrect partitioning of the total effect of early-life SES into "direct" and "indirect" components [ 57 - 59 ]. Thus, early SES → outcome studies which seek to determine the "direct," adjusted effect of early-life SES on CVD risk may incorrectly estimate this effect [ 60 ], leading to questions about the accuracy of such estimates. Some studies examined and qualitatively compared the effects of adjustment for different classes of CVD risk factors. Gliksman et al. (1995), for example, analyzed the association between father's occupational class and adult CVD in a series of models adjusting for different classes of potential mediators or confounders. This allowed for a more complete description of the impact of adult environment, behavior, and physiologic risk factors versus the latent impact of early life socioeconomic factors [ 39 ]. Studies utilizing more than one life course study design ("mixed-design studies") take a related approach and are considered in the Discussion section. Early SES → risk factor studies: overview Life course studies in the early SES → risk factor group evaluated the influence of early- and/or mid-life SES or living conditions on later-life behavioral or physiologic CVD risk factors [ 14 , 40 , 44 , 52 , 61 - 73 ]. These designs are often similar to those used in the early SES → outcome studies group except that CVD risk factors are the outcomes. The life course hypotheses typically examined, however, include both the latent effects and pathway conceptual models. The typical study design involves a statistical model where early-life SES measures predict the levels of CVD risk factors in adulthood. Early SES → risk factor studies: summary of results Table 3 outlines the 17 studies reviewed (see Additional file 2 for detailed summaries). Five of six studies reported associations between low early-life SES and little or no adult leisure-time physical activity [ 14 , 61 , 63 , 64 , 66 ], five of five found associations with high adult alcohol intake [ 52 , 61 , 63 , 68 , 73 ], and eight of twelve studies reported associations with higher smoking rates as study findings [ 14 , 44 , 52 , 63 , 65 - 68 ]. Nine of 11 studies reported (unadjusted) associations between lower early-life SES and elevated BMI or WHR [ 14 , 44 , 63 , 64 , 67 , 68 , 70 , 71 , 73 ]. The impact of a statistical adjustment for other CVD risk factors varied, including no impact [ 59 ], attenuation with a persistence of the significant effect [ 46 , 55 ], and strong attenuation with associations no longer apparent [ 54 , 78 ]. Table 3 Summary of SES – CVD life course studies using an early SES → risk factor design 1 st Author, year & reference number Study name Early life and/or adult SES measures Variables adjusted for other than age CVD risk factor(s) measured Arnesen 1985 [65] Tromso Heart Study Early-life: 4-level index of household economic conditions CVD RF's Cholesterol, SBP, glucose, BMI, smoking, more Notkola 1985 [40] East-West Study Early-life: 5-level index using father's occup & farm size; Adult: occup (6 groups) None Smoking, cholesterol, SBP Wadsworth 1985 [98] Medical Research Council National Survey of Health Study Early-life: index of father's occup & parents' edu; Adult: occup (RG) & employment CVD RF's SBP, DBP Braddon 1986 [71] British 1946 Birth Cohort Early-life: 2 social class indices of father's occup; Adult: occup (RG, 8 groups), edu (high/low) Adult SES, CVD RF's Obesity (BMI > 30.0 for men, > 29.1 for women) Peck 1994 [66] Swedish census cohort study Early-life: father's occup (7 groups); Adult: occup (7 groups) None Smoking, physical activity Blane 1996 [64] Collaborative Study Early-life: father's occup (RG, 4 groups); Adult: occup (RG, 4 groups) Adult SES DBP, physical activity, smoking, BMI, FEV1, more Lynch 1997 [99] Kuopio Study Child: SES index (3 groups); Adolescent: edu (3 groups); Adult: occup, income, possessions, more Energy intake Smoking, drinking, obesity, physical activity, diet Power 1997a [14] 1958 British Birth Cohort Child: father's occup (RG, 4 groups); At 23 years: occup, edu; At 33 years: occup None BMI (obesity) Power 1997b [90] 1958 British Birth Cohort Child: father's occup (RG, 4 groups); At 23 years: occup, edu; At 33 years: occup None Smoking, BMI (obesity) Davey Smith 1998 [44] Collaborative Study Early-life: father's occup (4 groups); Adult: occup (6 groups) None Smoking, DBP, cholesterol, BMI, FEV1 van de Mheen 1998 [63] Longitudinal Study, Netherlands Early-life: father's occup (6 groups); Adult: occup (6 groups) Adult SES BMI, smoking, alcohol, leisure physical activity Brunner 1999 [67] Whitehall II Study Early-life: father's occup (RG, 4 groups); Adult: occup (Civil Service grade, 4 groups) Adult SES Smoking, activity, HDL, BMI, fibrinogen, more Davey Smith 2002 [52] Collaborative Study Early-life: father's occup (mnl/non-mnl); Adult: occup (mnl/non-mnl) None Smoking, alcohol, area deprivation Lawlor 2002 [68] British Women's Heart Study Early-life: father's occup (RG, 6 groups); Adult: current occup (RG, 6 groups) Adult SES Insulin resistance, SBP, cholesterol, BMI, smoking, triglycerides, alcohol, more Poulton 2002 [73] Dunedin Multidisciplinary Study Early-life: parental occup (6 groups) at 0, 3, 5, 7, 9, 11, 13 & 15 years; Adult: current occup (3 groups) Adult SES BMI, WHR, SBP, smoking, alcohol, more Lawlor 2003 [69] British Women's Heart Study Early-life: father's longest occup (mnl / non-mnl); Adult: longest occup (RG, 6 groups) CVD RF's HOMA, SBP, HDL, triglycerides Parker 2003 [70] Newcastle 1000 Families Study Birth: father's occup & housing; 5 & 10 years: same, plus adverse life events; Adult: wage earner's occup None CMS, BMI, WHR, fasting insulin, triglycerides, HDL BMI = Body mass index; CMS = Central metabolic syndrome; DBP = Diastolic blood pressure; Edu = Education; FEV1 = Forced expiratory volume in 1 second; HOMA = Homeostasis model assessment score; Mnl = Manual occupational class; Non-mnl = Non-manual occupational class; Occup = Occupation; RF = Risk factor; RG = Registrar General's social class categories; SBP = Systolic blood pressure; WHR = Waist-to-hip ratio. Early SES → risk factor studies: methodologic issues As with the early SES → outcome studies, concerns that lack of adequate adjustment or over-adjustment may lead to biased estimates are germane to this group of studies. The issue of covariate adjustment is complex in these studies, given that many CVD risk factors are considered. Certain physiologic risk factors (e.g., insulin resistance, dyslipidemia) may be linked to early-life SES through latent physiologic effects of negative childhood exposures, continuous exposure to negative physiologic stimuli, or other physiologic or life course pathway mechanisms [ 14 , 63 , 70 , 74 ]. In contrast, the links between early-life SES and levels of behavioral risk factors (e.g., smoking, diet), and certain physiologic risk factors (e.g., BMI, hypertension), may operate principally through learned behaviors, behavioral responses to negative psychosocial stimuli, or other psychosocial life course pathway effects. The associations may develop through the operation of multiple physiologic and psychosocial pathways, making "correct" covariate adjustment challenging within a single model. Thus, statistical adjustment should take into account the specific life course pathways hypothesized to be operating [ 39 ]. Previous findings should be used to generate a priori hypotheses; nonetheless, there is considerable opportunity for inappropriate adjustments to be made. Social trajectory studies: overview Studies evaluating the social mobility life course model typically considered the impact of inter-generational or intra-generational social mobility on CVD and CVD risk factors [ 38 , 44 , 49 , 64 , 72 , 73 , 75 - 79 ]. Inter-generational mobility was usually determined by contrasting the participant's father's occupational SES to the participants'. Intra-generational SES was typically defined as a change in occupational SES from early adulthood to later adulthood. Six of 11 studies evaluated adult CVD risk factors as outcomes [ 49 , 64 , 72 , 73 , 75 , 77 ]; seven evaluated CVD mortality or CHD [ 38 , 44 , 75 - 79 ]. Four adjusted only for age [ 38 , 49 , 76 , 80 ]. Social trajectory studies: summary of results Table 4 outlines these studies (see Additional file 3 for greater detail). Of 10 studies carrying out statistical analyses [ 38 , 44 , 49 , 72 , 73 , 75 - 79 ], six did not report associations between upward or downward mobility and either elevated levels of CVD risk factors or increased CVD morbidity or mortality when compared to stable low-SES or high-SES trajectories [ 38 , 44 , 49 , 75 , 78 , 79 ]. Nine studies, however, reported the suggestion of inverse, although not always statistically significant, relationships between social mobility and a CVD-related outcome [ 38 , 44 , 72 , 73 , 75 - 79 ]. Two studies found marked differences in CVD mortality risk between upwardly mobile individuals and individuals maintaining the same SES across time. However, one study reported increased CVD risk among the upwardly mobile [ 77 ] and one reported decreased mortality risk [ 79 ]. Table 4 Summary of SES – CVD life course studies using a social trajectory design 1 st Author, year & reference number Study name SES measures Variables adjusted for other than age CVD risk factor(s) / outcomes measured Kaplan 1971 [76] Evans County Heart Study Early-life: father's occup (7 groups); Early Adult & Adult: occup & social class (5 classes) None MI, chronic IHD, AP, sudden death Gillum 1978 [77] Harvard Alumni Cohort Early-life: by parental occup (blue/white collar); Adult: assumed to be at least middle class A confounder summarizing score AP, HTN, fatal or non-fatal CHD or MI Burr 1980 [49] South Wales hospital cohort Father's occup (RG, 3 groups), father unemployed (> than 1 year), family size, current occup (RG, 3 groups) None Survived MI Wadsworth 1985 [72] British 1946 Birth Cohort Early-life: index of father's occup and parents' edu; Adult: occup (RG), employment & edu for women Smoking, edu, father's CVD, BMI, more SBP, DBP Faresjo 1994 [78] Swedish Study of Men Born in 1913 Early-life: father's SES (3 groups) by occup, social class; Adult: same; all mobility relative to child SES SBP, cholesterol, smoking MI Lynch 1994 [38] Kuopio Study Early-life: SES index (3 groups); Adult: current income (2 groups) None CVD mortality Blane 1996 [64] Collaborative Study Early-life: father's occup (RG, 4 groups); Adult: occup (RG, 4 groups) BMI, cholesterol, DBP, smoking, activity, FEV1 BMI, cholesterol, DBP, smoking, activity, FEV1 Hart 1998 [75] Collaborative Study Early-life: father's occup (mnl/non-mnl); Labor force entry: occup (mnl/non-mnl); Adult: occup (mnl/non-mnl), area deprivation index Smoking, DBP, cholesterol, FEV1, angina, ischemia CVD mortality & risk factors listed Davey Smith 1998 [44] Collaborative Study Early-life: father's occup: (RG, mnl/non-mnl); Adult: occup: (RG, mnl/non-mnl) Smoking, DBP, BMI, area deprivation, more Mortality from CHD & stroke Poulton 2002 [73] Dunedin Multidisciplinary Study Early-life: parental occup (6 groups) at 0, 3, 5, 7, 9, 11, 13 & 15 years; Adult: current occup (3 groups) Infant health index, gender, adult SES BMI, WHR, SBP, smoking, alcohol dependence Pensola 2003 [79] Finnish census cohort Early-life: father's occup (mnl vs. non-mnl); Adult: current occup (mnl vs. non-mnl) None CVD mortality AP = Angina pectoris; BMI = Body mass index; CHD = Coronary heart disease; DBP = Diastolic blood pressure; Edu = Education; FEV1 = Forced expiratory volume in 1 second; HTN = Hypertension; IHD = Ischemic heart disease; MI = Myocardial infarction; Mnl = Manual occupational class; Non-mnl = Non-manual occupational class; Occup = Occupation; RG = Registrar General's social class categories; SBP = Systolic blood pressure; WHR = Waist-to-hip ratio. Four studies examined differences in CVD risk between stable low-SES trajectory and stable high-SES trajectory individuals [ 38 , 44 , 75 , 79 ]. Three reported that individuals with stable low-SES trajectories had a greater CVD risk than stable high-SES trajectory individuals [ 44 , 75 , 79 ]; the fourth reported a marginally significantly greater risk [ 38 ]. These results are similar to those of most cumulative SES studies (described below) in that greater exposure to low SES was associated with increased CVD risk. Social trajectory studies: methodologic issues In these studies the unit of analysis is a trajectory, permitting the impact of change over time to be examined. However, the socioeconomic trajectories in most reviewed studies were limited to two time points, and groups compared tended to share the same SES at one of these time points. These similarities may partly explain why seven of ten studies did not report an association between social mobility and risk of CVD risk factors or events. Social trajectory studies incorporating SES at three or more time points allow for the analysis of more informative trajectories than studies evaluating SES at only two points [ 75 ]. Yet, even with large studies, analysis becomes cumbersome if more than two or three SES levels at three time points are measured. Additionally, the impact of uncommon (e.g. downward) trajectories are difficult to study due to the small numbers of individuals who typically comprise them. Cumulative SES studies: overview These studies tested the operation of the cumulative life course conceptual model, typically by summing the number of times participants experienced unfavorable SES situations during early, middle or later life, and creating SES indices representing the accumulation of these experiences [ 45 , 52 , 53 , 68 , 79 , 81 , 82 ]. For example, three of these studies summed the number of times a participant (or their parents during the participant's childhood) had been in a manual occupation to create an index of accumulated low-SES exposure [ 68 , 81 , 82 ]. Cumulative SES was also measured using indices of occupational class, socioeconomic categories, exposures to negative socioeconomic experiences/conditions, income, and housing conditions. [ 45 , 52 , 53 , 68 , 81 , 82 ]. Cumulative SES studies: summary of results Table 5 summarizes the seven cumulative SES studies (see Additional file 4 for greater detail). All authors reported that participants' cumulative life course exposure to low SES conditions was associated with increases in CVD outcome, supporting the cumulative life course SES hypothesis. Several studies indicated that cumulative SES was a more powerful predictor of CVD morbidity and/or mortality than adult or early-life SES alone [ 45 , 53 , 79 , 81 ]. In studies that adjusted for CVD risk factors, graded associations were attenuated but remained strong in two studies [ 81 , 82 ] and were greatly attenuated in another [ 45 ]. Table 5 Summary of SES – CVD life course studies using a cumulative SES design 1 st Author, year & reference number Study name Cumulative SES measure(s) Variables adjusted for other than age CVD risk factor(s) / outcomes measured Davey-Smith 1997 [82] Collaborative Study Sum of # of times at mnl vs. non-mnl SES using father's, own first, & own current occup class BMI, DBP, FEV1, cholesterol, smoking, AP, ischemia, more CVD, mortality, AP, smoking, BMI, DBP, cholesterol, more Heslop 2001 [81] Collaborative Study Sum of # of times at mnl vs. non-mnl SES using father's, own first, & own current class DBP, BMI, FEV1, cholesterol, activity, smoking, alcohol CVD mortality, DBP, BMI, FEV1, exercise, smoking, alcohol, more Wamala 2001 [45] Stockholm Study Sum of # of instances (0–6) of SES disadvantage (large family, born last, low edu, blue-collar/ housewife, economic hardship) Height, HDL, HTN, marriage, fibrinogen, obesity, smoking, more Cases: CHD event (acute MI, unstable/ recurrent AP) Davey-Smith 2002 [52] Collaborative Study Sum of # of risks (0–6): Father mnl SES, left edu at < 15 years, current mnl SES, smoking, high alcohol, high deprivation area None CVD mortality Lawlor 2002 [68] British Women's Heart Study Cross-classification of father's longest occup and current occup (RG, mnl/non-mnl) None Insulin resistance, HTN, smoking, triglycerides, LDL, HDL, BMI, more Claussen 2003 [53] Oslo Mortality Study Early life: Index of housing conditions (scored 0–7) Adulthood: standardized income (7 groups) None CVD mortality Pensola 2003 [79] Finnish census cohort Sum of # of times in mnl vs. non-mnl class, by father's occup & own occup at 30–34 None CVD mortality AP = Angina pectoris; BMI = Body mass index; CHD = Coronary heart disease; DBP = Diastolic blood pressure; Edu = Education; FEV1 = Forced expiratory volume in 1 second; HTN = Hypertension; MI = Myocardial infarction; Mnl = Manual occupational class; Non-mnl = Non-manual occupational class; Occup = Occupation; RG = Registrar General's social class categories. Davey-Smith et al. (2002) employed a unique cumulative SES measure, combining early and later-life occupational class experience with the CVD risk factors of smoking and heavy alcohol consumption [ 52 ]. They reported a marked, graded relationship between the number of negative SES exposures and risk of CVD mortality. Only a large cohort study of Norwegians reported a statistically significant supra-multiplicative interaction between early- and later-life SES on risk of CVD [ 53 ]. Another study reported marginally statistically significant supra-multiplicative interaction effects between early and later life [ 45 ]. Cumulative SES studies: methodologic issues The reported associations between cumulative SES and CVD risk are more consistent than those reported in the other life course study designs. However, three of the seven studies evaluated were based on the same cohort [ 52 , 81 , 82 ]. Cumulative life course SES variables include current SES, and therefore may be conflating the effect of current SES with that of SES over the life course. For example, in Pensola et al. (2003), social mobility analyses suggested that the observed gradient in CVD mortality associated with cumulative manual occupational class was driven primarily by the impact of current occupational status. Additionally, cumulative indices implicitly assume that a specific negative life experience or situation has the same impact regardless of when it occurs in an individual's lifetime, with no distinction made between the impact of childhood versus adult events on risk of disease [ 45 ]. Furthermore, some cumulative SES studies combined disparate measures into a single, lifetime index variable, or summed the number of times different negative events or exposures occur [ 45 , 52 ]. This approach assumes that different types of exposures have an equivalent impact on the risk of CVD. Evidence supporting these assumptions is not provided in these studies. Discussion General life course SES study issues Certain general limitations and assumptions of life course SES studies should be considered. First of all, SES is a theoretical construct operationalized using various measures (e.g., income, occupation, education) that tap into different components of this construct. Despite several proposed SES evaluation schemes [ 83 - 85 ], there is no overarching theory in the biomedical literature providing a rationale for the use of specific SES measures. Occupational status, the SES measure most commonly used in the studies reviewed, is assumed to be an adequate proxy of SES, but in fact represents only one component of SES. As associations of a given SES measure with CVD risk factors are not always consistent with those of other SES measures [ 47 , 60 ], it is important to consider that results may depend upon the proxy SES measure employed [ 64 , 81 ]. Most life course studies used retrospective cohorts or a case-control design, relying on participants' recall of early life SES. There has been little systematic evaluation of the validity of recalled early life circumstances or of the potential for such recall errors to bias associations. In studies directly comparing the impact of childhood and adult SES on adult CVD risk, greater error in childhood (vs. adult) SES measures may underestimate the true impact of child SES [ 86 , 87 ]. Additionally, selection bias due to either loss to follow-up or selective survival may distort findings. Studies using cohorts followed from birth or early life probably do not have these limitations [ 14 , 28 , 47 , 62 , 71 , 72 , 77 , 88 - 93 ] and may avoid the problem of changing occupational status and workplace functions across time if they establish SES scales at the time of measurement and update them as appropriate. As more studies analyze data from prospective cohorts starting in early life, concerns about unequal measurement error and the changing status of specific occupations should decrease [ 14 , 54 , 73 , 89 , 93 , 94 ]. Most studies were limited to white males. While the number of studies including women has increased in recent years, minority groups remain underrepresented in most life course studies. These groups may interact differently with the economic and/or social systems of the majority group; there are also established associations between minority status, low SES, and increased CVD risk [ 5 , 84 , 85 ]. The lack of information on the manner in which life course SES relates to adult CVD risk in minorities should be addressed. Support for life course effects on CVD risk The majority of the early SES → outcome group studies' results describe an inverse association between early life SES and risk of adult CHD, non-fatal MI, and CVD mortality. Similarly most early SES → risk factor studies suggest that low early-life SES negatively impacts levels of adult CVD risk factors. Most studies reported relationships between lower early-life SES and elevated alcohol intake, BMI/WHR, and insulin resistance, as well as decreased leisure physical activity. Together, the early SES → outcome and early SES → risk factor studies support the hypothesis that low early-life SES is inversely associated with adult risk of CVD. However, support for the hypothesis of a "direct effect" of early-life SES on risk of CVD mortality is equivocal as findings were less consistent after risk factor adjustment. Additionally, reported associations between early-life SES and adult health behaviors detract from the evidence supporting a latent effects life course model, as they suggest that CVD risk may be driven by learned behaviors. The operation of social chains of risk or life trajectories described by a pathway life course model are difficult to observe or test, since almost all early SES → risk factor studies consider only two points during the life course. However, significant associations between early-life SES and the adult risk factors suggest behavioral, psychosocial or environmental links that may be best explained as a pathway effect. As described below, the pathway model will need further examination through the use of series of linked studies. While nine of 10 social trajectory studies reported some relationship between inter- or intra-generational social mobility and CVD-related outcomes, only four were identified as providing evidence in favor of a social mobility hypothesis. Conversely, most studies comparing stable low- and high-SES trajectory groups reported markedly increased CVD mortality among the low-SES group, with many socially mobile groups having a CVD risk between the stable high and low groups. Determining whether these associations are a function of social mobility, or whether they are due to exposure to low SES at some point in the life course and not to mobility, is problematic. Future studies should focus on differentiating between the influences of cumulative SES and social mobility. Of the conceptual models discussed, the cumulative life course model was the most consistently supported. However, the weight of these findings is limited as much of the data came from one cohort [ 57 , 60 , 61 ]. Also, specific methodological concerns (e.g., equal weighting of life course periods, conflation of current and life-course SES) need to be considered when interpreting these findings. Future studies conducted in other cohorts that address these concerns will help clarify the viability of cumulative life course SES hypotheses. Directions for future research Researchers have noted the limitations of life course SES study designs examining only one potential pathway linking SES and CVD. Accordingly, recent studies tend to use a combination of SES measures, CVD risk factor and/or outcomes and life course study designs [ 40 , 44 , 45 , 52 , 53 , 68 ]. These mixed-design studies allow for a comparison of how well different life course SES conceptual models fit the patterns observed in the same data. Such studies are also less likely to overlook patterns of association in the data than studies which only evaluate one possible relationship between SES and CVD. As modern life course cohorts are followed from childhood, longer prospective studies or series of cross-sectional studies that evaluate associations between early-life experiences and risk factor levels at several time points may provide the opportunity to observe the operation of pathway life course effects. For example, two studies in the early SES → risk factor group [ 14 , 90 ] compared levels of CVD risk factors at two points in young adulthood (22 and 33 years of age). The authors observed an association between childhood SES and adult obesity, which was smaller for participants at 33 than at 23 years of age [ 14 ], suggesting a decreasing impact of early life SES on adult health with increasing age. Such information is only obtained through longer ongoing studies or series of interrelated studies. Individuals' socioeconomic environments may impact their CVD risk factor levels and CVD risk. Community socioeconomic conditions may influence individuals' ability to engage in leisure physical activity or eat a healthy diet [ 95 - 97 ]. Some life course studies have evaluated the community socioeconomic environment, primarily through the use of indicators of neighborhood deprivation [ 44 , 48 , 54 , 81 , 82 ]. Future studies should likewise consider the effect of physical and psychosocial environmental factors. Life course SES studies have generally failed to consider the length of exposure to the various socioeconomic conditions measured. As this may influence the impact of negative SES experiences on adult health, future cumulative life course studies may benefit from evaluating the effect of length of exposure into their indices. Conclusions The wide range of populations, analysis designs, exposures, and outcomes used in the life course studies reviewed precludes a simple, quantitative analysis of the impact of life course SES on CVD risk. Nevertheless, the results thus far modestly support the existence of life course SES effects on risk of adult CVD. The cumulative life course model is more consistently supported by extant studies than other models. However, the different methodologic issues of each study design make direct comparisons of the relative support for each conceptual model difficult. Analyses utilizing multiple life course designs within the same study offer the best approach to testing which theories best describe the links between life course SES and CVD risk. The inclusion of minority participants, different SES measures, and data from early and middle adulthood in large, prospective, mixed-design studies will allow more informed and generalizable statements to be made about the impact of life course SES on adult disease risk. While this area of research needs methodologic refinement, it offers a promising and informative perspective from which to understand the development of chronic disease. Competing interests The author(s) declare that they have no competing interests. Authors' contributions R. A. Pollitt conducted the literature search, reviewed and categorized the articles and had primary responsibility for writing the manuscript. K. M. Rose assisted in the categorization and summarization of the papers reviewed, and helped to revise the manuscript in response to the reviewer's comments. All authors participated in interpreting the studies' results and preparing the methodologic criticism, provided input on the various drafts, and 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 SES – CVD life course studies using an early SES → outcome design Click here for file Additional File 2 SES – CVD life course studies using an early SES → risk factor design Click here for file Additional File 3 SES – CVD life course studies using a social trajectory design Click here for file Additional File 4 SES – CVD life course studies using a cumulative SES design Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548689.xml

534795

Hand-carried ultrasound performed at bedside in cardiology inpatient setting – a comparative study with comprehensive echocardiography

Background Hand-carried ultrasound (HCU) devices have been demonstrated to improve the diagnosis of cardiac diseases over physical examination, and have the potential to broaden the versatility in ultrasound application. The role of these devices in the assessment of hospitalized patients is not completely established. In this study we sought to perform a direct comparison between bedside evaluation using HCU and comprehensive echocardiography (CE), in cardiology inpatient setting. Methods We studied 44 consecutive patients (mean age 54 ± 18 years, 25 men) who underwent bedside echocardiography using HCU and CE. HCU was performed by a cardiologist with level-2 training in the performance and interpretation of echocardiography, using two-dimensional imaging, color Doppler, and simple calliper measurements. CE was performed by an experienced echocardiographer (level-3 training) and considered as the gold standard. Results There were no significant differences in cardiac chamber dimensions and left ventricular ejection fraction determined by the two techniques. The agreement between HCU and CE for the detection of segmental wall motion abnormalities was 83% (Kappa = 0.58). There was good agreement for detecting significant mitral valve regurgitation (Kappa = 0.85), aortic regurgitation (kappa = 0.89), and tricuspid regurgitation (Kappa = 0.74). A complete evaluation of patients with stenotic and prosthetic dysfunctional valves, as well as pulmonary hypertension, was not possible using HCU due to its technical limitations in determining hemodynamic parameters. Conclusion Bedside evaluation using HCU is helpful for assessing cardiac chamber dimensions, left ventricular global and segmental function, and significant valvular regurgitation. However, it has limitations regarding hemodynamic assessment, an important issue in the cardiology inpatient setting.

Introduction Bedside echocardiography can bring important anatomical and hemodynamic information for the management of critically ill patients, and is often required in hospitalized patients for the assessment of left ventricular function. Standard echocardiographic equipments, while optimal, have large size and sometimes are difficult to maneuver in the emergency room or intensive care units. Recently, hand-carried ultrasound (HCU) devices have been demonstrated to broaden the versatility in ultrasound application. Due to their portability and low cost, HCU acts like a stethoscope, providing information beyond physical examination at the point-of-care [ 1 , 2 ]. Although it has been shown to improve the detection of cardiovascular abnormalities over the physical examination, its role in the assessment of hospitalized patients is not completely established [ 3 , 4 ]. This study was undertaken to compare the findings of the bedside evaluation using HCU to the comprehensive echocardiography (CE), in cardiology inpatient setting. Methods Patients We studied 44 consecutive hospitalized patients with cardiovascular disorders. Patients were included in the study when their referring physicians asked for a bedside evaluation with conventional echocardiography. In all patients, we performed the echocardiography with both HCU and CE within a maximal interval of 24 hours. The clinical characteristics of patient population are shown in Table 1 . Among these patients, 61% were in the cardiac ward, 27% in the emergency room, and 12% in the intensive care unit. This study was approved by our Institutional Ethical Committee and informed consent was obtained from all participants or their legal representatives. Table 1 Clinical characteristics Variables Age (years) 54 ± 18 Male gender 25 (57%) Cardiomyopathy 16 (36%) Acute coronary syndrome 10 (23%) Postoperative of cardiac surgery 9 (21%) Valvulopathy 5 (11%) Cardiogenic shock 3 (7%) Total atrioventricular conduction block 1 (2%) Data are mean ± SD or number (%) of patients. Study Protocol All patients underwent two echocardiographic evaluations. First, HCU was performed with the portable device OptiGo (Philips Medical Systems, Andover, Massachusetts, USA) and, consecutively, by a commercially available system (HDI 5000, Philips Medical Systems, Bothell, Washington, USA) equipped with a 4-2 MHz transducer and second-harmonic imaging. HCU was performed by one same cardiologist with level 2-training in the performance and interpretation of echocardiography according to the specifications of the American Society of Echocardiography [ 5 ], after a period of instructions about the HCU settings. CE was performed by one experienced echocardiographer with level-3 training [ 5 ], and was considered the gold standard. Each investigator was blinded to the results of the other examination. The final echocardiographic report was based on the results of CE. Imaging Analysis OptiGo is equipped with a 2.5 MHz phased-array transducer and operates on a rechargeable Lithium ion battery, which facilitates its use at bedside. HCU was performed using two-dimensional imaging, color Doppler flow mapping, and simple caliper measurements. Images were frozen and scrolled for review and the measurements were performed on-line. CE evaluation included two-dimensional with second-harmonic imaging, M-mode, and both spectral and color Doppler flow mapping. Images were recorded on videotape or digitalized. The aorta, left atrium, left ventricular end-diastolic and end-systolic diameters, as well as interventricular septal and posterior wall thickness were measured according to the recommendations of the American Society of Echocardiography [ 6 ]. The left ventricular ejection fraction (LVEF) was visually estimated, and a normal ventricular function was defined as LVEF ≥ 55%. According to the segmental wall motion analysis, patients were divided as those with segmental wall motion abnormality (WMA) and those without WMA. Pericardial effusions were classified as mild, moderate, or large effusion. Valve structure and function were analyzed. Dysfunctions were classified into mild, moderate, or severe degree according to the qualitative evaluation by HCU, and using both qualitative and quantitative parameters by CE [ 7 ]. A significant valvular regurgitation was defined in our study as those of moderate or severe degree [ 8 ]. On the other hand, non-significant valvular regurgitation was defined as those of no, trace, or mild degree. The estimation of transvalvar gradients and valvular areas in patients with prosthetic and stenotic valves, as well as the estimation of pulmonary artery pressure, were performed only by CE [ 7 ]. The intraobserver variability of CE findings was assessed in 15 randomly assigned patients, with analysis made at least 4 weeks apart. The interobserver agreement between the experienced echocardiographer and the level 2-trained cardiologist was assessed by the analysis of 15 CE recorded examinations. Statistical analysis Continuous data are expressed as mean ± one standard deviation (SD) and categorical data as proportions. Comparisons between groups for continuous variables were made using Student t test. Chi-square and Fisher Exact tests were used for categorical variables. Agreement between HCU and CE results were assessed by the Kappa statistics. Interobserver and interobserver variability was determined by intra-class correlation and linear regression. A two-tailed p value <0.05 was considered significant. Results Imaging analysis was feasible with HCU and CE in all patients. The percentage of patients with good, regular and bad image quality using CE were 80%, 18% and 2%, respectively. HCU had a lower percentage of patients with good image quality (59%), and a higher percentage of patients with regular (27%) and bad (14%) image quality (p < 0.05 versus CE), as shown in Figure 1 . Figure 1 Image quality by hand-carried ultrasound and comprehensive echocardiography. Image quality obtained by hand-carried ultrasound device (solid bars) and by comprehensive echocardiography with second-harmonic imaging (open bars). Comprehensive echocardiography had a higher percentage of patients with good quality than hand-carried ultrasound. * p < 0.05 between groups. Determination of cardiac chamber dimensions and left ventricular function There were no significant differences in cardiac chamber dimensions obtained by HCU and CE, except for the posterior wall thickness, which was lower when assessed by HCU (Table 2 ). Among the 44 studied patients, 24 (55%) had some degree of left ventricular dysfunction and 20 (45%) had normal left ventricular function. There was no difference between the LVEF estimated by CE and by HCU (Table 2 ). Table 2 Cardiac chamber measurements obtained by comprehensive echocardiography (CE) and by hand-carried ultrasound (HCU) Variables CE HCU Aorta (mm) 28.7 ± 4.1 27.8 ± 4.2 Left atrium (mm) 45.5 ± 7.7 44.0 ± 7.5 LVED (mm) 57.1 ± 11.2 54.9 ± 10.7 LVES (mm) 44.4 ± 15.5 42.8 ± 10.7 IVST (mm) 10.0 ± 2.4 9.4 ± 2.4 PWT (mm) 9.6 ± 1.8 8.7 ± 1.7* LVEF (%) 47 ± 16 44 ± 15 Data are mean ± SD. IVST = interventricular septal thickness; LVED = left ventricular end-diastolic diameter; LFEF = left ventricular ejection fraction; LVES = left ventricular end-systolic diameter; PWT = posterior wall thickness. * p < 0.05 compared to CE. The analysis of segmental wall motion was not possible by HCU in two patients, due to poor endocardial border delineation, and was deemed feasible in all patients using CE. In the remaining 42 patients, HCU correctly identified eight of the 11 patients with WMA by CE, and failed to identify three of them. These three false-negative results occurred in patients with bad image quality by HCU. On the other hand, among the 31 patients without WMA by CE, HCU correctly identified 27 patients, and had four false-positive results. Among these four cases, three patients had global left ventricular dysfunction and one had asynchronic movement of the interventricular septum. The agreement between HCU and CE for the detection of WMA was 83% (Kappa = 0.58; p < 0.001). CE identified nine patients with pericardial effusion, one patient with a moderate effusion and the remaining eight patients with mild effusions. HCU correctly detected and classified pericardial effusion in seven patients, and failed to diagnose two patients with mild effusions, both localized posterior to the left ventricle. Analysis of valvular dysfunction Significant mitral valve regurgitation was diagnosed by CE in 16 patients, and by HCU in 13 of them. The agreement between HCU and CE for detection of this abnormality was 93% (Kappa = 0.85; p < 0.001). Significant aortic valve regurgitation was detected in three patients by CE and in two by HCU, and the agreement between the two techniques was 98% (Kappa = 0.89; p < 0.001). Significant tricuspid regurgitation was detected by CE in 16 patients and by HCU in 12 of them. However, there was one additional case misdiagnosed by HCU as of significant degree. The agreement between HCU and CE for the detection of significant tricuspid regurgitation was 88% (Kappa = 0.74; p < 0.001). Four patients (9%) had aortic valve stenosis, two cases of moderate, one of mild, and the other one of severe degree. Qualitative analysis of valve structure by HCU was capable to identify aortic stenosis in two of these patients, but the severity of these lesions was not determined. There were six (14%) patients with prosthetic valves and eight (18%) patients with pulmonary hypertension, in which estimation of transvalvar gradients or valvular areas were performed only by CE. A complete evaluation of these patients was not possible using HCU due to its technical limitations in determining hemodynamic parameters. Intra and interobserver variability The intraobserver agreement for detection of pericardial effusion, WMA, and significant valvular dysfunction was 100%, and the interobserver agreement was 91%. There was an excellent correlation between LVEF estimated in the first and second evaluation by the same observer (r = 0.91; p < 0.05). The correlation between LVEF estimated by the experienced echocardiographer and the cardiologist with level 2-training in echocardiography was r = 0.88 (p < 0.05). Discussion The present study describes the value of bedside evaluation of cardiac patients using HCU in comparison to CE. Our results demonstrate that HCU may be used for the assessment of cardiac chamber dimensions, estimation of left ventricular function, and detection of WMA. Moreover, we found a good agreement between HCU and CE for detecting significant valvular regurgitation. However, HCU presents important limitations regarding the assessment of prosthetic and stenotic valves, as well as for the evaluation of patients with pulmonary hypertension, due to the lack of spectral Doppler. Bedside echocardiography is a frequently used diagnostic tool in the cardiology inpatient setting and can affect the patient management, direct further diagnostic work-up, and modify therapeutic decisions. The recent development of portable ultrasound devices has the potential to allow quick and easy-to-use echocardiography at the point-of-care, although its value in hospitalized patients was not completely defined. In patients with cardiovascular disorders, HCU has been shown to increase the diagnostic accuracy over physical examination when performed by cardiologists with level-2 training in echocardiography [ 1 ]. In the present study we confirmed the usefulness of these portable devices for estimating cardiac chamber dimensions and left ventricular function, which are frequent indications for bedside echocardiographic examination. We also demonstrated a good agreement between CE and HCU for detecting WMA. We would like to emphasize that CE with second-harmonic imaging was chosen as the gold standard for evaluation of segmental wall motion, since it has already been proven that this imaging modality ameliorates the endocardial border delineation [ 9 , 10 ]. Another point to be noted is that our population included a high proportion of patients with left ventricular global dysfunction, in whom detection of segmental abnormality can be somehow challenging by non-experienced observers. In at least three of the four false-positive results, the presence of global myocardial impairment due to cardiomyopathy could lead to a confounding diagnosis of segmental impairment. In the false-negative cases, the use of second-harmonic modality improved the quality of imaging, allowing for a better visualization of endocardial thickening and detection of segmental left ventricular dysfunction. Our results are in accordance with recently published data demonstrating that HCU was highly concordant with clinical diagnosis of acute coronary syndrome based on the analysis of wall motion by these portable devices [ 11 ]. However, when considering the full spectrum of abnormalities in hospitalized patients, previous reports demonstrated that hand-carried bedside echocardiography failed to quantify valvular regurgitation and also missed findings relevant to clinical questions in a significant number of patients. These studies concluded that HCU falls far short of standard echocardiography in evaluation of critically ill patients [ 3 , 12 ]. We do believe that the lack of spectral Doppler in HCU is an important limitation for evaluation of cardiac patients, since this technique provides unique hemodynamic information, especially in patients with prosthetic or stenotic valves, and pulmonary hypertension. In our study population, these clinical conditions occurred in a considerable number of cases, since we included patients in the pre and postoperative period of valvular surgery. Limitations In the present study, we did not evaluate the effect of HCU on patient management. Agreement between CE and HCU for detection of WMA was analyzed in 42 patients, since two patients were initially excluded because of inadequate acoustic window. The specific issue of agreement between HCU and CE according to image quality was not addressed in the present study because of the limited number of patients in each group. Conclusions and clinical implications We concluded that HCU is useful for bedside assessment of left ventricular global and segmental left ventricular function as well as for evaluation of significant valvular regurgitation. However, it has limitations regarding hemodynamic assessment, which is an important issue in the cardiology inpatient setting. Therefore, we emphasize that bedside echocardiography using HCU should be performed by cardiologist with at least level 2-training in echocardiography, and always complemented with CE when hemodynamic evaluation is necessary. Competing Interests The authors declare that they have no competing interests. Authors' contributions JMT and WMJ designed the study, did the selection and recruitment of the patients, and wrote the text. RRM and JMC performed the echocardiograms and participated in the subsequent analysis of data. JLA and JFR analyzed the statistics and revised this manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC534795.xml

509415

Gene Targeting Turns Mice into Long-Distance Runners

Have you ever noticed that long-distance runners and sprinters seem specially engineered for their sports? One's built for distance, the other speed. The ability to generate quick bursts of power or sustain long periods of exertion depends primarily on your muscle fiber type ratio (muscle cells are called fibers), which depends on your genes. To this extent, elite athletes are born, not made. No matter how hard you train or how many performance-enhancing drugs you take, if you're not blessed with the muscle composition of a sprinter, you're not going to break the 100-meter world record in your lifetime. (In case you'd like to try, that's 9.78 seconds for a man and 10.49 seconds for a woman.) Of course that doesn't prevent those at the highest levels from using the latest performance enhancer to get that extra 1% edge. But wait until trainers hear about the Marathon Mouse. A new study by Ronald Evans and colleagues provides evidence that endurance and running performance can be dramatically enhanced through genetic manipulation. Skeletal muscles come in two basic types: type I, or slow twitch, and type II, or fast twitch. Slow-twitch fibers rely on oxidative (aerobic) metabolism and have abundant mitochondria that generate the stable, long-lasting supplies of adenosine triphosphate, or ATP, needed for long distance. (For more on muscle fiber metabolism, see synopsis titled “A Skeletal Muscle Protein That Regulates Endurance”) Fast-twitch fibers, which produce ATP through anaerobic glycolysis, generate rapid, powerful contractions but fatigue easily. Top-flight sprinters have up to 80% type II fibers while long-distance runners have up to 90% type I fibers. Coach potatoes have about the same percentage of both. Endurance training can enhance the metabolic performance of muscle types, and now it appears that training can also induce conversion between fiber types. Specific changes in gene expression trigger this oxidative fiber transformation, but the transcription factor responsible for engineering this shift was unknown. Evans and colleagues suspected that a nuclear receptor called PPARδ—a major regulator of fat burning in fat tissue that is also prevalent in skeletal muscle—might be involved. To investigate this possibility, the authors genetically engineered mice to express an activated form of the PPARδ protein in skeletal muscle. Type I fibers normally express higher levels of PPARδ than type II fibers, and the transgenic mice showed much higher levels of the protein than their normal littermates. The transgenic mice also had much redder muscles than the controls—type I fibers have high levels of myoglobin, the red-pigmented protein that facilitates the movement of oxygen within muscle—and elevated levels of proteins associated with mitochondrial biogenesis and operation. A final line of evidence indicating a type I fiber switch was the elevated level of specialized type I contractile proteins and decreased level of specialized type II contractile proteins in the transgenic mice. Interestingly, these same results were seen when naturally occurring (endogenous) PPARδ levels were stimulated in the normal mice (with an orally active compound). This suggests that muscle fibers can be transformed into type I endurance fibers by simply activating the endogenous PPARδ pathway. In a weight-conscious world, oxidative fibers are thought to offer resistance against obesity since obese individuals have fewer type I fibers than average-weight individuals. Sure enough, transgenic mice fed a high-fat diet gained far less weight than normal mice fed the same diet, even in the absence of exercise. The transgenic mice had much smaller fat cells, which the authors attribute to enhanced oxidative capacity of the muscle tissue, and improved glucose tolerance. (Obese individuals lose the ability to metabolize glucose, which puts them at risk for diabetes.) But what about performance? Remarkably, the marathon mice ran about an hour longer than controls, showing dramatic improvement in both running time and distance—increases of 67% and 92%, respectively. Altogether, these results show that PPARδ drives the conversion of type I muscle fibers by activating pathways that enhance physical performance and protect against obesity. The finding that endurance and running capacity can be genetically manipulated suggests that muscle tissue is far more adaptable than previously thought. Maybe Olympiads can be made after all—but don't give up on training just yet. A full understanding of the molecular basis of muscle fiber determination, including the interactions between PPARδ and its regulatory components, awaits further study.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509415.xml

549030

Morning versus evening dosing of desloratadine in seasonal allergic rhinitis: a randomized controlled study [ISRCTN23032971].

Background A circadian rhythm of symptoms has been reported in allergic rhinitis and some studies have shown the dosing time of antihistamines to be of importance for optimizing symptom relief in this disease. The objective of this study was to examine the efficacy of morning vs. evening dosing of the antihistamine desloratadine at different time points during the day. Methods Patients ≥ 18 years, with seasonal allergic rhinitis received desloratadine 5 mg orally once daily in the morning (AM-group) or evening (PM-group) for two weeks. Rhinorrhea, nasal congestion, sneezing and eye symptoms were scored morning and evening. Wilcoxon rank sum and 2-way ANOVA test were used. Results Six-hundred and sixty-three patients were randomized; 336 in the AM-group; 327 in the PM-group. No statistically significant differences were seen between the AM and PM group at any time points. In the sub-groups with higher morning or evening total symptom score no difference in treatment efficacy was seen whether the dose was taken 12 or 24 hours before the higher score time. There was a circadian variation in baseline total symptom score; highest during daytime and lowest at night. The circadian variation in symptoms was reduced during treatment. This reduction was highest for daytime symptoms. Conclusions A circadian rhythm was seen for most symptoms being more pronounced during daytime. This was less apparent after treatment with desloratadine. No statistically significant difference in efficacy was seen whether desloratadine was given in the morning or in the evening. This gives the patients more flexibility in choosing dosing time.

Background Allergic rhinitis is a common illness, which affects approximately 15 % of the population [ 1 ] and has a large impact on the quality of life of the patients. In some studies the symptoms of allergic rhinitis have shown a circadian rhythm with morning symptoms being most prominent in a majority of patients [ 1 - 6 ]. Antihistamines are important medications in the treatment of allergic rhinitis. One should expect that the effect of an antihistamine is best near or shortly after peak serum level is attained. If this also coincides with the peak in allergy symptoms, an optimal treatment effect should be expected. In one study evening dosing of the antihistamine mequitazine (half-life of 38–45 hours and time to peak serum level about 6 hours) gave better symptom relief than morning dosing on morning symptoms [ 4 , 7 ]. Desloratadine has as mequitazine a rather long half life of 27 hours, and the time to peak serum level at about 3 hours [ 8 ]. Evening dosing of this antihistamine may be expected to give better symptom relief than morning dosing on peak morning symptoms. Some studies have also confirmed a circadian variation in efficacy of some antihistamines on histamine induced skin reactions [ 9 , 10 ]. The aim of this study was to examine the efficacy of the antihistamine desloratadine at different time points during the day and to evaluate whether the time of dosing of desloratadine has any impact on the treatment efficacy in seasonal allergic rhinitis (SAR). Methods This was a randomized, open label, parallel group, multicenter study of two weeks duration in patients with SAR during the birch or grass pollen season. Eighty one medical centers in the Nordic countries participated. The inclusion criteria were: patients 18 years or above with a minimum of two years history of SAR confirmed by either a positive skin prick test or a positive serologic allergen test to the relevant seasonal allergen; clinically symptomatic with SAR at baseline/inclusion with a minimum total nasal symptom score (rhinorrhea, congestion, itching and sneezing) of at least 6 and rhinorrhea being minimum 2 (moderate); willingness to adhere to dosing and visit schedule. Females of childbearing potential had to use medically accepted methods of birth control and written informed consent had to be obtained from all patients. The exclusion criteria were: pulmonary disease, perennial rhinitis, sinusitis, rhinitis medicamentosa, pollen desensitization during the last 6 months, respiratory tract infection within the last two weeks, structural nasal abnormalities (including polyps), use of oral, nasal, ocular decongestants, corticosteroids in any form (except mild dermatological group I corticosteroids allowed in only small areas), other antihistamines (oral or topical), any investigational drug during the last 30 days, pregnant or nursing females. The patients were randomized into one of two treatment groups with dosing of 5 mg desloratadine tablets either in the morning between 07 – 09 (AM-group) or evening between 19 – 21 (PM-group) in a 1:1 ratio. Randomizing was computer generated for the whole study population using SAS version 6.12 and performed in blocks of eight. Each subject unit (bottle with medication) was labelled with randomization number. Physicians in the different Nordic countries recruited the patients. They assigned the medication in consecutive order. The study was monitored by Schering-Plough. The following symptoms were assessed using a scale from 0 to 3 (0=none, 1=mild, 2=moderate, 3=severe): rhinorrhea, nasal congestion, sneezing, itching nose and eye symptoms (itching, burning, tearing, redness). These symptoms were recorded in a patient diary every morning (AM 12 hours reflective and AM last hour) and evening (PM 12 hours reflective and PM last hour) both at baseline and during the 2 weeks treatment period. Interference with sleep and interference with daily activity were also assessed by the patients every day using the same scale from 0 to 3. In addition, the number of hours spent outdoors was recorded. Visit 1 was at day 0 at the start of baseline, visit 2 after one week and visit 3 after two weeks. A wash-out period prior to Visit 1 was necessary if the patient had been on any drugs which could interfere with the study results (e.g. no other commonly used antihistamines allowed during the prior 10 days). Baseline symptoms were recorded in the evening at day 0 and the following morning (day 1) after which the patients started taking the study medication as randomized. A physical examination was performed at visit 1 and 3. All adverse events were recorded. The study period was from April 11 th 2001 to September 2 nd 2002. Pollen counts were not recorded. The primary objective was to evaluate the efficacy of 5 mg desloratadine taken orally once daily in the morning versus evening. The primary efficacy variable was the mean change from baseline for the AM last hour Total Symptom Score (TSS) over the 2 weeks treatment period. TSS is the sum of the individual symptom scores for the following symptoms that in prior studies [ 2 , 3 ] have shown a circadian rhythm: rhinorrhea, nasal stuffiness/congestion, sneezing and eye symptoms (maximum score 12). Since nasal itching had shown little circadian rhythm in these studies, this symptom was omitted from the TSS. AM last hour was chosen as primary time point since the symptoms had in the same studies shown to be worst in the morning. The study was designed to enrol 700 patients in order to have 600 evaluable patients. This sample size was chosen to detect with 90 % power and 5 % significance level, a difference between treatment groups of 0.6 units or more in mean change from baseline diary TSS, assuming a pooled SD of 2.25 units. In a study on morning vs. evening dosing of the antihistamine mequitazine the differences in dosing-time-related efficacy increased in the sub-group of patients having predominantly morning symptoms [ 4 ]. A sub-group analysis was therefore performed on patients with a higher TSS (≥ 1 point difference at baseline) in the morning (AM last hour) than in the evening (PM last hour) and patients with higher TSS in the evening than in the morning in this study. A comparison was then done on the treatment efficacy seen 12 hours and 24 hours after dosing (AM vs. PM dosing) in these patients. All patients receiving at least one dose of study drug and having at least one post dose registration were included in the efficacy analysis (intention-to-treat, ITT), and confirmatory analysis were based on evaluable patients with no protocol violations. Statistical analyses were made with 2-way ANOVA. For evaluation of response of therapy Wilcoxon rank sum test was used. Adverse events were tabulated. The study protocol and the patient informed consent form were approved by Ethics Committees and Health Authorities in each of the participating countries. Results Patients Six hundred and sixty-three patients were randomized at baseline; 336 to the AM-group and 327 to the PM-group. The two groups were comparable with respect to demographics and baseline characteristics (Tab 1). To assess the primary parameter 310 in the AM and 294 in the PM group fulfilled the criteria for ITT. Of the AM group 259 and of the PM group 254 patients completed the study without any violation. Mean baseline TSS varied between 4.64 and 6.10, a difference of 31%; highest during daytime (PM 12 hours reflective) and lowest at night (AM 12 hours reflective). The circadian variation at baseline was more evident for sneezing (around 60% difference between night and day), rhinorrhea and eye symptoms, less so for nasal itching and hardly noticeable for nasal congestion. Fig. 1 shows total and individual symptom scores at baseline and during two weeks treatment. The circadian variation was much less apparent during treatment with desloratadine (Fig. 1 ). Figure 1 Symptom scores. Total and individual symptom scores at baseline and over two weeks treatment period Baseline: ○ AM-group; Δ PM-group, Treatment: ● AM-group; ▲ PM-group 1 Max score = 12, 2 Max score = 3 Efficacy During the two weeks period the mean reduction in TSS ± SE for AM last hour (primary efficacy variable) was 1.63 ± 0.17 (30 %) for the AM-group and 1.80 ± 0.17 (35 %) for the PM-group. There was no statistically significant difference (ITT-analysis) between the groups at this time point (p = 0.456) or at any other time points. The reduction in TSS was highest (2.5 – 41%) for day time symptoms (PM previous 12 hours) and lowest at night. This was evident for all individual symptoms except for nasal congestion. In the subgroup analysis comparing TSS AM last hour and PM last hour at baseline, 32 % of the patients had more severe symptoms in the morning (≥ 1 point difference in TSS) than in the evening, and 37 % had more severe symptoms in the evening. Looking at these two sub-groups, no difference in treatment efficacy on TSS was seen 12 or 24 hours post dosing (Fig. 2 ). Figure 2 Sub-group Total Symptom Score. These sub-groups consists of patients with higher (one or more score points) morning TSS (AM last hour) than evening TSS (PM last hour) at baseline and of patients with higher evening TSS (PM last hour) than morning TSS (AM last hour) at baseline. There was no statistical significant difference in the treatment efficacy between the AM-group and the PM-group. According to their diaries the patients spent in average more than 3.5 hours outdoors daily. The score for the interference of SAR on the patients' sleep and daily activity at baseline and throughout the study is shown in Fig. 3 . Figure 3 Sleep and daily activity. The score for interference with sleep and daily activity at baseline and during treatment shows that there is a higher interference with daily activity at baseline and during treatment than with sleep. Safety The incidence of treatment related adverse events were comparable between the groups, 20 % in the AM-group and 18 % in the PM-group, headache being most frequent, 7 % and 4 % respectively. Discussion This study was randomized but without a placebo control. Since this study was a comparison between two different dosing times of the same medication, a placebo control was superfluous. The study was not blinded as there is no reason to believe that neither the patients nor the physicians should have a biased opinion as to the time of dosing. To blind such a study, the patients need to take study medication from different boxes in the morning and evening. However, this method was not used since this may complicate the study and impair patient compliance. A circadian rhythm has been found in many diseases, also in allergic rhinitis [ 1 - 6 ]. The effect of an antihistamine may be modulated [ 9 - 13 ] by variations in allergen exposure, hormonal activity, organ sensitivity and plasma concentration of the drug. In this study we have shown that desloratadine maintains its effect at different time points throughout the day and thus the effect appears unaffected by a modulating factor. The baseline period in this study lasted 24 hours which is the same as in the study on mequitazine [ 4 ] and other studies [ 16 , 17 ]. In some studies of the effect of antihistamines the baseline period has been longer [ 14 , 15 ]. It would have been difficult to keep patients in the Nordic countries off medication for more than one day in addition to any washout period during the pollen season. We do not believe that the duration of baseline influenced the results of this comparative study. The circadian rhythm at baseline found in this study with maximum symptoms during the day differs from some other studies [ 1 - 6 ] where more patients had the most severe symptoms in the morning. This difference may partly be due to patient selection. Patients with perennial rhinitis were excluded from our study. Thus indoor allergens do not influence symptom variation. The patients spent several hours outdoors during the day in the pollen season. It seems likely that this exposure would influence the symptoms. The circadian variation was not apparent during treatment, possibly because the suppression of symptoms by desloratadine is more observable when symptoms are most prominent. The best effect of mequitazine was obtained after evening dosing (12 hours before peak of symptoms) compared to morning dosing (24 hours before peak of symptoms). In our study no difference in treatment efficacy was seen 12 or 24 hours after dosing in the sub-group analysis of patients with higher baseline morning or evening TSS. Whatever the cause for this discrepancy between these two antihistamines, other antihistamines may show a variation in effect during the day not only on dermal symptoms [ 9 , 12 ] but also on nasal ones. Thus studies on the effect of other antihistamines in allergic rhinitis should be encouraged. The adverse events recorded were of a magnitude and nature as seen in other studies of desloratadine and other antihistamines [ 14 - 17 ]. Many patients have circadian variations in symptoms. The peak of symptoms can be at different time points from patient to patient. Individual dosing time of medication may improve symptom relief. Desloratadine, however, apparently shows no circadian variation in effect. Conclusions A circadian rhythm was seen for most SAR symptoms at baseline, being most distressing during daytime, possibly due to long outdoor exposure. This circadian variation is less apparent after treatment with desloratadine. No statistically significant difference in efficacy was seen whether desloratadine was given in the morning or in the evening. This gives the patients more flexibility in choosing dosing time. Competing interests The study was funded by Schering-Plough in the Nordic countries. None of the authors will gain financially from the publication. There are no patents pending. There are no other competing interests. Authors' contributions RH participated in the design of the protocol and is the main author of the article. KH participated in the design of the protocol and as investigator. OB participated as principal investigator in Sweden and enrolled most patients in the study. SF participated in the statistical analysis, drafted the tables and figures and participated in drafting the manuscript. TØ was project leader and participated in the design of the protocol, the statistical analysis and drafting of the manuscript. Table 1 Demographics Treatment groups Demographic Characteristics AM-group (n = 336) PM-group (n = 327) p-value Age (years) Mean (SD) 35.4 (11.0) 36.5 (12.0) 0.232 a Min-Max 18–69 18–75 Age Group (years) 18–29 115 (34.2 %) 113 (34.6 %) 30–39 113 (33.6 %) 95 (29.1 %) 40–49 70 (20.8 %) 66 (20.2 %) ≥ 50 38 (11.3 %) 53 (16.2 %) Sex Male (%) 152 (45) 163 (50) 0.235 b Female (%) 184 (55) 164 (50) Duration of SAR History (years) Mean (SD) 14.2 (9.2) 14.4 (10.3) 0.662 a Min-Max 2–50 1–55 c a t-test for differences between the means of the two treatment groups b χ 2 -test for the distribution between the two groups c One patient in the PM-group had only one year duration of SAR history although the inclusion criterion was 2 years

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549030.xml

509428

Two-stage normalization using background intensities in cDNA microarray data

Background In the microarray experiment, many undesirable systematic variations are commonly observed. Normalization is the process of removing such variation that affects the measured gene expression levels. Normalization plays an important role in the earlier stage of microarray data analysis. The subsequent analysis results are highly dependent on normalization. One major source of variation is the background intensities. Recently, some methods have been employed for correcting the background intensities. However, all these methods focus on defining signal intensities appropriately from foreground and background intensities in the image analysis. Although a number of normalization methods have been proposed, no systematic methods have been proposed using the background intensities in the normalization process. Results In this paper, we propose a two-stage method adjusting for the effect of background intensities in the normalization process. The first stage fits a regression model to adjust for the effect of background intensities and the second stage applies the usual normalization method such as a nonlinear LOWESS method to the background-adjusted intensities. In order to carry out the two-stage normalization method, we consider nine different background measures and investigate their performances in normalization. The performance of two-stage normalization is compared to those of global median normalization as well as intensity dependent nonlinear LOWESS normalization. We use the variability among the replicated slides to compare performance of normalization methods. Conclusions For the selected background measures, the proposed two-stage normalization method performs better than global or intensity dependent nonlinear LOWESS normalization method. Especially, when there is a strong relationship between the background intensity and the signal intensity, the proposed method performs much better. Regardless of background correction methods used in the image analysis, the proposed two-stage normalization method can be applicable as long as both signal intensity and background intensity are available.

Background cDNA microarrays consist of thousands of individual DNA sequences printed in a high density array on a glass slide. After being reverse-transcribed into cDNA and labeled using red (Cy5) and green (Cy3) fluorescent dyes, two target mRNA samples are hybridized with the arrayed DNA sequences or probes. Then, the relative abundance of these spotted DNA sequences can be measured. The ratio of the fluorescence intensity for each spot represents the relative abundance of the corresponding DNA sequence. In cDNA microarray experiments, there are many sources of systematic variation. Normalization is the process of removing such variation that affects the measured gene expression levels. The main idea of normalization is to adjust for artifact differences in intensity of the two labels. Such differences result from differences in affinity of the two labels for DNA, differences in amounts of sample and label used, differences in photomultiplier tube and laser voltage settings and differences in photon emission response to laser excitation. Although normalization alone cannot control all systematic variations, normalization plays an important role in the earlier stage of microarray data analysis. Many normalization methods have been proposed by using the statistical regression models. Kerr et al. [ 1 ] and Kerr et al. [ 2 ] suggested the ANOVA model approach. Wolfinger et al. [ 3 ] proposed a mixed effect model for normalization. Schadt et al. [ 4 ] proposed smoothing splines with generalized cross-validation (GCVSS). Kepler et al. [ 5 ] used a local polynomial regression to estimate the normalized expression levels as well as the expression level dependent error variance. Yang et al. [ 6 ] summarized a number of normalization methods for dual labeled microarrays such as global normalization and robust locally weighted scatter plot smoothing (LOWESS, Cleveland [ 7 ]). Workman et al. [ 8 ] proposed a robust nonlinear method for normalization using array signal distribution analysis and cubic splines. Wang et al. [ 9 ] suggested iterative normalization of cDNA microarray data to estimate normalization coefficients and to identify control gene set. Chen et al. [ 10 ] presented subset normalization to adjust for location biases combined with global normalization for intensity biases. After performing two dye cDNA microarray experiments, we get foreground and background intensities from red channel and green channel, respectively. Although a complex modeling approach can be used, the signal intensity is usually computed by subtracting the background intensity from the foreground intensity. Thus, the noise in the background intensity may have a large effect on the signal intensity. Several approaches have been proposed for decreasing the background noises in image analyses (Yang et al. [ 11 ] and Kim et al. [ 12 ]). Kim et al. [ 13 ] found out the influences of background intensities on signal intensities, and showed that background intensities could play an important role in normalization. Recently, some background correction methods have been proposed using Bayesian method or smoothing function rather than simple subtraction when defining signal intensity (Kooperberg et al. [ 14 ] and Edwards [ 15 ]. As pointed out by Kim et al. [ 13 ], the signal intensities need to be robust to the local background intensity. In general, the signal intensities tend to have some correlations with background intensities (Figure 1 ). We think it is important to reduce variation in signal intensities caused by the background intensities. However, no systematic methods have been proposed that use the background intensities in normalization. In order to make the effect of background intensities more robust to the signal intensities, we propose a new method so called 'two-stage normalization method' to adjust for the effect of the background intensities. The first stage fits a regression model to adjust for the effect of background, and the second stage applies the usual normalization method such as a nonlinear LOWESS method to the background-adjusted intensities obtained from the first stage. In order to perform the two-stage normalization method, we consider nine different background measures and investigate their performances in normalization. A detailed description on background measures is given in Methods section. Also, Methods section describes the proposed two-stage normalization methods. Results section describes the results from NCI 60 cDNA microarray experiment, which illustrates the effects of background intensities (Zhou et al. [ 16 ]). In addition, some comparative results are presented from cDNA microarrays of cortical stem cells of rat (Park et al. [ 17 ]) and those from kidney, liver, and testis cells from mice (Pritchard et al. [ 18 ]). The performance of two-stage normalization is compared to those of global normalization as well as intensity dependent nonlinear LOWESS normalization. We use the variability among the replicated slides to compare the performance of normalization methods. For certain selected background measures, the proposed two-stage normalization performs better than global or intensity dependent nonlinear normalization method. Finally, Conclusion section summarizes the concluding remarks. Methods We propose a two-stage normalization method for the cDNA microarray data analysis using background intensities. At the first stage, we adjust for the effect of background intensities on M ; at the second stage, we correct bias on M caused by other sources of systematic variation. Stage 1. Background normalization Let g fi and g bi be the means(or medians) of the i th foreground and background intensity of green channel, respectively; r fi and r bi be the corresponding means(or medians) of red channel, respectively. Then for each spot, we have two pairs of intensities: ( g fi , g bi ) and ( r fi , r bi ), i = 1,..., p , where p is the number of spots in a slide. (For simplicity, we omit the subscript and define A and M using notation of Yang et al. [ 6 ] as follows: M = [log( r f - r b ) - log( g f - g b )] = [log(R) - log(G)],     (2) In cDNA microarray experiments, there are red and green background intensities. It would be desirable to consider the background intensities that are more closely related with the signal intensities. We consider nine possible background measures from red channel, green channel, and both channels as follows: (a) Red channel Y 1 = log( r b ),     (3) Y 2 = log( r f / r b ),     (4) Y 3 = log( r f )/log( r b ),     (5) (b) Green channel Y 4 = log( g b ),     (6) Y 5 = log( g f / g b ),     (7) Y 6 = log( g f )/log( g b ),     (8) (c) Both channels For each category, there are three types of background measures. The first one is log-transformed background intensities. The second one is the log-transformed ratio of foreground and background intensities. The third one is the ratio of the log-transformed values of foreground and background intensities. Here we used log base 2, but any logarithmic base can be used as desired. Figure 2 shows the relationship between signal intensities and these background measures. At the first stage, we adjust for the effect of background intensities by fitting the usual nonlinear LOWESS curve. For simplicity, let Y k ( k = 1,...,9) be an appropriate background intensity defined by red channel (a), green channel (b), or two channels (c). Then, fit the nonlinear LOWESS curve as follows: M k = C (1) ( Y k ),     (12) M k (1) = M k - C (1) ( Y k ), for k = 1 ,..., 9 (13) where C (1) (·) represents the LOWESS curve and M k (1) is the residual from the curve. Note that M k (1) is the log-ratio of relative intensities after removing the effect of background intensities. For these ratios, we can perform the usual MA normalization at the second stage. Stage 2. MA normalization In the second stage, we perform the normalization process as follows: M k (1) = C (2) ( A ),     (14) M k (2) = M k (1) - C (2) ( A ),     (15) where C (2) (·) is the LOWESS curve and M k (2) is the residuals from the curve in the second stage. Note that at the second stage any normalization method can be applied including a simple global normalization method. Results Results of NCI 60 data We first apply the proposed two-stage normalization method to a microarray data set of the NCI 60 cancer cell lines. These cell lines derived from human tumors have been widely used for investigations on various drugs and molecular targets . The National Cancer Institute's Developmental Therapeutic Programs has been studying a large number of anti- cancer drug compounds and molecular targets on the 60 cancer cell lines (Weinstein et al. [ 19 ]). In particular, the NCI 60 microarray data have been frequently reanalyzed as an experimental model due to the inaccessibility to human tumor tissues for various studies on cancer. Using HCT116, one of the colon cell lines in the NCI 60 panel, Zhuo et al. [ 16 ] performed gene expression profile of dose- and time-dependent effects by the topoisomerase inhibitor I camptothecin compound (CPT). We here use a subset of the array data set consisting of ten slides. These slides were randomly selected to demonstrate the proposed method. Each slide contains 2,208 spotted clone sequences. We also apply global median normalization and intensity dependent nonlinear LOWESS normalization to this data set. From ten slides we choose one slide to illustrate the proposed method. Figure 2 shows the plots of M versus Y k , where M = log ( R/G ) and Y k , k = 1,...,9 are background measures described in Methods section. The correlation coefficients between M and Y k ' s ( k = 1,...,9) are 0.2025, 0.6238, 0.6256, -0.0184, 0.5065, 0.5096, 0.1291, 0.5707, and 0.5729, respectively. Background measures Y 2 and Y 3 tend to have higher correlations than others. Figure 3 shows the results of the first stage normalization. The plots of M (1) versus Y k , where M (1) is the residual in equation (13) demonstrate some reduction in variability, which can be seen more clearly by comparing Figure 2 with Figure 3 . Note that each correlation coefficient between M (1) and Y k have values close to zero. Using M (1) , we carry out the second stage normalization. Figure 4 shows MA plots. The first row shows the MA plots for original (before normalization), after global normalization, and after nonlinear LOWESS normalization, respectively (from left to right), The second row shows M (2) versus A plots after two-stage normalization for Y 1 , Y 2 , and Y 3 , respectively. The third row shows MA plots for Y 4 , Y 5 , and Y 6 respectively, and the bottom row shows MA plots for Y 7 , Y 8 , and Y 9 , respectively. We can see that the two-stage normalization methods using Y 2 and Y 3 have the effect of reducing the variability among Ms and perform better than global and non-linear LOWESS normalization methods. Comparative studies The goal of this study is to compare performances of normalization methods. We compare two-stage normalization to global median normalization and intensity dependent nonlinear LOWESS normalization. Following the idea of Park et al. [ 20 ], we use the variability among the replicated slides as comparison measures, which can be estimated by the mean square error (MSE). For each gene, we can calculate MSE l ( l = 1,..., number of gene) which is the variance estimator for each gene derived from replicated slides. The main idea is that the better the normalization method, the smaller the variation among the replicated observations. Here, we use three different sets of microarray data: colorectal cancer data of NCI 60 (Zhou, et al. [ 16 ]), cortical stem cells data (Park, et al. [ 17 ]), and mouse gene expression data (Pritchard et al. [ 18 ]). The goal of cortical stem cells study is to identify genes that are associated with neuronal differentiation of cortical stem cells. In this experiment, there are 3,840 genes in each slide from two experimental groups for comparison measured at six different time points (12 hrs, 1 day, 2 days, 3 days, 4 days, 5 days). All experiments were replicated three times, thus we have 36 slides for the analysis. The objective of mouse gene expression study of Pritchard et al. [ 18 ] is to assess natural differences in murine gene expression. A 5406-clone spotted cDNA microarray was used to measure transcript levels in the kidney, liver, and testis from each of 6 normal male C57BL6 mice. Experiments were replicated four times per each mouse organ, two red fluorescent dye sample and two green dye samples. Since there are three organs, we have three sets of microarrays. In each organ, there are 24 slides available for the analysis. In this comparative study, all five microarray data sets were used: colorectal cancer data set from NCI 60, cortical stem cells data set from Park et al. [ 17 ], and three organ data sets from Pritchard et al. [ 18 ]. Since results are similar among three organs, we only present the results of kidney. For simplicity, denote CCD for colorectal cancer data, SCD for stem cells data, and KD for kidney data, respectively. In this study, the performances of two-stage normalization using nine background measures are compared to global normalization and intensity dependent nonlinear LOWESS normalization. Figure 5 shows dot plots of log-transformed variance estimates for (a) CCD, (b) SCD, and (c) KD. Here each dot represents the mean value of the log-transformed MSEs for all genes. For all three different data sets, the global normalization reduces variability of the original data but the nonlinear LOWESS reduces variability much more. In general, these dot plots show that the two-stage normalization method using background intensities and the nonlinear normalization method have much smaller variabilities than those of global normalization. However, if we compare the two-stage normalization methods with the nonlinear normalization, the results differ depending on the background measures. That is, the background measures Y 2 and Y 3 in two-stage normalization methods always yield better performances than the nonlinear normalization method, while the other background measures yield comparable results to those of the nonlinear normalization. Thus, we suggest either Y 2 or Y 3 as background measures in the two-stage normalization. Conclusions In microarray studies, many undesirable systematic variations are commonly observed. Normalization becomes a standard process for removing some of the variation that affects the measured gene expression levels. One major source of variation is the background intensities. Recently, some methods have been employed for adjusting for the background intensities. However, all these methods focus on defining signal intensities appropriately from foreground and background intensities during the image analysis (Kooperberg et al. [ 14 ], Edwards [ 15 ]). Although a number of normalization methods have been proposed, no systematic methods have been proposed using the background intensities in the normalization process. In this paper, we propose two-stage normalization for adjusting for the effect of background intensities systematically. The motivating idea is that the noise caused by background intensities may increase the variability in signal intensities. Although we use the log-transformed ratio of two channels denoted by M in most subsequent analysis, the noise caused by background intensities may still remain in M even after normalization. The two-stage normalization may be quite effective especially when there is a high correlation between M and background measures such as Y 2 and Y 3 . Among nine background measures, we recommend two background measures Y 2 and Y 3 based on the results of the comparative studies. For these two background measures, we show that the two-stage normalization method always performs better than the global normalization methods and the nonlinear LOWESS normalization method. We wondered if the relative good performance of two-stage normalization using Y 2 and Y 3 is due to low intensities. We investigated this problem for NCI 60 data after removing spots with low intensities. The spots whose ratio of foreground and background intensities were smaller than 1.5 were removed in the analysis. This new data set also provided quite similar results. The main reasons why background measures Y 2 and Y 3 perform better than other backgrounds are as follows. The background fluorescence might be relatively strong in the Cy5 channel due to interaction between the slide substrate and the hybridization materials. This effect is weaker in the Cy3 channel. It might be also possible that the background fluorescence in the Cy5 channel inflates the values of r b without correspondingly inflating the values of r f . This means that for weakly-responding spots, the r f and r b values are similar. This produces very low values in M , Y 2 and Y 3 for these spots. Note values under 5 for log(R) but not for log (G) in Figure 1 . Also note downward outliers but not upward outliers for M in all panes in Figure 2 . These artefacts in M are partially corrected by regressing against Y 2 or Y 3 . However, the effectiveness of the proposed method for other data may depend on the background fluorescence artefacts. The comparison is based on the variability measures derived from the replicated microarray samples. These variability measures can be easily derived from any replicated microarray experiment. Although we have studied only a limited number of data sets, our findings are quite consistent. For these data sets, we also conduct a similar analysis to see the effect of print-tips in normalization process. The results were consistent to those in Figure 5 and not reported here. Background measures Y 2 and Y 3 yielded better results than other background measures. One major concern for normalization is about over-fitting which may cause overcorrection of real biological significance. In fact, every normalization method has a possibility of overcorrecting the data and removing some existing biological significance in the data. For the NCI60 data, for example, it is not easy to find out whether such negatively low expression genes are biologically significant genes or not. We investigate the possibility of overcorrecting by drawing the density plot of M : the original one (O), after the global normalization (G), after LOWESS normalization (L), and two-stage normalization using Y2 and Y3 (T2 and T3, respectively). The density plot is in Figure 6 . The distributions of M for T2 and T3 are quite similar to that of L. This simple empirical investigation shows that the proposed method does not always cause a bigger overcorrection than the LOWESS normalization. However, we must be careful for over-fitting which may overcorrects the biological findings of interest. In addition, we performed the hypothesis test for M equal to zero and counted the number of genes that have different expression level between two channels. The number of significant gene with different expression value is 909 for the original data before normalization, 326 for the data after global normalization, 302 for the data set after LOWESS normalization, and 297 for the data after two-stage normalization. The numbers of significant genes do not differ much among normalization methods. The proposed method can be applicable when both background and foreground intensities are available after the image analysis. Many image software provides both signal intensity as well as background intensity of spots. In most cases, the signal intensity is defined by simple subtraction such as ( r f - r b ) for the red channel. Our method was illustrated using the usual subtraction ( r f - r b ). Although our method starts with a most common approach based on the usual subtraction, it can be applied to any other models for defining spot intensity r = f( r f , r b ), where r is the signal intensity, as long as a background intensity is available. Our method can be applicable as long as the image analysis provides the signal intensity and background intensity. For example, our method using Y 2 builds on the relationship between r = f( r f , r b ) and r b . We recommend that experimentalists examine their data carefully and consider applying the two-stage normalization methods using Y 2 and Y 3 . The performance of the two-stage normalization method tends to depend on the correlations between background measures and M . That is, if there is a strong relationship between them, our method has a large effect. Thus, for the experimentalists it might be important to tell when to use the two-stage normalization method. In order to answer this question, we compute correlations between M and all background measures from Y 1 to Y 9 . Figure 7 shows the boxplots of correlation coefficients between M and Y k s for each data set. For example, Figure 7(a) shows the distribution of correlation coefficients from the NCI 60 data set. The correlation coefficients were computed for all ten slides. As expected, the correlations of Y 2 and Y 3 are relatively higher than those from other background measures. We also think that is why Y 2 and Y 3 have better performances than the other background measures. Figure 7 also shows that the median values of correlations of Y 2 and Y 3 are higher than 0.5 for both CCD and SCD, while those for KD are smaller than 0.5. Note that CCD and SCD have more reduction in variability than KD in the two-stage normalization. For lower correlations of Y k s do not reduce variability much compared to the usual LOWESS normalization.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509428.xml

526385

Road-traffic pollution and asthma – using modelled exposure assessment for routine public health surveillance

Asthma is a common disease and appears to be increasing in prevalence. There is evidence linking air pollution, including that from road-traffic, with asthma. Road traffic is also on the increase. Routine surveillance of the impact of road-traffic pollution on asthma, and other diseases, would be useful in informing local and national government policy in terms of managing the environmental health risk. Several methods for exposure assessment have been used in studies examining the association between asthma and road traffic pollution. These include comparing asthma prevalence in areas designated as high and low pollution areas, using distance from main roads as a proxy for exposure to road traffic pollution, using traffic counts to estimate exposure, using vehicular miles travelled and using modelling techniques. Although there are limitations to all these methods, the modelling approach has the advantage of incorporating several variables and may be used for prospective health impact assessment. The modelling approach is already in routine use in the United Kingdom in support of the government's strategy for air quality management. Combining information from such models with routinely collected health data would form the basis of a routine public health surveillance system. Such a system would facilitate prospective health impact assessment, enabling policy decisions concerned with road-traffic to be made with knowledge of the potential implications. It would also allow systematic monitoring of the health impacts when the policy decisions and plans have been implemented.

Introduction The prevalence of asthma is increasing and there is concern that the increase may in part be attributable to increasing road traffic related pollution. The concerns relate especially to childhood asthma. In this article, we set out the arguments for using modelled exposure assessment to create a surveillance system that will facilitate routine public health work, such as monitoring and health impact assessment. We first discuss the increasing prevalence of asthma and the effects of air pollution on asthma. We then set out the benefits of monitoring the link between outdoor air pollution related to road traffic and asthma and discuss methods of exposure assessment in some detail. We describe modelling air quality in the UK as an example and then address the implementation of a surveillance system. Increasing prevalence of asthma Asthma is a common disease. Figures from the International Study of Asthma and Allergies in Childhood suggest up to 25.9% of children in Oceania have ever had asthma [ 1 ]. Although less evidence is available with regard to the prevalence amongst adults, the European Community Respiratory Health Study which studied asthma prevalence throughout Europe, Australasia and the United States suggested asthma affects up to 11.9% of adults in Australia and 8.4% of adults in the UK [ 2 ]. Over recent years the prevalence of asthma appears to have been steadily increasing [ 3 ]. It has also been suggested by some that the severity of asthma is on the increase [ 4 ], although other studies do not confirm this suggestion [ 5 ]. Initially thought to be a disease of the western world, in recent years the incidence of asthma has also been shown to be increasing dramatically in less developed countries [ 6 ]. Much research has taken place to find the cause of such an increase but the reasons are not fully understood. Socio-economic status [ 7 ], ethnicity [ 8 ], allergen exposure [ 9 ], smoking [ 10 ], nutrition [ 11 ] and infection exposure [ 12 ] have all been considered as possible factors. Links have also been made to living in an urban as opposed to a rural area [ 13 ], raising speculation as to a possible effect of air pollution on the prevalence of the disease. Effects of air pollution on asthma Air pollution has been linked to morbidity and mortality of several diseases, including diverse conditions such as coronary heart disease [ 14 ] and Hodgkin's disease [ 15 ]. In terms of the effects on respiratory disease, exposure to air pollution has been linked to the aggravation of chronic respiratory symptoms and increased mortality from chronic obstructive pulmonary disease [ 16 ]. Over the 20 th century, air pollution increased greatly alongside the noted increase in the prevalence of asthma. Yet the last decades of the century have seen a considerable reduction in pollutants such as sulphur dioxide (SO 2 ) following a cut back on industrial emissions [ 17 ]. Levels of pollutants such as nitrogen dioxide (NO 2 ) however, still remain problematic due to the increasing number of vehicles on our roads. In fact almost 50% of NO 2 is thought to be produced by vehicles and much particulate matter is produced by diesel exhaust fumes [ 17 ]. Experimental studies have shown NO 2 exposure increases cell membrane permeability, decreases ciliary beat frequency [ 18 ] and increases the response of asthmatics to inhaled allergens [ 19 ], whilst exposure to diesel exhaust particles in mice has been shown to alter IgE antibody production [ 20 ]. Epidemiological investigations into the effects of these pollutants have suggested an association between pollutant levels and the exacerbation of asthma. Results from a study in Paris showed an increase of 100 μg/m 3 of NO 2 to be associated with a relative risk of 1.175 for asthma admission [ 21 ]. Emissions of nitrogen oxides (NO x ) have also been shown to influence emergency room visits in Israel [ 22 ]. Both particulate matter less than 10 μm in diameter (PM 10 ) and NO 2 appeared to consistently increase attendances at accident and emergency units with asthma in London [ 23 ]. With traffic emissions accounting for such a high proportion of these pollutants and with the volume of traffic on the increase, such a link between traffic-related pollution and asthma would be of importance. It has been suggested that pollutant exposure may induce asthma, aggravate asthma or increase the permeability of the airways to other allergens to which asthmatics are susceptible [ 17 ]. Any of these effects could potentially cause a significant increase in asthma morbidity if the level of traffic continues to rise. A number of studies have considered the association between asthma and road-traffic pollution specifically. Ciccone et al. [ 24 ] showed the odds ratios for asthma and a number of asthmatic symptoms to be increased in those exposed to heavy lorry traffic. Heavy traffic flow has also been shown to increase childhood asthma admissions [ 25 ]. Studies carried out within the UK and the United States have suggested those living within close proximity to a road are also at an increased risk of hospitalisation with asthma [ 25 , 26 ]. Monitoring the link between air pollution and asthma The evidence suggests that there is a link between air pollution and asthma but it is not conclusive. The increasing prevalence of asthma and the continuing increase in road traffic are both of concern. Monitoring the association between asthma and road traffic pollution would be useful for public health purposes, both in terms of surveillance and in terms of influencing policy. Policy implications might include routing of traffic, construction of bypasses, congestion reduction schemes, utilisation of non-fossil fuel cars and possibly even the location of schools. A monitoring system would use estimates of air pollution from road traffic which would be linked to data on asthma obtained from routine systems, such as hospital admissions, attendance at accident and emergency departments or primary care consultations or to data from periodic surveys on health, including asthma prevalence. Methods of exposure assessment A number of methods for exposure assessment have been used in studies examining the association between asthma and road traffic pollution and these are described below. These include comparing asthma prevalence in areas designated as high and low pollution areas, using distance from main roads as a proxy for exposure to road traffic pollution, using traffic counts to estimate exposure, using vehicular miles travelled and using modelling approaches which can take into account a number of variables. (i) High vs low pollution areas One method frequently used is estimating exposure levels of individuals based on assessing whether a residence is in a high or low pollution area. Many have estimated exposure status of individuals on the basis of whether or not they live on a street with heavy traffic, for example in the study conducted by Jedrychowski and Flak [ 27 ]. Another approach was used by Nicolai and v. Mutius [ 28 ] in their study of former East and West Germany. In this study, West Germany was classified as an area with high traffic-related NO 2 emissions and low SO 2 emissions from industry, whilst the opposite was said to be true of East Germany. The problem, particularly with this method, is that although the countrywide generalisation may hold true, it may not be true for each individual. Even when the measure is made for each individual, it should be remembered it may be subject to bias. It has been postulated that asthmatic individuals and their families may be more aware of the speculation over such a link between road-traffic and asthma and therefore may be more likely to report or consider heavy traffic to be associated with their symptoms [ 24 ]. (ii) Distance to roads Distance from roads has commonly been used as a proxy for road traffic exposure in a number of studies. Postcodes are georeferenced and may be used in a geographical information system (GIS) to calculate the distance from an individual's residence to a road, most often a main road, carrying over a certain volume of vehicles. In certain cases the distance from a child's school to a main road has been used instead. Examples of the use of this method are studies by Livingstone et al. [ 29 ] and Wilkinson et al. [ 30 ]. A number of authors have only considered individuals living within 1000 m of a main road as they felt traffic would be unlikely to influence pollution levels beyond this distance [ 31 ]. Indeed, with analyses using this method, effects of pollutants have often only shown an effect within a short distance from main roads. By using this method an assumption is made that all individuals living within a certain distance of a road are subjected to the same level of exposure, yet this is unlikely to be the case. Traffic on different roads varies, both in volume and in type, and meteorological conditions can alter dispersion of pollutants. It is well known that cold weather conditions trap air close to the ground, prolonging the duration of the time pollutants remain close to where they were produced [ 17 ]. (iii) Traffic counts Another popular method is considering traffic flow along the street of residence or one in close proximity, as used by English et al. [ 32 ]. In a similar way Wjst et al. [ 33 ] have investigated traffic flow around a child's school. The traffic count method has the advantage of being likely to be a more valid measure than distance to roads. It is worth however considering the daily movements of an individual. Throughout the day an individual travels between home and work or school experiencing a number of different exposure levels on the way. Recreational activities may also subject a person to different levels of exposure. Indeed even within the residential area, exposure may vary dependent on the time one spends indoors or outdoors. Indoor exposure to NO 2 may be high, with levels possibly higher than outdoors if a gas stove is used in the home [ 34 ]. Another point to consider is the type of traffic exposure. Emissions vary greatly between cars and trucks. Some have approached this by analysing data from different vehicles separately, suggesting truck pollution to be more detrimental to health than that from cars [ 31 ]. It could be suggested that as car and truck pollution varies, for example, trucks produce a lot more particulate matter consisting of diesel particles than cars, that perhaps one should consider the effects of particulates separately from those of NO 2 . This however, is not without difficulties, as if one is exposed to traffic there will be a combined effect from a cocktail of pollutants produced by both trucks and cars. (iv) Vehicle miles travelled Some authors (e.g. Lin et al. [ 26 ]) have attempted to combine both length of road and traffic counts as a measure referred to as vehicle miles travelled. It involves multiplying the length of a road in a specified area around the home by the traffic volume travelling along that section of road. Authors have varied in the selection of roads used in such analyses. This method may have an advantage over measuring traffic flow alone, as exposures may be more accurate. There is still however, no account taken of individuals moving between areas through the day or different topographical conditions. Some feel that buildings in the vicinity of one's home should be considered when looking at exposure to traffic pollution [ 35 ], due to their influence on pollution dispersion, as well as the presence of bus stops and distance to street crossings which may influence exposure [ 27 ]. (v) Modelling approach A number of studies have used modelling to estimate pollution exposure [ 35 , 36 ]. A model is capable of taking into account a whole range of factors that may affect exposure. As illustrated by Pershagen et al. [ 36 ] exposures both at home and at day-care centres or for others at school or work can be considered, with these being adjusted for the time spent in each location. Factors considered in the models used in the studies above have included vehicle type and density, presence and type of buildings on a street, meteorological conditions, street width and distance from house to middle of the street amongst other factors. Even within a model however, accounting for personal day-to-day exposures is still problematic. In order to take previous exposures into account a cohort study would be necessary [ 37 ]. Certainly if one is trying to account for the prevalence of a disease like asthma, knowing previous exposure levels prior to the onset of the disease is important. To do this one would need to look at the previous residences and day-to-day exposures of that person throughout their life. An alternative would be to use a personal monitoring system. Both these methods of assessing long-term exposure, however, would be very expensive. One could consider the use of monitoring stations already in place throughout cities. The problem with using such stations is that they are generally widely dispersed while pollution levels may vary substantially within short distances, e.g. exponential decline in the concentration of certain pollutants with increasing distance from busy roads [ 38 ]. Installing sufficient monitoring stations to adequately capture spatial variation in levels of pollution encountered over short distances, would be both impractical and expensive. Despite limitations, a model would appear to be the most practical way of assessing traffic-related exposure where routine surveillance is concerned. Information such as vehicle density, type of vehicle, risk of traffic congestion, presence of bus stops and street crossings, distance of residences to roads, street width, type of street, building presence and type and meteorological conditions (e.g. wind speed and direction, absolute temperature and temperature differences, global and gamma radiation) could be collected routinely for use in a variety of models for predicting exposure to NO 2 and PM 10 . The model could be used to estimate exposures on all the streets within a certain radius of the home or place of work as dispersion of pollutants from these streets may also be affecting the individual. In a sophisticated model it may be possible to make adjustment for the height of an individual's residency or place of work in high rise buildings to account for the vertical dispersion of pollutants. Such a system could also be used to estimate exposures at previous residences, work places or schools of an individual so that an assessment of lifelong exposure could be made as accurately and practically as possible. However, the latter might be too complicated for a routine monitoring system. Modelling air quality in the UK The UK Government's current policy on air quality within the UK is set out in the Air Quality Strategy for England, Scotland, Wales and Northern Ireland published in January 2000 pursuant to the requirements of Part IV of the Environment Act 1995. The Strategy sets out a framework for improving air quality and for ensuring that international commitments are met. It is designed to be an evolving process that is monitored and regularly reviewed. The Strategy sets standards and objectives for ten pollutants that have an adverse effect on human health, vegetation or ecosystems and target dates for achieving them. The standards generally set concentration limits above which sensitive members of the public (e.g. children, older people, people who are unwell) might experience adverse health effects. In early 2003 an Addendum to the Strategy was published introducing standards and objectives for a new pollutant and revising those for three others. The pollutants currently specified in the Strategy now include benzene, 1,3 butadiene, carbon monoxide, lead, NO 2 , PM 10 , SO 2 , ozone (O 3 ), NO x and polycyclic aromatic hydrocarbons. The predominant source of most of these pollutants is road traffic, but industrial and domestic sources are also contributors. The air quality standards or guideline limits are long-term benchmarks for ambient pollutant concentrations which represent negligible or zero risk to human health, based on medical and scientific evidence reviewed by the Expert Panel on Air Quality Standards (EPAQS) and the World Health Organization (WHO). For some pollutants, (e.g. NO 2 ), there is both an annual mean guideline limit and a short-term mean guideline limit. These reflect the varying impacts on human health of exposure to some pollutants over differing time periods, (e.g. temporary exposure on the pavement adjacent to a busy road compared with the exposure of residential properties adjacent to a road). The air quality objectives are medium-term policy-based targets set by the Government which take into account economic efficiency, practicability, technical feasibility and timescale. Some objectives are equal to the EPAQS or WHO recommended air quality standards and guideline limits, whereas others involve a margin of tolerance, i.e. a limited number of permitted exceedances of the standard over a given period. The Government has issued guidance to local authorities on how to conduct Reviews and Assessments required under the system of Local Air Quality Management (LAQM). The latest available guidance is Policy Guidance LAQM.PG(03) and Technical Guidance LAQM.TG(03). Air quality modelling is key to assessing the future potential for attainment, or not, of the objectives. Part IV of the Environment Act 1995 requires a local authority to designate an Air Quality Management Area (AQMA) covering any part of its administrative area where air quality objectives are not likely to be achieved by, or at any point beyond, the relevant objective's target date at locations where the general public might reasonably be exposed. These AQMAs have been determined by modelling future scenarios. For each AQMA the local authority has a duty to draw up an Air Quality Action Plan (AQAP) setting out the measures the authority intends to introduce to deliver improvements in local air quality in pursuit of the air quality objectives. Local authorities are not statutorily obliged to meet the objectives, but they must show that they are working towards them. As of June 2004, there were 120 designated AQMAs in the UK, with 80 AQAPs produced outlining how air quality would be tackled in these areas. Implementation of a surveillance system A routine surveillance system would be one which links modelled air quality data, such as those derived in the UK described above, with routinely collected health data. There are a number of issues regarding the technical aspects of linking spatial information on air quality with health information, typically carried out using GIS. These are discussed in detail elsewhere [ 39 ]. The information on air quality could be used in a statistical model and analysed alongside hospital admission, accident and emergency attendance and prevalence data for asthma, or any other conditions to which a link with air pollution has either been made or considered. The majority of studies examining the link between asthma and road-traffic pollution have concentrated on the prevalence of asthma, for example studies by Ciccone et al. [ 24 ], Jedrychowski and Flak [ 27 ], Livingstone et al. [ 29 ] and Nicolai and v. Mutius [ 28 ]. Prevalence data is typically gathered from health surveys, as information regarding disease prevalence is not generally available through routine recording. It would however be possible through the use of periodic health surveys to collect the relevant information required to analyse disease prevalence along with modelled pollution exposure data. The ISAAC investigators for example have designed a standardised questionnaire now being used throughout the world which is concerned with the prevalence of asthma amongst children. It considers those children suffering from asthma to be those who answer, or whose parents answer, that they have suffered from wheezing or whistling from the chest in the last twelve months [ 40 ]. Using a standardised questionnaire such as the ISSAC questionnaire for children would allow comparable information to be collected and compared both within and between countries. A few of the studies examining the link between asthma and road-traffic pollution have used information that is readily available. For example studies by English et al. [ 32 ] and Lin et al. [ 26 ] have looked at children being hospitalised with exacerbations of asthma. Such information is routinely recorded in hospitals, but tends to reflect disease severity. Relating such information to levels of air pollution is still important in determining the effects of pollution on asthma. A routine surveillance system recording spatial variation in pollutant levels would allow improved understanding of the link between road-traffic pollution and asthma, or indeed other diseases and could be used to help predict future health impact, particularly in cities and towns. The results of such assessment would allow local policy decisions concerning the routing of traffic around residential areas or schools and plans to reduce congestion to be made with knowledge of the implications of the decision on the health of the local population. It would also allow systematic monitoring of the health impacts when the policy decisions and plans have been implemented. We should point out here that to examine acute effects (i.e. on events such as admissions, general practice consultations, emergency room attendances) using daily time series analyses, monitoring data would be required as modelling would probably be too insensitive to detect daily variation in outdoor air pollution levels. However, time series analyses are complicated research methods that are not within the realm of routine public health practice. What we have argued for here is a surveillance system that looks at spatial variation in pollution and asthma which would highlight problem areas and could monitor the effects of interventions to reduce pollution in these areas. Conclusions We believe that a routine surveillance system which links modelled outdoor air pollution data to health data would provide a useful tool for facilitating routine environmental public health work. Such a system would be especially useful for monitoring the health effects of traffic related pollution and for aiding health impact assessment. Implementation of the system will require close collaboration between public health and environmental health departments, protocols for sharing data and investment in training to develop the necessary technical expertise to set up and maintain the surveillance system. Of particular importance will be the ability of high level management to interpret surveillance information within a wider policy context. Authors' contributions RM proposed the idea for the article. ECF wrote the first draft, supervised by RM. MD contributed the section on modelling air quality in the UK. RM edited subsequent drafts.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526385.xml

539299

Value of supplemental interventions to enhance the effectiveness of physical exercise during respiratory rehabilitation in COPD patients. A Systematic Review

Background There is a controversy about the additional benefit of various supplemental interventions used in clinical practice to further enhance the effectiveness of respiratory rehabilitation in patients with Chronic obstructive pulmonary disease (COPD). The aim of this research was to assess randomised controlled trials (RCTs) testing the additional benefit of supplemental interventions during respiratory rehabilitation in COPD patients. Methods Systematic review with literature searches in six electronic databases, extensive hand-searching and contacting of authors. Two reviewers selected independently eligible RCTs, rated the methodological quality and extracted the data, which were analyzed considering the minimal important difference of patient-important outcomes where possible. Findings We identified 20 RCTs whereof 18 provided sufficient data for analysis. The methodological quality was low and sample sizes were too small for most trials to produce meaningful results (median total sample size = 28). Data from five trials showed that supplemental oxygen during exercise did not have clinically meaningful effects on health-related quality of life while improvements of exercise capacity may be even larger for patients exercising on room air. RCTs of adding assisted ventilation, nutritional supplements or a number of anabolically acting drugs do not provide sufficient evidence for or against the use any of these supplemental interventions. Interpretation There is insufficient evidence for most supplemental interventions during respiratory rehabilitation to estimate their additional value, partly due to methodological shortcomings of included RCTs. Current data do not suggest benefit from supplemental oxygen during exercise, although the methodological quality of included trials limits conclusions. To appropriately assess any of the various supplemental interventions used in clinical practice, pragmatic trials on respiratory rehabilitation of COPD patients need to consider methodological aspects as well as appropriate sample sizes.

Introduction Chronic obstructive pulmonary disease (COPD) has a large impact on health-related quality of life (HRQL) and represents a major health burden in industrialized and developing countries [ 1 - 4 ]. A systematic review including 23 randomized controlled trials (RCTs) has shown that patients with COPD improve their HRQL and exercise capacity during respiratory rehabilitation[ 5 ]. Recent data on long-term outcomes after respiratory rehabilitation show reductions of exacerbations and hospitalizations [ 6 - 8 ]. Physical exercise is the central component of respiratory rehabilitation programs because it reverses peripheral muscle dysfunction[ 9 ], a highly prevalent comorbidity of COPD associated with increased risk of exacerbations and mortality[ 10 , 11 ]. While respiratory rehabilitation including physical exercise has become a cornerstone of COPD management [ 12 - 14 ], there is controversy about the additional value of several supplemental interventions to support exercise programs such as oxygen during exercise[ 15 ] or anabolically acting hormones[ 16 ]. Clinicians, who consider these supplemental interventions during respiratory rehabilitation programs, should know their benefits and downsides. They need evidence from RCTs directly comparing respiratory rehabilitation with or without supplements in order to carefully discuss these benefits and downsides with their patients. Therefore, we conducted a systematic review of pragmatic RCTs comparing the effects respiratory rehabilitation with and without any supplemental intervention to assess their added value in HRQL and exercise capacity improvement. Methods Identification of studies We performed electronic database searches in MEDLINE (Ovid version, New York, New York, from inception to May 2004), EMBASE (DataStar version, Cary, North Carolina from inception to December 2003), PEDRO (online version, University of Sydney, Australia, December 2003) and the Cochrane Central Register of Controlled Trials (Oxford, United Kingdom, 2003, Issue 4). We also searched the Science Citation Index database (Web of Science, Thomson ISI, Philadelphia, Pennsylvania) and the "related articles" function of PubMed (National Library of Medicine, Washington, Maryland) by entering all included studies. In addition, we hand searched the bibliographies of all included studies, of reviews on respiratory rehabilitation or physical exercise in patients with COPD that were identified in the literature search, as well as the Proceedings of the International Conferences of the American Thoracic Society and the congress of the European Respiratory Society to identify further relevant studies. We also contacted authors in the field to ask for published or unpublished data. Selection criteria We included RCTs investigating any supplemental intervention added to respiratory rehabilitation that included a standardized physical exercise program. We focused on standardized exercise programs because only these allow reproduction in clinical practice. A standardized exercise protocol was defined as the use of an identical exercise activity for all patients (e.g. treadmill walking or cycle ergometer training) at measurable exercise intensity (e.g. in Watts, metabolic equivalents or kilograms). We included studies if more than 90% of study participants patients had COPD according to the following criteria: (1) a clinical diagnosis of COPD, (2) irreversible airways obstruction and (3) one of the following: (a) best recorded FEV1/FVC ratio of individual patients < 0.7; (b) best recorded FEV1 of individual patients < 70% of predicted value. We considered the following outcome measures: HRQL as measured by generic (e.g. SF-36) or disease-specific (e.g. St. George Respiratory Questionnaire) questionnaires, symptom scales, functional exercise capacity as measured walk tests and results from cardiopulmonary exercise testing. We did not apply any language restrictions. We excluded studies that compared any exercise program versus usual care (i.e. no exercise) or studies that used unstandardised exercise protocols (e.g. home exercise programs). Data extraction and quality assessment The bibliographic details of all retrieved articles were stored in a Reference Manager file. We removed duplicate records resulting from the various database searches. Two members of the review team independently scrutinized the titles and abstracts of all identified citations (see figure 1 ). We ordered the full text of any article that was deemed potentially eligible by one of the reviewers. The two reviewers then evaluated the full text of all retrieved papers, made a decision on in- or exclusion and discussed the decisions. Any disagreement was resolved by consensus with close attention to the inclusion and exclusion criteria. Final decisions on papers were recorded in the Reference Manager file and bibliographic details as well as the reasons for exclusion. We recorded the initial degree of agreement between the reviewers and corrected discordant scores based on obvious errors. We resolved discordant scores based on real differences in interpretation through consensus. Figure 1 Study flow from identification to final inclusion of studies. Details about study patients, interventions and outcome measures as well as the results were extracted onto a predefined data form. We pilot tested the data forms using five studies with high likelihood for inclusion. Two reviewers independently evaluated the methodological quality of included trials reported in full reports using a detailed list of quality items assessing components of internal validity[ 17 ] (table 3, see Additional file 3 ). We also contacted the authors of the primary studies to obtain missing information. Data synthesis and interpretation We summarized the results of the data extraction and assessment of study validity in structured tables to allow looking at the variation in patient characteristics, interventions, outcome measures, study quality and results. In addition, we used forest plots to compare results across the trials. If appropriate we planned to explore sources of heterogeneity (i.e. differences between studies) using multivariable regression models (study level meta-regression analysis) where clinical and methodological items would act as explanatory variables. No pooling was undertaken in the presence of significant source heterogeneity. Whenever possible, for each outcome, estimates and confidence limits was related to its minimal important difference[ 18 ]. We assessed whether the estimates and 95% confidence limits for the difference between study groups exceeded the minimal important difference (for the Six-minute walk distance ± 50 meters[ 19 ], Chronic Respiratory Questionnaire ± 0.5 points[ 20 ] and St. George Respiratory Questionnaire ± 4 points)[ 21 ]. Data were analyzed using STATA (version 8.2, Stata Corp., College Station, Texas). Results Study selection Figure 1 shows the study selection process and agreement on study inclusion. Main reasons for study exclusion (Appendix, see Additional file 6 ) were that patients did not have an exercise programme but only exercise testing with or without oxygen (n = 12), studies were not RCTs (n = 8) and that the control group had no exercise programme (n = 5). We excluded only one study because of an undefined exercise programme. We excluded two trials[ 22 , 23 ] from the analysis because the abstract provided little information and the authors did not provide further details. Initially, we excluded another abstract, but since this trial was published in the meantime[ 24 ], we could include it in the analysis. Quality assessment Table 3 (see Additional file 3 ) shows a detailed assessment of the methodological quality of the included trials. Interrater agreement for all items of the quality assessment was 87% (chance corrected agreement: κ = 0.76). In general, most included trials scored poorly on the checklist used. Important methodological aspects that bear on the validity such as blinding of outcome assessors were not or just partially addressed in most trials. Supplemental oxygen during exercise The characteristics of the five trials on supplemental oxygen [ 25 - 29 ] are summarized in table 1 (see Additional file 1 ) and the results shown in figures 2 and 3 . There was a trend towards larger improvements of HRQL and exercise duration in constant work rate tests in the groups with oxygen, but patients exercising on room air had larger improvements of the walking distance. Emtner[ 25 ] reported that the use of oxygen enabled patients to exercise at higher intensity (mean 62 Watt [SD 19] corresponding to 138% of baseline maximum exercise capacity) compared with patients on room air (52 Watt [SD 22] corresponding to 96% of baseline maximum exercise capacity, p < 0.01 for difference between groups). In the trial by Rooyackers[ 28 ], patients achieved mean exercise intensities corresponding to 124% of maximum exercise capacity in the group with oxygen and 114% of maximum exercise capacity in the group without oxygen (p = 0.12). Two trials reported on safety of exercise with oxygen or room air. Rooyackers[ 28 ] assessed whether oxygen prevented the development of pulmonary hypertension. The investigators did not find any differences between groups in resting mean pulmonary artery pressure measured by Doppler echocardiography. Waddell[ 29 ] did not find significant CO 2 retention during walking tests despite high oxygen flow of 5 l/minute. Figure 2 Effect of supplemental oxygen on health-related quality of life. The forest plot shows the results from three trials comparing physical exercise with and without oxygen, separately for each domains of the Chronic Respiratory Questionnaire (CRQ). The x-axis represents the difference in change scores between study groups with negative values favoring exercise on room air and positive values favoring exercise with supplemental oxygen. A difference of 0 means that both study groups changed to the same amount. Boxes with 95% confidence intervals represent point estimates for the difference between the CRQ change scores (from baseline to follow-up) of the study groups. Dotted lines represent the minimal important difference of the CRQ (change of 0.5). On the right of the forest plot, point estimates for differences between groups and 95% confidence intervals are shown. Figure 3 Effect of supplemental oxygen on exercise capacity. The forest plot shows the results from five trials comparing respiratory rehabilitation with and without oxygen. Walking tests, incremental and constant work rate exercise tests were used to assess the additional effect of supplemental oxygen during exercise. The x-axis represents the difference in change scores between study groups with negative values favoring exercise on room air and positive values favoring exercise with supplemental oxygen. A difference of 0 means that both study groups changed to the same amount. Boxes with 95% confidence intervals represent point estimates for the difference between the walking distance and maximum exercise capacity change scores (from baseline to follow-up) of the study groups. Dotted lines represent the minimal important difference of the six-minute walking distance (53 meters). On the right of the forest plot, point estimates for differences between groups and 95% confidence intervals are shown. Assisted ventilation Two trials[ 30 , 31 ] evaluated proportional assist ventilation during exercise and did not find an additional benefit (tables 1 and 4, see Additional files 1 and 4 ). Only 50%[ 30 ] and 71.4%[ 31 ] of patients exercising with positive pressure ventilation and 67%[ 30 ] and 60%[ 31 ] exercising without positive pressure ventilation completed these trials. Garrod[ 32 ] assessed the benefit of overnight non-invasive positive pressure ventilation at home during the training period. They found a statistically significant improvement of the walking distance for patients assigned to overnight non-invasive positive pressure ventilation. HRQL improvements also tended to be larger for patients with ventilation, but the difference reached only statistical significance for the fatigue domain and total score of the CRQ. Johnson[ 33 ] evaluated the effect of ventilation and Heliox during exercise on exercise duration and intensity. They found a small, but statistically not significant increase in exercise duration and intensity for patients exercising with ventilation and Heliox. Patient satisfaction for overall condition, exercise capability and breathing ability measured with global ratings of change did not differ significantly between groups (exact data not available). In this trial, 73.3% of patients with ventilation, 90.9% of patients with Heliox and 84.6% of patients without a supplement completed the trial. Nutritional supplements We identified two RCTs that assess the additional benefit of nutritional supplements during respiratory rehabilitation (table 2, see Additional file 2 )[ 34 , 35 ]. Steiner[ 35 ] did not find statistically significant differences for HRQL and exercise capacity (table 5, see Additional file 5 ). In a subgroup of patients with a BMI>19 kg/m 2 (22 in group with supplement and 30 in group with placebo) the difference between groups was 27 meters (95% CI 1–53) in the incremental and 121 seconds (95% CI -44–286) in the endurance shuttle walk test. Patients with the carbohydrate-rich diet increased their body weight compared to the placebo group by 1.23 kg (95% CI 0.42–2.05), which occurred mainly because of an increase of the fat mass (difference between groups 1.46 kg, 95% CI 0.65–2.27). There was a dropout rate of 40% in the group with and of 16% in the group without carbohydrate-rich diet. Another RCT[ 34 ] found not significant differences between patients supplemented with an additional fat-rich diet, but did not report the results in detail and could not provide these data for our review. Compared to placebo, non-depleted patients increased their body weight by 1.5 kg (95% CI 0.4–2.6) when receiving a fat-rich diet and by 1.6 kg (95% CI 0.39–2.81) when receiving a fat-rich diet plus anabolic steroids. Anabolic steroids Creutzberg[ 36 ] (table 2, see Additional file 2 ) found that only patients receiving nandrolone improved their HRQL, whereas patients following the respiratory rehabilitation program without nandrolone did not change. This trend was consistent for all domains of the St. George Respiratory Questionnaire, but only statistically significant for the symptom domain (table 5, see Additional file 5 ). For the subgroup of patients receiving maintenance treatment with oral glucocorticosteroids, patients with nandrolone improved their maximum exercise capacity significantly more. Isometric leg strength and isokinetic legwork improved in both groups, but did not differ significantly between groups. There was a trend in erythropoetic parameters towards an increase of erythrocyte count, hematocrit and hemoglobin in patients treated with nandrolone compared to those treated with placebo. No changes in blood pressures and any androgenic effects or fluid retention were registered in either group. Casaburi[ 24 ] assessed the additional benefit of testosterone for male COPD patients with low testosterone levels who followed a strength exercise program (table 2, see Additional file 2 ). The group with testosterone had larger increases in exercise capacity and muscle strength, but none of the differences reached statistical significance (table 5, see Additional file 5 ). Total lean mass increased and total fat mass decreased more in patients with supplemental testosterone, but differences between groups were not significantly different (mean difference in changes between groups in lean mass 3.09 kg, p > 0.05, and total fat mass -1.28 kg, p > 0.05). Casaburi found, like Creutzberg[ 36 ], differences in hemoglobin changes between groups (mean difference in hemoglobin change between the testosterone and placebo group 1.4 g/dL, p < 0.05). They observed neither adverse events nor any differences in most safety measures between groups (prostate specific antigen, liver enzymes, alkaline phosphates, cholesterol and high-density lipoprotein cholesterol) between groups. Serum creatinine levels, however, increased in the testosterone group by 0.12 mg/dL and decreased in the placebo group by 0.05 mg/dL (difference between groups 0.17 mg/dL, p < 0.05). Tiotropium, Creatine, Coenzyme Q10, and growth hormone Casaburi[ 37 ] assessed the additional benefit of tiotropium in 47 patients and found a significantly increased exercise endurance time compared to patients who received placebo (n = 44, tables 2 and 5, see Additional files 2 and 5 ). Further results were not available. Four small RCTs evaluated the additional benefit of creatine[ 38 ], coenzyme Q10[ 39 ] and growth hormone[ 40 , 41 ] during respiratory rehabilitation, but did not find any additional benefit on respiratory or peripheral muscle function or HRQL (tables 2 and 5, see Additional files 2 and 5 ). Casaburi et al[ 41 ] reported that no adverse effects of growth hormone occurred. Discussion There are three main results from this systematic review. First, evidence suggests that supplemental oxygen during physical exercise does not provide a clinically relevant benefit. Second, the evidence for any other supplemental intervention is not strong enough to recommend or discourage their use in clinical practice and third, there were major methodological limitations in most trials that may explain some of the inconclusive findings. We discuss each of these results in turn. Cotes[ 42 ] reported in 1956 that oxygen increased exercise performance in patients with COPD. Since then, many investigators assessed the short-term effect of increased oxygen availability during exercise[ 15 ]. Some investigators argue that patients tolerate higher exercise intensities or longer exercise time with supplemental oxygen leading to larger training effects[ 43 , 44 ]. Others believe that only without oxygen, an adequate hypoxemic stimulus is provided for peripheral muscles to improve exercise capacity. The studies by Emtner[ 25 ] and Rooyackers[ 28 ] demonstrated that patients indeed tolerate higher exercise intensities if supplemented by oxygen. Mean differences on the CRQ domain scores, however, showed a slight but clinically not meaningful trend towards a benefit with oxygen supplementation (figure 2 ). The trial by Emtner[ 25 ] was the only one that showed a consistent trend towards a small benefit of oxygen on HRQL and exercise capacity. Across all studies, however, results from exercise testing were contradicting. Supplemental oxygen did prolong exercise duration in constant work rate tests, but led to considerably smaller improvements of functional exercise capacity (figure 3 ). It was hypothesized earlier that those patients with the highest oxygen desaturation during exercise would benefit most from supplemental oxygen[ 45 ]. The trials do not provide sufficient evidence for or against this hypothesis. There is limited evidence on the safety of oxygen during exercise and on the safety of exercise without oxygen in patients with desaturation. Clinicians may have concerns about training in hypoxemia because of adverse events and will encourage oxygen supplementation in patients with desaturation during exercise. In theory, oxygen carries the risk of CO 2 retention in COPD patients. The only trial reporting on CO 2 retention[ 29 ] did not observe significant differences of CO 2 levels during exercise tests with oxygen compared with exercise on room air. However, exercise tests may have been too short to assess the effect of CO 2 retention. Exercise is a risk indicator for unmasking latent pulmonary hypertension[ 46 ], but supplemental oxygen may reduce this risk by decreasing the sympathetic tone and the respiratory rate allowing for less end-expiratory pressure[ 47 ]. Rooyackers[ 28 ] did not find any differences in resting mean pulmonary artery pressure between patients with and without oxygen. However, patients stopped exercising when oxygen saturation fell below 90% so that the risk of the exercise program under hypoxemic conditions on the development of pulmonary hypertensions could not be studied. Several studies found a positive acute effect of oxygen during exercise testing on exercise capacity and a number of physiologic mechanisms for the effects of oxygen have been proposed [ 48 - 50 ]. However, these results on the short-term benefit of oxygen during exercise testing seem not to translate into improvements of clinically relevant outcomes during exercise programs. Current data do not suggest benefit from the use of oxygen during exercise to enhance training effects (figure 3 ), but show some benefit in terms of HRQL (figure 2 ) Given the limited methodological quality of trials, any conclusions are vague. The general use of oxygen is only justified, if larger trials of good quality show its benefit on clinically relevant outcomes. The mechanisms of the effects of oxygen during exercise are still insufficiently understood and call for more basic research[ 15 ]. Assisted ventilation also aims at increasing oxygen availability during exercise, but the trials indicated no additional benefit. An exception may represent overnight non-invasive positive pressure ventilation. This treatment may improve quality of sleep as well as daytime gas levels and respiratory muscle function thereby providing a better milieu (pH, PaO 2 , PaCO 2 ) for peripheral muscle function. One trial[ 32 ] found statistically significant improvements of functional exercise capacity and also large improvements of HRQL (mean differences between groups 0.45 to 0.85 in CRQ domain scores, table 4, see Additional file 4 ) with additional non-invasive positive pressure ventilation. These results support the hypothesis formulated by authors of a recent meta-analysis showing that nocturnal non-invasive positive pressure ventilation alone has no effect on exercise capacity and HRQL, but may be beneficial as an adjunct to respiratory rehabilitation[ 51 ]. The eight trials that assessed various supplemental interventions during rehabilitation produced inconclusive results that do not allow recommendations for clinical practice yet. An important result of this systematic review with implications for future research is the low methodological quality and small sample sizes. For example, the majority of trials did not consider stratification for important prognostic factors such as exercise capacity[ 52 ] for randomization. In some trials there were baseline imbalances between groups, for example in terms of exercise capacity[ 27 , 28 , 32 , 33 , 40 ]. The influence of these imbalances on the results was not investigated in any of the trials. Concealment of random allocation and blinding of treatment providers or outcome assessors was also not addressed in most trials. Sample sizes were small except in three trials[ 34 , 35 , 37 ]. Pragmatic trials comparing active interventions, as included in this systematic review, are very useful for clinical practice when clinicians are confronted with the choice between interventions[ 53 ]. However, small sample sizes are problematic in pragmatic trials for at least two reasons: First, differences between study groups tend to be smaller in pragmatic trials than in trials comparing an active intervention with placebo or a sham intervention. Figure 4 shows the results and 95% confidence intervals of a trial comparing respiratory rehabilitation with usual care and of a trial comparing respiratory rehabilitation with and without a supplemental intervention with different sample sizes. It illustrates the importance of sufficient sample sizes in pragmatic trials by showing that for pragmatic RCTs in respiratory rehabilitation, in which widely established patient-important outcomes such as HRQL are used, sample sizes of up to 40 per group will produce imprecise results (large confidence intervals). This imprecision hinders interpretation. Another reason for sufficient samples sizes is that in pragmatic trials patient profiles are usually more variable than in explanatory trials reflecting the wide patient spectrum encountered in clinical practice[ 53 ]. The greater variability in patient profiles carries, on one side, a greater risk for confounding and, on the other side, subgroup analyses will be important to assess whether the effects differ between patient subgroups (effect modification). Subgroup analyses based on prognostically important patient characteristics will provide more differentiated evaluations than one mean for the whole study group, but they require sufficient sample for well-balanced intervention groups. Figure 4 Sample size and interpretation of randomized controlled trials in respiratory rehabilitation. Forest plot with simulated results from two trials with varying sample size, in which the CRQ was used. Boxes with 95% confidence intervals represent point estimates for the difference between CRQ change scores (from baseline to follow-up) of the study groups. Dotted lines represent the minimal important difference of the CRQ (change of 0.5). Trial 1 shows the results from a typical explanatory trial comparing respiratory rehabilitation and no respiratory rehabilitation (usual care) with differences in CRQ change scores around 0.75[5]. Because of the large effect, trial results are interpretable also with imprecise results. Trial 2 shows the results from a pragmatic trial assessing the additional effect of a supplemental intervention (for example oxygen during exercise). The difference between groups is 0.3 and sample size must be large (80 per group) to produce results that are precise enough to allow interpretation. We propose that investigators consider the following aspects in future pragmatic trials on respiratory rehabilitation: First, preliminary sample size considerations should be based on realistic estimates for expected differences between groups, which are typically smaller than in trials without active comparators. To better understand what these sample sizes mean, 95% confidence intervals around the predicted point estimate can be calculated as shown in figure 4 . This approach will help to better foresee the consequence of different sample sizes on interpretation of the data[ 54 ]. Second, COPD patients represent a heterogeneous group and stratification for prognostically important variables should be considered to avoid baseline imbalances that bear on outcomes[ 55 ], as seen in some trials included in this review[ 27 , 28 , 32 , 33 , 40 ]. Third, more attention needs to be paid to general requirements for RCTs of high quality like method of randomisation, concealment of random allocation and blinding of those who assess treatment effects. The strengths of our systematic review include the broad literature search including several databases and extensive hand searching for trials with direct comparisons of interventions that are used in clinical practice. In addition, we contacted authors for additional data and received them from the majority of investigators. This greatly enhanced the informativeness of included studies and thereby of this review. A weakness of this review includes the discussion that is limited to the best-investigated supplements because of the number of interventions included in this review. However, the aim of this review was to analyze current evidence from a meta-epidemiological perspective not giving to much emphasis to single studies. Some may criticize that we did not pool the results from trials on supplemental oxygen during exercise using meta-analysis. However, desaturation or no desaturation during exercise was an important inclusion criterion in four of the five trials and investigators wanted to learn about the effect of supplemental oxygen in these subgroups, in particular. Therefore, we considered the patient profiles of these trials to be too different to provide meaningful pooled estimates. Instead, we provided forest plots (figures 2 and 3 ) to show the individual studies' point estimates and 95% confidence intervals for the CRQ domains and the exercise tests to allow comparisons across studies. In conclusion, data for most supplemental interventions during respiratory rehabilitation are inconclusive. Oxygen during exercise does not seem to provide a patient-important additional benefit for COPD patients during a respiratory rehabilitation, but methodological shortcomings of the trials on supplemental oxygen do not allow conclusive answers. Future trials should pay careful attention to the methodological trial design and to sufficient sample sizes. Abbreviations COPD: Chronic obstructive pulmonary disease RCT: Randomised controlled trial HRQL: Health-related quality of life CRQ: Chronic Respiratory Questionnaire Conflict of interest The authors declare that they have no competing interests. Contributions Protocol writing: Puhan, Bachmann, Schunemann Acquisition of data: Puhan, Bachmann Analysis and interpretation of data: Puhan, Bachmann, Schunemann, Frey Drafting of manuscript: Puhan, Bachmann Critical revision of manuscript for important intellectual content: Puhan, Bachmann, Schunemann, Frey Funding Helmut Horten Foundation Lucas M. Bachmann: Swiss National Science Foundation Research Fellow (PROSPER programme) Supplementary Material Additional File 3 Table 3: Internal validity of included studies Click here for file Additional File 6 Appendix: The appendix lists all studies that were excluded after full text assessment. The full reference and the reason for exclusion are given. Click here for file Additional File 1 Table 1: Characteristics of randomised controlled trials investigating supplemental oxygen and assisted ventilation Click here for file Additional File 4 Table 4: Effect of assisted ventilation on HRQL and exercise capacity Click here for file Additional File 2 Table 2: Characteristics of randomised controlled trials investigating drug and nutritional supplements Click here for file Additional File 5 Table 5: Effect of drug and nutritional interventions on HRQL and exercise capacity Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539299.xml

548306

Selection and characterization of a promoter for expression of single-copy recombinant genes in Gram-positive bacteria

Background In the past ten years there has been a growing interest in engineering Gram-positive bacteria for biotechnological applications, including vaccine delivery and production of recombinant proteins. Usually, bacteria are manipulated using plasmid expression vectors. The major limitation of this approach is due to the fact that recombinant plasmids are often lost from the bacterial culture upon removal of antibiotic selection. We have developed a genetic system based on suicide vectors on conjugative transposons allowing stable integration of recombinant DNA into the chromosome of transformable and non-transformable Gram-positive bacteria. Results The aim of this work was to select a strong chromosomal promoter from Streptococcus gordonii to improve this genetic system making it suitable for expression of single-copy recombinant genes. To achieve this task, a promoterless gene encoding a chloramphenicol acetyltransferase ( cat ), was randomly integrated into the S. gordonii chromosome and transformants were selected for chloramphenicol resistance. Three out of eighteen chloramphenicol resistant transformants selected exhibited 100% stability of the phenotype and only one of them, GP215, carried the cat gene integrated as a single copy. A DNA fragment of 600 base pairs exhibiting promoter activity was isolated from GP215 and sequenced. The 5' end of its corresponding mRNA was determined by primer extention analysis and the putative -10 and a -35 regions were identified. To study the possibility of using this promoter (PP) for single copy heterologous gene expression, we created transcriptional fusions of PP with genes encoding surface recombinant proteins in a vector capable of integrating into the conjugative transposon Tn 916 . Surface recombinant proteins whose expression was controlled by the PP promoter were detected in Tn916-containing strains of S. gordonii and Bacillus subtilis after single copy chromosomal integration of the recombinant insertion vectors into the resident Tn 916 . The surface recombinant protein synthesized under the control of PP was also detected in Enterococcus faecalis after conjugal transfer of a recombinant Tn 916 containing the transcriptional fusion. Conclusion We isolated and characterized a S. gordonii chromosomal promoter. We demonstrated that this promoter can be used to direct expression of heterologous genes in different Gram-positive bacteria, when integrated in a single copy into the chromosome.

Background In the past ten years there has been a growing interest in engineering Gram-positive bacteria for biotechnological applications, including vaccine delivery. [ 1 - 4 ], and in situ production of anti-infective protectants [ 5 ] and microbicides [ 6 ]. A common approach to genetic manipulation of bacteria is based on the use of plasmid expression vectors since these recombinant molecules can be introduced into bacterial cells by a variety of genetic techniques such as natural transformation, artificial transformation, transduction, conjugative mobilization, and electroporation [ 7 - 9 ]. However, the major limitation of this approach is due to the fact that recombinant plasmids are often lost from the bacterial culture upon removal of antibiotic selection. Certainly, this has consequences when using recombinant bacteria in vivo where their replication occurs in the absence of selection. An alternative approach is to integrate recombinant DNA molecules into the bacterial chromosome since this method allows increased in vivo stability of the genetic constructs. Therefore a lot of efforts have focused on the development of efficient expression systems based on chromosomal integration of expression cassettes [ 10 , 11 ]. Naturally transformable bacteria represent a convenient model, since heterologous DNA can be easily integrated into their chromosomes, whereas genetic manipulation of non-transformable bacteria is more difficult and relies mainly on electroporation and conjugative mobilization of foreign DNA molecules. We have previously described a genetic system based on conjugative transposons allowing stable integration of recombinant DNA into the chromosome of transformable and non-transformable streptococci [ 12 , 13 ]. A series of transposon insertion vectors containing two regions of homology with Tn 916 [ 14 ] have been created in order to manipulate both naturally transformable and non-transformable Gram-positive bacteria carrying Tn 916 [ 12 ]. The aim of this work was to select a strong promoter to improve this genetic system making it suitable for expression of single-copy recombinant genes in a broad spectrum of Gram-positive bacteria. Results and discussion Promoter selection by chromosomal integration To select resident promoters from the genome of Streptococcus gordonii , we performed a random ligation of streptococcal DNA to a promoterless cat gene, conferring resistance to chloramphenicol (Cm). The ligation mixture was used to transform the naturally transformable S. gordonii «Challis» strain V288 and transformants were selected for Cm resistance. Chromosomal DNA flanking the promoterless cat gene provided the homology for the random integration of cat into the chromosome during transformation (Fig. 1 ). 71 Cm-resistant (Cm R ) transformants were isolated, presumably as a result of transcriptional fusions of streptococcal promoters to the promoterless cat gene. Eighteen Cm R transformants were selected for further characterization. The strategy commonly used to select promoters is based on cloning random chromosomal DNA fragments in a promoter probe vector upstream of a promoterless reporter gene. However, integrating the promoterless reporter gene ( cat ) directly into the streptococcal chromosome, allowed us to select resident chromosomal promoters expressing cat after in vivo transcriptional fusion at a single locus on the chromosome. This is preferable when looking for promoters to express heterologous genes integrated into the chromosome in a single copy. In vivo analysis of promoter strength was determined in the eighteen selected transformants by measuring the minimum inhibitory concentration (MIC) of Cm. Fifteen transformants exhibited a MIC of 16 μg/ml, whereas the remaining three transformants exhibited a MIC of 8 μg/ml. The fifteen strains with higher MIC were tested for the stability of the chloramphenicol-resistance phenotype: after 50 generations of growth without selection, bacterial cultures were plated on non-selective medium and at least 200 colonies were picked and tested for Cm resistance. The stability of the resistance phenotype varied considerably among the different transformants (data not shown). Three strains (GP214, GP215 and GP216) showed a 100% stability and were chosen for further analysis. Cloning of a promoter from S. gordonii The structure of the integrated cat gene in GP214, GP215 and GP216 was analyzed by Southern blot. Only GP215 showed to have a single cat copy integrated into the chromosome, whereas in GP214 and GP216 integration occurred at multiple sites (data not shown). In order to clone the regions flanking cat integration site in GP215, the chromosome of this strain was cut with Taq I whose recognition sequence is absent inside the cat gene. The derivative fragments were ligated to pBLUESCRIPT, and the ligation mixture was used to transform Escherichia coli cells; transformants were then selected for Cm R . All transformants analyzed for plasmid content showed to carry a plasmid of the same size. One of these transformants (GP334) was selected for further analysis and the transforming plasmid was named pVMB5. By restriction analysis we showed that pVMB5 contains a 2.2 kb Taq I insert where a 600 base pairs (bp) streptococcal DNA fragment was cloned upstream of the cat gene. This DNA fragment was stably maintained in E. coli , where it retained its promoter activity, conferring Cm R to E. coli . This is of considerable interest since it is known that very often cloning streptococcal promoters in a high copy number plasmid results in the failure of that plasmid to be established [ 15 ]. Sequence analysis The streptococcal DNA upstream of the cat gene in pVMB5 was sequenced (GenBank accession number: U74080). Analysis of the sequence revealed the presence of an open reading frame (ORF1), preceded by a typical ribosome binding site (RBS) (Fig. 2 ). A BLAST search with the partial sequenced genome of S. gordonii showed that ORF1 represents the 5'-end of a gene encoding the first 48 amino acids (aa) of an uncharacterized protein conserved in bacteria and whose function is unknown. As indicated in figure 2 , this truncated streptococcal protein is translationally fused to the N-terminus of CAT. Sequence analysis also revealed that the first 122 bp of the cloned streptococcal fragment belong to the 3'-end of a gene encoding a putative acetyltransferase (ORF2). To obtain information about the chromosomal region containing the 600 bp streptococcal DNA fragment cloned in pVMB5, we looked throughout the whole contig sequence and we found out that upstream ORF2, and partially overlapping with it, there is an ORF encoding a DltD horthologue ( dltD ), a protein involved in D-alanine incorporation into lipoteichoic acid (LTA) [ 16 ], whereas downstream ORF1, there is a gene encoding a Tmp7 transmembrane protein (Fig. 2 ). Identification of the transcriptional start site To identify the sequences responsible for the observed promoter activity, the 5'-end of the cat- specific mRNA was mapped by primer extension with a specific primer. Total RNA was isolated from E. coli GP334 (containing pVMB5) and S. gordonii GP215. In both strains, the position of the 5'-end of the mRNA was located at the same purine residue at position 423 of the 600 bp region of streptococcal DNA, 26 nucleotides upstream of the ORF1 translational start site (Fig. 2 and 3 ). Putative -35 and -10 sequences closely resembling the consensus E. coli σ 70 and Bacillus subtilis σ 43 [ 17 ] recognition sequences TTGACA (-35) and TATAAT (-10) could be identified. The -35 region was TTGCAA, and the -10 region was TAGAAT. The spacing between the -35 and -10 region was 17 bp and was thus similar to the spacing in B. subtilis (17 to 19 bp) and E. coli (16 to 18 bp) promoters. Moreover, a TG nucleotide pair was found 1 bp upstream of the -10 region; such a structure is typical of Gram-positive bacteria promoters [ 18 ]. Based on this information, we concluded that indeed we isolated a streptococcal promoter which was designated PP. The spacing between the 5'-end of the mRNA and the -10 hexanucleotide was only 2 bp, which is unusually short. A second putative -10 region could be identified in the TATGAT hexanucleotide (Fig. 2 ), whose distance from the 5' end of the mRNA is 7 bp. However, since this sequence is not preceded by the TG nucleotide pair typical of Gram-positive promoters, and is separated from the -35 region only by 12 bp, we suppose that this -10 region of PP is probably not active. Insertion vectors to express heterologous proteins We have previously developed a genetic system based on the use of Streptococcus pyogenes surface fibrillar M6 protein, as a partner for the construction of translational fusions to deliver foreign proteins on the surface of S. gordonii [ 10 ]. To determine the possibility of using PP for chromosomal single copy heterologous gene expression in Gram-positive bacteria, we first generated a transcriptional fusion of PP with a promoterless gene encoding the M6 protein ( emm6 ) in pSMB47, a suicide vector capable of integrating heterologous DNA into the conjugative transposon Tn 916 via homologous recombinantion [ 12 ]. The recombinant plasmid was named pSMB139 (see Methods) (Fig. 4 ). To construct a Tn 916 insertion vector that could be used to express translational fusions between M6 and any heterologous proteins under the PP promoter control, a 900 bp Avr II- Hind III fragment internal to emm6 in pSMB139 was replaced by a 390 bp Avr II- Hind III fragment of pSMB55 containing a multiple cloning site [ 10 ] (Fig. 4 ). The resulting vector, named pSMB148 (Fig. 4 ), allows to create translational fusions of heterologous proteins between the first 122 N-terminal aa and the last 140 aa of the M6 protein, which provides sequences necessary for cell wall anchoring. pSMB148 was used to create two derivative vectors in which the emm6 gene was fused respectively with a DNA sequence encoding 339 aa of the chicken ovalbumin (OVA) (pSMB156), and a DNA sequence encoding 458 aa of the tetanus toxin fragment C (TTFC) (pSMB288) (see Methods). A schematic representation of the recombinant proteins is shown in Fig. 5A . Expression of recombinant proteins in Gram-positive bacteria S. gordonii , B. subtilis , and Enterococcus faecalis were three Gram-positive hosts used to analyze the capability of PP to direct transcription of the heterologous genes expressing M6, M6/TTFC and M6/OVA recombinant proteins. Expression of M6 in S. gordonii pSMB139, bearing a transcriptional fusion of PP with emm6 , was introduced by natural transformation in S. gordonii GP201, a strain with a single copy of Tn 916 integrated into the chromosome. One of the transformants (GP1241), in which the integrative suicide vector drove the integration of the PP- emm6 fusion into Tn 916 , was isolated and analyzed for M6 protein expression. Envelope fractions (containing surface associated proteins) prepared from equal amounts of cells grown to mid-log, early and late stationary phase, were analyzed by Western-blotting with an anti-M6 monoclonal antibody. Multiple bands could be detected in fractions of cells grown to mid-log and early stationary phase, whereas no band was detected in the fraction of cultures grown to late stationary phase (Fig. 6A ). The intensity of the signal was higher during the mid-log growth phase suggesting that either PP is more active during exponential growth or that the M6 protein is being degraded during stationary phase. The presence of multiple reactive bands of molecular masses close to the hypothetical size of M6 (predicted molecular weight, 49 kDa) is probably due to the fact that coiled-coil proteins like M6 run at aberrant sizes on denaturing gels [ 19 ]. Expression of M6/TTFC in B. subtilis Competent cells of B. subtilis GP800.2, containing one copy of Tn 916 integrated into the chromosome, were transformed with the insertion vector pSMB288, bearing the transcriptional fusion PP- emm6/ttfc . One transformant, GP848, was chosen for further studies. A culture of GP848 was grown to mid-exponential phase and analyzed by Western-blotting for the presence of recombinant M6/TTFC. As shown in Fig. 6B , two reactive bands could be detected in the envelope fraction. The lower band, indicated by an arrow, corresponds to the mature protein (predicted molecular weight, 82 kDa), while the upper band probably represents an unprocessed form (predicted molecular weight, 86.4 kDa). Expression of M6/OVA in E. faecalis Using a previously described genetic system [ 12 ], we constructed a derivative of the E. faecalis strain OG1SS [ 20 ] expressing the recombinant M6/OVA protein under PP control. pSMB156, bearing the PP- emm6/ova fusion, was first introduced in the Tn 916 containing B. subtilis GP800.2 by natural transformation, to obtain a recombinant conjugative transposon containing the transcriptional fusion. The recombinant transposon was then transferred by conjugation into E. faecalis OG1SS. Transconjugants were detected at a frequency of 4 × 10 -10 transconjugants/recipient. One of them (GP431) was analyzed for cell-surface expression of M6/OVA by flow-cytometric analysis using an anti-ovalbumin polyclonal antibody. The presence of recombinant M6/OVA on the surface of GP431 was clearly demonstrated by the increase of the fluorescence intensity in this strain, as compared to the parental control OG1SS (Fig. 7 ). Conclusions We have isolated and characterized a promoter from the chromosome of S. gordonii , and demonstrated that it can be used to direct expression of heterologous genes in different Gram-positive bacteria when integrated in a single copy into the chromosome. This promoter, together with the genetic system based on suicide vectors able to integrate into conjugative transposons, represents a useful tool for the stable manipulation of a broad spectrum of Gram-positive bacteria. Methods Bacterial strains, plasmids and growth conditions All strains and plasmids used in this work are listed in Table 1 . E. coli strains DH5α and HB101 were cultured in Luria-Bertani (LB) broth. For maintenance of plasmids, ampicillin (100 μg/ml), chloramphenicol (20 μg/ml) or erythromycin (100 μg/ml) was added to the growth medium. Streptococcal strains were cultured in Brain Heart Infusion medium (BHI, Difco) or Tryptic Soy Broth (TSB, Difco) in the presence of chloramphenicol (5 μg/ml), erythromicin (100 μg/ml) or streptomycin (500 μg/ml) whenever required. Transformation of naturally competent cells of S. gordonii V288 and GP201, scoring and genetic analysis of transformants was carried out as already described [ 21 , 22 ]. B. subtilis strains were grown in LB broth with erythromicin (3 μg/ml) when it was required. Competent cells of B. subtilis GP800.2 were prepared and transformed according to described procedures [ 23 ]. Agar (1.5%) was added to LB, BHI or TSB to obtain solid media. All cultures were incubated at 37°C. DNA manipulation Total DNA preparation of S. gordonii was performed as previously described [ 21 ]. Plasmid DNA was prepared using the Qiagen Plasmid Kit (Qiagen) according to the manufacturer's instructions. All recombinant techniques were performed following standard procedures [ 24 ], using E. coli HB101 or DH5α as a host. DNA restriction enzymes were obtained from Roche and used according to the manufacturer's instructions. Promoter selection by chromosomal integration A 1.6 kb Hind III- Bam HI fragment of plasmid pKT [ 25 ], containing a cat gene, was ligated with random chromosomal DNA fragments of S. gordonii V288 previously cut with both Hind III and Bam HI. The initiation codon of cat was 34 bp downstream of the Hind III site, therefore Hind III cleavage would leave cat promoterless and preceded by an intact ribosome binding site. The ligation mixture was introduced in S. gordonii V288 and transformants were selected for Cm R . MIC determination The MIC of chloramphenicol for S. gordonii was determined following standard procedures [ 26 ]. Construction of recombinant vectors A 1648 bp fragment containing the emm6 gene (encoding M6, a fibrillar surface protein of S. pyogenes ) was amplified by PCR from plasmid pVMB20 [ 22 ] using the oligonucleotides 5'-AT GGATCC AT CATATG GCTAAAAATAACACGAAT-3' (upstream primer, containing a Bam HI site and introducing a Nde I site at the ATG translation initiation codon) and 5'-GCAT GTCGAC CATAATCATTAAATGTATCTCAT-3' (downstream primers containing a Sal I site). This 1648 bp PCR fragment was digested with Bam HI and Sal I and cloned in pVA891 [ 27 ] previously digested with Bam HIand Sal I, resulting in plasmid pSMB89. A 451 bp region containing the streptococcal promoter PP was amplified from pVMB5 using the following primers: 5'-CGA GGATCC TTTA ATCGAT ACTCATG-3' (upstream primer, containing a Bam HI and a Cla I site) and 5'-CCG CATATG GTTCTCCTTTTTATTTGT-3' (downstream primers containing a Nde I site). After digestion with Bam HI and Nde I, this PCR product was inserted between the Bam HI and Nde I site of pSMB89 to obtain a transcriptional fusion of PP with emm6 . The resulting plasmid was named pSMB128. This plasmid was first cut with Bam HI, treated with Klenow enzyme to generate blunt ends, and finally cut with Sal I to obtain a 2.0 kbfragment containing PP- emm6 fusion. This fragment was gel-purified and ligated to the suicide integrative plasmid pSMB47 [ 12 ] previously cut with Hind III, treated with Klenow enzyme to generate blunt ends, and finally cut with Sal I. The resulting plasmid was named pSMB139 (Fig. 4 ). To introduce a multiple cloning site in the emm6 gene contained in pSMB139, the 900 bp Avr II- Hind III fragment internal to emm6 was replaced with the 390 bp Avr II- Hind III fragment of emm6 from pSMB55 [ 10 ] (Fig. 4 ). The resulting plasmid was named pSMB148. A 1016 bp DNA region encoding 339 aa of the chicken ovalbumin (OVA) (Gene Bank accession number: V00383) was amplified with the following primers: 5'-CT AGATCT GACAGCA CCAGGACAC-3' (upstream primer containing a Bgl II site) and 5'-TA AAGCTT TAGGGG AAACACATCTG-3' (downstream primer containing a Hind III site). After digestion with Bgl II and Hind III, this segment was introduced in pSMB148 previously digested with Bgl II and Hind III. The resulting plasmid, named pSMB156, contained a translational fusion of M6 with OVA. To create a translational fusion of M6 with the tetanus toxin fragment C (TTFC) a 1374 bp Bgl II- Hind III fragment encoding 458 aa of TTFC was isolated from pSMB158 [ 28 ] and cloned between the Bgl II and Hind III sites of pSMB148. The resulting plasmid was named pSMB288. Western-blot analysis Preparation of S. gordonii and B. subtilis cell envelope fractions (representing the protoplast surface containing the cell membrane together with cell wall fragments associated to the protoplasts) was performed as already described [ 29 , 30 ]. The monoclonal antibody 10B6 [ 31 ] diluted 1:1000 was used to detect the presence of M6 protein. M6/TTFC fusion protein was visualized with an anti-TTFC rabbit serum (Calbiochem-Novabiochem Corporation) diluted 1:1000. Flow-cytometric analysis Flow-cytometric analysis of E. faecalis was performed as already described [ 28 , 32 ] using an anti-OVA rabbit serum diluted 1:300 [ 29 ]. DNA sequence determination The promoter containing fragment cloned in pVMB5 was sequenced by dideoxy chain termination method [ 33 ] as already described [ 34 ]. Denatured plasmid DNA was used as template. RNA isolation Total RNA was isolated from a 50 ml cell culture of S. gordonii and E. coli grown to late exponential phase (OD 590 ≌ 0.5). Cells were harvested by centrifugation at 6000 × g at 4°C and lysed according to the following procedures. E. coli cells were first resuspended in hot (100°C) lysis buffer (50 mM Tris/HCl pH8, 1 mM EDTA, 1% SDS) and lysed by boiling the suspension for 5 minutes. S. gordonii cells were resuspended in lysozyme buffer (25 mM Tris/HCl (pH8), 10 mM EDTA, 50 mM glucose) and subjected to three cycles of freezing in liquid nitrogen and thawing at 52°C. After incubating with 0.2 mg/ml of lysozyme for 30 min at 37°C, one volume of hot (100°C) lysis buffer (100 mM Tris/HCl (pH8), 2 mM EDTA, 2% SDS) was added, and complete lysis was obtained by boiling cells for 5 minutes. After boiling, all lysates were cooled on ice for 5 min and total RNA was purified using the SV Total RNA Isolation System (Promega). Primer-extention analysis Primer extention analysis was performed with a synthetic oligonucleotide 5'-GTTCTTTACGATGCC-3' (position 47 to 61 relative to the initiation of the cat gene). Two pmol of the oligonucleotide, labeled with [γ- 32 P] ATP (3000 Ci/mmol, Amersham) using T 4 polynucleotide kinase (New England Biolabs), were precipitated with 10 μg of RNA and the pellet was resuspended in 8 μl of Moloney Murine Leukemia Virus (M-MuLV) Reverse Transcriptase buffer (50 mM Tris/HCl (pH8.3), 8 mM MgCl 2 , 10 mM DTT). The mixture was heated at 65°C for 3 min, cooled rapidly at -80°C for 1 min and then transfered on ice until it was completely thawed. 1 μl of a 3.75 mM deoxynucleoside triphosphate solution and 10 U of M-MuLV Reverse Transcriptase (New England Biolabs) were added to the RNA-primer hybrid. The reaction mixture was incubated at 48°C for 30 min and terminated with 10 ml of stop solution (95% formamide, 20 mM EDTA pH8.0, 0.05% bromophenol blue, and 0.05% xylene cyanol FF). The reverse transcriptase reactions were analyzed by electrophoresis on a 6% polyacrylamide-7 M urea gel with sequencing reaction obtained with the same primer used as size standards. Conjugation Conjugation experiments were performed on solid media as previously described [ 12 ]. Authors' contributions RP, characterization of promoter, engineering of S. gordonii , writing of manuscript. TM, engineering of B. subtilis and E. faecalis MRO, participation in experimental work, data evaluation RM, participation in experimental work, data evaluation, writing of manuscript GP, design and coordination of the study, data evaluation, direct supervision of experimental work, writing of manuscript All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548306.xml

549031

Responsiveness: a reinvention of the wheel?

Background Since the mid eighties, responsiveness is considered to be a separate property of health status questionnaires distinct from reliability and validity. The aim of the study was to assess the strength of the relationship between internal consistency reliability, referring to an instrument's sensitivity to differences in health status among subjects at one point in time, and responsiveness referring to sensitivity to health status changes over time. Methods We used three different datasets comprising the scores of patients on the Barthel, the SIP and the GO-QoL instruments at two points in time. The internal consistency was reduced stepwise by removing the item that contributed most to a scale's reliability. We calculated the responsiveness expressed by the Standardized Response Mean (SRM) on each set of remaining items. The strength of the relationship between the thus obtained internal consistency coefficients and SRMs was quantified by Spearman rank correlation coefficients. Results Strong to perfect correlations (0.90 – 1.00) was found between internal consistency coefficients and SRMs for all instruments indicating, that the two can be used interchangeably. Conclusion The results contradict the conviction that responsiveness is a separate psychometric property. The internal consistency coefficient adequately reflects an instrument's potential sensitivity to changes over time.

Background Responsiveness, a concept introduced in the mid-eighties by bio-medical researchers, is considered to be an essential measurement property of health status questionnaires, distinct from reliability and validity [ 1 ]. However, it can be questioned whether an instrument's sensitivity to differences between health status changes over time, which refers to responsiveness, is different from an instrument's sensitivity to differences in health among subjects at one point in time, which refers to the psychometric concept of parallel forms reliability from the framework of classical test theory [ 2 ]. A number of theorists have argued that responsiveness is not a separate psychometric attribute of health status instruments, but merely some form of construct validity [ 3 ]. The aim of the study is to provide empirical evidence of this notion by investigating the relationship between instrument responsiveness and the traditional psychometric concept of parallel forms reliability as embodied by the internal consistency coefficient [ 2 , 3 ]. Methods We used three datasets comprising the scores of patients on three widely used health status instruments on two moments in time in order to assess health changes. The first dataset was from a randomized clinical trial investigating the effects of arm and leg rehabilitation training on the functional recovery of stroke survivors using the 10-item Barthel (basic) activities of daily living scale [ 4 ]. Patients (n = 89) were rated one week and 12 weeks after stroke. A Barthel scale score ranges from zero to 20 points with higher scores indicating more independent functioning. The second dataset comprised the scores on the 45-item physical component of the Sickness Impact Profile (SIP) of 227 patients with myocardial infarction [ 5 ](on average 2 year interval between assessments), 120 patients with stroke [ 6 ] (3 year interval between assessments) and 141 patients scheduled for a carotid endartectomy surgical procedure [ 7 ] (3 months time interval). The SIP physical items are scored on a dichotomous scale with 1 point for each endorsed item statement. The scale ranges from 0 to 45 points with higher scores indicating higher levels of sickness related dysfunction. The third dataset contained the scores of 164 patients with Graves' ophthalmopathy scored on the 8-item psychosocial dimension of the Graves' ophthalmopathy quality of life (GO-QOL) instrument [ 8 ]. The GO-QOL scale items are scored on a 1 to 3 point rating scale. Overall scores are transformed to a 0–100 scale with higher scores indicating better psychosocial functioning. Patients completed the instrument before and three or six months after radiotherapy or eye surgery. Only subjects with no missing values at baseline or follow-up were included in the analysis. The Barthel dataset had no missing values, the SIP datasets had 13 % (16/120), 28% (64/227) and 0.7% (1/141) missing values respectively and the GO-QOL had 0.6% (1/164) missing values. For the SIP datasets, there was a mean deterioration in health (1 point), for the Barthel and GO-QOL scales patients were improved at follow-up (Table 1 ). Table 1 Score statistics and reliabilities (α) at baseline and follow-up. Barthel N = 89 SIP N = 407 GO-QOL N = 163 Baseline score (SD, IQR 1 ) 7.98 (4.52, 5–12) 5.12 (6.28, 1–7) 59.32 (24.64, 44–81) Follow-up score (SD, IQR) 14.25 (5.06, 10–19) 6.13 (7.57, 0–9) 65.22 (24.17, 50–81) Mean change score (SD) 2 6.27 (3.21) 1.01 (4.36) 5.90 (17.13) SRM 3 1.95 0.23 0.34 Chronbach's α time 1 0.86 0.92 0.83 Chronbach's α time 2 0.89 0.94 0.84 1) Distribution of change scores was approximately normal for all three datasets 2) IQR = interquartile range 3) SRM = Standardized Response Mean (see statistical analysis section) Statistical analysis The analysis aimed to assess the strength of the relationship between internal consistency reliability (Cronbach α or Kuder-Richarson-20) reflecting sensitivity to differences in health status among patients [ 3 ], and the Standardized Response Mean effect size (SRM) indicating an instrument's sensitivity to change. The SRM is calculated as the mean of the change scores divided by the standard deviation of the change scores [ 3 , 9 ]. In a stepwise procedure, we reduced the baseline internal consistency by removing the item contributing most to the internal consistency coefficient until 0.60 was reached, which was considered as the minimum standard for reliability [ 10 ]. For the 45-item SIP physical scale, two items were removed at every step. For the other instruments, one item was removed at each step. Using the remaining items at each item reduction step, we calculated the SRM. The thus decreasing internal consistency coefficients and associated SRMs were plotted and the strength of the relationship was calculated using Spearman rank correlation coefficients. All analyses were performed with SPSS 11.0, a commercially available software package. Results Figures 1 , 2 and 3 show the spearman rank correlations between the internal consistency coefficients and the SRM using the scores on the Barthel, the SIP and the GO-QOL respectively. The spearman rank correlations ranged between 0.90 for the Barthel index to 1.00 for the GO-QoL indicating strong to perfect relations between internal consistency and responsiveness. Figure 1 Barthel – Relation between the internal consistency reliability (alpha) and the SRM (Spearman's r = 0.90) Figure 2 SIP – Relation between the internal consistency reliability (alpha) and the SRM (Spearman's r = 0.99) Figure 3 GO-QOL – Relation between the internal consistency reliability (alpha) and the SRM (Spearman's r = 1.00) Discussion Our results contradict the conviction that responsiveness is a separate psychometric property of health scales. Internal consistency reliability, reflecting a scale's sensitivity to cross-sectional differences in health, closely coincided with the instruments' sensitivity to change as measured with the standardized response mean. Our results also reflect what is already known within the framework of classical test theory. A test score cannot correlate more highly with any other variable than its own true score [ 2 ]. This implies that the maximum correlation between an observed test score and any other variable, i.e. its validity, is the square root of its reliability [ 2 ]. Thus, the more reliable a test, the more potential for validity, in this case responsiveness, there exists. We used nested versions of the same test, which are highly correlated with each other, to illustrate this phenomenon. It is likely, however, that the results will also apply with different instruments measuring similar health constructs that are highly inter-correlated. It should also be noted that the results apply to one-dimensional psychometric scales and not to instruments containing so-called "causal" variables, for example disease symptoms [ 3 ] since these instruments are not strictly one-dimensional. We used the SRM effect size that uses the standard deviation the change scores and therefore includes all information about the changes on the selected instruments. The results can not generalized to alternative effect sizes such as Cohen's effect size or Guyatt's responsiveness statistic [ 1 ] because these largely depend on the variability of scores at baseline or the variability in scores obtained from a separate, not improved, sample. In a frequently cited paper, Guyatt et al. [ 1 ] made the distinction between discriminative instruments, whose purpose it is to measure differences between subjects and evaluative instruments, designed to examine change over time. This in contrast to most of the scales used in clinical medicine (blood pressure, cardiac output), which are assumed to work well in both discriminative and evaluative roles. To corroborate his arguments, he used the hypothetical example of two health status instruments designed to evaluate therapeutic interventions in patients with chronic lung disease that were presented to the same patient sample (Table 2 ). "Evaluative" instrument A showing poor test-retest reliability because of small between subject score variability but excellent responsiveness, and "discriminative" instrument B with excellent reliability because of large between-subject score variability and poor responsiveness. From Table 2 , however, it can be seen that this representation of instrument behaviour in clinical research is logically inconsistent, since it does not explain how two instruments, both measuring the same health construct show such divergent score distributions at baseline. According to instrument A the sample is highly homogeneous, while it is highly heterogeneous according to instrument B. In Appendix 1 (see additional file 1 ), we show that the above representation is not impossible, but highly unlikely since it occurs only in extreme situations. Table 2 Representation of the scores on "evaluative" instrument A and "discriminative" instrument B in a randomized clinical trial [1] Instrument A Time 1 Time 2 Intervention Time 3 Difference score Exercise test Result Subject 1 8 9 Verum 15 +6 Much improved Subject 2 9 8 “” 15 +7 Much improved Subject 3 8 9 “” 15 +6 Much improved Subject 4 9 8 “” 15 +7 Much improved Subject 5 8 9 Placebo 8 -1 Unchanged Subject 6 9 8 “” 9 +1 Unchanged Subject 7 8 9 “” 8 -1 Unchanged Subject 8 9 8 “” 9 +1 Unchanged Instrument B Time 1 Time 2 Time 3 Difference score Exercise test Result Subject 1 5 5 Verum 5 0 Much improved Subject 2 9 9 “” 9 0 Much improved Subject 3 13 13 “” 13 0 Much improved Subject 4 17 17 “” 17 0 Much improved Subject 5 5 5 Placebo 5 0 Unchanged Subject 6 9 9 “” 9 0 Unchanged Subject 7 13 13 “” 13 0 Unchanged Subject 8 17 17 “” 17 0 During the past 20 years, clinimetric research has resulted in about 25 definitions and 30 measures of instrument responsiveness, sometimes referred to as sensitivity to change or longitudinal validity [ 11 ]. Moreover, it is evaluated in literally hundreds of published papers on the validation of health status instruments. Our results show that responsiveness, as measured with the SRM, mirrors the traditional concept of parallel test reliability as embodied by the internal consistency coefficient. When comparing instruments measuring similar health constructs, an instrument sensitive to health differences among subjects is also likely to be sensitive to therapy-induced change as well. However, further empirical data will be needed to confirm the relationship between internal consistency and responsiveness, e.g., by reviewing studies in which health status instruments were compared on their responsiveness. Authors' contributions Robert Lindeboom conceived the idea for the manuscript, Mirjam Sprangers and Robert Lindeboom wrote the manuscript, Koos Zwinderman provided the mathematics and rewrote parts of the manuscript in earlier drafts Supplementary Material Additional File 1 Appendix 1: Is it possible to have one high reliable scale with low 'responsiveness', and a low reliable scale with high 'responsiveness' measuring the same construct? Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549031.xml

526391

Resistant starch consumption promotes lipid oxidation

Background Although the effects of resistant starch (RS) on postprandial glycemia and insulinemia have been extensively studied, little is known about the impact of RS on fat metabolism. This study examines the relationship between the RS content of a meal and postprandial/post-absorbative fat oxidation. Results 12 subjects consumed meals containing 0%, 2.7%, 5.4%, and 10.7% RS (as a percentage of total carbohydrate). Blood samples were taken and analyzed for glucose, insulin, triacylglycerol (TAG) and free fatty acid (FFA) concentrations. Respiratory quotient was measured hourly. The 0%, 5.4%, and 10.7% meals contained 50 μCi [1- 14 C]-triolein with breath samples collected hourly following the meal, and gluteal fat biopsies obtained at 0 and 24 h. RS, regardless of dose, had no effect on fasting or postprandial insulin, glucose, FFA or TAG concentration, nor on meal fat storage. However, data from indirect calorimetry and oxidation of [1- 14 C]-triolein to 14 CO 2 showed that addition of 5.4% RS to the diet significantly increased fat oxidation. In fact, postprandial oxidation of [1- 14 C]-triolein was 23% greater with the 5.4% RS meal than the 0% meal (p = 0.0062). Conclusions These data indicate that replacement of 5.4% of total dietary carbohydrate with RS significantly increased post-prandial lipid oxidation and therefore could decrease fat accumulation in the long-term.

Background Resistant starch (RS) is any starch that is not digested in the small intestine but passes to the large bowel for fermentation [ 1 ]. Retrograded amylose (a linear polymer of glucose residues linked by α(1→4) bonds; RS1), such as cooked and cooled starchy foods like pasta salad, and native starch granules (RS2), such as those found in high-amylose maize starch and bananas, are the major components of dietary RS. Calories from RS that are undigested in the small intestine can be salvaged by fermentation to short-chain fatty acids (SCFA; acetate, butyrate, proprionate) by the microflora of the large bowel. Fermentation of RS in the large bowel gives rise to increased production of SCFA which is reflected in higher epithelial and portal concentrations. SCFA concentration in the periphery, however, is very low and therefore difficult to measure accurately so any increase in production of SCFA in response to RS consumption may not be detectable in the peripheral circulation. Acute human studies describe variable postprandial glycemic and/or insulinemic responses to RS ingestion. In general, it is accepted that RS consumption lowers postprandial glucose concentrations marginally and postprandial insulin concentrations markedly. Many groups report a decrease in postprandial glycemic or insulinemic responses to RS ingestion relative to digestible starch (DS) consumption [ 2 - 7 ], whereas some report no change [ 8 - 11 ]. It is important to note that the fat content of the diet has a significant impact on the glycemic response to a meal and some meal tests contained no fat or the fat content of the meal varied among the different RS diets making results from these studies difficult to interpret [ 2 - 4 ]. Also, there are many sources of RS, such as beans, high amylose corn starch, and potatoes, which possess different physicochemical properties. So, the source of RS can influence the glycemic/insulinemic response to RS ingestion. Many studies have examined the relationship between RS ingestion and postprandial metabolite and hormone concentrations. Fewer studies have documented the effect of RS on lipid metabolism. In humans, five weeks of RS feeding lowered fasting cholesterol and triglyceride concentrations and postprandial plasma insulin concentrations relative to digestible starch (DS) feeding [ 12 , 13 ]. It has also been reported that chronic RS feeding in rats causes a decrease in adipocyte cell size relative to DS feeding [ 14 , 15 ]. In addition, expression of fatty acid synthase was lower in rats fed a RS-based diet than in those fed a DS-based diet [ 16 ]. Taken together, these studies provide evidence that RS intake has an effect upon the activity of key lipogenic enzymes and adipocyte morphology. Thus, it seems that the effects of this carbohydrate subtype on lipid metabolism should be carefully examined in human studies. It is possible that strong physical association between RS and dietary lipid may slow the absorption, and thereby increase the oxidation, of dietary lipid. Currently, there is no evidence pertaining to the dose-response relationship for RS ingestion (as part of a mixed meal) and postprandial glycemia, insulinemia, fat oxidation, or meal fat storage. It is important that these parameters be defined before designing and conducting long-term, prospective RS feeding studies. Results No difference in fasting or postprandial insulin, glucose, FFA, or triglyceride concentration was observed between any of the RS doses examined (Figure 1 ). Figure 1 Circulating glucose (a, b), insulin (c, d), free fatty acid (e), and triglyceride (f) concentrations in response to the RS content of a breakfast meal. Serum glucose and insulin measurements were conducted on 12 healthy adults. Data is presented as mean ± SEM. Overall, the dose of RS in the meal had a significant influence on ΔRQ (respiratory quotient) values (F-test, 0.04; Figure 2 ). This overall effect was due to a significantly lower ΔRQ at the 5.4% RS dose than the 0% (p = 0.02) or 10.7% (p = 0.009) RS doses, indicating an increase in fat oxidation in response to the 5.4% RS meal relative to the 0% and 10.7% RS doses (Figure 2 ). ΔRQ was significantly lower for the 5.4% RS meal than 0% RS meal at 120, 240, 300 and 360 minutes (p = 0.05, 0.03, 0.02 and 0.04, respectively) whereas significant differences occurred at 120, 180, 240, 300 and 360 minutes (p = 0.01, 0.01, 0.005, 0.02, and 0.03, respectively) for the 5.4% RS versus 10.7% RS meals. These data are reflected in total macronutrient oxidation rates (Figure 3 ), which show a significant increase in the amount of fat oxidized at the 5.4% RS dose relative to the 0% RS meal, with a concomitant decrease in total carbohydrate oxidation. Figure 2 Respiratory quotient (RQ; change from baseline) in response to RS content of a breakfast meal. Respiratory gas exchange measurements were conducted on 12 healthy adults using the ventilated hood method. Data is presented as mean ± SEM. * p < 0.05 for a difference from the 0% meal at the same time point. # p < 0.03 for a difference with the 10.7% meal at the same time point. Figure 3 Total fat (a) and carbohydrate (b) oxidation in response to RS content of a breakfast meal. Macronutrient oxidation, assessed via indirect calorimetry and calculated from non-protein RQ, was measured in 12 healthy adults. Data is presented as mean ± SEM. * p ≤ 0.003 for a difference from the 0% and 10.7% RS meals. Similarly, the oxidation of [ 14 C]-triolein to 14 CO 2 was different between RS doses (F-test, 0.0005). Meal fat oxidation at the 5.4% RS dose was significantly higher than both the 0% (p = 0.0062) and 10.7% doses (p < 0.0001). Separate tests at 6 h or 24 h following the test meal gave comparable results (Figure 4a ). Taken together, these independent measurements of fat oxidation (indirect calorimetry, oxidation of [ 14 C]-triolein to 14 CO 2 ) suggest that the inclusion of 5.4% RS in the meal elevated postprandial fat oxidation. Unexpectedly, this effect was lost if the dose was increased to 10.7% RS. Figure 4 Meal fat oxidation (a) and storage (b) in response to RS content of a breakfast meal. Meal fat oxidation, assessed via measurement of 14 CO 2 in expired air, and meal fat storage in gluteal adipose tissue was measured in 12 healthy adults. Data is presented as mean ± SEM. * p ≤ 0.0006 for a difference from the 0% and 10.7% RS meals at the same time point. FFM, fat free mass. There was a trend for fat storage from the test meal, as assessed by incorporation of 14 C into gluteal adipose tissue, to be lower for the 5.4% RS meal than all other meals, although this effect did not reach statistical significance (Figure 4b ). Discussion This study demonstrated that the addition of RS to a mixed meal, balanced for total fat and fiber content, had no effect on postprandial glucose, insulin, FFA, or triglyceride excursions. However, meals containing a moderate amount of RS caused an increase in fat oxidation as measured by both indirect calorimetry and the production of 14 CO 2 from a 14 C-triglyceride tracer. Unexpectedly, the dose-response relationship between RS content of the diet and fat oxidation was not linear. Although this result is difficult to explain in the current context, it emphasizes the need for careful selection of RS dose in prospective feeding studies. There was no difference in postprandial glucose (Figure 1a ), FFA (Figure 1e ), triglyceride (Figure 1f ), or insulin (Figure 1c ) concentrations at any RS dose examined. This concurs with data from other acute human studies using complete, mixed meals which showed no difference in postprandial glycemia/insulinemia in response to RS content of the diet [ 8 - 11 ]. Although this seems contrary to the general perception that RS ingestion reduces postprandial insulinemia and glycemia, many of the studies indicating this did not balance test diets for total fat and/or fiber content [ 17 ]. However, in the current study all diets were carefully matched for total fat and fiber content. This an important distinction between this and other studies as fiber has extensively been shown to reduce postprandial glycemia/insulinemia and increasing the RS content of the diet intrinsically increases the total fiber content. Also, dietary fat can have potent effects on the accessibility of dietary carbohydrate to digestive enzymes and on the rate of gastric emptying/gut motility. Thus, the glucose- and insulin-lowering effects of RS that have been observed in other studies may be due to changes in fiber and/or fat between test meals which have been extensively shown to lower postprandial glycemic and insulinemic responses. So, the balanced conditions used in the meal tests for the study described herein, which included baked products and processed foods as part of a complete, mixed meal, balanced for total fat and fiber content, could account for the lack of difference in insulinemia and glycemia in response to increased RS content in the diet. Both indirect calorimetry and 14 C-tracer data indicate that there was an increase in fat oxidation between the 0% and 5.4% RS doses (Figures 2 , 3 , and 4a ). This increase in total and meal fat oxidation in response to the 5.4% RS meal is not driven by disparate responses amongst subjects as 11 of the 12 subjects studied showed the greatest fat oxidation in response to the 5.4% RS meal, relative to the 0% and 10.7% RS meals (see Figure S1, Additional File 1 , for individual responses). Tracer data showed that the addition of 5.4% of RS to a meal increased meal fat oxidation by more than 20% over the 6 h and 24 h post-meal ingestion period (Figure 4a ). The increase in fat oxidation at 6 h accounted for approximately one-half of the total increase over 24 h, indicating that the increase in meal fat oxidation in response to a single meal containing 5.4% RS is a prolonged, sustained effect. In addition, comparison of total and meal fat oxidation (Figures 3a and 4a ) indicates that endogenous fat stores were the predominant source of fat utilized for energy, contributing approximately 80% of the total fat oxidized, with a much lower contribution from ingested meal fat. Figure 3 shows that this increase in fat oxidation at the 5.4% RS dose is accompanied by a relative reduction in carbohydrate oxidation (does not reach statistical significance). The increase in fat oxidation at the 5.4% RS dose relative to the 0% dose was not driven by any disparity in circulating glucose, insulin or FFA concentration (Figure 1 ; see Figures S2, S3, S4, Additional Files 2 , 3 , 4 , respectively, for individual subject responses) nor by a difference in available carbohydrate between the 0% and 5.4% RS meals. If decreased carbohydrate availability was responsible for the observed increase in fat oxidation, the 10.7% RS meal, which has the least available carbohydrate, would show the greatest increase in fat oxidation. However, there was no difference in fat oxidation between the 0% and 10.7% RS meals. Thus, carbohydrate availability cannot be a contributing factor to the increase in fat oxidation observed at the 5.4% dose of RS. It is possible that this increase may be due to an increase in circulating SCFAs from the fermentation of RS reaching the large bowel. The observed increase in fat oxidation is not due to oxidation of these SCFAs per se as it was measured directly from conversion of 14 C-labeled meal fat to 14 CO 2 (Figure 3a ). Such a measurement would not detect any increase in SCFA oxidation. Rather, it may be that the metabolic effects of increased SCFA production cause an increase in fat oxidation. RS consumption has been shown to alter the acetate:butyrate:propionate ratio compared to fermentation of non-starch polysaccharides [ 29 ]. In particular, the amount of butyrate is substantially elevated in response to RS fermentation [ 30 , 31 ]. In humans fed a low or high RS diet for three days, the concentration of excreted SCFA rose from 20 mmol/d to 33 mmol/d, respectively [ 19 ]. This increase in total SCFA concentration was caused by a doubling of the acetate and butyrate content changing the acetate:butyrate:propionate ratio from 12:3:3 to 21:6:4 in response to the low and high RS diets, respectively. In vitro data from isolated animal tissues provide convincing evidence for the role of SCFAs in carbohydrate and lipid metabolism [ 26 , 32 - 34 ]. Acetate and/or butyrate have been shown to decrease glycogenolysis and glycolysis in isolated rat and sheep hepatocytes [ 35 - 37 ]. So, it is plausible that the fermentation of RS from the 5.4% RS diet increases the net production of SCFAs which inhibit glycolysis in the liver. In this scenario, the liver, deprived of carbohydrate-derived acetyl CoA would be more reliant on fat-derived acetyl CoA as a fuel source, thereby contributing to an overall increase in fat oxidation [ 17 ]. This possibility needs to be investigated in future studies. No difference in fat oxidation was evident between the maximal 10.7% dose of RS and the 0% dose. This is an unexpected result that is difficult to explain. The loss of any effect on fat oxidation when the RS dose in the meal was increased to 10.7% may occur because this dose is at the threshold of the starch's properties as RS. That is, at the 10.7% dose of RS, the starch may not be completely fermented in the large bowel thereby causing a loss of energy from the diet via the feces. If this is the case, the strong physical association between RS and dietary lipid may cause excretion of lipid and therefore, less dietary fat to be available for oxidation at the 10.7% dose. Indeed, it has previously been shown that intake of high-amylose maize starch, such as that used in this study, caused an increased number of bowel actions per day [ 18 ]. RS has also been shown to decrease colonic transit time and, as more RS enters the large bowel, more starch is also excreted [ 19 , 20 ]. This indicates that, at higher levels of RS consumption, only a portion of the RS can be fermented and the remainder passes through the colon as an insoluble fiber. Furthermore, if indeed RS at the 10.7% dose is being excreted as insoluble fiber, less fermentation and SCFA production would be occurring. As SCFA are hypothesized to be the cause of the observed increase in fat oxidation in response to the 5.4% RS meal, this would have a large impact on the fat oxidation potential of the 10.7% RS diet. The hypothesis that RS is acting like dietary fiber and being excreted can be tested by measuring the amount of fat excreted in the feces. As this outcome was not predicted, fecal samples were not collected from subjects during this study. It is important to consider that it is difficult to add 10.7% RS to a standard diet without the use of specially designed foods and/or without significantly increasing caloric intake. Therefore, this level would be difficult to attain in a free-living situation and the lower doses used in this study are more reflective of predicted levels if normal, starchy foods in the diet were to be replaced with commercially available RS products. In addition, not all biological processes display linear dose-response curves. Dose-response curves can vary from sigmoidal to 'U'-shaped curves for processes as diverse as drug absorption/clearance [ 21 ], low dose radiation effects on cells [ 22 ], DNA repair following double-strand breaks [ 23 ], and metabolic parameters. Metabolic processes that are non-linear functions include the level of illuminance and plasma melatonin levels [ 24 ], caffeine intake versus plasma caffeine metabolite concentrations [ 25 ], allergen exposure (concentration) and histamine response [ 26 ], zinc-stimulated histamine release from mast cells [ 27 ], and fructose-1,6-diphosphate metabolism in cardiomyocytes [ 28 ]. Thus, it is possible that the lack of any effect on fat oxidation at the 10.7% RS dose may indicate that the relationship between RS intake and fat oxidation is indeed a 'U'-shaped curve. However, more RS doses between 5.4% and 12% must be tested to accurately define the shape of this dose response curve. It must be noted that the calculation of oxidation of [ 14 C]-triolein via measurement of 14 CO 2 did not take into account the dilution of tracer in vivo due to the incorporation of labeled carbons into intermediates of the TCA cycle and endogenous bicarbonate pools. Generally, an acetate correction factor is used to account for this effect. In this study, subjects consumed all four test meals under the same conditions and it was assumed that there was no difference in tracer recovery between tests. Also, these TCA intermediate and bicarbonate pools were not pre-labeled prior to the ingestion of the label in the meal which would cause a total underestimation of total fat oxidation. Therefore, the rate of fat oxidation calculated from 14 CO 2 recovery in the breath was probably underestimated in all subjects but remains valid to compare differences between test meals. There was a trend towards a decrease in gluteal fat storage at the 5.4% RS dose relative to all other doses (Figure 4b ). Again, the dose-response curve for this parameter was not linear, lending credence to the idea that the dose-response curve for fat oxidation is actually U-shaped. Although the decrease in fat storage at the 5.4% RS dose did not reach statistical significance, it is intuitive that, given the magnitude of the increase in fat oxidation observed at this dose, there would be a reciprocal decrease in fat storage. However, there was high variability associated with the measure of meal fat storage indicating that more subjects may be needed to decrease the standard deviation and, hence, detect any significant meal affect. Conclusion This study is the first to identify that addition of 5.4% RS to a single meal can cause a significant increase in total and meal fat oxidation in healthy individuals relative to a 0% RS diet over the postprandial/postabsorptive period (24 h). This discovery was verified using two different methods, indirect calorimetry and the oxidation of [ 14 C]-triolein to 14 CO 2 , to measure in vivo fat oxidation. This increase in fat oxidation was accompanied by a concomitant decrease in carbohydrate oxidation and fat storage, although these parameters did not reach statistical significance. Further, the magnitude of the increase in fat oxidation indicates that this effect is biologically relevant and could be important for preventing fat accumulation in the long term by effecting total fat balance under chronic feeding conditions. Finally, this study revealed that there may be a maximal effect of RS addition to the diet and that the addition of RS over this threshold confers no metabolic benefit or change from a 0% RS meal. Methods Subjects 12 healthy adults, 7 male and 5 female, participated in the present study. This study was approved by the Colorado Multiple Institution Review Board, in compliance with the Helsinki Declaration, and full written consent was obtained from all subjects. To participate, subjects were required to be between 28 and 45 years of age, have normal glucose tolerance (as judged via response to an oral glucose tolerance test; fasting glucose concentration < 6 mM, postprandial glucose concentration not higher than 9 mM), moderate level of physical activity (no more than 4 one-hour bouts of planned physical activity per week), and a BMI between 20 and 28. All female subjects were taking oral contraceptive pills or progesterone injections and were tested during the early follicular phase of the menstrual cycle. All subjects underwent dual energy X-ray absorptiometry (DEXA; Lunar Radiation Corp, Madison WI) for analysis of body composition. As a group, subjects were 33 ± 5 years of age, 1.7 ± 0.07 m tall, weighed 75 ± 11 kg, had a BMI of 24.7 ± 2.4, total fat mass of 18.3 ± 5.0 kg (mean ± SD), and a fasting RQ of 0.750 ± 0.023 (mean ± SEM). Diet Subjects received four meals differing only in resistant starch (RS) content in random order, approximately four weeks apart. Test meals contained either 0%, 2.7%, 5.4%, or 10.7% RS as a percentage of total dietary carbohydrate. All added RS was in the form of high-amylose maize starch, or RS2. High-amylose maize starch was chosen as it has the unique property of a very high gelatinisation temperature which allows it to maintain its granular structure during and after the processing conditions used to manufacture the foods being consumed in this study [ 38 ]. All meals were isocaloric, accounting for 30% of the subject's daily energy needs as measured by indirect calorimetry prior to study commencement (RMR × daily activity factor of 1.49). The composition of the test diet was 55% carbohydrate, 15% protein, and 30% fat as a percentage of total energy (Table 1 ). All meals were matched for total dietary fiber content and liquid volume (250 ml). Table 1 Composition of test breakfasts. All values are based on a hypothetical subject who requires 8374 kJ (2000 kcal) per day. RS content (% total carbohydrate) 0 2.7 5.4 10.7 RS content (g) 1 0 g 2.5 g 5 g 10 g Total energy (kJ) 2508 2506 2500 2506 Carbohydrate (g) 93.8 93.3 92.9 93.0 Protein (g) 22.7 22.6 23.0 23.0 Fat (g) 17.0 16.8 16.9 16.9 Total sugars (g) 45.6 45.2 45.7 45.1 Total Fiber (g) 9.4 9.3 9.5 9.5 Liquid volume (mL) 250 250 250 250 Foods consumed (g) Canned spaghetti 197 58 *RS Canned spaghetti 147 218 216 Parmesan cheese 10 8 8 12 Margarine 4 3 2 2 Butter 2 1 1 Milk (2% fat) 250 250 210 *Up & Go breakfast drink 40 250 Bread 38 44 36 *Banana muffin 43 Strawberries 203 162 123 Grapes 80 93 *Fruit fingers 15 16 Sugar, white 10 * Denotes foods with added RS. 1 Absolute RS inclusion varied according to the energy needs of the subject so that RS content always remained the same fraction of total dietary carbohydrate, namely 0%, 2.7%, 5.4%, and 10.7% for the 0 g, 2.5 g, 5 g, and 10 g meals, respectively. For example, a subject who had a daily caloric need of 9421 kJ (2250 kcal) would receive meals containing 0 g, 2.7 g, 5.4 g, and 10.8 g RS. 2 Energy and macronutrient values were determined using the USDA Nutrient Database for standard foods and from information supplied by the manufacturer for foods with added RS. Note that energy values calculated from the carbohydrate, fat, and protein content of study foods using the 4-9-4 kcal/g factor method differ from reported energy values due to use of the Atwater system. Three days prior to each test day, subjects received a standardized lead-in diet, equivalent to daily energy needs as judged by indirect calorimetry and of the same macronutrient composition as the test diet with no added RS, to ensure that they were in energy balance. All food for these three days was provided by the General Clinical Research Center (GCRC) on an outpatient basis. Subjects were instructed to eat all of the food/drink provided and not to consume any other foods. Non-caloric beverages could be consumed during the three day lead-in diet. Protocol Following an overnight fast (12 h), subjects were admitted to the GCRC and an intravenous catheter was placed for the purposes of drawing blood. The test meal began at 0 min (0800 h) with all food/drink fully consumed within 15 min. Blood samples were taken at 0, 30, 60, 90, 120, 180, 240, 300, and 360 min following meal ingestion and analyzed for glucose, insulin, triacylglycerol (TAG) and free fatty acid (FFA) concentrations. Respiratory quotient (RQ) was measured at hourly intervals after ingestion of the meal via gas collection under a ventilated plexiglass hood for 15 min (Sensormedics 2900 metabolic cart). All urine produced between 0 and 360 min was collected and analyzed for nitrogen content by the GCRC Core Laboratory to facilitate calculation of non-protein RQ. In three of the test meals (0%, 5.4%, and 10.7% RS meals), the bread product in the test meal was spiked with 50 μCi [1- 14 C]-triolein (glycerol tri [1- 14 C]oleate; Amersham Pharmacia Biotech, Amersham, UK) suspended in olive oil and the tests were conducted as 24 h inpatient stays at the GCRC. The fat tracer was fed as a triglyceride (glycerol tri [1- 14 C]oleate) rather than a FFA (eg. [1- 14 C]oleate) in order to reflect any change in the absorption of triglyceride FFA which might be due to a strong physical association with RS thereby slowing absorption. At hourly intervals following the meal, then at 8, 10, 12, 14 and 24 hours, breath samples were collected via exhalation through a tube with a one-way valve into scintillation vials containing 2 mmol benzethonium hydroxide (to trap 2 mmol CO 2 ), 1 ml methanol, and 1 mg phenolpthalene as a pH indicator. Gluteal fat biopsies were collected by aspiration through a 14 g stainless steel needle at baseline and 24 h after ingestion of the test meal. All breath and fat samples were assayed for the presence of 14 C (as described below). For these 24 h tests, subjects received 30% of daily energy needs at each of breakfast, lunch, and dinner, with the remaining 10% of calories received in an evening snack. The timing of meals/snacks was kept constant over all tests. All food was provided by the GCRC on an inpatient basis and the macronutrient content of each meal was the same as that of the test meal. Only the test breakfast contained RS during these 24 h tests, all other meals were composed of standard, commercially available products. Analyses All glucose, FFA, and TAG assays were conducted by the GCRC Core Laboratory using an automated Cobas Mira Plus (Roche Diagnostics, Basel, Switzerland). Serum insulin measurements were also performed by the GCRC Core Laboratory using a human insulin RIA kit (Linco, St. Louis, USA). Fat samples, frozen in liquid nitrogen and stored at -80°C until processing, were incubated in 450 μl Solvable (Packard Bioscience, Groningen, Netherlands) at 50°C for 12 h before the addition of 100 μl 30% (v/v) hydrogen peroxide (for sample bleaching). Fat samples were counted in Aquasol (Packard Bioscience, Groningen, Netherlands) whereas breath samples were counted in Scintisafe 30% (Fisher Chemical, New Jersey) using a Beckman LS6500 scintillation counter (Beckman Instuments, Fullerton, CA). After scintillant was added, all samples were kept in the dark at room temperature for 48 h before being counted to reduce chemiluminescence. Calculations Calculation of total fat and carbohydrate oxidation Formulae used to calculate non protein RQ and subsequent estimations of carbohydrate and fat oxidation were based on the derivations described by Jéquier et al . ([ 39 ]). Calculation of ΔRQ ΔRQ = RQ t - RQ baseline where t is sample time (min). Calculation of meal fat storage from biopsy data μg fat stored/g fat tissue = (dpm 24h /g tissue weight) - (dpm baseline /g tissue weight) × 1/specific activity μg fat stored/whole body = μg fat stored/g fat tissue × total fat mass (from DEXA) Calculation of 14 C-triolein oxidation counts from sample (dpm/mol CO 2 )/vCO 2 (min.ml) = (dpm t - dpm background ) × 1/vCO 2 = dpm.mol CO 2 / min.ml dpm/min = dpm.mol CO 2 / min.ml × 0.446 (as 1 ml CO 2 = 0.446 mol) g fat oxidized = AUC(dpm/min) × 1/specific activity where vCO 2 is the rate of CO 2 production as assessed during indirect calorimetry. t is sample time (min). AUC is the incremental area under the curve. Analysis All statistical analyses were performed using the statistical analysis software SAS, version 8.1 (SAS OnlineDoc, 2000) with a significance level of p = 0.05 and p = 0.01 for interaction terms. All results are presented as mean ± SEM, except for subject characteristics which are described as mean ± SD. To investigate each of the outcomes (glucose, insulin, FFA, TAG, RQ, meal fat oxidation, and meal fat storage) we used a mixed model with fixed effect terms for RS DOSE, TIME and the interaction of the two, RS DOSE*TIME. Subjects were included as random effects. The interaction term was not significant for any of the outcomes tested so an additive model was used to test the overall effect of RS DOSE and the differences between doses. To test the effects of RS DOSE at different TIMES, a model that included RS DOSE, TIME and RS DOSE*TIME was used. The repeated measures nature of the study design was taken into account by using the covariance structures available in SAS PROC MIXED. For example, measurements within a subject are assumed to be more highly correlated than between subjects, and within a particular treatment, within a subject, the measurements are assumed to be more correlated. Measurements closer in time to one another were modeled with an autoregressive, or AR(1) covariance structure. Abbreviation List RS, resistant starch; DS, digestible starch; TAG, triacylglycerol; FFA, free fatty acid; FFM, fat free mass; SCFA, short-chain fatty acids; GCRC, General Clinical Research Center; RQ, respiratory quotient Competing interests Janine Higgins and Ian Brown are listed as inventors on RS patents filed by Penford Australia Limited. Both Drs. Higgins and Brown are listed as inventors on these patents as they have intellectual property ownership of some of data used in these but receive no financial benefit. Authors' Contributions JH conceived of the study design and was responsible for overall study coordination, conducting patient visits, data analysis, and manuscript preparation. DH was responsible for patient scheduling, day-to-day study coordination, conducting patient visits, and data entry. WD contributed to the study design and manuscript preparation, and conducted patient physical examinations and fat biopsies. IB contributed to the study design, selection of RS foods, and assisted with manuscript preparation. MB conducted all statistical analysis. DB contributed to the study design and manuscript preparation, and conducted patient visits, patient physical examinations and fat biopsies. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Individual meal (a) and total fat oxidation (b) in response to the RS content of a test breakfast. Meal fat oxidation, assessed via measurement of 14 CO 2 in expired air, and total fat oxidation, assessed via indirect calorimetry and calculated from non-protein RQ, and was measured in 12 healthy adults. Click here for file Additional File 2 Individual area under the glucose curve vs. meal (a) and total fat oxidation (b) in response to a test breakfast. Meal fat oxidation, assessed via measurement of 14 CO 2 in expired air, and total fat oxidation, assessed via indirect calorimetry and calculated from non-protein RQ, and was measured in 12 healthy adults. Data from all three test meals (0%, 5.4%, and 10.7% RS) is shown. The relationship between area under the glucose curve and fat oxidation remains the same (i.e. no relationship) when represented as individual doses or, as in this plot, for all doses (see Figure S3). Click here for file Additional File 3 Individual area under the glucose curve vs. meal fat oxidation in response to a 0% (a) or 5.4% (b) RS test breakfast. Meal fat oxidation, assessed via measurement of 14 CO 2 in expired air, and total fat oxidation, assessed via indirect calorimetry and calculated from non-protein RQ, and was measured in healthy adults. Data from individual test meals is shown. Click here for file Additional File 4 Individual area under the insulin curve vs. meal (a) and total fat oxidation (b) in response to a test breakfast. Meal fat oxidation, assessed via measurement of 14 CO 2 in expired air, and total fat oxidation, assessed via indirect calorimetry and calculated from non-protein RQ, and was measured in 12 healthy adults. Data from all three test meals (0%, 5.4%, and 10.7% RS) is shown. (Document type: Powerpoint, PPT) Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526391.xml

522875

Microarray analysis of microRNA expression in the developing mammalian brain

A microarray technology suitable for analyzing the expression of microRNAs and of other small RNAs was used to determine the microRNA expression profile during mouse-brain development and observed a temporal wave of gene expression of sequential classes of microRNAs.

Background MicroRNAs constitute a large class of small regulatory RNAs [ 1 ]. Their mechanism of action and the scope of their biological roles are beginning to be understood. The first two microRNAs were discovered as the products of heterochronic genes that control developmental timing in Caenorhabditis elegans [ 2 ]. In heterochronic mutants, the timing of specific developmental events in several tissues is altered relative to the timing of events in other tissues. These defects result from temporal transformations in the fates of specific cells; that is, certain cells acquire fates normally expressed by cells at other developmental stages. The molecular characterization of the heterochronic gene lin-4 led to the surprising discovery that this gene encodes a 21-nucleotide non-coding RNA that regulates the translation of lin-14 mRNA through base-pairing with the lin-14 3' UTR [ 3 , 4 ]. A second heterochronic gene, let-7 , encodes another small non-coding RNA that is conserved in flies and mammals [ 5 ]. Biochemical and bioinformatic approaches have identified many genes that encode microRNAs in C. elegans , plants, Drosophila melanogaster and mammals [ 6 - 18 ]. Like the lin-4 and let-7 genes, other microRNAs encode 21-25-nucleotide RNAs derived from longer transcripts that are predicted to form stem-loop structures. More than 200 microRNAs are encoded by the human genome [ 8 , 14 ]. The biological roles of microRNAs are poorly understood. In C. elegans , lin-4 and let-7 act in developmental timing, and the microRNA lsy-6 controls neuronal asymmetry [ 19 ]. In Drosophila , the microRNAs bantam and mir-14 act in the regulation of cell growth and cell death [ 20 , 21 ]. The mouse microRNA miR-181 has been implicated in the modulation of hematopoietic differentiation, and other mammalian microRNAs have been suggested to play roles in cancer [ 22 , 23 ]. Mature microRNAs are excised from a stem-loop precursor that itself can be transcribed as part of a longer primary RNA (pri-miRNA) [ 24 ]. The pri-miRNA appears to be processed by the RNAse Drosha in the nucleus, cleaving the RNA at the base of the stem-loop [ 25 ]. This cut defines one end of the microRNA. The precursor microRNA is then exported by Ran-GTP and Exportin-5 to the cytoplasm, where it is further processed by the RNAse Dicer [ 26 , 27 ]. Dicer recognizes the stem portion of the microRNA and cleaves both strands about 22 nucleotides from the base of the stem [ 25 ]. The two strands in the resulting double-stranded (ds) RNA are differentially stable, and the mature microRNA resides on the strand that is more stable [ 28 , 29 ]. Mature microRNAs can be found associated with the proteins eIF2C2 (an Argonaute-like protein), Gemin2 and Gemin3 and are thought to act in a protein-RNA complex with these and maybe other proteins [ 17 , 30 ]. The animal microRNAs studied so far act by reducing the levels of proteins from genes that encode mRNAs with sites partially complementary to microRNAs in their 3' UTRs [ 4 , 31 ]. The mechanism responsible is not understood in detail [ 32 ]. In contrast, although some plant microRNAs with partially complementary target sites also act by preventing translation, the majority studied so far cause the cleavage of target mRNAs at sites perfectly complementary to the microRNAs [ 33 - 38 ]. Determining spatial and temporal patterns of microRNA expression should yield insight into the biological functions of microRNAs. As the number of microRNAs identified has increased rapidly, the need for a method that allows for the parallel detection of microRNA expression has become apparent. Recent studies used a dot-blot technique to study 44 mouse microRNAs and northern blotting analysis to study 119 microRNAs from mouse and human organs [ 39 , 40 ]. In this study we cloned microRNAs from rat and monkey brains, developed a microRNA labeling method and used a microarray to monitor expression of microRNAs during mouse brain development. We determined the temporal expression pattern of 138 microRNAs in the mouse brain and found that the levels of 66 microRNAs changed significantly during development. We identified sets of genes with similar expression patterns, including genes that peaked in expression at different stages of development. More generally, the microRNA microarray we have developed can be used to determine the expression of all known microRNAs simultaneously under any set of experimental conditions or constraints. Results and discussion Identification of microRNAs from developing rat and monkey brains To analyze microRNAs expressed in the developing mammalian brain, we cloned small 18-26-nucleotide RNAs from the neocortex and hippocampus of a 12-day postnatal rat ( Rattus norvegicus ) and from the cerebral wall of a 114-day-old fetal rhesus monkey ( Macaca mulatta ) (Table 1 ). In both species, by these stages most neurons have been generated and have begun synaptogenesis [ 41 , 42 ]. We identified a total of 1,451 sequences, 413 of which correspond to microRNA sequences on the basis of their potential to generate stem-loop precursors as predicted from corresponding sequences in the rat and/or human genomes. In all cases but one, the microRNAs we identified corresponded to known microRNAs from other species and defined 68 unique microRNAs (Table 1 and Additional data file 1). One of these microRNAs is new: it differs in sequence from any microRNA previously described and is conserved in the mouse and human genomes. We named this new microRNA rno-miR-421 (Figure 1 and Additional data file 2). As observed in similar studies, in addition to microRNAs a number of candidate small RNAs that do not fulfill all criteria of a microRNA were cloned (Additional data file 3) [ 9 , 43 ]. Of the 52 rat microRNA sequences we cloned, 27 had previously been cloned from rat primary cortical neurons [ 11 ]. For 21 of the 52 microRNAs from rat and 14 of the 40 microRNAs from monkey we isolated only a single clone, indicating that our surveys are not saturated. By contrast, we isolated microRNA miR-124a 19 times from rat and 97 times from monkey. Mouse miR-124a as well as miR-128, miR-101 and miR-132 have been reported to be expressed specifically in brain [ 15 ]. We found that rat miR-138 also was expressed only in brain (Additional data file 4). MicroRNA microarrays for the study of temporal and spatial patterns of microRNA expression Previous analyses of microRNA expression have relied on dot blots, northern blots and cloning strategies [ 8 , 11 - 14 , 18 , 39 , 40 ]. A highly scalable approach using a microarray would facilitate the analysis of microRNA expression patterns for a large number of samples and is feasible now that many mammalian microRNAs have been identified. We arrayed 138 oligonucleotides complementary to microRNAs (probes) corresponding to the 68 mammalian microRNAs we isolated from rat and monkey brains, to 70 mammalian microRNAs isolated by others from a variety of mouse tissues and mammalian cell lines, and to predicted microRNAs. In addition, we included a set of control probes as well as 19 probes corresponding to presumptive small RNAs that we and others identified but that do not satisfy all the criteria for a microRNA (see below and Additional data file 5). Probes had a free amine group at the 5' terminus and were printed onto amine-binding glass slides and covalently linked to the glass surface. All probes were printed in quadruplicate (Additional data file 5). We developed a method for preparing microRNA samples for microarray analysis. Several methods for mRNA sample labeling for microarray analysis have been described [ 44 - 47 ], but none is suitable for labeling RNAs as small as microRNAs. To fluorescently label small RNAs we adapted strategies for RNA ligation and reverse transcription PCR (RT-PCR) devised for microRNA cloning [ 12 - 14 ]. Briefly, 18-26-nucleotide RNAs were size-selected from total RNA using denaturing polyacrylamide gel electrophoresis (PAGE), oligonucleotide linkers were attached to the 5' and 3' ends of the small RNAs and the resulting ligation products were used as templates for an RT-PCR reaction with 10 cycles of amplification. The sense-strand PCR primer had a Cy3 fluorophore attached to its 5' end, thereby fluorescently labeling the sense strand of the PCR product. The PCR product was denatured and then hybridized to the microarray. As in microarray analysis, the labeled sample used for hybridization is referred to as the target. Significant biases in amplification, a problem when amplifying heterogeneously sized mRNAs, are less likely in the case of microRNAs because of their short uniform lengths. MicroRNA cloning frequencies obtained using a similar amplification strategy correlate well with expression levels as assayed by quantitative northern blots [ 7 ]. Because RNA is amplified before hybridization, relatively low amounts of starting material may be used with this method [ 8 , 11 - 14 , 18 , 39 , 40 ]. We optimized the conditions for hybridization to our microarray. The small sizes of microRNAs leave little opportunity for oligonucleotide (array probe) design to achieve homogeneous probe-target melting temperatures. Differences in melting temperatures are expected to result in greater nonspecific binding if hybridizations are performed at low temperatures (to allow the detection of probe-target pairs with low melting temperatures) and in less specific binding if hybridizations are performed at high temperatures (to specifically detect probe-target pairs with high melting temperatures). To assess this issue we included control probes with two internal mismatches on the microarray for a subset of the microRNA probes (Additional data file 5). We tested a range of hybridization temperatures, and, on the basis of the signal of microRNA probes versus control probes, we determined that a hybridization temperature of 50°C was a reasonable compromise between sensitivity and specificity (data not shown). Even at 50°C, specificity as assayed by comparing microarray spot signal intensities from matched and mismatched probes varied among the microRNAs assayed. As expected, specificity at 50°C was negatively correlated with calculated melting temperatures (Figure 2a ). In all cases the cumulative signal from 10 hybridizations for the mismatched probe was equal to or lower than that for the microRNA probe, but differences in the ratio of the matched to mismatched probe signal ranged widely (Figure 2a ). Given these data, we do not expect the microRNA microarray to distinguish reliably between microRNAs that have only one or a few mismatches. This limitation is alleviated somewhat by the fact that for most microRNAs that have been identified the most closely related paralogs differ by five mismatches or more (Figure 2b ). The signal from a mismatched control probe is likely to be caused by cross-hybridization with the microRNA for which it was designed, as other control probes corresponding to unrelated mRNA subsequences or synthetic probes that do not correspond to known microRNAs did not show signals above background (Additional data file 5). Microarray results for closely related microRNAs should be interpreted with caution, as differences in the apparent expression of a given microRNA could be dampened or exaggerated depending on the expression of the paralogs (Figure 2a ). To determine the detection range of the microarray, we synthesized three artificial RNAs with the characteristics of microRNAs. These RNAs were phosphorylated RNA oligonucleotides of 20-23 bases; their sequences were chosen at random and were without any significant sequence similarity to known mammalian microRNAs (see Additional data file 5 for details). We titrated these RNAs into total mouse RNA samples, labeled them and hybridized them to a microarray that in addition to microRNA probes included probes corresponding to these three RNAs, called syn1, syn2 and syn3. Figure 2c shows the correlation between the amount of the RNAs and the microarray signal intensities. For comparison, the background signal for the array is also shown. All three RNAs were reliably detected at levels as low as 0.1 fmoles. The dynamic range of the array was from 0.1 fmoles to at least 10 fmoles, or two orders of magnitude. Analysis of microRNA expression during mouse brain development We isolated small RNAs from mice at five developmental stages: embryonic days 12.5 and 17.5 (E12.5 and E17.5), postnatal days 4 and 18 (P4 and P18) and 4-month-old adults. E12.5-E17.5 spans a period of major neuronal proliferation and migration in the mouse brain, in particular the birth and subsequent migration of most neurons in the ventricular zone epithelium [ 48 ]. Between postnatal days P4 and P18, major sensory inputs are established. For example, eye opening occurs around P13 and is thought to result in activity-dependent neuronal remodeling [ 49 ]. We purified and size-selected RNA from whole mouse brains. For each sample, the products of four independent RNA amplifications based on two independent RNA ligations were hybridized to the array. A detailed description of our analysis of the microarray data is presented in Additional data file 5. Of the 138 microRNAs and 19 small RNAs represented by the probe set, 116 (74%) were expressed robustly (more than 75-fold over the level of background controls) at least at one time point. Of these, 83 (71%) changed significantly during the period surveyed (analysis of variance, ANOVA, p < 0.001) and 66 (57%) changed more than twofold. Of the microRNAs we cloned from rat and monkey and for which probes against the corresponding mouse homologs were present on the microarray, we detected 97% robustly. We grouped microRNAs that changed more than twofold in expression during the period analyzed using a hierarchical clustering algorithm (Figures 3a , 4 ) [ 50 ]. A group of microRNAs peaked at each of the developmental time points. The signal from 34 of the 66 probes that changed more than twofold peaked in the fetus (E12.5 and E17.5), suggesting roles in early development (Figure 4a ). Nine and eleven microRNAs peaked during the neonate (P4) and juvenile (P18) stages, respectively. Twelve microRNAs had the highest signals at the adult stage (Figure 4b ). These data indicate that murine brain development involves a wave of expression of sequential classes of microRNAs (Figure 3a ). We also grouped the developmental time points according to their microRNA expression pattern using hierarchical clustering. We found that samples from stages that are developmentally proximal had the most similar microRNA expression patterns (Figure 3b ), indicating that a microRNA expression profile can be a marker of developmental stage. Examination of the temporal clusters revealed that probes with similar sequences showed correlated expression, as exemplified by miR-181a, miR-181b, miR-181c, smallRNA-12 (Figure 4a ) and miR-29a, miR-29b and miR-29c (Figure 4b ), respectively. Given our observation that the microRNA microarray can detect mismatched sequences, it is possible that this correlation among closely related family members is an artifact of hybridization. We found that four of the 66 RNAs that changed more than twofold were small RNAs rather than microRNAs. The temporal regulation of these small RNAs indicates that they may play a role during development. Several mouse microRNAs are clustered closely in the genome, suggesting that they might be expressed from a single precursor transcript or at least share promoter/enhancer elements. We searched all known microRNA clusters in the mouse genome to attempt to identify coordinately controlled clustered microRNAs. We sought clusters with the following features: first, the clustered microRNAs are not all members of the same family; second, the microRNAs have no or few paralogs; and third, the microRNAs are detected robustly on our microarray and their expression changes significantly during the timecourse studied. The mir-17 cluster on chromosome 14 fulfills all these criteria. Figure 4c shows that the expression of all six microRNAs in this cluster is indeed highly co-regulated. Validation of microarray results using northern blots To validate our microarry results, we performed northern blots of eight microRNAs that were robustly expressed at least at one point during development according to our microarray data. The relative changes of microRNA expression assayed using microarray analysis and northern blots were consistent (Figure 5 ). For example, on a northern blot miR-29b was almost undetectable at the embryonic and P4 stages but appeared at P18 and was strongly expressed in the adult. The microarray data showed a similar pattern. In only a few cases did there seem to be discrepancies; for example, relative levels of expression of miR-138 at P4 compared to adult differed between the northern blots and the microarrays. As is the case for mRNAs, small differences may be seen between the methods and northern blot analysis is superior to microarrays for quantitative analysis [ 51 ]. Nonetheless, microarrays offer a high-throughput method that generally captures changes in microRNA expression. Conclusions Here we describe the development of a microarray technology for profiling the expression of microRNAs and other small RNAs and apply this technology to the developing mammalian brain. Recently, Krichevsky et al. described the temporal expression of 44 microRNAs during mouse brain development [ 39 ]. Their study used a dot-blot array approach and direct labeling of microRNAs using radioactivity instead of a glass microarray and RT-PCR/fluorescent labeling, as we used in our study. Despite differences in sample selection as well as in the number of microRNAs analyzed, there is good agreement between the overlapping aspects of the two datasets. Our strategy has the potential to be highly scalable, allowing high-throughput analysis of samples with limiting starting material. MicroRNA microarrays offer a new tool that should facilitate studies of the biological roles of microRNAs. We speculate that some of the developmentally regulated microRNAs we describe in this report play roles in the control of mammalian brain development, possibly by controlling developmental timing, by analogy to the roles of the lin-4 and let-7 microRNAs in C. elegans . Materials and methods MicroRNA cloning We isolated RNAs and cloned microRNAs from R. norvegicus and M. mulatta using methods described previously [ 13 ], except that the samples were not dephosphorylated during the cloning procedure. Microarray printing and hybridization Microarray probes were oligonucleotides (named EAM followed by a number) with sequences complementary to microRNAs. Each probe was modified with a free amino group linked to its 5' terminus through a 6-carbon spacer (IDT) and was printed onto amine-binding slides (CodeLink, Amersham Biosciences). Control probes contained two internal mismatches resulting in either C-to-G or T-to-A changes (Additional data file 6). Printing and hybridization were done using the protocols from the slide manufacturer with the following modifications: the oligonucleotide concentration for printing was 20 μM in 150 mM sodium phosphate pH 8.5, and hybridization was at 50°C for 6 h. Printing was done using a MicroGrid TAS II arrayer (BioRobotics) at 50% humidity. Sample and probe preparation Whole brains from three to eight C57BL/6 mice were pooled. Starting with 250 μg of total RNA for each time point, 18-26-nucleotide RNA was purified on denaturing PAGE gels. The samples were divided, and the following cloning steps were done independently twice for each time point. 3' and 5' adaptor oligonucleotides were ligated to 18-26-nucleotide RNA followed by reverse transcription, essentially as described for microRNA cloning [ 13 ]. Briefly, a RNA-DNA hybrid 5'-pUUUaaccgcgaattccagt-idT-3' (Dharmacon: X, RNA; x, DNA; p, phosphate; idT, inverted [3'-3' bond] deoxythymidine) was ligated to the 3' end and 5'-acggaattcctcactAAA-3' (Dharmacon) was ligated to the 5' end. The ligation products were divided into two aliquots, and the following steps were done independently twice for each time point. Ligation products were reverse transcribed and amplified by 10 rounds of PCR (40 sec at 94°C, 30 sec at 50°C, 30 sec at 72°C). For PCR, the oligonucleotides used were: oligo1 5'-Cy3-ACGGAATTCCTCACTAAA-3' and oligo2 5'-TACTGGAATTCGCGGTTAA-3'. The PCR product was precipitated, washed and resuspended in hybridization buffer (5× SSC, 0.1% SDS, 0.1 mg/ml sheared denatured salmon sperm DNA). Data acquisition and analysis Microarray slides were scanned using an arrayWoRx biochip reader (Applied Precision), and primary data were analyzed using the Digital Genome System suite (MolecularWare) and Spotfire DecisionSite (Spotfire). Cluster analysis was performed using the CLUSTER/TreeView software [ 50 ]. For details concerning microarray data analysis see Additional data file 5. The predicted stem-loop RNA structures were generated using the mfold (version 3.1) software [ 52 ]. Northern blots Northern blots were performed as described [ 14 ]. Twenty micrograms of total RNA were loaded per lane. A probe for the mouse U6 snRNA (5'-TGTGCTGCCGAAGCGAGCAC-3') was used as loading control. The probes for the northern blots had the same sequences as the corresponding EAM oligonucleotides printed on the microarray (see Additional data file 6). The blots were stripped by boiling for 5 min in distilled water and reprobed up to four times. The probes used were: EAM119 (miR-29b), EAM125 (miR-138), EAM224 (miR-17-5p), EAM234 (miR-199a), EAM131 (miR-92), EAM109 (miR-7), EAM150 (miR-9) and EAM103 (miR-124a). Additional data files The following additional data files are available with the online version of this article. A file (Additional data file 1 ) with details of rat microRNA precursors: using the assembly of the rat genome [ 55 ] we identified candidate genomic locations for all of our rat microRNAs that have orthologs in the mouse but that have not been described previously for the rat. An alignment of the top BLAST hit of each mouse microRNA precursor sequence (The miRNA Registry, Release 3.2) against the rat genome sequence (public release draft genome assembly, version 3.1) is shown in this file. In addition, predicted precursor secondary structures are shown for each rat microRNA gene for which the precursor sequence differs from that of the corresponding mouse precursor. We used the mfold algorithm to make secondary structure predictions [ 52 ]. A file (Additional data file 2 ) containing precursor sequences and secondary structure predictions for the novel microRNA miR-421. An alignment of the predicted precursor sequences from human and mouse of the novel microRNA miR-421, which we identified from rat, is shown. The cloned sequence (corresponding to the mature microRNA) is shown in bold. The single mismatch is indicated by an asterisk (*). The accession number for each sequence is given. The predicted secondary structures for the corresponding genes from mouse and human are also shown. The human and mouse genomic sequences for candidate miR-421 precursors are identical. The sequence of the mature microRNA is colored in red. The residue in the rat genomic sequence that is different from the mouse and human genomic sequences is indicated. We used the mfold algorithm to make secondary structure predictions [ 52 ]. We have not been able to detect miR-421 in mouse brain using northern blots. A file (Additional data file 3 ) with details of other small RNAs cloned from rat and monkey brains. We cloned 13 small RNAs that do not satisfy all criteria to be considered microRNAs. One, small RNA-1 from monkey, is present in the mouse genome and has a predicted stem-loop precursor sequence characteristic of microRNAs. However, the predicted stem-loop ends on the final base of the microRNA, which is not typical for microRNAs. To be conservative, we refer to this RNA as a small non-coding RNA. A northern blot for smallRNA-1 revealed a high molecular weight band that may represent a precursor RNA (data not shown). Since there is no perfect match to smallRNA-1 in the current release of the human sequence, a presumptive precursor based on mouse genomic sequence is shown. The cloned sequence is highlighted in red. The other small RNAs are not predicted to form stem-loop structures. A file (Additional data file 4 ) showing rno-miR-138 brain specific expression. A northern blot shows that rno-miR-138 expression was restricted to brain. The probe was identical to EAM125. Total RNA isolated from various adult rat tissues (Ambion) was size-separated on denaturing PAGE (12 μg per lane), transferred to a nylon membrane and used for hybridization. Equal loading was verified using a probe for U6 snRNA (data not shown). A file (Additional data file 5 ) with details of microarray design and data analysis. And a file (Additional data file 6 ) with a summary of the microRNA microarray data. Oligonucleotide sequences correspond to probes on the array. MicroRNA names were obtained from the miRNA registry (Release 3.2) or if not available these were obtained from NCBI. Probes named smallRNA-1 through -13 correspond to unique small RNAs that we cloned but that did not correspond to known microRNAs and did not have perfect matches in the current release of the rat genome sequence. Column A indicates whether the probe is complementary to a microRNA or to one of the small RNAs we cloned ("-" = no, "+" = yes). Column B indicates whether the probe is complementary to a mouse microRNA. Oligonucleotides with a "-" in column A were either controls or sequences that were submitted to public databases as microRNAs and later found not to encode microRNAs. In a few cases we printed probes that represented the same microRNA twice, but we analyzed the data from only one of these probes. The probes we did not analyze had no labels in columns A, B, and C. Melting temperatures were calculated using the nearest neighbors method [ 53 ]. Data for the five time points of mouse brain development (E12.5, E17.5, P4, P18 and adult) are shown. Microarray data were derived as described (Additional data file 5 ). Briefly, data correspond to mean spot intensities averaged over quadruplicates (on each array) and four independent hybridizations. SEM refers to the standard error of the mean. microRNAs labeled with % are not in the current release of the miRNA registry but are deposited in NCBI and described elsewhere [ 8 , 15 , 18 ]. An Excel file (Additional data file 7 ) with the primary microarray data corresponding to the E12.5 time point; an Excel file (Additional data file 8 ) with the primary microarray data corresponding to the E17.5 time point; an Excel file (Additional data file 9 ) with the primary microarray data corresponding to the P4 time point; an Excel file (Additional data file 10 ) with the primary microarray data corresponding to the P18 time point; an Excel file (Additional data file 11 ) with the primary microarray data corresponding to the Adult time point. All data in Additional files 7-11 were exported from the Digital Genome System suite (MolecularWare). The primary microarray data will also be submitted to the Gene Expression Omnibus (GEO) database [ 56 ]. Supplementary Material Additional data file 1 A file with details of rat microRNA precursors Click here for additional data file Additional data file 2 A file containing precursor sequences and secondary structure predictions for the novel microRNA miR-421 Click here for additional data file Additional data file 3 A file with details of other small RNAs cloned from rat and monkey brains Click here for additional data file Additional data file 4 A file showing rno-miR-138 brain specific expression Click here for additional data file Additional data file 5 A file with details of microarray design and data analysis Click here for additional data file Additional data file 6 A file with a summary of the microRNA microarray data Click here for additional data file Additional data file 7 An Excel file with the primary microarray data corresponding to the E12.5 time point Click here for additional data file Additional data file 8 An Excel file with the primary microarray data corresponding to the E17.5 time point Click here for additional data file Additional data file 9 an Excel file with the primary microarray data corresponding to the P4 time point Click here for additional data file Additional data file 10 An Excel file with the primary microarray data corresponding to the P18 time point Click here for additional data file Additional data file 11 An Excel file with the primary microarray data corresponding to the Adult time point Click here for additional data file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC522875.xml

517830

The Genome of a Methane-Loving Bacterium

Mention greenhouse gases to most people and they're apt to think of carbon dioxide, fossil fuels, and big cars. Though carbon dioxide emissions are the major source of greenhouse gases, methane is far more effective at trapping heat in the atmosphere. Like increasing carbon dioxide levels, rising levels of atmospheric methane have been attributed to human activity, mostly in the form of landfills, natural gas and oil processing (about 60%), domesticated livestock (cattle account for about 75% of livestock contributions), and rice fields (up to 29% of total emissions). Methylococcus capsulatus cultured in the presence of a high concentration of copper (Image: Anne Fjellbirkeland) Ruminants—from cows and water buffalo to llamas and vicunas—emit methane gas as a natural by-product of their digestive process, which confers a unique ability to digest cellulose. Ruminants don't digest cellulose directly, but depend on a variety of microbes living in their rumen (main stomach) to do it for them. These microbes ferment cellulose, breaking it down into products the ruminant can digest. During this process, some microbes—bacteria called methanogens—produce methane, which ruminants expel by eructation (otherwise known as belching). Luckily, there are microbes, called methanotrophs, that consume methane. A type of aerobic bacteria, methanotrophs oxidize methane as an energy and carbon source using the enzyme methane monooxygenase. They've been found in soils, landfills, sediments, hotsprings, and peat bogs, among other environments. Methanotrophs have been the subject of increasing interest because they can use methane as a sole source of carbon and energy—which means they play a major role in global carbon cycles—and could dramatically reduce biologically generated methane emissions. They've also been the focus of bioremediation efforts aimed at environmental decontamination. And now, with the publication of the first complete genome sequence of a methanotroph, such efforts will be all the easier. In this issue of PLoS Biology, a multidisciplinary team spanning the fields of genomics, bioinformatics, microbiology, evolutionary biology, and molecular biology report the complete genome sequence of Methylococcus capsulatus and shed light on the metabolism and biology of this ubiquitous microbe. Contained in a single, circular molecule, the M. capsulatus genome comprises about 3.3 million base pairs—which is about average for a free-living bacterium—with an estimated 3,120 genes. The genome appears well-equipped to meet the specialized needs of this methanotroph, with what appear to be multiple pathways involved in the metabolism of methane and duplications of genes that code for methane monooxygenases, which are essential for the first step of methane oxidation. Ward et al. also found evidence of “genomic redundancy” in methane oxidation pathways, suggesting that M. capsulatus employs different pathways depending on the availability of molecules needed to sustain cellular activities. Most surprising, the researchers note, was evidence that this methane specialist can turn into a sort of metabolic generalist—with a capacity to use sugars, hydrogen, and sulfur—and appears able to survive reduced oxygen levels. These genome-based hypotheses will require experimental validation, the authors note, but could have important implications for M. capsulatus ecology—including what environments might be amenable to methanotroph-mediated bioremediation. The genomes of important microbial players in the carbon cycle—including microbes involved in photosynthesis and methanogenesis (methane production)—have already been sequenced. With the addition of a sequenced methanotroph genome, scientists can systematically investigate different genes and regulatory elements to better understand how these methane consumers fit into this global cycle. The M. capsulatus genome provides a platform for investigating the details of methanotroph biology and its potential as a biotech workhorse. It may also guide efforts to harness this bacterium's penchant for methane to reduce global greenhouse gas emissions, to degrade chlorinated hydrocarbons and other persistent pollutants, and to produce protein for animal feed.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC517830.xml

539273

Teaching statistics to medical students using problem-based learning: the Australian experience

Background Problem-based learning (PBL) is gaining popularity as a teaching method in UK medical schools, but statistics and research methods are not being included in this teaching. There are great disadvantages in omitting statistics and research methods from the main teaching. PBL is well established in Australian medical schools. The Australian experience in teaching statistics and research methods in curricula based on problem-based learning may provide guidance for other countries, such as the UK, where this method is being introduced. Methods All Australian medical schools using PBL were visited, with two exceptions. Teachers of statistics and medical education specialists were interviewed. For schools which were not visited, information was obtained by email. Results No Australian medical school taught statistics and research methods in a totally integrated way, as part of general PBL teaching. In some schools, statistical material was integrated but taught separately, using different tutors. In one school, PBL was used only for 'public health' related subjects. In some, a parallel course using more traditional techniques was given alongside the PBL teaching of other material. This model was less successful than the others. Conclusions There are several difficulties in implementing an integrated approach. However, not integrating is detrimental to statistics and research methods teaching, which is of particular concern in the age of evidence-based medicine. Some possible ways forward are suggested.

Background Problem-based learning (PBL) is a method of teaching and learning which is widely used in the education of medical students [ 1 , 2 ]. In problem-based learning, students working in a small group are presented with a problem, typically a description of a patient presentation. They decide what features of the problem are outside their present knowledge and divide these topics between them. They then research their topics using library and internet material and report back to the next small group tutorial with their findings. In a problem-based learning curriculum, this is the principal method of learning. More traditional methods, such as lectures and practical exercises, provide background and support material. An example of a PBL teaching problem is given in the Appendix. This project was conceived at a meeting of statisticians from UK medical schools, which is held every year at the Burwalls conference centre, University of Bristol. In 2002, one of the topics for discussion was problem-based learning (PBL). Several things became apparent during this discussion. First, most of the statisticians present had no idea what problem-based learning was. Second, in the few UK medical schools where PBL was in use, statistics and research methods were not being taught through this medium. Statisticians in some of these schools felt excluded. Third, the new medical schools which were starting up in the UK were adopting PBL as their method of teaching. As far as I know, only four of the older UK medical schools have adopted a PBL curriculum at the time of writing: Liverpool, Manchester, Glasgow, and St. George's. St. George's is a special case, as it has two medical courses. The five-year undergraduate entry course is taught in the traditional way, with a small 'case-based learning' element added on. The new four-year Graduate Entry Programme (GEP) is based on the McMaster model, a PBL course which takes graduates in any subject [ 2 ]. In none of these courses are statistics and research methods taught as an integral part of the PBL. At St. George's, for example, there is a parallel, non-PBL, seminar-based course. This attempts to match the illustrative material to the case of the week, but is taught separately from that case. New medical schools, such as Anglia, Peninsular, and Hull-York, are being set up as PBL courses [ 3 ]. This is particularly suitable for courses based in more than one university, as in Peninsular (Plymouth and Exeter) and Hull-York. Problems can be set by teachers in either centre, and presented to students in both. The adoption of PBL in new medical schools may indicate the possibility that PBL will become more widely adopted in UK medical schools. If this were to happen, there would be a real danger of statistics and research methods being marginalised in the medical curriculum. It is difficult enough to persuade medical students of the importance of these topics and it would be even more so if it were taught outside the mainstream of the course. More importantly, in the era of evidence-based medicine these topics should be central to the medical curriculum, not on the sidelines. Methods During a sabbatical visit to Australia, I spent one to three weeks each at most of the medical schools. I interviewed educationalists and statisticians at each of these to ask how statistics and research methods were taught to medical students in their problem-based learning curricula. I had no prior knowledge of how this might be done, so the interviews were open in nature, rather than structured. I enquired about statistics as a subject in its own right and the wider principles of research, critical reading of research, and evidence-based medicine. I then synthesised the information collected to provide a picture of the current situation. I visited the following universities with medical schools: the University of Western Australia, Perth, Flinders University, Adelaide, Monash University, Melbourne, the University of Melbourne, the Australian National University, Canberra (medical school about to commence), the University of Sydney, the University of Newcastle, and the University of Queensland, Brisbane. I omitted the University of Adelaide, the University of New South Wales, and James Cook University, Townsville, but was able to get information from them by email. I identified potential informants from the university website and emailed them as follows: 'I am interested in the teaching of statistics to medical students. Last year I visited several medical schools in Australia, but unfortunately I did not manage to visit [your university] on this trip, a serious omission. I have written a report on my experiences, which is available on one of my websites at . I am particularly interested in how statistics is taught in problem-based learning programmes. I am presenting this material at the forthcoming International Biometrics Conference in Cairns and preparing a paper for possible publication. I would be very interested to learn how these matters are ordered in [your university] . . . Are you the right person to ask? If not, could you suggest someone who would be?' I obtained helpful replies from all three universities. Results I went to Australia in search of the fully integrated teaching of statistics and research methods as part of the PBL tutorials. I did not find it anywhere. I did find three different models: • Material integrated but separately taught, • A parallel course, • PBL used only for 'public health' related subjects. In addition, one school (New South Wales) did not use problem-based learning and one (James Cook) taught virtually no statistics. More information about the individual medical schools is available [see Additional file 1 ]. Material integrated but separately taught This approach was used or planned at the University of Sydney, the University of Melbourne, and the Australian National University. However, of these only the University of Sydney had actually put this into practice. The University of Melbourne and the Australian National University were about to implement what was essentially the Sydney model. In this model, statistics and research methods are taught by PBL and the PBL triggers are integrated with the PBL problems for other parts of the course, but the material is not taught in the same tutorials or by the same tutors as anatomy, biochemistry and physiology. There will be a separate set of triggers for the statistics, etc., presented in a separate tutorial, and by separate tutors. I asked why the main PBL tutors could not do this. As I understood it, the function of a PBL tutor is to facilitate and guide the group, not to impart knowledge. I had no problem, at least that I was aware of, in acting as PBL tutor when students were working with triggers designed to elicit questions about anatomy, biochemistry, and physiology, subjects of which I know virtually nothing. Besides, many of these tutors must routinely read journals which bristle with P values, t tests, correlation coefficients, etc. They must be familiar with the terms, if nothing else. Answers to this included: • the tutors themselves refused to do it, • in the early years of the course we need expert tutors, • many tutors are still rooted in the old paradigm and are reluctant to embrace EBM and it was the view of my informants that experience round the world shows that it is difficult to teach EBM principles. The team at the University of Sydney were on the whole positive about their course. A parallel course This approach was used at Monash University, the University of Queensland, Flinders University, the University of Newcastle, and the University of Adelaide. In this approach, a non-PBL course is given separately from the main PBL course. This may consist of any combination of lectures, seminars, practicals, web pages, or text handouts. Usually there is an attempt to link this to the PBL cases by using examples related to the case of the week. For example, the case of the week might be asthma and the parallel course could include a critique of a paper reporting a trial of a treatment for asthma. There are several problems with this approach to teaching statistics and research methods. As noted in the Background, the subject may seem peripheral to the main thrust of the medicine course. Student feedback tends to give a much lower approval to parallel courses than to the main PBL teaching. Finally, teaching is dependent on the cases chosen by the PBL teachers, who may change the cases or reorder them at little or no notice. This can make statistics teaching, which is much more dependent on the order of presentation than most subjects in the medical curriculum, extremely difficult. Most people involved in these courses were unhappy with them, the exception being a group project in the Flinders course. PBL used only for 'public health' related subjects This approach was used at the University of Western Australia. This was an unusual model, found at only one university. PBL teaching had been initiated by an enthusiast, after a period spent at McMaster University. She was a member of the public health group and persuaded her colleagues to introduce PBL. However, only about one third of the course is taught this way, anatomy, biochemistry and physiology are taught traditionally. The consequence of this approach is that the tutors are drawn from the population medicine area and so are quite happy to teach statistics, research methods, and EBM. The triggers can be chosen as population-oriented problems, rather than being restricted to the patient case. People I spoke to were very positive about this course, not surprisingly as they saw themselves as the educational leaders in their institution. Discussion I did not to find anywhere in Australia a single instance of truly integrated teaching of statistics and research methods through PBL. I have to conclude from this that there are considerable difficulties. I did, however, find an instance of unintegrated PBL teaching, at the University of Western Australia, where anatomy, biochemistry and physiology have continued to be taught in a more traditional manner. What are these difficulties? • If tutors are drawn mainly from the laboratory disciplines, they may be unsympathetic to the population and clinical foci of EBM and its core subjects. We need strong advocates to persuade them of the importance of these topics. • If tutors are drawn from the clinical staff, they may be unsympathetic to the idea of EBM as a core activity. Converting the existing clinical teachers to the new paradigm of EBM may take a generation, but this will be a problem whatever teaching method we use. • Tutors may be ignorant of the principles and details of statistics. This may be true. My own experience as a PBL tutor has been that lack of knowledge has seldom been a problem. It is not the function of a PBL tutor either to impart knowledge or to explain concepts. Persuading tutors of this should be part of their general PBL training. • The changing patterns of cases as the course develops is a particular problem for statistics. This is undoubtedly true. • The nature of the subject does not lend itself to PBL. It would certainly be difficult to deduce the calculations required for a t test from a problem based on a patient with asthma. However, we must consider what we actually want to teach. I do not think there is much point in teaching undergraduate medical students to analyse data, even using computers. What we need to teach them is how to understand research publications, the evidence on which we hope their future evidence-based practice will be based. We can certainly do this using triggers such as published papers, as the University of Western Australia has demonstrated. • The patient case is not suited to teaching these subjects. This is true, though I do not think it is beyond us to devise cases which lead to research and evidence questions. However, we must persuade our colleagues that the patient vignette is not the only type of problem which we can use. We should not see only the difficulties, however, but also the opportunities. If we can integrate statistics and research methods into PBL, there may be great advantages. One consequence of integration would be that the subject would not be marginalised or seen as separate. It would be just one aspect of medicine. Indeed, one of the features of a PBL course is that the distinctions between the different subjects should blur. A second advantage would be that students would be learning statistics and research methods in the contexts in which we hope that they will apply them: the interpretation of clinical data and the assessment of research evidence. The relevance of the subject should be very clear and during their professional careers they would be more likely to be able to recall and use this material when needed. So how can we do it? There must be an advocate for statistics and research methods at the start of the preparations for PBL, who is able to argue convincingly for the inclusion of these subjects. They must form part of the matrix of topics and problems. This is not an easy task and I have been told of great difficulty experienced at this level, as other teachers sought to exclude these subjects so that their own discipline could have more time. This is a natural human reaction, of course. I think that statistics and research methods are central to medical education, and would argue for more of them, at the expense of some anatomy if necessary. We must get away from the idea that the problem must be a patient vignette. There are some statistical topics which can be covered quite conveniently in this way, such as measurement error, coefficients of variation, reference intervals, sensitivity and specificity, etc. After all, when my GP looks at my serum cholesterol on his practice computer, it has by the side of it a 95% reference interval. However, problems could equally be a published paper. We use these routinely in the teaching of statistics, research methods, and critical appraisal. We could use a paper as a trigger for non-research methods topics. For example, a paper on an asthma trial could trigger questions about asthma as well as about randomisation. This may lead to too many questions being raised by the trigger. Another possibility would be to have such a trigger immediately following a case vignette problem on the disease in question. The patient vignette would raise the questions on anatomy, biochemistry, pharmacology, etc., associated with the disease. The following problem using a research paper would then raise only the research methodological questions. The fixed resource sessions in this week would then be devoted to these. We may also, for variety, link research publications to the patient vignette by devices such as newspaper articles which the patient presents to the physician (e.g. reporting a trial), or say that in a patient problem the clinician has already found a Cochrane review. Fixed resource sessions could include lectures, but I am very reluctant to suggest formal lectures in statistics and research methods for medical students. I would prefer to offer open question and answer sessions, where students can ask the statistician lecturer to explain anything they are unsure of. For example, if the PBL trigger is a randomised controlled clinical trial with results presented in terms of P values, we might be expecting this trigger to lead to questions such as 'why randomize?', 'what does P < 0.05 mean?', and 'should patients be told they are in a trial?'. Students should have made some attempt to answer these before the fixed resource session. Conclusions There are several difficulties in implementing an integrated approach. However, not integrating is detrimental to statistics and research methods teaching, which is of particular concern in the age of evidence-based medicine. I remain optimistic that we can incorporate statistics and research methods into a fully problem-based curriculum. I think that if we do not, medicine will be the poorer for it. Competing interests The author(s) declare that they have no competing interests. Authors' contributions J M Bland is the sole author. Appendix The following is a typical example drawn from the teaching at St. George's Hospital Medical School. The students work in a small group with a tutor to act as facilitator. The students are given the following information: 'You are a member of the on-duty Trauma Team in the Accident and Emergency department one Monday evening, when you are alerted to the imminent arrival of an "RTA (road traffic accident) patient". As she is being wheeled in to the Emergency Room, the paramedic accompanying her reports on the circumstances: 'The patient is called Janet Phillips, she appears to be a medical student in her early 20s, and was on a Pelican crossing when she was struck by a car which had gone out of control. Janet had been thrown some distance from the car by the force of the impact. She is confused, in shock, has superficial lacerations to her face and severe pain in her pelvis and legs. 'At the accident site her airway was cleared, oxygen administered, a traction splint applied to her right leg, and she was immobilised on a long spinal board. 'The driver had only minor injuries but his breath smelled strongly of alcohol, and he was now in police custody. 'In the Accident and Emergency Department Janet is given a pelvic clamp to arrest her internal bleeding.' The students are told that they should address four general themes in their discussion: basic and clinical sciences, patient and doctor, community and public health, and professional development. In the first tutorial, they are asked to work through the following steps: 1. Clarify any terms and concepts in the scenario with which they are not familiar. 2. Define the problem(s) and issue(s) raised by the scenario. 3. Analyse the problems and issues, seeking explanations or hypotheses. 4. Agree on specific questions (learning objectives) for each of the four general themes to which you need answers. They then work individually on tackling the questions agreed upon. In thesubsequent tutorial they discuss their findings and decide how the problems and issues raised by the scenario could be resolved. They may then be given further information about the patient, which might raise further topics for their research. When a problem has raised all the points required, a fresh problem is presented to the students. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 'The Australian Universities visited'. This provides a descriprtion of teaching at the universities visited as part of this project. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539273.xml

514493

Protein Helps Orchestrate Cells' Fluid Uptake

You are what you eat and drink. Steak can sit in your stomach or orange juice wind through your intestines, but they only become part of your body once they're taken up by your cells. First, foods must be reduced to a soup of proteins, fats, sugars, and so on. But even then, getting these materials into a cell isn't as simple as sticking them in your mouth. For one, there's the membrane enclosing a cell. Simply puncturing a hole in the membrane would spill the cell's contents, harming or killing the cell. Instead, all eukaryotes—organisms whose cells have nuclei—use a carefully orchestrated process called endocytosis to bring materials into their cells. Eukaryotic cells first form cavities in their cell membrane that surround nearby particles or fluid. These pockets seal shut and bud off into the cell to form small membrane-bound sacs called vesicles. When taking in fluids, eukaryotic cells use two distinct mechanisms—to take tiny sips or huge gulps. With one process, called pinocytosis, cells continually form small pockets in the cell membrane that enclose small droplets of fluid in vesicles called pinosomes. These newly formed vesicles, called early endosomes, bud off from the membrane and fuse with other early endosomes. In one form of pinocytosis, the vesicles are encaged by a protein called clathrin that tightly constrains their size. These carriers incorporate membrane constituents (for example, growth factors) with very high selectivity. In macropinocytosis, on the other hand, large ruffles in the membrane engulf mass quantities of fluid in vesicles known as macropinosomes. Beyond taking in nutrients, these processes are essential to the function of many organs—from the brain, where nerve cells receive other cells' chemical signals by pinocytosis, to the kidney, where cells use macropinocytosis to take in waste fluids for processing. Macropinocytosis is also relevant to cancer cells; it has long been known that oncogenes dramatically induce this endocytic process, affecting the signaling status of these cells. But compared with other types of endocytosis, molecular biologists know surprisingly little of the mechanisms behind macropinocytosis. They do know that the Rab5 protein—an enzyme that coordinates a complex network of other proteins, called effectors—is crucial for both pinocytosis and macropinocytosis. Now, as reported in this issue of PLoS Biology , Marino Zerial and colleagues have found a new protein, which they named Rabankyrin-5, that forms a further link between these two mechanisms for fluid uptake. The protein is necessary for macropinocytosis, and its levels control the rate of this process. In addition, Rabankyrin-5 helps regulate endosome trafficking and coordinates this mechanism with macropinocytosis. In two commonly used human and mouse cell lines, the researchers found the protein Rabankyrin-5 along with Rab5 on both types of pinosomes, early endosomes and macropinosomes. The early endosomes usually fuse with one another inside the cell, but when the researchers blocked Rabankyrin-5 activity, this fusion fell sharply. Suppressing Rabankyrin-5 activity also stifled macropinocytosis; overexpressing the effector, on the other hand, sent macropinocytosis into overdrive. The researchers also looked at endocytosis in mouse kidney and canine kidney cell lines. Inside the kidney, fluid-carrying ducts are lined with epithelial cells that take up liquids through their exposed surface. The researchers found Rabankyrin-5 predominately on vesicles at this surface, and as in the other experiments, overexpression of the protein promoted macropinocytosis. Together, these findings suggest Rabankyrin-5 plays a role in regulating this form of fluid uptake and plays a role in kidney function. The discovery of Rabankyrin-5 involvement in macropinocytosis also has implications for other physiological and pathological mechanisms such as the immune system response, defense against pathogens, and hyperactivation of signaling pathways in cancer cells. Rabankyrin-5 contains various regions that bind other proteins and also lipids found in cell membranes, suggesting the protein plays a mechanical role in forming vesicles. The protein also has regions found on other proteins that are involved in signaling and development, so it may help direct vesicles' traffic within the cell. The protein also has regions characteristic of proteins involved in clathrin-dependent endocytosis, which fits with the researchers' finding that Rabankyrin-5 affects pinocytosis. Rabankyrin-5 (green) colocalizes with rhodamine-conjugated EGF on macropinosomes after growth factor stimulation All told, Rabankyrin-5 appears to form a bridge between two distinct mechanisms, pinocytosis and macropinocytosis, that cells use to take in fluids. While the details of how Rabankyrin-5 functions are still unclear, these findings give researchers a new handle for grasping how macropinocytosis works and how cells control when and how much they drink in their surroundings.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514493.xml

534794

Effect of 50% ethanolic extract of Syzygium aromaticum (L.) Merr. & Perry. (clove) on sexual behaviour of normal male rats

Background The flower bud of Syzygium aromaticum (L.) Merr. & Perry. (clove) has been used in Unani medicine since ancient times for the treatment of male sexual disorders. The present study is aimed to investigate the effect of 50% ethanolic extract of clove on general mating behaviour, libido, potency along with its likely gastric ulceration and adverse effects on sexually normal male albino rats. Methods The suspension of the extract was administered orally at the dose of 100, 250, and 500 mg / kg, to different groups of male rats (n = 6) once a day for seven days. The female albino rats involved in mating were made receptive by hormonal treatment. The general mating behaviour, libido and potency were determined and compared with the standard reference drug sildenafil citrate. The probable gastric ulceration and adverse effects of the extract were also evaluated. Results Oral administration of the extract significantly increased the Mounting Frequency, Intromission Frequency; Intromission Latency, Erections; Quick Flips, Long Flips as well as aggregate of penile reflexes and caused significant reduction in the Mounting Latency and Post Ejaculatory Interval. The most appreciable effect of the extract was observed at the dose of 500 mg/kg. The test drug was also found to be devoid of any conspicuous gastric ulceration and adverse effects. Conclusion The results indicated that the 50% ethanolic extract of clove produced a significant and sustained increase in the sexual activity of normal male rats, without any conspicuous gastric ulceration and adverse effects. Thus, the resultant aphrodisiac effectivity of the extract lends support to the claims for its traditional usage in sexual disorders.

Background Clove is the dried flower bud of Syzygium aromaticum (L.) Merr. & Perry. (Family: Myrtaceae) an evergreen tree 10–20 m in height indigenous to India, Indonesia, Zanzibar, Mauritius and Sri Lanka [ 1 ]. It is one of the most important drugs used in indigenous medicine in India, especially in Unani medicine. Clove is reported as aphrodisiac [ 2 ], stomachic [ 3 , 4 ], carminative, antispasmodic [ 5 , 6 ]. It is reported to be useful in conceiving in high doses and act as a contraceptive in low doses [ 7 ] and useful in cataract [ 8 ]. Clove is also reported to have anticarcinogenic property [ 9 ]. It inhibits platelet aggregation and alters arachidonic acid metabolism in human platelets [ 10 ]. It posseses antiviral activity against Herpes simplex [ 11 ]. Phytochemical studies indicate that the clove contains free eugenol, eugenol acetate, caryophyllene, sesquetrepene ester [ 12 ], phenyl propanoid [ 13 ], β caryophyllene [ 14 ], eugenol and acetyle eugenol [ 15 ]. Eugenol, the major constituent, inhibits lipid peroxidation and maintains activities of enzyme superoxide dismutase, catalase, glutathione peroxidase-6 phosphate dehydrogenase [ 16 ], and has also been reported to have vasodilatory [ 17 ] and smooth muscle relaxant property [ 18 ]. Phyto chemical study of the test drug that was carried out according to the methods described by Jenkin et al [ 19 ], showed that it contains alkaloids, amino acids, flavonoids, proteins, sterols, reducing sugar, tannins and phenols. However, clove, or its known compound had not been scientifically studied for their effect on sexual function. Earlier we carried out a preliminary study of hydro alcoholic extract (50%) of clove using only mounting frequency and mating performance as the marker for sexual function in normal male mice, and the results from our study demonstrated the aphrodisiac activity and safety from short term toxicity of the test drug [ 20 ]. The present study thus, is aimed to investigate the aphrodisiac effect of 50% ethanolic extract of clove in detail, using multiple parameters along with its probable gastric ulceration and adverse effects in sexually normal male albino rats. The doses used in the study were selected according to the Freirich [ 21 ], multiplying the Unani clinical doses reported in standard Unani text [ 2 ] by the conversion factor of 7. Methods Plant material and extraction The authenticated dried flower bud of S. aromaticum (clove) was procured from the market (Delhi, India). A voucher (S721) sample was kept for further reference. The clove was crushed to coarse powder and sieved through No. 20 mesh size. The extraction was carried out by mixing the powdered clove with 1:3 w/v in 50% ethanol v/v by Soxhlet apparatus for 6 h. The extract was filtered and the solvent from the filtrate was removed by rotary evaporator under reduced pressure and low temperature. The yield of extract was 10.40% w/w in terms of dried starting material. It was yellowish and of pleasant smell. The extract was preserved in a refrigerator. Chemicals used Sildenafil citrate was purchased from Zydus Cadila, (Ahmadabad, India). Other drugs used were ethinyl oestradiol (Infar Limited, Calcutta, India), progesterone (Sun Pharmaceutical Industries Limited, Mumbai, India) and 5% xylocane ointment (Astra IDL Limited, Bangalore, India) Animals Twelve weeks old male and female albino rats of wistar strain weighing 350–400 g and 225–275 g respectively, were used for the study. They were housed singly in separate standard cages and maintained under standard laboratory conditions (temperature 24–28°C, relative humidity 60–70%, 12 h light-dark cycle) with free access to solid pellet diet (Gold Mohar, Lipton-India) and water ad libitum throughout the study except during the experiment. The study design was approved by the ethical committee of the Department for animal care and use. Drug preparation Since clove in Unani medicine is orally administered, therefore, the extract of clove was suspended in distilled water using Tween 80 (1%) for oral administration. Sildenafil citrate and ethinyl oestradiol were also suspended in distilled water using Tween 80 (1%) separately, for oral use. Progesterone was dissolved in olive oil for subcutaneous injection. All the drug solutions were prepared just before administration. Mating behaviour test The test was carried out by the methods of Dewsbury and Davis Jr [ 22 ] and Szechtman et al [ 23 ], modified by Amin et al [ 24 ]. Healthy and sexually experienced male albino rats (350–400 g) that were showing brisk sexual activity were selected for the study. They were divided into 5 groups of 6 animals each and kept singly in separate cages during the experiment. Group 1 represented the control group, which received 10 ml/kg of distilled water orally. Groups 2–4 received suspension of the extract of clove orally at the doses of 100, 250 and 500 mg/kg, respectively, daily for 7 days at 18:00 h. Group 5 served as standard and given suspension of sildenafil citrate orally at the dose of 5 mg/kg, 1 h prior to the commencement of the experiment. Since the male animals should not be tested in unfamiliar circumstances the animals were brought to the laboratory and exposed to dim light (in 1 w fluorescent tube in a laboratory of 14' × 14') at the stipulated time of testing daily for 6 days before the experiment. The female animals were artificially brought into oestrus (heat) [ 25 ] by the Szechtman et al method (as the female rats allow mating only during the estrus phase) They were administered suspension of ethinyl oestradiol orally at the dose of 100 μg/animal 48 h prior to the pairing plus progesterone injected subcutaneously, at the dose of 1 mg/animal 6 h before the experiment. The receptivity of the female animals was confirmed before the test by exposing them to male animals, other than the control, test and standard animals. The most receptive females were selected for the study. The experiment was carried out on the 7 th day after commencement of the treatment of the male animals. The experiment was conducted at 20:00 h in the same laboratory and under the light of same intensity. The receptive female animals were introduced into the cages of male animals with 1 female to 1 male. The observation for mating behaviour was immediately commenced and continued for first 2 mating series. The test was terminated if the male failed to evince sexual interest. If the female did not show receptivity she was replaced by another artificially warmed female. The occurrence of events and phases of mating were called out to be recorded on audio-cassette as soon as they appeared. Their disappearance was also called out and recorded. Later, the frequencies and phases were determined from cassette transcriptions: number of mounts before ejaculation or Mounting Frequency (MF), number of intromission before ejaculation or Intromission Frequency (IF), time from the introduction of female into the cage of the male upto the first mount or Mounting Latency (ML), time from the introduction of the female up to the first intromission by the male or Intromission Latency (IL), time from the first intromission of a series upto the ejaculation or Ejaculatory Latency (EL), and time from the first ejaculation upto the next intromission by the male or Post Ejaculatory Interval (PEI). In the second mating series only the EL was recorded. The values for the observed parameters of the control, test and standard animals were statistically analysed by using one-way analysis of variance (ANOVA) method. Test for libido The test was carried out by the method of Davidson [ 26 ], modified by Amin et al [ 24 ]. Sexually experienced male albino rats were divided into 5 groups of 6 animals each and kept singly in separate cages during the experiment. Group 1 represented the control group, which received 10 ml/kg of distilled water orally. Groups 2–4 received suspension of the extract orally at the doses of 100, 250 and 500 mg/kg, respectively, once a day in the evening (18:00 h) for 7 days. Group 5 served as standard and given suspension of sildenafil citrate orally at the dose of 5 mg/kg, 1 h prior to the commencement of the experiment. The female rats were made receptive by hormonal treatment and all the animals were accustomed to the testing condition as previously mentioned in mating behaviour test. The animals were observed for the Mounting Frequency (MF) on the evening of 7 th day at 20:00 h. The penis was exposed by retracting the sheath and 5% xylocaine ointment was applied 30, 15 and 5 min before starting observations. Each animal was placed individually in a cage and the receptive female rat was placed in the same cage. The number of mountings was noted. The animals were also observed for intromission and ejaculation. The MF in control, test and standard animals was statistically analysed by employing one-way analysis of variance (ANOVA) method. Test for potency The effect of the test drug was studied according to the methods described by Hart and Haugen [ 27 ] and Hart [ 28 ], modified by Amin et al [ 24 ]. The male animals were divided into 5 groups of 6 animals each and kept singly in separate cages during the experiment. Group 1 represented the control group, which received 10 ml/kg of distilled water orally. Groups 2–4 received suspension of the test drug orally at the doses of 100, 250 and 500 mg/kg, respectively, daily for 7 days. Group 5 received a suspension of sildenafil citrate orally at the dose of 5 mg/kg, 1 h before the commencement of the experiment. On the 8 th day, the test for penile reflexes was carried out by placing the animal on its back in a glass cylinder partial restraint. The preputial sheath was pushed behind the glans by means of thumb and index finger and held in this manner for a period of 15 min. Such stimulation elicits a cluster of genital reflexes. The following components were recorded: Erections (E), Quick Flips (QF), and Long Flips (LF). The frequency of these parameters observed in control, test and standard groups was statistically analysed by using one-way analysis of variance (ANOVA) method. Test for ulcerogenecity The male animals (350–400 g) were divided into 4 groups of 6 animals each. Group 1 represented the control group, which received 10 ml/kg of distilled water. Groups 2–3 received suspension of the extract orally at the doses of 100, 250 and 500 mg/kg, respectively, daily for 7 days. After the treatment, on 8 th day all the animals were killed and the stomach was then incised along the greater curvature and washed carefully with physiological saline. Any gastric lesions were observed immediately using a magnifying glass. The number of erosions per stomach was assessed for severity, according to the score of Cioli et al [ 29 ] : (o) absence of lesion, vasodilation or up to 3 pin point ulcers; (1) more than 3 pin point ulcers, (2) from 1 to 5 small ulcers (< 2 mm); (3) more than 5 small ulcers (< 2 mm), (4) 1 or more giant ulcers. Evaluation of gastric damage was carried out by two observers, who followed the same evaluation criteria. Adverse effects All treated rats were observed at least once daily for any overt sign of toxicity (salivation, rhinorrhoea, lachrymation, ptosis, writhing, convulsions and tremors), stress (erection of fur and exophalmia) and changes in behaviour (such as spontaneous movement in the cage, climbing, cleaning of face). In addition food and water intake were noted. Statistical analysis The significance of difference between the means was determined by one-way analysis of variance (ANOVA) with post-hoc't' test. P value <0.05 was considered as significant. Results The data obtained with the mating behaviour test indicated that the clove-extract at the dose of 500 mg/kg increased the Mounting Frequency (MF) (P < 0.0001), Intromission Frequency (IF) (P < 0.001) ; Ejaculatory latency in first series (EL 1 ) (P < 0.0001) and decreased Mounting Latency (ML) (P < 0.0001), Intromission Latency (IL) (P < 0.01) in a significant manner. The dose 250 mg / kg of the extract significantly increased the MF (P < 0.001), IF (P < 0.05) and significantly reduced the ML (P < 0.001), EL 1 (P < 0.05); Post Ejaculatory Interval (PEI) (P < 0.001) and did not significantly alter the IL (P < 0.61), Ejaculatory Latency in second series (EL 2 ) (P < 0.41). Whereas the dose of the test drug at 100 mg/kg significantly increased the MF (P < 0.05), PEI (P < 0.05) but did not significantly affect the IF (P < 0.61), ML (P < 0.61), IL (P < 0.61); EL 1 (P < 0.61); EL 2 (P < 0.81). However, the standard drug increased the MF (P < 0.0001), IF (P < 0.0001); EL 1 (P < 0.0001), EL 2 (P < 0.0001); PEI (P < 0.0001) as well as decreased ML ((P < 0.0001) and IL (P < 0.0001) in a highly significant manner (Table 1 ). Table 1 Effect of 50% ethanolic extract of clove ( S. aromaticum ) on mating behaviour in male rats Mean Frequency ± SEM Parameters Control (10 ml/kg) Clove (100 mg/kg) Clove (250 mg/kg) Clove (500 mg/kg) Sildenafil citrate (5 mg/kg) Mounting Frequency (MF) 11.50 ± 1.22 13.20 ± 1.47* 19.00 ± 1.10*** 31.80 ± 2.32**** 48.70 ± 2.34**** Intromission Frequency (IF) 5.50 ± 1.22 5.83 ± 1.17 NS 6.67 ± 0.81* 8.17 ± 1.47*** 24.70 ± 0.81**** Mounting Latency (ML, in sec) 35.30 ± 1.51 36.70 ± 1.21 NS 31.30 ± 1.63*** 24.30 ± 1.63**** 11.70 ± 1.37**** Intromission Latency (IL, in sec) 40.00 ± 5.29 41.30 ± 3.89 NS 38.50 ± 3.89 NS 21.00 ± 3.83** 15.00 ± 0.89**** Ejaculatory Latency in first series (EL 1, in sec) 198.00 ± 0.98 201.00 ± 13.00 NS 208.00 ± 13.70* 233.50 ± 12.00**** 344.50 ± 12.00**** Ejaculatory Latency in second series (EL 2 , in sec) 297.33 ± 8.10 295.50 ± 11.70 NS 311.33 ± 9.67 NS 343.16 ± 7.70**** 398.16 ± 13.50**** Post Ejaculatory Interval (PEI, in sec) 364.00 ± 12.22 343.16 ± 7.70* 309.33 ± 10.90**** 217.33 ± 8.96**** 99.00 ± 5.68**** Tabular values are mean ± SEM, n = 6 (number of animals in each group); significant difference from control, NS: Not significant. * P < 0.05, ** P < 0.01; *** P < 0.001, **** P < 0.0001. The test for libido showed that the extract at the dose of 500 mg/kg increased the Mounting Frequency (MF) in a significant manner (P < 0.001). The extract at the doses of 100 mg/kg and 250 mg/kg did not significantly alter the MF(P < 0.24, P < 0.75 respectively). The standard drug strikingly increased the MF (P < 0.0001). Intromission and Ejaculation were found absent in control, test and standard groups (Table 2 ). Table 2 Effect of 50% ethanolic extract of clove ( S. aromaticum ) on Mounting Frequency (test for libido) in male rats Mean Frequency ± SEM Parameters Control (10 ml/kg) Clove (100 mg/kg) Clove (250 mg/kg) Clove (500 mg/kg) Sildenafil citrate (5 mg/kg) Mounting Frequency (MF) 6.17 ± 0.98 6.33 ± 0.81NS 7.17 ± 1.72 NS 11.20 ± 1.17* 23.00 ± 2.17** Intromission Frequency (IF) Nil Nil Nil Nil Nil Ejaculation (EJ) Absent Absent Absent Absent Absent Tabular values are mean ± SEM, n = 6 n (number of animals in each group); significant difference from control, NS: Not significant. *P < 0.001,** P < 0.0001 The test for potency exhibited that the higher dose (500 mg/kg) of the test drug significantly increased the frequency of Erections (E) (P < 0.0001), Quick Flips (QF) (P < 0.0001), Long Flips (LF) (P < 0.0001) as well as the aggregate of these penile reflexes (TPR) (P < 0.0001). The extract at the dose of 250 mg/kg significantly increased the E (P < 0.01), LF (P < 0.05) and TPR (P < 0.05) but comparatively less than the higher dose of the extract and standard drug, and did not significantly increase the QF (P < 0.10). whereas, the test drug at the dose of 100 mg/kg did not alter the E (P < 0.78), QF (P < 0.78); LF (P < 0.44) and TPR (P < 0.44) in a significant manner (Table 3 ). Table 3 Effect of 50% ethanolic extract of clove ( S. aromaticum ) on Penile reflexes (test for potency) Mean Frequency ± SEM Parameters Control (10 ml/kg) Clove (100 mg/kg) Clove (250 mg/kg) Clove (500 mg/kg) Sildenafil citrate (5 mg/kg) Erections (E) 7.67 ± 1.63 8.33 ± 1.21 NS 10.50 ± 1.76** 13.80 ± 0.98*** 19.00 ± 2.64*** Quick Flips (QF) 5.17 ± 0.75 5.33 ± 1.21 NS 6.00 ± 1.22 NS 9.67 ± 1.37*** 17.30 ± 4.13*** Long Flips (LF) 2.17 ± 1.17 3.17 ± 1.50 NS 4.00 ± 1.55* 7.50 ± 1.22*** 12.00 ± 2.26*** Total Penile Reflexes (TPR) 15.01 ± 3.55 16.83 ± 7.21 NS 20.50 ± 4.53* 30.97 ± 3.50*** 48.30 ± 9.03*** Tabular values are mean ± SEM, n = 6 (number of animals in each group); significant difference from control, NS: Not significant. * P < 0.05, ** P < 0.01; *** P < 0.0001. Seven days treatment of low and high doses of the extract caused no significant ulceration in gastric mucosa of albino rats. Moreover, there were neither treatment related defects nor overt clinical signs of toxicity, stress or changes in behaviour and appearance evident. The food and water intake of all test drug treated rats remained similar to those of the control group. Discussion In the present study, clove ( S. aromaticum ) was tested in animal experimentation for its effect on sexual behaviour, and sildenafil citrate was used as the standard referent. The study showed that the 50% ethanolic extract of clove possesses significant sexual function enhancing activity as observed in sexual behaviour tests. Mating behaviour test revealed that the test drug significantly increased the Mounting Frequency (MF) and Intromission Frequency (IF) as compared to control but less than that of the standard drug. The (MF) and (IF) are considered as the indices of both libido and potency. So, this is an indication that the test drug possesses a sexual function improving effect. The premature ejaculation is one of the important causes of sexual dysfunction, so the assessment of Ejaculatory Latency in first series (EL 1 ) and in second series (EL 2 ) was studied. The test drug significantly increased the EL 1 and EL 2 as compared to control animals, whereas a highly significant increase was observed with the standard drug. The test drug was found to produce a significant reduction in the Mounting Latency (ML) and Intromission Latency (IL) as compared to control while a highly significant decrease was found in ML of animals treated with sildenafil citrate. This is also an evidence of the sexual function improving effect of the clove. The Post Ejaculatory Interval (PEI) is considered as an index of potency and libido, and also a parameter of the rate of recovery from exhaustion after first series of mating. PEI was found significantly decreased with clove-extract and also with the standard drug. The test drug decreased PEI either by enhancing the potency and libido or by producing lesser exhaustion in the first series of mating or by both. The effect of the test drug on libido was also evaluated by testing the MF after genital anaesthetization which does away with the reinforcing effect of genital sensation thus affording the study of pure libido or intrinsic sexual desire. During the experiment the extract produced a significant increase in the MF of sexually normal male rats whereas the efficacy of the standard drug, as expected, was found to be highly significant. The MF was much reduced in control, test and standard animals in comparison with the MF of corresponding groups in mating behaviour test where the penis had not been anaesthetized. None of the control, test and standard animals were observed to show Intromission or Ejaculation because their occurrence depends upon local genital sensation which was obstructed due to anaesthetization. The effect on potency was also evaluated by testing the effect of the drug on the frequency of penile reflexes namely Erections (E), Quick Flips (QF), and Long Flips (LF). The test drug significantly increased the frequency of all the components of penile reflexes (E, QF, & LF) in the test animals as compared to control group but comparatively lesser than the standard drug. The aggregate of penile reflexes (TPR) was also found increased in the animals treated with the extract and sildenafil citrate. Therefore, the experiment revealed that the test drug produced a marked increase in potency. The spices are reported to produce an increase in gastric acid secretion by a cholinergic mechanism [ 30 ], and so, their use for sexual invigoration may cause gastric ulceration and other adverse effects. Therefore, ulcerogenic and other adverse effects of the extract were also evaluated. The results of this evaluation were negative. This suggests that the short term use of clove for this purpose is apparently safe. The results of the present study clearly proved that the clove is endowed with sexual function improving activity. This is in consonance with our earlier study showing sexual function improving effect of the test drug in male mice. However, the established drug i.e. sildenafil citrate exhibited, as expected, tremendous activity. With regard to the mechanism of the test drug, it is difficult to interprete the mechanism involved in potentiation of sexual function. The drugs induced changes in neurotransmitter levels or their action at cellular levels could change sexual behaviour [ 31 ]. Hence, the increased sexual function could be due to the nervous stimulant action of the test drug [ 32 ]. Further, phyto chemical study of the extract indicated that it contains sterols and phenol. Thus, the resultant aphrodisiac effectivity of the test drug might also be attributed to sterols or phenolic compounds. Moreover, research should be aimed at isolating the active principle(s) responsible for aphrodisiac activity and the mechanism by which the drug enhances sexual function. In addition, to discover the applied effective concentration or dosages of the extract, more studies are also required. Conclusions The present results indicated that the 50% ethanolic extract of clove possesses potent aphrodisiac activity in normal male albino rats without any gastric ulceration and adverse effects and provided scientific evidence in favour of the claims made in Unani medicine that the clove is clinically useful as sexual invigorator in males. Competing interests The author(s) declare that they have no competing interests. Authors' contributions T-Supervised the design and coordinatrion of the study. SA-Practically conducted the design of the study. AL-Participated and performed the statistical analysis IA-Participated in the drafting of 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/PMC534794.xml

526384

Infectious reservoir of Plasmodium infection in Mae Hong Son Province, north-west Thailand

Background It was unknown whether the main reservoir of Plasmodium falciparum and Plasmodium vivax, which infects mosquitoes in Thailand, was (a) in people feeling sufficiently ill with malaria to come to a clinic or (b) in people who had remained in their home villages with some fever symptoms or with none. Methods Mass surveys were carried out in Thai villages to identify people with Plasmodium infections and with fever. Malaria patients were also located at a clinic which served these villages. Adults from both sources whose blood slides registered positive for Plasmodium spp. were requested to allow laboratory-bred Anopheles minimus to feed on them. Seven to nine days after the blood feeds the mosquitoes were dissected and checked for presence of oocysts. Results and Discussion There were higher rates of Plasmodium infection among people in the villages with fever than without fever and much higher rates of infection among clinic patients than among people who had remained in the villages. People with malarial infections identified via the clinic and the village surveys could infect mosquitoes, especially, but not only, if their blood slides showed visible gametocytes. Because only a very small minority of the village populations were visiting the clinic on any one day, assessment indicated that the main reservoir of infection was not primarily among clinic patients but among those in the villages, especially those feeling feverish. Conclusions Efficient use of an anti-gametocyte drug to suppress the parasite reservoir in a population requires that it be given, not just to clinic patients, but to infected people located by mass surveys of the villages, especially those feeling feverish.

Introduction An estimate of the infectiousness of the Plasmodium reservoir to mosquitoes is of interest in understanding the epidemiology of malaria and its changes after application of certain types of control measures. Different approaches have been used to investigate this, including direct feeding of mosquitoes through the skin of human subjects [ 1 - 5 ] or feeding through an artificial membrane [ 6 - 13 ]. The present study was designed to determine whether the reservoir of infection of vectors was mainly in people ill enough to go to a clinic or whether it was mainly in those with slight or no malarial symptoms who have remained in their villages. Patients and Methods Subjects This study was approved by the Thai Ethics Committee and the Ethics Committee of London School of Hygiene and Tropical Medicine. Subjects were recruited from mass blood surveys in villages and from a clinic in Muang District, Mae Hong Son Province, which is in north-western Thailand. Anti-malarial drug use is tightly controlled in Thailand and these drugs are generally available on prescription. Self-medication is, therefore, unusual in this area. Mass blood surveys were conducted twice a year. After slides were examined for parasites, the people whose slides were positive and were more than 14 years of age were invited to take part in this study. The Thai ethical committee did not allow direct feeding of mosquitoes on children aged less than 15 years. Adults positive for malaria parasites were informed about the purpose of the study and, if they agreed, a consent form was signed and patients were interviewed. Human subjects for the study were also enrolled from the Vector-borne Disease Control Unit No. 8 (VBDU: described as "Clinic" throughout this paper). In this Clinic, blood smears were taken by the clinic staff and stained with Giemsa to check for malaria. Patients who had presented with clinical malaria and/or parasitaemia and were more than 14 years of age were informed about the purpose of the study and asked to participate in it. If they agreed and signed the consent form, then the mosquito feeding was performed. Mosquito feeding and dissection Before feeding, patients' arms were cleaned with 70% alcohol. Then, 50 laboratory-bred female Anopheles minimus species A (aged 4–6 days), which had been starved for 9–12 hr, were allowed to feed on their arms for 30 minutes. After feeding was completed, anti-malarial drugs were given to all the human subjects by a malaria worker. Two hours after feeding the unengorged mosquitoes were removed from the cups using a sucking tube and destroyed, leaving only fully engorged mosquitoes in the cup. The mosquitoes were brought back to the Chiang Mai insectarium and maintained at 25 to 27 C and 70–80% relative humidity with permanent access to sucrose solution and without any further blood meals. Mosquitoes at 7–9 days post-feed were anaesthetized and the wings and legs were removed. Midguts were dissected on glass slides in a drop of 0.85% NaCl and examined at a 40× magnification. The number of oocysts present on the midgut to each mosquito was counted and recorded individually. Quality control of blood slide data Asexual parasitaemia and gametocytaemia were quantified as the number of aseuxal forms/200 white blood cells on a thick film. Throughout the period of the study, 20% of the negative blood smears and all positive slides which had been examined by the microscopist of the Clinic were chosen randomly and re-examined by a team from the Office of Vector-borne Disease Control No.2, Chiang Mai. All cases of positive slides without visible gametocytes and which led to oocyst production were re-examined by an expert team from the London School of Hygiene and Tropical Medicine to ascertain whether a few gametocytes might have been present but were initially missed. Analyses The database from mosquito feeding included 1) densities of gametocytes and trophozoites, 2) number of infected mosquitoes, and 3) mean number of oocysts per infected mosquito. The chi-square test or Fisher's exact test was used to examine the significance of differences in the proportion of people in various categories who transmitted malaria to the mosquitoes (i.e. yielded at least 1 oocyst). Regressions on the natural log of gametocyte density were also computed. Stata statistical software (version 6) was used for analysis. Results The upper part of Table 1 shows the results associated with 5,227 blood smears from children and adults in village surveys. 10.7% (561/5,227) of subjects reported fever within the previous seven days but the corresponding rate for slide positives was much higher (53.8%, 28/52). In children (age <15 yrs), 60% (9/15) of slide positives reported fever. Among 37 slide positive adults (age >14 years), 19 (51.4%) reported fever. Among these 19 cases, nine were infected with Plasmodium falciparum and 10 with Plasmodium vivax. The prevalence of infection of both species from the village surveys was higher in children (15/623 = 2.4%) as compared to adults (37/4604 = 0.8%) (χ 2 = 12.8, P < 0.001). The lower part of Table 1 shows data from the Clinic where 101 patients aged >14 years were interviewed; 85.2% (86/101) reported fever. Among those, 39 were infected with P. falciparum and 47 with P. vivax. Table 1 Data from village surveys and the clinic on reported fever and malaria infection. Slide reading Fever reported Fever not reported Fisher's exact tests of association of fever with infection Village surveys, age <15 yrs P. falciparum +ve 2 (3.2%) 3 (0.54%) P = 0.082 P. vivax +ve 7(11.1%) 3 (0.54%) P = 0.000007 Total slides 63 560 Village surveys, age >14 yrs P. falciparum +ve 9 (1.8%) 10 (0.24%) P = 0.00006 P. vivax +ve 10 (2.0%) 8 (0.19%) P = 0.000003 Total slides 498 4106 Clinic (age >14 yrs) P. falciparum +ve 39 (45.35%) 10 (66.67%) χ 2 = 0.39, P = 0.53 P. vivax +ve 47 (54.65%) 5 (33.33%) χ 2 = 0.14, P = 0.51 Total slides 86 15 As shown in Table 2 , at least one mosquito became infected with oocysts in the batches fed on 10/28 people from the village infected with either species (35.7%). The proportion of individual patients from the Clinic who infected at least one mosquito was 43.4% (40/92). The difference between these two rates was not significant (χ 2 = 0.26, p = 0.6). The probability of infection of mosquitoes from people infected with P. falciparum was significantly less than from people infected with P. vivax (24.7% versus 57.1%, χ 2 = 11.6, P < 0.001). 77.1% (44/57) of P. falciparum cases reported fever. Among those, only 27.3% could infect mosquitoes. For P. vivax 84.1% (53/63) reported fever; among those 56.6% could infect mosquitoes. More than half of the infections from whose blood no mosquitoes developed oocysts had parasite density > 3,960/μl. There was no association of probability of mosquito infection with parasite density (χ 2 = 1.3, P > 0.05). Table 2 Results of feeding mosquitoes on human blood No. people from whose blood some mosquitoes developed oocysts No. people from whose blood no mosquitoes developed oocysts Significance of differences P. falciparum Villages 3 (23.1%) 10 (76.9%) Fisher not sig. Clinic 11 (25.0%) 33 (75.0%) Observable gametocytes 6 (46.2%) 7 (53.8%) Fisher P = 0.06 No observable gametocytes 8 (18.2%) 36 (81.8%) Fever reported 12 (27.3%) 32 (72.7%) Fisher P = 0.32 Fever not reported 2 (15.4%) 11 (84.6%) P. vivax Villages 7 (46.7%) 8 (53.3%) χ 2 = 0.41, P = 0.52 Clinic 29 (60.4%) 19 (39.6%) Observable gametocytes 31 (68.9%) 14 (31.1%) χ 2 = 7.3, P = 0.007 No observable gametocytes 5 (27.8%) 13 (72.2%) Fever reported 30 (56.6%) 23 (43.4%) Fisher P = 0.56 Fever not reported 6 (60%) 4 (40%) Parasite density ( P. falciparum and P. vivax ) <3961/μl 21 (35.9%) 38 (64.4%) χ 2 = 1.30, P = 0.25 >3960/μl 29 (47.5%) 32 (52.5%) Considering each species separately, the probability of infectiousness to mosquitoes was significantly related to being observably gametocyte positive in P. vivax (P= 0.007), but the relationship was only of borderline significance for P. falciparum (Fisher's exact test, P = 0.06). Approximately 50% (7/13) of the P. falciparum cases with observable gametocytes failed to infect mosquitoes and 30% (14/45) of the P. vivax cases with observable gametocytes failed to infect. However, the difference was not significant (Fisher's exact test, P= 0.19). Approximately 21% (13/62) with no observable gametocyte of either species could infect mosquitoes. 58 of the infections with either species had observable gametocytes (11 cases from the village survey and 47 from the Clinic). Feeds on 37 of these 58 gametocyte carriers led to oocyst production. The regression of percent of mosquitoes infected on the natural log of gametocyte density was not significant ( t = 0.87, df= 36, P= 0.39). Similarly, there was not a significant association of mean oocyst load per infected mosquito on the natural log of gametocyte density ( t = 0.87, df = 36, P= 0.38). Discussion The results from the experiments with mosquitoes showed the infectiousness of subjects from the village surveys as well as from the Clinic. These indicated that some symptomatic and asymptomatic infections of each species could infect mosquitoes. The differences in percentages of infection in different studies [ 2 , 4 , 14 ] might be explained by several possible factors. First, the method of feeding: a recent study comparing the infectivity of gametocyte carriers to mosquitoes, using membrane and direct feeding, found significantly higher proportions of mosquitoes infected and higher oocyst burdens in mosquitoes fed directly on the skin [ 15 ]. Conversely, Vanderberg [ 16 ] reported that infectivity in mosquitoes fed through a membrane usually equaled or exceeded infections by direct methods. However, most studies gave better results for direct feeding than membrane feeding. Thus studies such as the present one using direct feeding may provide a more reliable estimate of the infectious reservoir. Second, recruited subjects: in several studies mosquitoes were fed on individuals selected randomly and not on the basis of gametocytaemia [ 5 , 10 , 14 ]. But in other studies mosquitoes were fed on selected gametocyte carriers [ 4 , 17 ] or parasitaemic cases with or without gametocytes (as in the present study). Another relevant factor is variability in mosquito populations [ 18 , 19 ], such as mosquito size [ 20 - 22 ]. The number of blood meals may also affect the infection rate [ 8 , 23 ]. A careful comparison of Anopheles dirus , An. minimus and Anopheles maculatus infectivity in relation to size and blood-feeding behaviour would be of interest. Moreover, variability in susceptibility between different mosquito colonies is possible [ 4 , 24 ]. The results in the present study showed that some cases with undetectable gametocytes could infect mosquitoes. This apparent anomaly is presumably at least partly due to larger volumes of blood in mosquito bloodmeals than are observed on slides, so that sufficient gametocytes to infect a mosquito may have been below the level of detection on blood slides. An attempt was made to see if gametocyte densities are higher in bloodmeals than finger pricks. However, it was found that gametocyte density was not significantly higher (and actually appeared to be lower) in the blood taken up by mosquitoes than in the blood from finger pricks. In some cases high densities of gametocytes were not infectious and similar results have been reported in most studies assessing human malarial infectivity to mosquitoes [ 3 , 4 , 11 , 24 , 25 ]. It has been suggested that the prevalence of gametocyte carriers is not a good indication of the infectiousness of a population to mosquitoes [ 10 , 25 - 27 ]. It is clear that, as the Ethical committee only authorized us to request people over 14 years of age to take part in our experiments, this excluded a sector of the reservoir population, the children, from the study. It is also clear that the occurrence of transmission-blocking immunity and prior histories of the subjects taking anti-malarial drugs might have influenced whether they infected mosquitoes. No data on these factors was collected and so no comment can be made on their possible influence on infectivity of people's blood to mosquitoes. However, the fact that availability of anti-malarial drugs is much more tightly controlled in Thailand than in most malarious countries should be emphasized. From the results of the study, an attempt can be made to estimate the number of adults in the catchment area of the Clinic who were reservoirs of infection. On the basis of the number of patients visiting the Clinic per day and the catchment population from which the patients come (Figure 1 ), it was concluded that the main reservoir of infection for mosquitoes was not in adult patients feeling ill enough to be motivated to come to the Clinic. Among villagers, occurrence of fever is a strong indicator of likely malarial infection (Table 1 ). Thus, fever is an indicator of likelihood of being part of the infectious reservoir for mosquitoes. However, because there are far fewer people who are feverish than those who are not, the numbers of people in the infectious reservoir who are, or are not, feverish do not differ greatly (Fig. 1 ). The calculations in Figure 1 do not take into account the fact that the infected people found in the villages will remain infected for several days, whereas a new group of about nine people go to the Clinic every day. It would be useful to know for how many days people remain infectious to mosquitoes, but ethically one cannot leave detected infections untreated in order to test this. The data from Figure 1 indicate that directing an anti-gametocyte drug only to the clinic patients would be ineffective. Instead the drug would have to be targeted at the village populations after mass surveys for parasitaemia. Feverishness would assist to some extent in helping to identify people most likely to be infected. In view of the current interest in anti-gametocyte drugs [ 28 - 30 ], these data may be of use in deciding how such drugs would have to be targeted to have an impact on transmission. Figure 1 Diagram to show location of the reservoirs of infection (* data from the surveys, ** data from the Vector-borne Disease Control Clinic No. 8, Mae Hong Son Province) Authors' contributions Aree Pethleart led the data collection team in the field and laboratory, analysed data and drafted the paper; Somsak Prajakwong provided research facilities; Wannapa Suwonkerd organized mosquito rearing and dissection; Boontawee Corthong was in charge of Clinic data and provided facilities for village surveys; Roger Webber was responsible for early planning of the work; and Christopher Curtis supervised data analysis and edited the draft.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526384.xml

539298

On-call work and health: a review

Many professions in the fields of engineering, aviation and medicine employ this form of scheduling. However, on-call work has received significantly less research attention than other work patterns such as shift work and overtime hours. This paper reviews the current body of peer-reviewed, published research conducted on the health effects of on-call work The health effects studies done in the area of on-call work are limited to mental health, job stress, sleep disturbances and personal safety. The reviewed research suggests that on-call work scheduling can pose a risk to health, although there are critical gaps in the literature.

Background The question of whether work hours and schedules affect people's health has been reviewed for a range of work patterns including shift work and overtime. Research in these areas indicates that shift work, and in particular night work can interrupt sleep patterns [ 1 ], aggravate existing medical conditions and increase the risk of cardiovascular, gastrointestinal, and reproductive dysfunctions [ 2 - 4 ]. However, the health effects of on-call work, where workers are called to work either between regular hours or during set on-call periods, has not merited as much attention. This form of work scheduling occurs in a variety of diverse occupations, for example medical technologists, doctors, ship engineers, utility workers, electrical technicians, tug boat pilots, midwives, information technologists, media personnel and junior airline pilots. For many of these professions being on-call is not an option, but rather a component of the job. This form of scheduling is often used to provide 24 hour coverage, 7 days a week, for facilities such as hospitals and laboratories, where emergencies require personnel to immediately deal with critical situations and where the volume of evening and weekend work does not necessitate full shift coverage. Having employees on-call, even if they are being paid a stipend for their call time, is often seen as less expensive for employers than providing full shift coverage during off-peak hours [ 5 ]. While on-call work scheduling may be less expensive, it is not without human costs. On-call employees must plan their lives and the lives of their families around a call schedule. This often means limiting behaviours and obliges employees to restrict their on-call time to activities that would not interfere with their ability to work. The unpredictability of the call scheduling can also generate a great deal of stress, as home life is interrupted and workers are required to "change hats" to shift to their professional roles at any time during call. These limitations and interferences present unique challenges for on-call workers that are not encountered by those working set schedules or even people with rotating shifts. It is thus not surprising that researchers have found that on-call work patterns can have a major influence on employees' lifestyles and their interactions with family members and friends [ 6 ]. However, in addition to the impact on lifestyle and relationships, on-call work patterns may impact the health of employees. Within the limited literature that has explored on-call work, there exists some pertinent findings concerning the impact of on-call for an employee's physical and psychological health, and social relationships, which this review seeks to bring together. Specific attention has been devoted to the areas of stress, sleep, mental health and personal safety. Types of on-call work The implementation of on-call schedules varies. For many occupations, workers leave their place of employment and are placed "on-call" on evenings and weekends, which means they can be called back to work during these periods. For many professions this form of scheduling is a normal component of the occupation, for example, marine pilots can spend up to 60% of their working time on-call. However, for a limited number of occupations such as airline pilots, on-call hours are reduced with seniority. Generally, but not always, employees are compensated monetarily for the period of call, usually with a stipend which is less than their hourly rate. When on-call employees are usually expected to restrict their use of alcohol and limit distance or travel time from the work-site. The on-call experience of these workers includes aspects of interruption, either of sleep or family or social life, and often includes an element of uncertainty as to the time of call or the occurrence of the call. Other forms of on-call include work done by junior doctors during their medical training. Medical residents spend periods of time "on-call" at a hospital, where space may be provided for them to sleep. This form of on-call work is distinct because workers remain at work to undertake their call duty. During these periods, residents often put in 30–36 hour shifts with little to no sleep [ 7 ], resulting in a combination that is both a night shift and an overtime shift. Because of the intensive demands placed on medical residents during their apprenticeship, this group has received a fair amount of research attention. This has been particularly so in the 1990s as the rigors of this period in junior doctors' training has come under much scrutiny both in the UK and in the US. New working regulations have been introduced in an attempt to deal with what is considered, by many, to be harsh and unacceptable working conditions. The debate over and outcome of these interventions continues [ 8 - 10 ]. This review focuses on the health effects of on-call work in which an employee spends a period of time on-call outside of their workplace and/or their regular working hours. The research on medical residents has been excluded because the medical resident experience is distinctly different from that of other professions where on-call is utilized. However, research on medical residents is used to illustrate findings from other work-related areas when appropriate. Methods This review explores the published literature referring to on-call work patterns and health. For the purpose of this review, the on-call period may be formal (e.g. a person is designated as being on-call for the weekend or overnight) or informal (emergency call back during a crisis). Search terms for this review included "on-call" and "work schedule tolerance". The terms "stand-by" and "night call" are used by some professions to describe on-call type work, and were also used as search parameters. This literature review was undertaken on journal articles included in databases up to December 2000. Database searches were performed on the following electronic sources: 1) OVID Databases: including Medline (1966–2000) and Current Contents (1996–2000). 2) Canadian Centre for Occupational Health and Safety Database 3) Cambridge Abstracts (Environmental Science and Pollution Management) 1981–2000. 4) PsycInfo (1989–2000). 5) Web of Science: including Science Citation Index and Social Science Citation Index (1989–2000). A manual review of the references generated from the computer-search was also done. Articles were excluded from the review if they were not original research, were not written in English or focused on medical residents experiences with on-call work. Two reviewers read through each of the eligible research papers independently. Results In total, 24 papers met the search criteria. Eight (8) were excluded as they focused on the impact of on-call work patterns on patient's health and not on the health of workers. The remaining sixteen studies were used for this review. The results are divided into four health-related sections; 1) Stress, 2) Sleep, 3) Mental Health and 4) Personal Safety. On-Call Work and Stress Of the five studies pertaining to on-call and stress uncovered in this review, all focus on the General Practitioners (GPs) as their subject. In these studies, the relationship between on-call work and stress was measured through self-report and perceived stress. Three of the studies were part of a major UK study carried out from 1989 to 1998 [ 11 - 13 ]. In the early 1990s the British health care system experienced considerable financial and administrative restructuring. This large study was conducted at different points in time to determine GP's satisfaction with the changes in their workplace. GPs were randomly selected throughout Britain in 1987, 1990 and 1998 to fill out postal questionnaires. The studies yielded sample sizes of 1817, 917, and 999 respectively, representing rather low response rates of 48%, 67% and 47%. However, the authors' assessment of all three samples found that they tended to be fairly representative of the larger population of GPs in the country [ 13 ]. In the first two studies, GPs ranked working on-call at night as one of the top two most stressful aspects of their work situation [ 11 , 12 ]. However results from the third study in 1998 revealed that night call was no longer a major source of stress, dropping to 12 th in a ranking of 14 major stressors. The authors believe this reduction in the level of stress from on-call work could be explained by the introduction of GP co-operatives in the mid 1990s for the management of out-of-hours calls. This cooperative system allowed GPs to either do their own calls or share them with a cooperative formed by 10 or more doctors. The cooperatives gave GPs greater flexibility for how and where they saw their patients and how they implemented 24-hour care and appear to have successfully reduced the stress of night visits for GPs. Indeed, night visit stress went from being one of two top stressors for GPs in 1987 and 1990 to being one of the least stressful issues by 1998. The authors also posit that this "success may also explain the reported reduction by 1998 in stress attributable to disturbance of home/family life" [[ 13 ] pg. 370]. The fourth study also dealt with the changes in the British health care system, in particular the introduction of partial shifts to decrease long on-call periods [ 14 ]. A small sample of GPs'(n = 14 and 12) were surveyed about their stress levels before and after the new system was in place. Doctors' stress levels were significantly reduced, particularly in relation to their mental well-being and their job satisfaction. The fifth study on GPs and stress was a qualitative analysis of 25 GPs and their spouses in Manchester [ 15 ]. This research found that for male GPs, the uncertainty of being on-call caused them to be unhappy. Some doctors spoke quite frankly about how night calls could "perturb family life and wreck personal intimacy" [[ 15 ] p. 158). The uncertainty of their on-call commitments also contributed to the male GPs' unhappiness. Female GPs were stressed by factors other than on-call, including time pressure, role conflict and work overload. They were also concerned about how their work schedule decreased the amount of time they spent with their children. These marked differences between how male and female doctors experience the stress of on-call work signals the importance of examining gender as a variable in this research. Other studies have revealed that the amount of time spent on-call varies between male and female doctors, but no clear pattern has emerged [ 16 , 17 ]. It has been hypothesized that female doctors who work reduced on-call hours do so because of the dual role they must play as both worker and care-giver [ 17 , 18 ]. Research conducted in other professions support the idea that work patterns, particularly night shifts, can increase stress in workers and have a negative impact on family life. Working late afternoon and evening shifts has been related to increased stress for both workers and their families [ 2 ]. Variable shifts have been shown to cause more stress than regular shifts [ 19 ] and working more than 50 hours per week is associated with increased job stress [ 20 ]. Many on-call workers regularly experience variation in their work patterns, as well as being expected to work at night, and undertake greater than normal hours when called in. On-Call Work and Sleep Besides stress, the interruption of sleep is another major component of on-call work, particularly for those who work nights on-call and in professions that deal with emergencies that occur at all hours. Three studies have dealt specifically with the sleeping patterns and problems experienced by train and ship engineers and transplant coordinators, all of whom regularly work on-call. The first study researched the on-call sleep patterns of 198 train engineers using prospective activity logs over a 14-day period in the United States [ 21 ]. It was determined that those working on-call had greater difficulty falling asleep and staying asleep while on-call versus when they were not on-call. Train engineers working on-call also had a greater number of days where there was less than 24 hours between the on-set of their work shifts. These engineers reported more sleep-related problems that those with at least 24 hours between the on-set of their shifts. The researchers also explored how sleeping was impacted when it was undertaken in different locations. They found that train engineers sleep varied when at home versus "away". (Engineers can finish a shift away from home, and have "away" terminals where they can sleep.) The researchers compared the amount and quality of sleep engineers had while both "at home" and "away" and found that engineers on-call slept less at home than they did "away". The authors attribute the difference to the presence of family and social obligations in the home that conflicted with the workers' ability to sleep while working on-call. However, the authors note that the response rate of this study was low, only 25% of the sampled population of approximately 800. The authors caution readers to remain critical of their findings, because their sample may be biased towards those who generally have difficulty sleeping. An analysis of the final study group did find that the responding sample reflected the age and gender distributions of the larger population, factors that the authors suggest indicate robustness even with the low response rate. The second study of on-call and sleep explored the sleeping patterns of 53 predominantly female organ transplant coordinators in the UK, using a postal questionnaire [ 22 ]. This research determined that not only was sleep affected when people worked on-call (51% occasionally had difficulty and 6% frequently had difficulty falling asleep) but that the effects carried over to time off call as well. Sixty-eight percent of the sample reported that the time they spent on-call negatively influenced their off-call lives. Workers pointed out that after being on-call they often had to spend additional time catching up on sleep. They also complained that on-call work left them too tired to undertake social and home activities. But although the workers complained about being fatigued at home, this was not correlated with days absent from work. The authors suggest that this finding may be the result of transplant coordinators "guilt" around placing an extra burden on a co-worker if they were absent. Another possible explanation was the overall satisfaction of the type of work being done by the coordinators, a factor which may decrease their willingness to take time off. The third study, conducted on a small sample (n = 5) of ship engineers in Sweden, measured sleep during on-call periods using electroencephalogram (EEG) and electrocardiogram (ECG) recordings and subjective ratings. [ 23 ]. This research found, like the others, that the sleep quality and quantity of the ship engineers was affected by the interruptions of being on-call. In their subjective assessments, the engineers reported being more drowsy during the day after being on-call, a finding similar to that of the transplant coordinators. But, the authors also found that the apprehension associated with the possibility of being awakened for call duty also negatively impacted sleep. On-call sleep registered less slow wave sleep (SWS) and rapid eye movement (REM) and a higher heart rate than when workers were testing during their normal sleep. Many of these conditions occurred prior to being awakened for call duty. Earlier research by the same authors examined the sleep patterns of Swedish merchant marines at sea. This population also found it difficult to fall asleep on nights when they were on watch. The anticipation of alarms that would wake them up was seen as an obstacle that prevented workers from relaxing enough to allow for normal sleep patterns to develop [ 24 ]. The impacts of sleep loss on job performance remain unclear and controversial. For example, research on the cognitive performance in sleep deprived medical residents has produced mixed results [ 25 - 27 ]. However, research on anaesthetists found that 86% reported fatigue related errors [ 28 ]. Job performance and fatigue have also been studied in relation to age, a factor not explored in the on-call studies. Significant changes were found between younger and older shift workers, with younger workers better able to maintain performance across day and night shifts and older shift workers prone to more sleep disruption [ 29 ]. Work-related fatigue has been related to an increase in car accidents. A review of traffic accidents determined that falling asleep while driving accounted for a major proportion of accidents while driving under monotonous conditions [ 30 ]. This finding has been corroborated with research done on medical residents working long night shifts. Seventy-five percent of accidents incurred by a population of emergency medicine residents happened after working a night shift [ 31 ]. In this study, the number of motor vehicle accidents and near misses was positively correlated to the number of nights worked per month. A similar study done on paediatric residents indicated that residents fell asleep at the wheel significantly more than other professionals, with 90% of these events occurring after a night on-call [ 32 ]. On Call Work and Mental Health Six studies were found that examined the impact of on-call work schedules on mental health. All of these studies used self-reported questionnaires and/or mood diaries. Five studies were conducted on GPs in the UK and one examined gas and electrical employees in France. Two surveys were conducted by Chambers et al. [ 33 , 34 ] on GPs in Staffordshire, UK. The first survey, conducted in 1994 (n = 704), was designed to research the factors predictive of anxiety and depression in GPs [ 33 ]. The study determined that working one or more nights on-call per week was significantly predictive of anxiety. Other factors predictive of anxiety were depression and three or more weekdays feeling exhausted or stressed. Males and females showed no significant differences in anxiety or depression determinants. The second survey conducted by Chambers et al in 1996 (n = 620) employed the Hospital Anxiety and Depression scale to assess the mental health of GPs [ 34 ]. It was determined that both anxiety and depression were associated with the amount of on-call duties undertaken. Findings revealed that both anxiety and depression increased with the frequency of time spent on-call per month. Again, the results were the same for both male and female GPs, and the authors conclude that GPs' mental ill health is associated with workload, of which on-call is a major factor. A third survey done on GPs in Leeds in 1993 was designed to determine the psychological symptoms and sources of stress among 268 GPs [ 35 ]. This survey used the UK General Health Questionnaire as well as qualitative questions regarding mental health and workload. Problems with physical and mental health were significantly associated with several aspects of workload, including the amount of time spent on-call per month. The study also found that those GPs who spent more time on-call each month were more likely to feel their work affected their physical health. Males and females reported differences in the sources of their stress, with females showing greater job satisfaction than males. The authors suggest that this finding may be due to the fact that, for this study population, female doctors worked fewer hours and spent significantly fewer nights on-call [ 34 ]. The fourth study in this area surveyed mental health and job stress on 414 GPs in England in 1992 [ 36 ]. This research determined that interruptions, a category which included taking night calls, remaining alert on-call, 24 hour patient responsibility and telephone interruption of family life, was a predictive factor for decreased mental health, depression and somatic anxiety. These factors were similar for men and women, although their contribution to each condition varied by gender. A pilot study of 44 male and female volunteer GPs using cognitive behavioural diaries assessed self-reported emotional states recorded in conjunction with hourly activities over 2 days [ 37 ]. Doctors' moods were significantly lowered when on-call as compared to off-call. Doctors on-call also had significantly increased tension and frustration. The main reported cause of dissatisfaction was the uncertain nature of the doctors working hours [ 37 ]. The sixth study examined male gas and electrical employees working in France [ 38 ]. Employees who worked on-call (n = 145) were assessed for health status and psychological problems and were compared to those not working on-call (n = 195). Workers were also questioned about the impact of their job on their family life. Although no particular mental or health disorder was found to be more frequent in the on-call group, the psychological equilibrium of the on-call workers was significantly worse than the comparison group. On-call workers also reported significantly worse global-well being and indicated significantly higher levels of social disturbance. On-call workers reported that their family and social life were acutely disturbed and they were significantly less likely to be involved in clubs or take on outside responsibilities. The research conducted on GPs in the UK supports a negative role of on-call work related to mental health. However, the results from the gas and electrical workers do not reflect the same findings from the research on GPs. This may be the result of either a difference in study methodology or a difference that is profession-specific. On-call gas and electrical workers did experience psychological disruption and the lack of significant diagnostic findings may be a function of other factors, such as self-selection, in this profession, where those most affected opt out early on. The on-call gas and electrical workers experience of family and social life disruption does mirror the experiences of doctors and transplant coordinators as discussed previously [ 13 , 15 , 22 ]. On-call Work and Personal Security Working on-call often necessitates leaving home alone, at night, to attend work, conditions that can jeopardize personal safety. Unfortunately, there is only sparse data regarding this issue. A study done in the north west of England, in a hospital where the on-call sleeping quarters were separate from the hospital found that 40% of anaesthetists feared for their safety while walking through hospital grounds at night [ 39 ]. In medical professions, patients can also present a danger to those working on-call. Doctors have cited fear of violence from night call visits as a significant stressor [ 11 ]. A study of 327 nurses in remote areas who worked on-call found increased incidences of violent acts perpetrated by patients, particularly in smaller communities [ 40 ]. This study found that working on-call increased the number of incidents ranging from verbal abuse to property crime and physical assault compared to working regular shifts. This issue has only been peripherally studied and further attention needs to be given to personal safety, particularly when being called in at night. Discussion What emerges from this review is the limited research that has been done in the area of on-call work. Preliminary work done in the areas of stress and mental health suggests that on-call work may play a role in increasing stress and decreasing mental well-being. The three studies that examined sleep indicate that on-call work does decrease the quality and quantity of sleep for workers and can leave people feeling fatigued for periods after their on-call work. The current body of literature on the health effects of on-call work is limited in part due to the narrow range of professions studied. The majority of research done to date has been on general practitioners. It is reasonable to assume that the effects of on-call will vary across occupations, given the host of other factors that can influence occupational health. However, the degree to which this variation exists might only be determined by examining a wider occupational base. The need to undertake more on-call research across a greater variety of occupational groups is suggested given that this form of work scheduling touches many occupations, and given that on-call work is estimated to continue to increase in many sectors in the future [ 6 ]. There is also an obvious lack of research focusing on the impact of on-call shifts on psychosocial factors. Given the very disruptive and limiting nature of on-call schedules, it would not be surprising that workers' family and social life suffer due to this type of scheduling. The results of the research addressing gender (discussed above) do suggest, albeit indirectly, that such social and familial impacts may be significant. However, without more research, it is not possible to determine the magnitude of these effects, nor the relative importance compared to other factors such as physiological responses. More rigorous methodological designs are needed for future research in the area of on-call work and health. The current research is predominantly cross-sectional in nature, a factor that makes it difficult to determine causality. Only two studies employed external comparison groups [ 21 , 38 ] and only a limited number have measured effects in workers on-call versus off-call (own-controls) [ 22 , 23 ]. Additionally, most of the measurement has been subjective in nature and often the operationalization of on-call work is not clear. In the GP studies, on-call is generally measured as the "number of nights spent on-call" either per week or per month. Some attempt is made to measure the amount of sleep during these periods, but there is little refinement of factors such as whether the subject were actually called in to work and for how long. Additionally, little attention has been paid to the amount of time worked or sleep obtained prior to the on-call shifts or factors such as second jobs or outside work, variables that may confound the outcomes. Other factors, such as age and personality type, that have been shown to be significant variables in other areas of work scheduling [ 41 , 42 ] also need to be explored. Attention also needs to be paid to the possible self-selection of workers out of on-call professions or adaptive strategies that workers may employ to cope with on-call (such as the sharing of on-call shifts). More controlled research that includes both subjective and objective measures would provide better evidence regarding the effects of on-call work. Future research on the health effects of on-call work also needs to examine the role of gender, not only from a physiological standpoint, (e.g. reproductive issues), but also from a psychosocial perspective. Many of the articles reviewed above indicate differences in how males and females experience the stress of on-call work [ 11 , 15 , 17 , 36 ]. Research in other work-related areas suggests that males and females cope differently with the impact of job schedules [ 43 - 45 ]. While gender may be a factor that directly mediates health effects, it may also be an indirect measure of other phenomena such as the division of labour outside of the workplace. More careful research is needed to illuminate the role gender may play in the effects of on-call work. The range of health effects studied in relation to on-call work has to date been inadequate. Health conditions such as cardiovascular disease, reproductive problems, gastrointestinal issues and overall mortality need to be explored as has been done in conjunction with work patterns such as overtime and shift work [ 41 , 45 ]. Factors such as personal safety and car accidents have only briefly been touched upon, and merit more attention. Conclusions While the results of this review are limited, initial research in this area suggest that being on-call can have negative impacts on workers' sleep patterns, mental health and personal life. Further research in this area is required to provide a clear picture of the risks of this form of work scheduling. List of abbreviations UK, United Kingdom US, United States of America Competing interests The authors declare that they have no competing interests. Authors' contributions AMN designed the research project, carried out the literature search, reviewed articles and drafted the manuscript. JSB reviewed articles and edited the manuscript. Both authors approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539298.xml