| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Endocrinology, Metabolism, and Molecular Medicine (M.U., A.D.), Northwestern University Medical School, Chicago, Illinois 60611; Department of Genetics (A.W., K.G.E., R.S.S.) and Center for Research on Reproduction and Women’s Health and Department of Obstetrics and Gynecology (J.F.S.), University of Pennsylvania, Philadelphia, Pennsylvania 19104; Endocrine Section, First Department of Internal Medicine (E.D.-K.), Athens University School of Medicine, Athens, Greece; and Department of Obstetrics and Gynecology (R.S.L.), Pennsylvania State University, Hershey, Pennsylvania 17033
Address all correspondence and requests for reprints to: Dr. Margrit Urbanek, Division of Endocrinology, Metabolism, and Molecular Medicine, Northwestern University School of Medicine, 303 East Chicago Avenue, Tarry 15-717, Chicago, Illinois 60611. E-mail: m-urbanek{at}northwestern.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of this study was to duplicate a previous study that localized a PCOS susceptibility region to chromosome 19p13.2 and to narrow the susceptibility region.
Design: This study was designed to test for genetic linkage and association between PCOS and short tandem repeat polymorphisms in 367 families, by analysis of linkage and family-based association.
Setting: The study was conducted at academic medical centers.
Patients or Other Participants: We studied 367 families of predominantly European origin with at least one PCOS patient. Families included 107 affected sibling (sister) pairs (ASPs) in 83 families, and 390 trios with both parents and an affected daughter. The data set comprises two independent groups. Set 1 consists of 44 ASPs and 163 trios. Set 2 consists of 63 ASPs and 227 trios.
Intervention(s): The intervention was the drawing of blood for DNA extraction.
Main Outcome Measure: We employed measures of evidence for linkage and association between PCOS and 19 STRs.
Results: Linkage with PCOS was observed over a broad region of chromosome 19p13.2. The strongest evidence for association was observed with D19S884 (
2=11.85; nominal P < 0.0006; permutation P = 0.034) and duplicated our earlier findings.
Conclusions: The present analysis suggests that a PCOS susceptibility locus maps very close to D19S884. Additional studies that systematically characterize DNA sequence variation in the immediate area of D19S884 are required to identify the PCOS susceptibility variant.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Several studies have found evidence of familial aggregation of PCOS, supporting a genetic contribution to its etiology (16, 17, 18), and more than 50 candidate genes have been considered or studied (reviewed in Ref.19). However, there is no general agreement on mode of inheritance, and PCOS is usually considered not to have a monogenic basis. Furthermore, it has not been possible to establish conclusively the role of any particular gene or region. The lack of progress probably reflects in part the difficulties recognized in analysis of complex genetic diseases, including heterogeneity of the PCOS phenotype, the likely contribution of multiple genes, and the uncertain role of the environment. In our initial analysis, we used affected sibling (sister) pairs (ASPs) to test 37 candidate genes for linkage with PCOS. We used the transmission/disequilibrium test (TDT) to test the same genes for the combined presence of linkage and association (linkage disequilibrium, LD) with PCOS (20, 21). Among the markers tested in the initial study, the strongest evidence for LD was seen at D19S884 (nominal P = 0.004), a dinucleotide repeat marker closely linked to the insulin receptor gene (INSR) on chromosome 19p13.2. These initial findings have been followed up by several other investigators, testing for association by means of case-control studies. Tucci et al. (22) tested 10 of the STRs we described (20) for association with PCOS in a group of 85 Caucasian PCOS patients and 87 age-matched Caucasian control women; they found evidence for association only with D19S884 (P = 0.006). (In this and other published studies, because the number of CA repeats in the associated allele was not given, it is not clear whether the associated allele is the same as the one we reported, now known to contain 17 CA repeats.) Another study (23) tested for association between PCOS and D19S884 in samples of Italian and Spanish women. When analyzed separately, the two samples showed nonsignificant differences (P > 0.05), in opposite directions, between cases and controls. In the combined data, the difference was also not significant (P > 0.05).
When material from parent-offspring trios is available for study, the association testing can be carried out with the TDT. In the present study, we have used this family-based test in a large sample of families to investigate the role of chromosome 19p13.2 in the etiology of PCOS in greater detail. To reduce as much as possible some of the problems often encountered with complex genetic diseases, we have also ensured that uniform diagnostic criteria were used throughout, and employed a narrow and objective definition of "affected." We first repeated our original studies of linkage and association in an independent set of 217 families and used the TDT for the analysis of LD and confirmed the association. The combined sample of 367 PCOS families is the largest studied to date. Then we refined the mapping of the putative susceptibility determinant by testing eight additional STRs located within 500 kb of D19S884. These studies provide strong, reproducible evidence for a PCOS susceptibility locus mapping to chromosome 19p13.2, at or near the dinucleotide repeat marker D19S884.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
We studied 1542 individuals belonging to 367 families with at least one PCOS patient. Most (339) families were of European origin; 13 Hispanic, six African-American, two American-Indian, and seven of unknown ethnic origin. Both parents were available in 308 families. One or neither parent was available in 51 and eight families, respectively. There were 107 ASPs in 83 families, and 390 trios with both parents and an affected daughter. The data set is comprised of two independent groups of families: set 1 and set 2. Set 1 corresponds to the families used in our initial candidate gene screen (20). With some additional sisters ascertained for the present study, set 1 consists of 150 families and includes 44 ASPs and 163 trios. Set 2 consists of 217 families, including 63 ASPs and 227 trios. The study was approved by the appropriate institutional review boards, and informed consent was obtained from the subjects. Ascertainment and phenotyping of individuals in sets 1 and 2 was identical.
Phenotyping
The diagnostic criteria followed in this manuscript meet both the National Institute of Child Health and Human Development (24) and Rotterdam (25) criteria and are described by Legro et al. (18). To qualify as a PCOS index case, a woman had six or fewer menses per year in addition to HA defined by either total testosterone (T) greater than 58 ng/dl (2 nmol/liter) and/or non-SHBG-bound T (unbound T) greater than 15 ng/dl (0.5 nmol/liter). These levels are greater than two SDs above the mean value that we have established in reproductively normal women aged 18–40 yr in the early follicular phase of the menstrual cycle (18). Other causes of anovulation and HA were excluded by the following laboratory tests (18, 24). All PCOS probands had prolactin levels 25 ng/ml or less. Nonclassical congenital adrenal hyperplasia was excluded with a morning 17-hydroxy progesterone level less than 3 ng/ml or an ACTH-stimulated 17-hydroxy progesterone level less than 1000 ng/dl. Ovarian ultrasound status was not considered for the diagnosis.
HA is a salient and unambiguous biochemical feature of PCOS and is significantly increased among sisters of PCOS patients, even in the absence of irregular menses (18, 26). For genetic analysis, therefore, female relatives of index cases were considered affected if they had elevated androgen levels, regardless of their menstrual status (18). As described before (20), we used "PCOS/HA" to describe this combined category. Women were considered unaffected if they had normal circulating androgen levels, had regular menstrual cycles (menses every 27–35 d) (18), and were not taking any confounding medications, e.g. oral contraceptives or insulin-sensitizing agents. The vast majority (89%) of affected women have the diagnosis of complete PCOS. All 367 probands and 55 of 107 affected sisters have PCOS, whereas 45 of the sisters are hyperandrogenic with regular menses (menses every 27–35 d) and seven sisters are hyperandrogenic with intermediate menses (menses every 35–60 d). Women not of reproductive age and those not fulfilling the criteria for affected or unaffected phenotypes were assigned the phenotype "unknown" (18). Because no male phenotype corresponding to PCOS is known, all men in the study were assigned the phenotype "unknown".
Genotyping
Genotypes were determined at 19 short tandem repeat (STR) polymorphisms mapping to the INSR region of chromosome 19p (Table 1
and Fig. 1). Fourteen of these markers were found in the GenBank database. The remaining five (STR1, STR2, STR5, STR6, STR8) were identified by electronically screening 250 kb of genomic sequence on each side of D19S884 with either (CA)12 or (GT)12.
|
|
Statistical analysis
We tested for linkage and LD between polymorphic markers and PCOS/HA in our families using identity by descent (IBD) methods in ASPs (27) and the TDT (28), respectively. These analyses were carried out as described previously (20) and were checked by hand for D19S884. Briefly, we calculated percent IBD for each ASP and adjusted for different sibship sizes using the method of Suarez and Hodge (29). When necessary, parental genotypes were reconstructed from the genotypes of unaffected siblings or siblings with unknown phenotype. The genotypes of unaffected siblings used for parental genotype reconstruction were not included in any additional analyses. The TDT analysis was carried out on 17 STRs mapping in the region with the strongest evidence of linkage to PCOS (Fig. 2). We restricted analysis in most cases to the 100 alleles with 50 or more possible transmissions (transmitted plus not transmitted); exceptions are noted. Except where noted, the probability values reported are nominal (not corrected for multiple tests). However, the large number of tests leads to a type I error rate much greater than the nominal significance level. We used the PERM1 function of the GENEHUNTER (http://www.fhcrc.org/science/labs/kruglyak/downloads) version 2.1 suite of programs (30) to assess the statistical significance of the TDT findings by a permutation test. Briefly, the PERM1 function was used to generate 10,000 new data sets by randomly switching the transmission status from each informative parent for all 17 STRs. The TDT was calculated for all alleles, and the corresponding empirical significance level for the actual TDT findings was determined. The permutation analysis assumed that the 17 STRs segregate independently. However, some of the markers tested may be in LD, so the assumption of independence may overcorrect for multiple testing and true probability values may be smaller than those derived from permutation.
|
It is formally possible that preferential transmission detected by the TDT is the result of segregation distortion associated with a marker allele, not specific to affected offspring, undermining the conclusions drawn from the TDT. To test the null hypothesis of equal transmission of alleles in the general population, we genotyped the markers that showed the strongest evidence for association with PCOS/HA (D19S884 and D19S922) in a set of 40 CEPH (Centre d’Etude du Polymorphisme Humain) families consisting of 388 individuals and we tested each allele for unequal transmission. Both parents were typed in most families, whereas eight families were missing one parent’s genotype. The PCOS/HA status was unknown and for the TDT analysis we counted transmissions from heterozygous parents to all offspring.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
We describe the ASP analysis very briefly. In both set 1 and set 2 families, there was nominally significant evidence (P < 0.05) for linkage of PCOS to the INSR region shown in Fig. 1
. In set 1, the maximum IBD (66%) occurred at D19S905 (2 = 5.02; 54 transmissions from a heterozygous parent, P < 0.025). In set 2, the maximum IBD (64%) occurred at INSR (2 = 4.25; 53 informative transmissions, P < 0.039). We consider this finding modest follow-up support for the set 1 results with regard to linkage. For the combined data, the maximum IBD was 61%, located at D19S865 (130 informative transmissions, 2 = 6.32, P = 0.012). Results for linkage (data not shown) for all markers in Fig. 1 show that IBD is approximately 60% (59–61%) over a central region bounded by INSR (telomeric) and D19S840 (centromeric), a span of 6.6 Mb. In this region, all markers show evidence for linkage with PCOS at or beyond the nominal P = 0.05 significance level. Outside this region, the markers show no significant elevation of IBD; we found IBD of 55% at D19S216 (telomeric), and 51% and 48% at D19S410 and D19S212, respectively (centromeric). The evidence described above for linkage with PCOS/HA on chromosome 19p13.2 (P = 0.012) is stronger than that with any of the other 18 independent candidate regions tested in our complete set of ASP families (data not shown). Note that this probability value is not adjusted for the 19 independent regions. It should be noted that approximately 48% of the affected sisters for the linkage analysis are hyperandrogenic but do not fulfill the criteria of six or fewer menses per year. However, as described by Legro et al. (18), we believe that the fundamental underlying genetic determinants are the same for PCOS and HA in PCOS families.
Analysis by TDT.
Like the IBD results, the TDT results are strengthened by the finding of similar results in set 1 and set 2 data separately. For the separate set 1 and set 2 analyses, alleles with 30 or more informative transmissions were included, instead of 50 required elsewhere; because the set 2 data were available for follow-up, we chose to decrease the chance of false negative findings, even at the risk of initially increasing false positives. In set 1, the strongest evidence for LD with PCOS/HA was seen at D19S884, where three alleles (A5, A7, and A8) were nominally significant (Table 2). Only two other markers had alleles (D19S216 A8 and INSR A13) nominally significant in the 150 families of set 1. In the 217 set 2 families, there were four markers with evidence significant beyond the P = 0.05 level (Table 2). The only significant TDT effect from set 1 that was replicated in set 2 (P < 0.01) was for D19S884 A8 (Table 2). D19S884 A8 has a frequency of 0.21 in 152 Caucasian CEPH parental chromosomes and consists of 17 CA repeats.
|
). By far the strongest evidence for association was with D19S884 A8 (2 = 11.85, P < 0.00058). After Bonferroni correction for having tested 100 alleles, the probability value for D19S884 A8 (P = 0.058) did not quite achieve conventional significance. However, in our permutation analysis, 338 of 10,000 tests had a maximum 2 value larger than 11.85 (P = 0.034). None of the other four markers shows strong evidence in both sets 1 and 2, and the evidence for these in the combined data is much weaker than that observed for D19S884 A8 (Table 2). Finally, the degree of transmission distortion that we observe with D19S884 A8 is remarkably consistent in the two data sets: 54 of 88 transmissions (61.4%) in set 1, 91 of 149 transmissions (61.1%) in set 2, and 145 of 237 transmissions (61.2%) in combined data. The sisters with HA and regular cycles contribute only a few transmissions to the association study. For instance, for D19S884 A8 the HA sisters added only 13 informative transmissions to the analysis. The TDT results are almost identical for the PCOS-only data set (n = 224; 2=11.16) and the complete data set (n = 237; 2=11.85). Excess transmissions were essentially identical in all three data sets (HA = 61.5%, PCOS = 61.2%, all = 61.2%).
Additional mapping of D19S884 region.
Figure 2 shows the results of refining our analysis from the STRs shown in Fig. 2A to a smaller region centered on D19S884 (see Figs. 1B and 2B). Although our focus was originally on INSR, the strongest TDT evidence for association with PCOS/HA is with D19S884, located approximately 800 kb centromeric to INSR. Because disequilibrium, the theoretical foundation for the TDT, is not generally observed over such large distances (33, 34, 35), it is unlikely that the association with D19S884 is due to a causal variant at INSR itself. We therefore tested alleles of eight STRs mapping to within 500 kb of D19S884 for association with the phenotype PCOS/HA in the total data set. The results are shown in Fig. 2B. The strongest evidence for association is still observed with D19S884 A8.
PDT analysis.
The PDT (31, 32) was used to provide a valid test for association in multiplex families. To eliminate the testing of multiple alleles, the D19S884 genotype data were reduced to a biallelic system (A8 and non-A8). Using the PDT, the association between D19S884 A8 and PCOS remains significant (2 = 5.41, P = 0.020).
Transmission of D19S884 A8 in controls without PCOS.
To determine whether D19S884 A8 is preferentially transmitted even in the absence of PCOS/HA, we genotyped 40 CEPH (control) families with no known disease phenotypes. None of the alleles at D19S884 or D19S922 showed significant evidence for transmission distortion in control samples (Table 3), after correction for testing 18 alleles. The results provide additional support for the conclusion that the association with D19S884 A8 is specific to the phenotype PCOS/HA.
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The present study was designed to minimize these problems. Susceptibility genes for complex diseases are expected to have relatively weak effects (36, 37). The 367 PCOS families in this study constitute the largest sample in any published genetic analysis of PCOS. Such a large sample can detect a relatively weak effect and reduce the sampling variation inevitable in small studies. Our approach allowed us to reduce potential heterogeneity within the sample in several ways. All individuals were diagnosed in the same manner and, with the exception of six families that include 16 affected daughters, all hormone assays were carried out by one central laboratory. This ensured that the phenotype assignment was consistent across the sample. A narrow phenotype, PCOS/HA, was used to reduce phenotypic heterogeneity. Although this phenotype does not span the entire spectrum of PCOS characteristics, it is expected to identify a subset of PCOS families that is more homogeneous than PCOS patients as a whole. Our sample is predominantly of European descent (>90%). Removal of the non-European families does not change the results of our analyses (data not shown).
D19S884 maps to chromosome 19p13.2 approximately 800 kb centromeric to the INSR. One plausible candidate PCOS susceptibility gene mapping to this interval is the gene for resistin located 420 kb telomeric to D19S884. Resistin is a protein hormone secreted by adipocytes that plays a role in insulin resistance and insulin action. For these reasons, we tested variation at the resistin gene for association with PCOS. However, we found no significant evidence for association with PCOS (38).
Genes that code for three proteins are known to map to within 100 kb of D19S884: ELAVL1, a ubiquitously expressed mRNA-binding protein; CCL25, a thymus-expressed chemokine; and FBN3, the third member of the fibrillin family of extracellular matrix proteins. Although none of these genes is an obvious candidate gene for PCOS, of course susceptibility genes cannot always be predicted a priori by their function, and (with some ingenuity) an argument can be made for each of these (ELAVL1, CCL25, FBN3). It is possible that association with PCOS detected at D19S884 is a proxy for association with variation at one of these genes, which plays a causal role. On the other hand, it is also possible that D19S884 is associated with a nearby regulatory element acting on a much more distant gene, such as INSR. Regulatory elements acting over distances even larger than 800 kb have recently been identified (39, 40). A final possibility is that D19S884 is itself the susceptibility locus. D19S884 is a dinucleotide repeat located in an intron of the FBN3 gene, 105 bp 3' to exon 55. Although STRs like D19S884 are generally believed to be essentially nonfunctional, several have been shown to be determinants of transcriptional activity or splicing. D19S884 itself may therefore regulate the transcription of a nearby gene or even of INSR. Although FBN3 is at first glance not a very plausible candidate gene for PCOS, it is possible that D19S884 regulates posttranscriptional processing of the FBN3 mRNA.
In the absence of a molecular genetic explanation for susceptibility, perhaps the most convincing evidence that a particular variant or gene contributes to a complex disease is that the findings are replicated; in PCOS, such findings have been difficult to obtain. The present results in independent families duplicate our original observations from a family-based screen of PCOS candidate genes (20). Consistent findings in two large independent samples using different genetic analyses (ASP analysis and TDT) constitute very strong evidence for a PCOS susceptibility locus mapping to chromosome 19p13.2. The present analysis of nearby markers suggests that susceptibility maps very close to the dinucleotide repeat marker D19S884. Additional studies will systematically characterize sequence variation in the region surrounding D19S884 and its role in PCOS.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Current address for A.W.: Department of Biostatistics, University of Michigan School of Public Health, Ann Arbor, Michigan 48109.
First Published Online August 9, 2005
1 A.D. and R.S.S. contributed equally to this work. 
Abbreviations: ASP, Affected sibling pair; CEPH, Centre d’Etude du Polymorphisme Humain; HA, hyperandrogenemia; IBD, identity by descent; INSR, insulin receptor gene; LD, linkage disequilibrium; PCOS, polycystic ovary syndrome; PDT, The Pedigree Disequilibrium Test; STR, short tandem repeat; TDT, transmission/disequilibrium test.
Received March 21, 2005.
Accepted August 2, 2005.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Pro12Ala polymorphism is associated with decreased risk of type 2 diabetes. Nat Genet 26:76–80[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  M.O. Goodarzi, J.F. Maher, J. Cui, X. Guo, K.D. Taylor, and R. Azziz FEM1A and FEM1B: novel candidate genes for polycystic ovary syndrome Hum. Reprod., December 1, 2008; 23(12): 2842 - 2849. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Simoni, C.B. Tempfer, B. Destenaves, and B.C.J.M. Fauser Functional genetic polymorphisms and female reproductive disorders: Part I: polycystic ovary syndrome and ovarian response Hum. Reprod. Update, September 1, 2008; 14(5): 459 - 484. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.O. Goodarzi, N. Xu, J. Cui, X. Guo, Y.I. Chen, and R. Azziz Small glutamine-rich tetratricopeptide repeat-containing protein alpha (SGTA), a candidate gene for polycystic ovary syndrome Hum. Reprod., May 1, 2008; 23(5): 1214 - 1219. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Kevenaar, J. S. E. Laven, S. L. Fong, A. G. Uitterlinden, F. H. de Jong, A. P. N. Themmen, and J. A. Visser A Functional Anti-Mullerian Hormone Gene Polymorphism Is Associated with Follicle Number and Androgen Levels in Polycystic Ovary Syndrome Patients J. Clin. Endocrinol. Metab., April 1, 2008; 93(4): 1310 - 1316. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Corton, J. I. Botella-Carretero, J. A. Lopez, E. Camafeita, J. L. San Millan, H. F. Escobar-Morreale, and B. Peral Proteomic analysis of human omental adipose tissue in the polycystic ovary syndrome using two-dimensional difference gel electrophoresis and mass spectrometry Hum. Reprod., March 1, 2008; 23(3): 651 - 661. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Urbanek, S. Sam, R. S. Legro, and A. Dunaif Identification of a Polycystic Ovary Syndrome Susceptibility Variant in Fibrillin-3 and Association with a Metabolic Phenotype J. Clin. Endocrinol. Metab., November 1, 2007; 92(11): 4191 - 4198. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Vigil, P. Contreras, J. L. Alvarado, A. Godoy, A. M. Salgado, and M. E. Cortes Evidence of subpopulations with different levels of insulin resistance in women with polycystic ovary syndrome Hum. Reprod., November 1, 2007; 22(11): 2974 - 2980. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. T. Bloomgarden Gut Hormones, Obesity, Polycystic Ovarian Syndrome, Malignancy, and Lipodystrophy Syndromes Diabetes Care, July 1, 2007; 30(7): 1934 - 1939. [Full Text] [PDF]
|
|
|
|
|
|
D. R. Stewart, B. A. Dombroski, M. Urbanek, W. Ankener, K. G. Ewens, J. R. Wood, R. S. Legro, J. F. Strauss III, A. Dunaif, and R. S. Spielman Fine Mapping of Genetic Susceptibility to Polycystic Ovary Syndrome on Chromosome 19p13.2 and Tests for Regulatory Activity J. Clin. Endocrinol. Metab., October 1, 2006; 91(10): 4112 - 4117. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. C. Florez Editorial: Genetic Susceptibility for Polycystic Ovary Syndrome on Chromosome 19: Advances in the Genetic Dissection of Complex Reproductive Traits J. Clin. Endocrinol. Metab., December 1, 2005; 90(12): 6732 - 6734. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
