| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Medicine, Division of Medical Sciences, Institute of Biomedical Research, University of Birmingham, Edgbaston, Birmingham B15 2TT, United Kingdom
Address all correspondence and requests for reprints to: Professor S. C. L. Gough, Department of Medicine, Division of Medical Sciences, Institute of Biomedical Research, University of Birmingham, Edgbaston, Birmingham, B15 2TT, United Kingdom. E-mail: s.c.gough{at}bham.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
B inhibitors SUMO-4 and MAP3K7IP2 and four of which (fcrl3_3, fcrl3_4, fcrl3_5, and fcrl3_6) modulate nuclear factor-B binding and production of the B cell surface molecule FCRL3, have been reported to be associated with a number of autoimmune diseases. Objective: The aim of this study was to investigate the association of these polymorphisms with disease in a large UK Caucasian Graves’ disease (GD) data set.
Design: The study was a case-control association study of six polymorphisms.
Setting: The study was conducted at a UK academic department of medicine.
Patients or Participants: Study population included 1056 GD patients and 864 controls.
Interventions: There were no interventions.
Main Outcome Measures: Tests for association with disease were measured.
Results: No association with disease was found for the M55V single-nucleotide polymorphism (SNP). Association was, however, found between GD and the 001Msp SNP [odds ratio (OR) 1.19 (95% confidence interval [CI]) 1.03–1.37], fcrl3_3 SNP [OR 1.17 (95% CI 1.02–1.34)], fcrl3_5 SNP [OR 1.18 (95% CI 1.04–1.35)], and fcrl3_6 SNP [OR 1.20 (95% CI 1.05–1.36)]. The 001Msp SNP was found to be associated with the presence of TSH receptor autoantibodies [OR 1.75 (95% CI 1.09–2.79)].
Conclusion: Functional evidence suggests that the 001Msp, fcrl3_3, fcrl3_5, and fcrl3_6 SNPs could cause changes in B cell signaling and activation pathways that could account for their association with GD. Further replication in independent data sets and fine mapping of the surrounding gene regions are needed to confirm the magnitude of the effect and location of the etiological variant(s) present within these gene regions.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The nuclear factor-B (NF-B) signaling pathway may also be involved in the development of disease. This pathway not only regulates many apoptotic and cell proliferation molecules but also plays a key role in both the innate immune response, through up-regulation of antiinflammatory factors, antimicrobial peptides, and components of the complement system, and in the adaptive response, by controlling the expression of components of the antigen presentation and T and B cell recognition pathways (6). Due to the pivotal role played by NF-B in the immune response, much interest has centered around this molecule in AID and its potential use as a therapeutic target (7).
Fine mapping of the IDDM5 region of linkage to T1D demonstrated evidence for association of two single-nucleotide polymorphism (SNPs) (001Msp and M55V) within inhibitors of the NF-B signaling pathway, MAP3K7IP2 (also known as TAB2) and SUMO-4, respectively (8, 9). However, studies of T1D patients of differing ethnic origin have produced conflicting results, with different allelic associations being observed that cannot be attributed to ethnicity alone, leaving the role of the these loci in some doubt (8, 9, 10, 11, 12, 13, 14). In addition, Kochi et al. (15) recently narrowed down association on chromosome 1q23 in Japanese RA patients to within two blocks of linkage disequilibrium (LD) encompassing the five known Fc receptor-like (FCRL) genes. This association was found to be attributed to four SNPs within the FCRL3 gene (fcrl3_3, fcrl3_4, fcrl3_5, and fcrl3_6). Of these four SNPs, only the fcrl3_3 C-T SNP, present at position –169 within the promoter, was found to affect NF-B binding to FCRL3 and therefore FCRL3 expression on B cells (15). Association between the fcrl3_3 SNP was also detected within a Japanese GD data set (15). The aim of this study was to attempt to replicate the association of two SNPs within MAP3K7IP2 and SUMO-4 and four SNPs within FCRL3 that play a key role in NF-B signaling pathways and two of which have been shown to have functional roles, in a UK Caucasian GD data set of sufficient size to have greater than 99% power to exclude an effect of odds ratio (OR) of 1.5 with P = 0.05.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Unrelated white Caucasian patients of UK origin with GD were recruited from large thyroid clinics in Birmingham, Walsall, Bournemouth, and Exeter, United Kingdom, as previously described (16). Patients were defined as having GD by the presence of biochemical hyperthyroidism together with either the presence of dysthyroid eye disease or two of the following criteria: diffuse goiter, a significant titer of microsomal, thyroglobulin, or TSH receptor (TSHR) autoantibodies (16). Microsomal and thyroglobulin antibodies were measured by gelatin particle agglutination (SERODIA_ATG, Fujirebio, Inc., Tokyo, Japan), and a titer of 1:100 was considered significant for both assays. TSHR autoantibody status was determined by a radioactive inhibition method (RSR Ltd., Cardiff, UK). Serum free T4 concentrations were measured with the Bayer ACS 180 and Bayer Advia Centaur system (Newbury, Berkshire, UK; normal range 9–20 pmol/liter). Matched control subjects with no family history of AID were bled at several sites: Blood Transfusion Services (Birmingham and Oxford, UK), Birmingham Heartlands Hospital, and the Queen Elizabeth Hospital (Birmingham, UK).
In total, DNA was obtained from 1056 GD patients and 864 control subjects for the population-based case-control study. All ethnically and gender-matched control subjects had tests of thyroid function and autoantibody status, and any showing evidence of subclinical autoimmune thyroid disease were removed from the study before genotyping. All subjects gave informed written consent and the project was approved by the local research ethics committee.
Genotyping: PCR/restriction fragment length polymorphism
DNA was extracted from 10 ml whole blood using the Nucleon Bacc III kit (Tepnel Life Sciences, Manchester, UK). Genotyping of the 001Msp SNP was performed using 20 ng genomic DNA in a 25-µl reaction containing 2.5 mM MgCl2, 50 ng/µl each primer (Sigma Genosys, Haverhill, UK), and 1 U Taq polymerase (Bioline, London, UK) giving rise to a 117-bp product. Amplification was performed in an Tetrad (MJ Research, Waltham, MA) for 35 cycles at an annealing temperature of 43 C using primers: forward, 5'-CCACCTCTTTAACTAA-3', and reverse, 5'-GTCAGCAACAGAACTCCG-3'. A mismatch in the reverse primer creates a restriction site for MspI in the presence of the C allele. Digestion of 5 µl of PCR product with 5 U of MspI enzyme at 37 C for 2 h resulted in an uncut band of 117 bp in the presence of the T allele and two bands (99 and 18 bp) in the presence of the C allele. Fragments were resolved on a 3% agarose gel and genotypes assigned accordingly.
Genotyping of the M55V SNP was performed by methods previously published (17). Briefly, 100 ng genomic DNA were amplified in a 25-µl reaction containing 2 mM MgCl2, 50 ng/µl each primer (Sigma Genosys), and 1 U Taq polymerase (Bioline), giving rise to a 185-bp product. Amplification was performed in an MJ Research Tetrad for 30 cycles at an annealing temperature of 60 C using primers: forward, 5'-ATTGTGAACCACGGGGATTGTTA-3', and reverse, 5'-CAGCGTTCTGGAGTAAAGAAG-3'. A mismatch at the 3' end of the forward primer creates a restriction site for MseI in the presence of the A allele. Digestion of 5 µl of PCR product with 5 U of MseI enzyme at 37 C overnight resulted in an uncut band of 185 bp in the presence of the G allele and two bands (162 and 23 bp) in the presence of the A allele. Fragments were resolved on a 3.8% agarose gel and genotypes assigned accordingly.
Genotyping of the CTLA-4 CT60 SNP, HLA DRB1 exon 2 position ß74, and PTPN22 1858 C/T SNP was as reported previously (1, 3, 4).
Genotyping: allele discrimination
The fcrl3_3 SNP (rs7528684), fcrl3_4 SNP (rs11264799), fcrl3_5 SNP (rs945635), and fcrl3_6 SNP (rs3761959) assays were purchased from Applied Biosystems (Foster City, CA) and were genotyped on an ABI7900 HT using Taqman (Applied Biosystems) genotyping technologies as previously described (15).
Statistical and LD analysis
Analysis of case-control data was performed using the 
2 test within the MINITAB statistical package (release 14.1, 1972–2003, Minitab Inc., State College, PA), and P < 0.05 was considered significant. ORs with 95% confidence intervals were calculated by the method of Woolf with Haldane’s modification for small numbers, where appropriate (18). Power and Hardy Weinberg equilibrium calculations were performed using Excel (Microsoft Office Excel, Redmond, CA). LD was analyzed using the pairwise LD measure D' and haplotype blocks constructed with the use of the computer program Haploview (19), using the default algorithm for generating haplotype blocks based on methods established by Gabriel et al. (20). A D' prime value of 1 = complete LD, D' value greater than 0.8 = strong LD, D' value 0.2–0.8 = incomplete LD, and D' less than 0.2 = negligible LD (21).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
No differences in allele or genotype frequencies of the M55V SNP were observed between cases with GD and controls in our cohort with more than 99% power to exclude an effect if present at OR of 1.5 and P < 0.05 (Table 1
). The 001Msp SNP was associated with GD, producing an OR of 1.19 [95% confidence interval (CI) 1.03–1.37] (Table 1). Association of the fcrl3_5 and fcrl3_6 SNPs was detected, producing OR of 1.18 [95% CI 1.04–1.35] and OR of 1.20 [95% CI 1.05–1.36] (Table 1), respectively. Association of the fcrl3_3 SNP was also detected (P = 0.024), producing an OR of 1.17 [95% CI 1.02–1.34] (Table 1), although only a trend toward significance could be detected for the genotypes (P = 0.068). A trend toward significance was seen for the fcrl3_4 SNP (P = 0.059); however, this was not replicated within the genotypes (P = 0.132). All the cases and controls were in Hardy Weinberg equilibrium.
|
Association of 001Msp was also found in the TSHR autoantibody-positive subgroup, compared with the TSHR autoantibody-negative subgroup (Table 2) producing an OR of 1.75 [95% CI 1.09–2.79]. No association among the fcrl3_3, fcrl3_5, and fcrl3_6 SNPs and TSHR autoantibody status was detected (Table 2). No correlation among the 001Msp, fcrl3_3, fcrl3–5, and fcrl3_6 SNPs and either microsomal or thyroglobulin autoantibody status was detected, and when these SNPs were compared with autoantibody status (in which presence of TSHR, thyroglobulin, or microsomal autoantibodies was considered as indicative of autoantibody involvement), no association was detected (data not shown). When the 001Msp, fcrl3_3, fcrl3_5, and fcrl3_6 SNPs were also correlated with severity of ophthalmopathy as determined by the NOSPECS (no signs or symptoms; only signs or symptoms limited to lid retraction; soft tissue involvement; proptosis; extraoccular muscle involvement; corneal involvement; sight loss) score (NOSPECS < 2 vs. NOSPECS 
2), age of onset of GD (
30 yr, representing early onset of GD vs. 31 yr, representing the normal age of onset for GD), biochemical severity as indicated by serum free T4 concentration at time of presentation before treatment was undertaken (serum free T4 concentrations of <40 pmol/liter or 40 pmol/liter), and the presence of goiter, no association was detected (data not shown).
|
Because the fcrl3_3 SNP was positively correlated with HLA-DRB1 status in the Japanese population (15), a preliminary interaction analysis was performed to determine whether fcrl3_3, fcrl3_5, fcrl3_6, and 001Msp interacted with the preestablished association of HLA-DRB1 in UK subjects with GD, with statistical interaction between the presence or absence of arginine at position ß74, the most associated SNP within the HLA-DRB1 exon 2-encoded domain being tested (with the presence of arginine at position ß74 found to be highly associated, P = 1.2 x 10–12, with GD within this data set) (3). Interaction was also investigated with the two other established susceptibility loci, the most strongly associated SNP (CT60 G-A SNP) within CTLA-4 (1) and the PTPN22 1858 C-T SNP (4, 17). The distribution of the fcrl3_3, fcrl3_5, fcrl3_6, and 001Msp SNPs was tested within the GD patients with the associated allele (in which, for the purpose of this analysis, both homozygous and heterozygous subjects for the associated allele of interest were classed as associated) vs. the GD patients without the associated allele (in which only those homozygous for not having the GD-associated allele were classed as nonassociated). No interactions could be detected among the fcrl3_3, fcrl3_5, fcrl3_6, and 001Msp SNPs and the HLA-DRB1 exon 2 ß74 and PTPN22 1858 SNPs (P = 0.996–0.207) (Table 3). A trend toward significant interaction between the fcrl3_3 and fcrl3_6 SNPs and the CTLA-4 CT60 SNP (P = 0.061 and P = 0.056, respectively) was detected (Table 3); however, this was not replicated within the alleles (P = 0.526 and P = 0.132, respectively; data not shown). No interactions between the fcrl3_5 and 001Msp SNPs and the CTLA-4 CT60 SNP were detected (P = 0.784–0.273) (Table 3).
|
LD was assessed among all four FCRL3 SNPs (including the fcrl3_4 SNP that showed only a trend toward significance) genotyped within our data set. Haplotype analysis [using the Gabriel et al. (20) method to define haplotype block structure] demonstrated evidence of one LD block within FCRL3 composed of the fcrl3_3, fcrl3_4, fcrl3_5, and fcrl3_6 SNPs (D' > 0.99). This is in contrast to findings in the Japanese population, in which fcrl3_3, fcrl3_5, and fcrl3_6 were shown to be in strong LD, with the fcrl3_4 SNP showing weaker LD (15).
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
B inhibitors, MAP3K7IP2 and SUMO-4, and four SNPs that affects NF-B promoter binding in the recently discovered FCRL3 gene with GD. No association of the M55V SNP was detected, despite the use of a data set of sufficient power to detect the previously reported effect. Our negative finding supports results from a smaller independent UK Caucasian data set of subjects with GD that also failed to detect association with the M55V SNP (22). In the present study, however, association was detected among the 001Msp, fcrl3_3, fcrl3_5, and fcrl3_6 SNPs and GD, albeit at a lower level than originally reported. When correlated with GD phenotype, 001Msp was also found to be associated with the presence of TSHR autoantibodies in those with GD. The identification of susceptibility loci for common diseases, such as GD, is proving to be a much greater task than many researchers originally predicted. With the increasing realization that individual loci are likely to confer ORs for the development of disease of significantly less than 1.5, the need for DNA resources with many hundreds and indeed thousands of subjects is now accepted. In the current study, it was estimated that sufficient power was available with 1919 subjects to detect the previously reported effects of the susceptibility loci under investigation. The ORs detected for 001Msp, fcrl3_3, fcrl3_5, and fcrl3_6 were, however, significantly less than expected, suggesting that either these results represent a false positive or that the contribution of these loci to the development of disease is much less than the earlier studies indicated. It is well recognized that a marked publication bias toward positive association studies does exist, in which the observed size of the genetic effect in the first positive report is biased upward (23), making replication in subsequent studies a more difficult task. Whereas too much emphasis has previously been placed on marginally significant results, which should be attributed to chance alone, a small effect producing an OR of less than 1.2 cannot be excluded in the current study despite the size of the data set used. In particular, the replication of association of three SNPs within FCRL3, all at a much lower level than previously reported (15), suggests that this study has detected a small effect for this gene with GD.
Because NF-B is a stimulator of so many immunological genes, regulation of its activity is paramount. NF-B activity is regulated through inhibitor of NF-B (IB) binding, thereby sequestering it in the cytoplasm (6, 24). In response to extracellular stimuli, IB kinase (IKK) is activated, rapidly targeting IBs for polyubiquitination and degradation, leading to NF-B release (6, 24). Once released, NF-B translocates into the nucleus and binds to DNA and along with accessory molecules leads to transcription of specific genes (6). MAP3K7IP2 is involved in the IKK signaling cascade in the IL-1 signaling pathway. When IL-1 binds to its receptor, MyD88 and IL-1 receptor-associated kinases are recruited to the receptor. The IL-1 receptor-associated kinase then dissociates from the receptor and interacts with TNF receptor-associated factor-6. Another molecule, TAK1, also interacts with TNF-receptor-associated factor 6 in a process that is facilitated by MAP3K7IP2, which enables the IKK cascade to occur, resulting in the release of NF-B (25). MAP3K7IP2 also plays a similar role in the TNF
signaling pathway (26, 27). Due to the important role of IL-1 and TNF signaling pathways in B cell activation and replication, dysregulation of NF-B could lead to increased TNF and IL-1 production, which as potent B cell activators, could lead to greater B cell activation (6). Association of the 001Msp SNP with TSHR autoantibody status in GD could be explained by the ability of this polymorphism, or one in LD with it, to induce changes in MAP3K7IP2 function within the TNF and IL-1 activation pathways leading to B cell dysregulation and increased autoantibody production.
Although the FCRL3 molecule is not involved in NF-B release, its gene is transcribed by NF-B. The susceptible C allele of the fcrl3_3 SNP has been shown to alter FCRL3 expression, through NF-B promoter binding, leading to higher FCRL3 expression on B cells (15). The FCRL3 molecules contain both immunotyrosine activation motifs and immunotyrosine inhibitory motifs, enabling them to both activate and inhibit signaling pathways and are believed to be involved in regulating cellular signaling thresholds (28, 29). It has been suggested that increased FCRL3 expression could lead to dysregulated B cell activation, which could contribute to AID onset. No functional role has to date been determined for the fcrl3_5 or fcrl3_6 SNPs.
Both SNP frequency and LD block structure have been shown to vary between Oriental and European populations. Differences detected between the original study and our own could be due to differences between the Japanese and UK populations. This study highlights these differences by showing that within the UK Caucasian population, fcrl3_3, fcrl3_4, fcrl3_5, and fcrl3_6 were all in strong LD, whereas in the Japanese population, fcrl3_4 showed weaker LD with the remaining three SNPs. Also, we found the G allele of the fcrl3_5 SNP associated with GD within our population, whereas Kochi et al. (15) reported association of the C allele of fcrl3_5 with RA in the Japanese population. Differences in association could also be due to differences between RA and GD. Although both diseases have been shown to share genetic susceptibility loci, including the HLA class II region, the associations at these regions can vary between diseases. For example, at the HLA class II locus, in RA the shared epitope region encompassing DRB1 exon 2 amino acid positions ß70-ß74 has been associated with disease (2), whereas in GD a series of nine DRB1 exon 2-encoded amino acids, including ß71 and ß74 present within the shared epitope region, have been associated with GD (3). This could help to explain the differences in association of the fcrl3_3 SNP with HLA-DRB1 status in the Japanese and UK Caucasian populations.
Although association among the fcrl3_3, fcrl3_5, fcrl3_6, and 001Msp SNPs with GD has been found, it is not possible to conclude that these loci represent primary etiological variants. Although the fcrl3_3 SNP has been shown to have functional consequences on NF-B binding, the M55V SUMO-4 SNP was also shown to have a functional role in NF-B inhibition (9), and the results generated here and previously (22) have shown that this SNP does not contribute to GD in the UK Caucasian population. This highlights the importance of determining the etiological variant present within a region of association before functional analysis is undertaken. Within the IDDM5 region, containing the 001Msp SNP, only 18 SNPs encompassing 40 kb were originally investigated (9). Because LD can span much greater distances than this, denser sequence coverage across this region is needed. Although 724 SNPs were used to cover the Fc receptor region on chromosome 1q23 and LD blocks defined (15), because both allele frequency and LD block structure vary among differing ethnic groups, a similar SNP coverage may be needed in UK Caucasians to define the levels of LD and association(s) across this region.
In conclusion, functional SNPs involved in NF-B signaling may contribute to the development of GD. Future genetic and immunological studies will determine the role of this important pathway.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online December 29, 2005
1 M.J.S. and J.M.H. contributed equally to this work. 
Abbreviations: AID, Autoimmune disease; CI, confidence interval; FCRL, Fc receptor-like; GD, Graves’ disease; HLA, human leukocyte antigen; IB, inhibitor of NF-B; IKK, IB kinase; LD, linkage disequilibrium; NF-B, nuclear factor-B; OR, odds ratio; RA, rheumatoid arthritis; SNP, single-nucleotide polymorphism; T1D, type 1 diabetes; TSHR, TSH receptor.
Received July 22, 2005.
Accepted December 15, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
B activation. Mol Immunol 41:701–714[CrossRef][Medline]B pathway as a potential target for autoimmune disease therapy. Curr Pharm Des 10:2827–2837[CrossRef][Medline]B modifier, is associated with type 1 diabetes. Nat Genet 36:837–841[CrossRef][Medline]B and the risk of type 1 diabetes. Genes Immunol 6:231–235B regulation in the immune system. Nat Rev Immunol 2:725–734[CrossRef][Medline]B pathway through binding to polyubiquitin chains. Mol Cell 15:535–548[CrossRef][Medline]B signaling: flipping the switch with polyubiquitin chains. Curr Biol 14:R898–R900This article has been cited by other articles:
 |
|
 |
|
  Y. Kochi, K. Myouzen, R. Yamada, A. Suzuki, T. Kurosaki, Y. Nakamura, and K. Yamamoto FCRL3, an Autoimmune Susceptibility Gene, Has Inhibitory Potential on B-Cell Receptor-Mediated Signaling J. Immunol., November 1, 2009; 183(9): 5502 - 5510. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. J. Brand, J. C. Barrett, M. J. Simmonds, P. R. Newby, C. J. McCabe, C. K. Bruce, B. Kysela, J. D. Carr-Smith, T. Brix, P. J. Hunt, et al. Association of the thyroid stimulating hormone receptor gene (TSHR) with Graves' disease Hum. Mol. Genet., May 1, 2009; 18(9): 1704 - 1713. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Simmonds, J. M.M. Howson, J. M. Heward, J. Carr-Smith, J. A. Franklyn, J. A. Todd, and S. C.L. Gough A novel and major association of HLA-C in Graves' disease that eclipses the classical HLA-DRB1 effect Hum. Mol. Genet., September 15, 2007; 16(18): 2149 - 2153. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. J. Owen, H. Kelly, J. A. Eden, M. E. Merriman, S. H. S. Pearce, and T. R. Merriman Analysis of the Fc Receptor-Like-3 (FCRL3) Locus in Caucasians with Autoimmune Disorders Suggests a Complex Pattern of Disease Association J. Clin. Endocrinol. Metab., March 1, 2007; 92(3): 1106 - 1111. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Manji, J. D. Carr-Smith, K. Boelaert, A. Allahabadia, M. Armitage, V. K. Chatterjee, J. H. Lazarus, S. H. S. Pearce, B. Vaidya, S. C. Gough, et al. Influences of Age, Gender, Smoking, and Family History on Autoimmune Thyroid Disease Phenotype J. Clin. Endocrinol. Metab., December 1, 2006; 91(12): 4873 - 4880. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. A. Aliesky, P. N. Pichurin, C.-R. Chen, R. W. Williams, B. Rapoport, and S. M. McLachlan Probing the Genetic Basis for Thyrotropin Receptor Antibodies and Hyperthyroidism in Immunized CXB Recombinant Inbred Mice Endocrinology, June 1, 2006; 147(6): 2789 - 2800. [Abstract] [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 |
