| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Atherosclerosis Research Unit, Department of Medicine Solna, Karolinska Institutet, Karolinska University Hospital, SE-17176 Stockholm, Sweden
Address all correspondence and requests for reprints to: Dr. Ferdinand M. van't Hooft, Cardiovascular Genetics Group, Centre for Molecular Medicine, L8:02, Karolinska University Hospital Solna, SE-17176 Stockholm, Sweden. E-mail: Ferdinand.vant.Hooft{at}ki.se.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: Here, we analyzed the relative significance of INSIG1 and INSIG2 in human liver and adipocyte metabolism, and defined a novel, functional polymorphism in the promoter of INSIG2 associated with body mass index.
Research Methods: Variations in gene expression of different human tissues, of hepatoma cells exposed to INSIG1 and INSIG2 gene silencing probes, and of differentiating 3T3-L1 adipocytes were determined by real-time quantitative PCR. The functional significance of a novel polymorphism in the promoter of INSIG2 was analyzed using in vitro methods and gene expression analysis of human adipose tissue, whereas the phenotype associated with this polymorphism was studied in two cohorts of middle-aged men.
Results: Gene expression analysis of 17 human tissues demonstrated that INSIG1 is highly expressed in the liver, whereas INSIG2 is ubiquitously expressed. Gene silencing experiments confirmed that INSIG1, but not INSIG2, regulates the expression of sterol regulatory element-binding proteins target genes in human hepatoma cells. In contrast, adipocyte differentiation of 3T3-L1 cells was associated with a 13-fold increase in expression of INSIG2. Significant relationships between the INSIG2–102G/A polymorphism and body mass index were observed in two cohorts of middle-aged men (ANOVA P = 0.017 and 0.044, respectively). In vitro studies and analysis of allele-specific expression in human adipose tissue substantiated the functional significance of the INSIG2–102G/A polymorphism.
Conclusion: INSIG2 is involved in adipocyte metabolism and body weight regulation.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
INSIG2 is a close homolog of INSIG1. Both INSIG proteins bind sterol regulatory element-binding protein (SREBP) cleavage-activating protein (SCAP) in the endoplasmatic reticulum, thereby inhibiting the release of the SCAP-SREBP complex and preventing the proteolytic processing of the three SREBP isoforms (designated SREBP1a, SREBP1c, and SREBP2) into active transcription factors (for review, see Ref. 6). Experimental studies demonstrated that the two INSIG proteins fulfill complementary roles in the regulation of the SREBP pathway. This feature was underlined by data from a recent knockout study that showed that INSIG1 and INSIG2 knockout mice appear grossly normal, whereas only the combined disruption of both INSIG genes was associated with a phenotype related to disturbances of the SREBP pathway (7). So far, the major differences between INSIG1 and INSIG2 relate to the regulation of their expression, but it appears that changes in expression of INSIG1 and INSIG2 are usually reciprocal in nature, i.e. decrease in concentration of one INSIG protein is compensated for by an increase in the other INSIG protein.
The pioneering work of Brown and Goldstein (6, 7) has highlighted the role of the INSIGs in the SREBP pathway, and consequently in the regulation of cholesterol and fatty acid metabolism. Nevertheless, several recent studies indicate that INSIG1 may have additional functions (reviewed in Ref. 6). For example, elegant studies by DeBose-Boyd and colleagues demonstrated that INSIG1 is involved in the degradation of 3-hydroxy-3-methylglytaryl coenzyme A reductase (HMGCR), a process that is independent of the INSIG1/SCAP interaction and activation of the SREBP pathway). In addition, it was reported that INSIG1 is involved in diet-induced obesity in rodents (8), presumably due to inhibition by INSIG1 of lipogenesis in adipocytes and of differentiation of preadipocytes (9). Moreover, we recently provided human evidence for the involvement of INSIG1 in the regulation of plasma glucose concentration (10). Together, these studies underline the multifunctional role of INSIG1.
In contrast, little is known regarding specific metabolic roles of INSIG2. It was shown that overexpression of INSIG2 reduces lipogenesis in the livers of obese Zucker diabetic fatty rats, but the same effect was observed for INSIG1 (11). INSIG2 has also been implicated as a susceptibility gene for plasma cholesterol levels in mice (12), but these studies in rodents have not been corroborated by human studies. Nevertheless, the recent report (5) of a relationship between BMI and the rs7566605 polymorphism in the vicinity of INSIG2 clearly suggests that INSIG2 may have metabolic roles distinct from INSIG1 and that this property may be tissue specific. Against this background we have evaluated the relative significance of INSIG1 and INSIG2 in different tissues, with special emphasis on a possible involvement of INSIG2 in human adipocyte metabolism.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A biobank based on blood samples from two cohorts of middle-aged men was used for the present study. The first cohort consisted of 634 healthy 50-yr-old men (13), and the second cohort consisted of 956 middle-aged male survivors of myocardial infarction and age-matched healthy control subjects (14). A standardized protocol, approved by the Ethics Committee of the Karolinska University Hospital, regulated the blood collection, anthropometric measurements, and the basic biochemical analysis (13, 14), and all subjects gave informed consent to their participation. Plasma leptin concentration was determined by ELISA (R&D Systems, Inc., Minneapolis, MN). A total of 40 gluteal sc adipose tissue biopsies, obtained from a separate cohort of healthy men (15), was used for gene expression analysis.
Polymorphism detection
The ABI3100 capillary sequencer (Applied Biosystems, Foster City, CA) was used for direct sequencing of INSIG2. We included all exons, more than 200-bp flanking introns, approximately 650 bp of the proximal promoter and approximately 1300 bp of the 3'-untranslated region. Sections of approximately 900 bp were amplified and purified with the QIAquick PCR purification kit (QIAGEN, Inc., Valencia, CA). These sections were used as templates for sequencing using nested primers. Primer sequences are provided in supplemental Table 1
, which is published as supplemental data on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org.
|
The –102G/A promoter polymorphism was genotyped (1.25 mM MgCl2, annealing temperature 59 C, and 4% dimethylsulfoxide) using the forward primer 5'-GGTCAGCAAACAACAGCAGA-3 and reverse primer 5'-GTACCCCCTACCGCCTCTT-3', followed by restriction digestion with AvaII. The rs12986752 polymorphism was genotyped (1.25 mM MgCl2, annealing temperature 59 C) using the forward primer 5'-biotinGCTCCTTTCGGCTACGAGAT-3' and reverse primer 5'-CTCCTCCACTCCCACAACTC-3', followed by pyrosequence analysis of the amplified fragments using the nested primer 5'-CCAGGCAGCCGTAGGTAATG-3'. Polymorphisms rs7566605, rs7589375, rs13393332, and rs1559509 were genotyped using TaqMan assays (Applied Biosystems).
Small interfering RNA (siRNA) probes
siRNA oligonucleotides specific for INSIG1 and INSIG2 were designed using the siRNA Target Finder program (www.ambion.com). siRNA oligonucleotides were synthesized and annealed by a commercial supplier (Sigma-Proligo, Sigma-Aldrich, St. Louis, MO). The forward sequence of the INSIG1 siRNA probe was 5'-UUACCAACCCGCAGUUUUGTT-3', corresponding to the sequences 1326–1347 of INSIG1 mRNA. The forward sequence of INSIG2 siRNA was 5'-UGGGAAACAUUGGUCACATT-3', corresponding to the sequence 818–839 of INSIG2 mRNA. A siRNA probe described by Elbashir et al. (16) with the forward sequence 5'-CGUACGCGGAAUACUUCGATT-3, targeting the pGL2 plasmid vector, was used as negative control.
Cell culture and transfection procedures
Human hepatoma Huh7 and HepG2 cells were cultured in DMEM (Life Technologies, Inc., Gaithersburg, MD) supplemented with 10% fetal bovine serum (FBS), 50 U/ml penicillin, and 50 µg/ml streptomycin. Subconfluent hepatoma cells were preincubated with FBS-free medium for 30 min. The cells were subsequently transfected with 50 nM gene-specific siRNA or with negative control siRNA using Lipofectamin 2000 (Invitrogen Corp., Carlsbad, CA) as a transfection agent. The transfection medium was removed after 5-h incubation, and the cells were cultured in FBS-free medium for another 48 h.
Mouse 3T3-L1 preadipocytes (obtained from American Type Culture Collection, Manassas, VA) were grown in high-glucose DMEM, supplemented with 10% FBS (Life Technologies), 110 mg/liter pyruvate, 50 U/ml penicillin, and 50 µg/ml streptomycin (all from Sigma-Aldrich). Adipocyte differentiation was induced in 2-d post-confluent cells by addition of 1.7 µM insulin, 0.5 µM dexamethasone, and 0.5 mM 3-isobutyl-1-methylxanthine (all from Sigma-Aldrich) to the medium.
Western blot analysis
Cells were rinsed twice with PBS and lysed on ice with radioimmunoprecipitation assay buffer. The extracted proteins (25 µg) were subjected to 10% SDS-PAGE and electrophoretically transferred onto a Hybond P PVDF membrane (Amersham Biosciences Inc., Piscataway, NJ). A rabbit antihuman SREBP1 (H-160) sc-8984 polyclonal antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA; final dilution 1:200) was used as a primary antibody, and horseradish peroxidase-conjugated antirabbit immunoglobulin (Santa Cruz Biotechnology; final dilution 1:5000) was used as the secondary reagent.
Real-time quantitative PCR
Total cellular RNA was isolated from the hepatoma cells using the RNeasy mini kit (QIAGEN). RNA samples from different human tissues were purchased from Ambion, Inc. (Austin, TX) (FirstChoice Human Total RNA Survey Panel). RNA was quantified spectrophotometrically, and the RNA concentrations of all samples were adjusted to 0.1 µg/µl. Total cDNA was synthesized from 0.5 µg RNA in a polymerase reaction using oligoT primer (2.5 pmol/µl final concentration), deoxynucleotide triphosphates (0.5 mM final concentration), and SuperScriptII Rnase H– reverse transcriptase (Invitrogen). RNA from gluteal sc adipose tissue biopsies was isolated, and total cDNA was synthesized as described by Gertow et al. (15). All assays and reagents for real-time quantitative PCR were obtained from Applied Biosystems. All assays were performed using the ABI Prism 7000 sequence detection system according to the manufacturer's protocol, using five-point standard curves generated from 10-fold dilutions of purified PCR products. The purified PCR products were quantified using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). The results were expressed in arbitrary units related to the levels of 18S or TATA-box binding protein (adipose tissue) mRNA expression for normalization.
EMSA
Nuclear extracts were prepared according to Alksnis et al. (17). Incubation for EMSA was conducted as described (18), and the reaction products were analyzed by 7% (wt/vol) PAGE. Nonradioactive competitor DNAs, either identical, or of the opposite allelic variant, or of nonspecific origin, were added in 100-fold excess of the labeled DNA.
Statistical methods
Distribution of continuous variables in groups was expressed as mean ± SEM. Logarithmic transformation was performed on all skewed variables to obtain a normal distribution before statistical computations, and significance testing were performed. Differences in continuous variables according to INSIG2 genotype were tested by one-way ANOVA.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The expression levels of INSIG1 and INSIG2 in different human tissues were analyzed using real-time PCR. INSIG1 was highly expressed in the liver, whereas relatively low expression levels were observed in extrahepatic tissues (Fig. 1A). In contrast, INSIG2 showed a more ubiquitous expression pattern (Fig. 1B). As shown in Fig. 1C, the INSIG1/INSIG2 mRNA ratio in the liver was 3.51 ± 0.43, whereas the INSIG1/INSIG2 mRNA ratio for all extrahepatic tissues combined was 0.57 ± 0.10.
|
The significance of INSIG1 and INSIG2 in the regulation of expression of SREBP target genes was analyzed in Huh7 and HepG2 human hepatoma cells. Huh7 cells exhibited expression levels for INSIG1 and INSIG2 comparable to human liver tissue, with an INSIG1/INSIG2 mRNA ratio of 2.84 ± 0.53 (n = 8). In contrast, HepG2 cells had extremely low expression of INSIG1, whereas the expression of INSIG2 was comparable to INSIG2 expression in human liver. As a consequence, the INSIG1/INSIG2 mRNA ratio in HepG2 cells was less than 0.01.
The effects of INSIG1 and INSIG2 on the regulation of the SREBP target genes were evaluated using siRNA gene silencing. siRNA inhibition of INSIG1 and INSIG2 in Huh7 cells decreased the INSIG1 and INSIG2 mRNA levels by 74 ± 5% (n = 8) and 76 ± 1% (n = 8), respectively, whereas siRNA inhibition of INSIG2 in HepG2 cells decreased the INSIG2 mRNA concentration by 81 ± 3% (n = 3). No compensatory increases in the nonsilenced INSIG mRNA levels were observed. The expressions of all eight SREBP target genes (ACLY, ACSS2, FASN, FDPS, HMGCR, HMGCS1, LDLR, and PCSK9) analyzed in this study were significantly increased after siRNA inhibition of INSIG1 in Huh7 cells, whereas no increased expressions of the SREBP target genes were found after siRNA inhibition of INSIG2 in either Huh7 cells or HepG2 cells (Fig. 2A).
|
, a significant increase in the concentration of the 68-kDa mature SREBP1 protein was observed after INSIG1 siRNA inhibition. Significance of INSIG2 in adipogenesis
The putative roles of INSIG1 and INSIG2 in adipocyte differentiation of 3T3-L1 preadipocytes were analyzed. The expected changes in morphology (accumulation of lipid droplets), cellular triglyceride concentration, and expression of marker genes for adipocyte differentiation (Pparg and Fabp4, shown in Fig. 3; Cebpa, Dlk1, and Fasn, not shown) were observed after induction of differentiation. The relative changes in expression of Insig1 and Insig2 were analyzed by real-time PCR. As shown in Fig. 3, a marked increase in expression of Insig2 and a modest increase in expression of Insig1 were observed during adipocyte differentiation. The Insig1/Insig2 mRNA ratio at the end of the 10-d differentiation experiment was 0.84 ± 0.17 (n = 3), comparable to the INSIG1/INSIG2 mRNA ratio of 0.73 ± 0.16 in human adipose tissue (Fig. 1).
|
INSIG2 was analyzed in search for a functional polymorphism responsible for the relationship between the rs7566605 polymorphism situated 10-kb upstream of the transcription start site of INSIG2 (5) and BMI. To this end we sequenced the exons and intron-exon borders of INSIG2 in DNA samples from 18 subjects. Six exons were identified, comparable to exons 1b and 2–6 described for mouse INSIG2 (19). No evidence was found for the expression of a human equivalent of mouse exon 1a (data not shown). As shown in Fig. 4, no polymorphisms were discovered in the exons. Three polymorphisms, two in the 3'-untranslated region and one in intron 5, had minor allele frequencies less than 10% and were not further analyzed. One novel polymorphism was uncovered in the putative promoter of INSIG2, designated –102G/A, whereas another promoter polymorphism, designated –231G/A, is also known as rs12986752. Human-mouse homology analysis of the INSIG2 promoter demonstrated a high degree of conservation, especially in the immediate vicinity of the –102G/A polymorphism (data not shown).
|
). All polymorphisms were in Hardy-Weinberg equilibrium. Complete allelic association was observed for the rs7589375, rs13393332, rs1559509, and rs12986752 polymorphisms (hereafter designated 4-SNPs). The minor allele frequencies of the 4-SNPs, rs7566605, and –102G/A polymorphisms were 22, 36, and 10%, respectively. Pairwise D' and r2 values between the 4-SNPs and rs7566605, between the 4-SNPs and the –102G/A polymorphism, and between rs7566605 and the –102G/A polymorphism were 0.84 and 0.10, 0.81 and 0.02, 0.96 and 0.04, respectively.
Significant relationships were observed between the –102G/A polymorphism and BMI in two cohorts of middle-aged men (Table 1). The associations were graded, i.e. subjects homozygous for the –102A allele had greater decreases in BMI compared with subjects who were heterozygous for the –102A allele. A comparable relationship was observed between the –102G/A polymorphism and plasma leptin concentration in the first cohort, but this association was abolished after adjustment for BMI (data not shown). Weak associations were also observed between the rs7566605 polymorphism and BMI or plasma leptin concentration, and between the 4-SNPs and BMI and leptin concentration, but these relationships did not reach the conventional level of statistical significance. Haplotype analysis revealed that the relationship with BMI and plasma leptin concentration was best explained by the –102G/A polymorphism (data not shown), indicating that the –102/A polymorphism is most likely of physiological significance. Of note, no relationships between the –102G/A polymorphism and biochemical parameters related to lipid and glucose metabolism were found (Table 1).
Functional analysis of the –102G/A polymorphism
The physiological significance of the –102G/A polymorphism was analyzed using EMSA. The binding characteristics of a 21-bp DNA fragment containing either the –102G or the –102A site of the INSIG2 promoter were evaluated using nuclear extracts derived from 3T3-L1 cells. As shown in Fig. 5A, increased concentrations of a major DNA-protein complex was observed with increased nuclear extract concentrations. However, the DNA-protein complex was present at considerably higher concentrations when the –102G fragment was compared with the –102A fragment. Quantification of the protein-DNA complex in three independent experiments corroborated the higher band intensities for the DNA fragment containing the –102G allele compared with the –102A allele (Fig. 5B). A transcription factor database search was performed, but no suitable candidate transcription factor with a consensus binding site at the position of the –102G/A polymorphism was uncovered.
|
). The intensity of the protein-DNA complex indicated by the arrow in Fig. 5C was quantified in four independent experiments (Fig. 5D). As shown in Fig. 5, C and D, a 100-fold excess of unlabeled –102G fragment substantially reduced the interaction of the labeled –102G fragment with nuclear proteins. In contrast, no effect was observed with a 100-fold excess of unlabeled –102A fragment or unrelated probe on the interaction of labeled –102G fragment with the nuclear proteins. The EMSA studies demonstrated that the –102G/A polymorphism affects the binding of nuclear factor(s) to the promoter of INSIG2, presumably influencing the rate of transcription of INSIG2. To test this hypothesis under in vivo conditions, the relationship between the –102G/A polymorphism and INSIG2 mRNA concentration was analyzed in sc fat biopsies obtained from healthy, middle-aged men. INSIG2 mRNA concentrations tended to be lower for subjects heterozygous for the –102A allele (33.5 ± 9.0 arbitrary units, n = 11) compared with individuals homozygous for the –102G allele (27.7 ± 7.5 arbitrary units, n = 29), but this difference did not reach the conventional level of statistical significance (P = 0.07).
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Recently, Herbert et al. (5) reported an association between the rs7566605 polymorphism and BMI. A meta-analysis of six cohorts involving in total 9881 subjects underlined the significance of this observation [odds ratio for obesity 1.22 (95% confidence interval 1.05–1.42); P = 0.008]. This association was reproduced in several but not all cohorts (20, 21, 22, 23, 24). In the present study, we found similar trends for relationships between the rs7566605 polymorphism and BMI, but these relationships did not reach the conventional level of statistical significance. Nevertheless, this observation encouraged us to search for the functional polymorphism responsible for the observed relationship with BMI. In this study we describe a novel polymorphism, a G to A substitution at position –102 in the promoter of INSIG2. In EMSA studies a distinct difference in the binding of nuclear factors was observed between fragments containing either the –102G or the –102A allele. Moreover, subjects heterozygous for the –102A allele had lower INSIG2 mRNA concentrations in adipose tissue compared with individuals homozygous for the –102G allele. Together, it appears that the –102G/A polymorphism affects the binding of nuclear factors to the promoter of INSIG2, resulting in changes in the expression of INSIG2 in adipose tissue, ultimately leading to changes in adipocyte metabolism. Thus, we propose that the –102G/A polymorphism is the functional polymorphism in INSIG2, responsible for the observed associations with BMI.
The independent replication in two cohorts of the relationship between INSIG2 and BMI, as first reported by Herbert et al. (5), underlines the putative role of INSIG2 in adipocyte metabolism. Little is known regarding the function of INSIG2 in adipocyte metabolism, but it seems reasonable to assume that the SREBP pathway mediates the effect(s) of INSIG2. As reviewed by Spiegelman and colleagues (25), SREBP1 plays an important role in adipogenesis, as analyzed in 3T3 cells. It can be expected that reduced INSIG2 concentration in adipocytes will enhance the concentration of the nuclear active form of SREBP1, increasing the expression of genes involved in adipogenesis. In this scenario, overexpression of a constitutively active SREBP1 is expected to promote fat cell development. Adipocyte hypertrophy, increased fatty acid secretion, and fatty liver were, indeed, characteristic features of mice overexpressing SREBP1a in adipose tissue (26), whereas overexpression of SREBP1c in adipose tissue resulted in the opposite phenotype: lipodystrophy (27). As discussed by Horton et al. (26), the dramatically different phenotypes that resulted from overexpressing the two closely related SREBP1 isoforms in adipose tissue raise the possibility that SREBP1a and SREBP1c have tissue-specific transcriptional activation properties. A more detailed analysis of the INSIG2 downstream target genes will be required to resolve the mechanism of action of INSIG2 in adipose tissue.
The human hepatoma cell lines HepG2 and Huh7 have been used extensively as model systems for the analysis of hepatocyte metabolism. Therefore, the profound differences in expression of INSIG1 and INSIG2 in these hepatoma cells were surprising. Unfortunately, antibodies capable of detecting endogenous INSIG1 and INSIG2 in human and mouse cells are not available (28) (Chernogubova, E., unpublished observation). Therefore, for the moment it seems reasonable to assume that INSIG mRNA levels in hepatoma cells provide a rough indication of endogenous INSIG1 and INSIG2 concentrations. We exploited the extreme differences in INSIG1/INSIG2 mRNA ratios in Huh7 and HepG2 cells to analyze the effects of INSIG2 siRNA silencing on the expression of SREBP target genes. INSIG1 silencing significantly increased the protein concentration of the mature form of SREBP1 and enhanced the expression of SREBP target genes in Huh7 cells. No effect of INSIG2 silencing on the expression of SREBP target genes was observed. This observation was not due to a compensatory increase in expression of INSIG1, but it cannot be excluded that the relatively high INSIG1 level in Huh7 cells masked the effects of siRNA-mediated reduction in INSIG2. However, in HepG2 cells, characterized by extremely low INSIG1 levels, again no effect of INSIG2 silencing on the expression of the SREBP target genes was observed. As a whole, these studies in human hepatoma cells indicate that INSIG1 is of key importance for the regulation of SREBP target genes in human liver, whereas it appears that INSIG2 does not play a role in this process. It is conceivable that the opposite mechanism operates in adipose tissue, i.e. that variation in INSIG2 concentration primarily regulates the expression of SREBP target genes, whereas INSIG1 is not involved in this process.
In summary, this study provided independent confirmation of an association between genetic variation in INSIG2 and BMI, and identified a functional polymorphism that could be responsible for this relationship. In addition, other lines of evidence in support of a significant role of INSIG2 in adipocyte metabolism were presented. It is expected that a more detailed analysis of the downstream target genes of INSIG2 may uncover the molecular mechanism that is responsible for the significant role of INSIG2 in adipocyte metabolism.
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to declare.
First Published Online March 4, 2008
Abbreviations: BMI, Body mass index; FBS, fetal bovine serum; HMGCR, 3-hydroxy-3-methylglytaryl coenzyme A reductase; INSIG, insulin-induced gene; SCAP, SREBP cleavage-activating protein; siRNA, small interfering RNA; SNP, single nucleotide polymorphism; 4-SNPs, rs7589375, rs13393332, rs1559509, and rs12986752; SREBP, sterol regulatory element-binding protein.
Received August 17, 2007.
Accepted February 25, 2008.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
