Expression in Human -Cells and Adipose Tissue and Strong Association W
http://www.100md.com
《糖尿病学杂志》
1 CNRS 8090 Institute of Biology, Pasteur Institute, Lille, France
2 INSERM U780-IFR69, Villejuif, France
3 Corbeil Hospital, Corbeil, France
4 INSERM ERIT-M 0106, Lille, France
5 Necker Hospital, INSERM E363, Paris, France
6 Atherosclerosis Department, Pasteur Institute, Lille, France
7 INSERM U545, Lille, France
8 University of Lille 2, Lille, France
9 Department of Metabolic Research, Clinica Universitaria de Navarra, Pamplona, Spain
10 Section of Genomic Medicine, Hammersmith Hospital, Imperial College, London, U.K
FBAT, family-based association test; GLP-1, glucagon-like peptide 1; PAR, population-attributable risk; SNP, single nucleotide polymorphism; TDT, transmission disequilibrium test
ABSTRACT
Recently, the transcription factor 7-like 2 (TCF7L2) gene has been associated with type 2 diabetes in subjects of European origin in the DeCode study. We genotyped the two most associated variants (rs7903146 and rs12255372) in 2,367 French type 2 diabetic subjects and in 2,499 control subjects. Both the T-allele of rs7903146 and the T-allele of rs12255372 significantly increase type 2 diabetes risk with an allelic odds ratio (OR) of 1.69 (95% CI 1.55–1.83) (P = 6.0 x 10–35) and 1.60 (1.47–1.74) (P = 7.6 x 10–28), respectively. In nonobese type 2 diabetic subjects (BMI <30 kg/m2, n = 1,346), the ORs increased to 1.89 (1.72–2.09) (P = 2.1 x 10–38) and 1.79 (1.62–1.97) (P = 5.7 x 10–31), respectively. The rs7903146 T at-risk allele associates with decreased BMI and earlier age at diagnosis in the type 2 diabetic subjects (P = 8.0 x 10–3 and P = 3.8 x 10–4, respectively), which is supported by quantitative family-based association tests. TCF7L2 is expressed in most human tissues, including mature pancreatic -cells, with the exception of the skeletal muscle. In the subcutaneous and omental fat from obese type 2 diabetic subjects, TCF7L2 expression significantly decreased compared with obese normoglycemic individuals. During rat fetal -cell differentiation, TCF7L2 expression pattern mimics the key marker NGN3 (neurogenin 3), suggesting a role in islet development. These data provide evidence that TCF7L2 is a major determinant of type 2 diabetes risk in European populations and suggests that this transcription factor plays a key role in glucose homeostasis.
To date, positional cloning for type 2 diabetes genes has not been very successful in detecting type 2 diabetes putative susceptibility genes. Although several positive reports have emerged (1,2), few have been consistently replicated (3). Several studies have shown linkage between type 2 diabetes and chromosome 10q in Mexican-American, French, English, and Icelandic populations (4–7) with a strong association between type 2 diabetes and variation in the transcription factor 7-like 2 (TCF7L2) gene found in Icelandic, Danish, and American populations (8). Little is known about the physiological implication of this transcription factor in glucose homeostasis. It has been suggested that intestinal proglucagon gene expression may be regulated by the Wnt/TCF7L2 pathway in enteroendocrine cells (9). Thus, TCF7L2 variants may modify type 2 diabetes susceptibility through modulation of glucagon-like peptide-1 (GLP-1) secretion. Because French type 2 diabetic families exhibit evidence for linkage in a chromosome 10q region encompassing the TCF7L2 locus, we assessed the contribution of TCF7L2 genetic variation to type 2 diabetes genetic risk. Tissue profiling analyses were also performed in rodents and humans to help understand the regulation of TCF7L2 and its involvement in the physiology of type 2 diabetes.
Among the previously reported TCF7L2 variants, the association between the T-allele of rs7903146 and the T- allele of rs12255372 have been replicated (8). These two variants were genotyped in 2,367 French type 2 diabetic subjects and in 2,499 normoglycemic nonobese control middle-aged individuals (fasting glucose <5.6 mmol/l and BMI <30 kg/m2) (Table 1). We found a strong association between type 2 diabetes and the two variants in our case-control study (Table 2). The genotypic distributions for both variants were in Hardy-Weinberg equilibrium, and the T-alleles of rs7903146 and rs12255372 were both found to significantly increase type 2 diabetes risk with an allelic odds ratio (OR) of 1.69 (95% CI 1.55–1.83) (P = 6.0 x 10–35) and 1.60 (1.47–1.74) (P = 7.6 x 10–28), respectively. Testing for the two-variant haplotype did not produce a better result than the two variants analyzed individually. Analyzing the 1,346 nonobese type 2 diabetic subjects (BMI <30 kg/m2) by themselves increases the allelic OR to 1.89 (1.72–2.09) (P = 2.1 x 10–38) and 1.79 (1.62–1.97) (P = 5.7 x 10–31), respectively.
As the two variants exhibit strong linkage disequilibrium (r2 = 0.77, D' = 0.89), further analyses focused on rs7903146 (the single nucleotide polymorphism [SNP] with the strongest association to type 2 diabetes). The population-attributable risk (PAR) for rs7903146 was estimated at 31% using our case-control sample. When the allelic OR is estimated in the nonobese subjects versus the control subjects (BMI <30 kg/m2), the attributable risk reaches 37.7%. The T-allele frequency was 30.5% in control individuals and 42.5% in type 2 diabetic subjects. We tested departure from the additive model in a logistic regression framework, where an index variable identifies heterozygous individuals. No significant deviation was observed (P = 0.74), supporting the hypothesis of a risk increase proportional to the number of variant alleles (genotypic OR = 1.66 for C/C vs. C/T and genotypic OR = 2.86 for C/C vs. T/T). In this study, there is a discrepancy in sex between type 2 diabetic and control subjects, but no statistical heterogeneity was found between male and female ORs (P = 0.32). The adjusted allelic OR, using the Mantel-Haenszel approach, was 1.70 (1.56–1.85) (P = 3.2 x 10–35), very close to the allelic OR observed with the pooled data. There is evidence of an association of the at-risk allele with lower BMI in the type 2 diabetic group (30.40 ± 6.30 kg/m2 for C/C, 30.01 ± 5.71 kg/m2 for C/T, and 29.29 ± 5.75 kg/m2 for T/T; P = 8.0 x 10–3), and T- allele carriers were 2.4 years younger at type 2 diabetes diagnosis (48.78 ± 9.79 years of age for C/C, 47.62 ± 10.29 years of age for C/T, and 46.42 ± 10.14 years of age for T/T; P = 3.8 x 10–4). There is no evidence of an association with the variant and type 2 diabetes complications such as history of coronary heart disease or severe retinopathy or nephropathy (data not shown). In the individuals without diabetes, there was no evidence of an association between the T at-risk allele and homeostasis model assessment–derived -cell function (fasting insulinemia [μU/ml] x 20)/(fasting glycemia [mmol/l] – 3.5), homeostasis model assessment for insulin resistance {[fasting insulinemia (μU/ml) x fasting glycemia (mmol/l)]/22.5}, fasting insulinemia, fasting glycemia, total cholesterol, LDL cholesterol, HDL cholesterol, glycated hemoglobin, BMI, or triglycerides at baseline and during the 9-year follow-up (online appendix Table 1 [available at http://diabetes.diabetesjournals.org]). There is a trend for a decrease in fasting insulinemia between CC and TT individuals (39.17 ± 19.01 pmol/l for C/C, 38.08 ± 25.79 pmol/l for C/T, and 35.56 ± 19.64 pmol/l for T/T; P = 0.08).
We investigated the familial association with type 2 diabetes–associated traits in the 148 type 2 diabetes pedigrees, which showed evidence of linkage on chromosome 10q in 633 individuals, 432 with type 2 diabetes. We observed an overtransmission of the T at-risk allele in affected offspring (76.9% in 60 informative pedigrees, P = 1.8 x 10–4). The association with BMI and the age at type 2 diabetes diagnosis was confirmed by family-based association tests (FBATs) (P = 0.01 and P = 9.9 x 10–4, respectively). We determined that the TCF7L2 gene variant does not explain the linkage signal by testing the correlation between familial at-risk allele frequency and familial linkage score (nonparametric linkage) at different locations. This correlation was first tested at the marker position where the maximal logarithm of odds score is found (5) (D10S1655) and then at the marker closest to the TCF7L2 gene (D10S597). Lack of evidence of a correlation showed that the SNP does not explain the linkage.
Figure 1 shows that TCF7L2 is highly expressed in most human tissues, including mature pancreatic -cells, with the exception of the skeletal muscle. In human subcutaneous and omental adipose tissues, TCF7L2 expression is significantly impaired in obese subjects developing type 2 diabetes (P = 0.03 for omental fat and P = 0.02 for subcutaneous fat) (Fig. 2A). During rat fetal -cell differentiation, TCF7L2 expression pattern mimics neurogenin 3 (NGN3) (early-step key marker) (P < 0.02) but not paired box gene 6 (Pax6) (late-step key marker), suggesting a role in the gene expression cascade previously described in islet development (10) (Figs. 2A and B). Pancreatic TCF7L2 expression does not vary in NGN3 knock-out mice, possibly because it is expressed in other tissues, as well as the endocrine pancreas (data not shown).
DISCUSSION
Our data provides strong evidence that the transcription factor TCF7L2 is a major susceptibility gene for type 2 diabetes in the French population. This is supported by both case-control analysis and familial transmission disequilibrium test (TDT) performed on 4,866 subjects, as well as FBAT analyses that exclude obvious stratification biases. However, the FBAT analysis was not completely independent from the case-control study, since 126 unrelated type 2 diabetic subjects from the case-control study were from the 148 pedigrees. In addition to a significant association for the rs7903146 T at-risk allele (allelic OR = 1.7, P = 6.0 x 10–35), 31% of diabetes cases can be attributable to the presence of the T at-risk allele of the rs7903146 variant, which suggests that TCF7L2 is one of the strongest type 2 diabetes polygenes found to date. The contribution of TCF7L2 is even higher in nonobese (PAR = 37.7%). However, to calculate the PAR, the population genotypic frequencies were estimated by the control genotypic frequencies, and the PAR was probably overestimated due to clinically selected type 2 diabetic subjects. A more appropriate PAR estimation requires a larger general population analysis, which is currently underway. We found that the rs7903146 T at-risk allele modulates BMI in type 2 diabetic subjects (–1.1 kg/m2) and decreases the age at diagnosis by 2.4 years (further supported by quantitative FBAT). The strong association of rs7903146 T at-risk allele with type 2 diabetes in nonobese subjects makes it unlikely that the TCF7L2 diabetogenic effect directly involves insulin sensitivity or fat deposition.
The lack of association with the linkage peak suggests that other variants in the same region are also associated with type 2 diabetes, but as the linkage peak is weak (5), further studies are required in other pedigrees in regions of strong linkage in chromosome 10q (4,6,7). There was no evidence of an association with any metabolic parameter, but there was a trend for a decrease in fasting insulinemia in T at-risk allele carriers within the 2,499 control subjects (online appendix Table 1). This does not provide evidence of the gene effect on insulin secretion/sensitivity; therefore, further investigation in a sample where oral glucose tolerance is measured is required.
Recently, Yi et al. (9) found that TCF7L2 was not expressed in murine pancreatic islets but suggested that it may activate the proglucagon gene in enteroendocrine cells. In contrast, we report significant expression in human fluorescence-activated cell sorter–purified pancreatic -cells, suggesting that this transcription factor may be involved in -cell development and/or function. A putative role in -cell differentiation is suggested by the strong correlation between NGN3 and TCF7L2 expression during the early steps of rat pancreas development. In this regard, attenuated Wnt signaling might restrict pancreatic growth (11).
We found that the TCF7L2 gene is significantly expressed in human tissues that are vital for glucose homeostasis, including visceral and subcutaneous fat. Its expression is downregulated in obese subjects developing type 2 diabetes, but its relevance in physiology still requires evaluation. TCF7L2 expression needs to be measured in other tissues to assess the tissue specificity of such differential expression. Recent work suggests that the Calpain system is involved in the constitutive regulation of -catenin signaling functions (12), which raises the hypothesis of a potential interaction between TCF7L2 and Calpain-10.
In conclusion, our results not only replicate the Icelandic data but also provide strong evidence that TCF7L2 is a major determinant of diabetes risk in the European population. As previously suggested, an additive model of inheritance fits well with our data (8), which triples the risk of developing type 2 diabetes for homozygous T/T carriers (8.9% of type 2 diabetic subjects). Further work remains to be completed to evaluate the contribution of TCF7L2 in other ethnic groups and to elucidate the pathways controlled by this transcription factor. Its -cell expression and its downregulation with type 2 diabetes occurrence suggest that it plays an important role in glucose homeostasis.
RESEARCH DESIGN AND METHODS
The clinical characteristics of all French type 2 diabetic subjects are described in Table 1. Type 2 diabetic subjects were diagnosed with diabetes according to World Health Organization criteria or were currently being treated for diabetes. Altogether, 355 probands from families with type 2 diabetes were recruited by the Centre National de la Recherche Scientifique–Institut Pasteur Unit in Lille, and 2,012 type 2 diabetic subjects were patients of the Endocrinology-Diabetology Department at Corbeil-Essonnes Hospital. Two groups of healthy control subjects of French origin were analyzed (fasting glucose <5.6 mmol/l, age at exam >40 years, BMI <30 kg/m2), including 228 normoglycemic subjects from the family collection recruited in Lille and 2,271 normoglycemic nonobese control subjects from the DESIR (Angers) prospective population-based cohort (13) were used for replication. The nonobese subgroup of the type 2 diabetic subjects (n = 1,346, BMI <30 kg/m2) was analyzed to estimate the association of the variants with the nonobese type 2 diabetic subjects. Genotypic homogeneity was tested by 2 test, and no significant difference was found between control subjects (Angers vs. Lille) or between type 2 diabetic subjects (Corbeil-Essones vs. Lille). In this study, 8.8% of type 2 diabetic subjects have a Maghrebian origin. Analyses have been repeated with Maghrebian or European subjects only (online appendix Tables 2 and 3). Genotypic homogeneity was tested by 2 test, and no significant difference was found between type 2 diabetic subjects (Maghrebian vs. European). In addition, familial association tests were performed in 148 French pedigrees recruited by UMR8090 that had been previously scanned for linkage with type 2 diabetes (5). These families consisted of a total of 633 individuals, of whom 432 presented with diabetes (sex ratio male:female 198:234, mean age at diagnosis 49.5 ± 10.6 years, and mean BMI 27.9 ± 4.5 kg/m2), 72 were glucose intolerant (sex ratio 36:36, mean age at diagnosis 59.1 ± 9.6 years, and mean BMI 27.4 ± 4.7 kg/m2), and 129 were normoglycemic subjects (sex ratio 40:89 and mean BMI 25.1 ± 4.1 kg/m2). A total of 126 unrelated type 2 diabetic individuals from the case-control study were from the 148 pedigrees with type 2 diabetes. The genetic study was approved by the ethical committee of Hotel-Dieu in Corbeil-Essonnes and CHRU in Lille, and informed consent was obtained from all participants.
Genotyping methods.
High-throughput genotyping of SNPs rs12255372 and rs7903146 was performed using TaqMan SNP Genotyping Assays (Applied Biosystems, Foster City, CA). The PCR primers and TaqMan probes were designed by Primer Express and optimized according to the manufacturer’s protocol. There was a 98% genotyping success rate, and the genotyping error rate was assessed by sequencing 384 control and 384 type 2 diabetic individuals and by regenotyping a random 10% of subjects. No difference was found with the first genotyping results; therefore, the genotyping error rate was 0%.
TCF7L2 mRNA expression.
Human cDNAs from the human multiple tissue cDNA panel (BD Biosciences Clontech), fluorescence-activated cell sorter–purified pancreatic -cells (provided by the Human Pancreatic Cell Core Facility, University Hospital, Lille, France), and subcutaneous and omental adipocytes were used for mRNA expression analysis (Fig. 1). -Cell purity was confirmed by immunochemistry (98% insulin-positive cells) and PCR (absence of amplification with chymotrypsin primers, specific for exocrine cells, and presence of amplification with Pdx1 primers, specific for -cells). The forward primer used for the multiple tissue cDNA panel was 5'-ACCATGTCCACCCCCTCACGCCTCT-3', and the reverse was 5'-TGGGGCTTCTTCTTTTCTTCTTCCTTTTTG-3'. These two primers were high-performance liquid chromatography purified. PCR was performed in a 25-μl mixture containing 10 mmol/l Tris-HCl, 1.5 mmol/l MgCl2, 10 mmol/l of each dNTPs, 2.5 units Taq polymerase (Promega), 30 mmol/l of both forward and reverse primer, and 3 μl single-strand cDNA, using the hot-start PCR method of 95°C for 2 min, 45 cycles of 95°C for 30 s, 68°C for 2 min, 72°C for 2 min, and then 72°C for 10 min. PCR products were separated on 2% (wt/vol) agarose gel and visualized using ethidium bromide and ultraviolet transillumination, and -actin amplification was used as internal control. Omental and subcutaneous adipose tissue from normoglycemic (n = 3, all women, aged 44 ± 7 years, BMI 39.2 ± 3.8 kg/m2), glucose intolerant (n = 3, all women, aged 45 ± 7 years, BMI 45.3 ± 7.0 kg/m2), and type 2 diabetic (n = 3, two women and one man, aged 55.7 ± 5.8 years, BMI 49.0 ± 7.4 kg/m2) individuals were analyzed for TCF7L2 expression in HG-U133_Plus2 (Affymetrix microarray) (Fig. 2A). Affymetrix chip HG-U133_Plus2 contains 15 probes for TCF7L2 genes. For six probes, the signal is too low to detect any expression level. The probe "212761_at" was selected to represent TCF7L2 gene, since it has the higher signal and was located in the 3' untranslated region of the gene. Expressions were computed using the GeneChip Robust Multi-array Analysis method, and normalization was performed for each condition using the quantile method. Both methods are implemented using R language in the bioconductor package. TCF7L2, Pax6, and NGN3 gene expressions were assessed in a rat pancreas development model in RAE230 (Affymetrix microarray). Rat pancreas embryonic samples were cultured in triplicate (n = 3 different embryos from the same mother) during different time lapses (0, 1, 3, 5, and 7 days), mimicking the endocrine cell differentiation and proliferation (Fig. 2B). Pancreatic TCF7L2 expression was also measured in NGN3 knockout mice by TaqMan (14).
Statistical methods.
Tests for deviation from Hardy-Weinberg equilibrium and for association were performed with the De Finetti program (http://ihg.gsf.de/cgi-bin/hw/hwa1.pl). The PAR was based on a model-free hypothesis, with two level of exposition, heterozygous and homozygous variants. To calculate the PAR, the population genotypic frequencies were estimated by the control genotypic frequencies. A multivariate linear regression model taking into account age and sex was performed for BMI parameters. The same analysis was done with BMI adjustment for diabetes-onset parameters. Analysis on binary and quantitative traits was done by the TDT and quantitative TDT methods implemented in FBAT software. Evidence for association between identity-by-descent sharing and possession of the T at-risk allele was also investigated using the Genotype-IBD Sharing Test, which assigns family-specific weights based on the genotype of the affected family members and the model of interest (dominant, recessive, and additive) and tests for a positive correlation between this weight and the family-based identity-by-descent sharing as represented by the nonparametric linkage score. All reported P values are two tailed. SPSS 10.1 software was used for general statistical analyses.
ACKNOWLEDGMENTS
This work was partly supported by the French Governmental Agence Nationale de la Recherche and the charity Association Franaise des Diabetiques.
We thank Marianne Deweider and Frederic Allegaert for DNA bank management, Cecile Lecoeur and Stefan Gaget for help with phenotype databases, and Kirsten Ward for grammar and spelling improvement of the manuscript. We are indebted to all the subjects who participated in the study.
FOOTNOTES
Additional information for this article can be found in an online appendix at http://diabetes.diabetesjournals.org.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
REFERENCES
Horikawa Y, Oda N, Cox NJ, Li X, Orho-Melander M, Hara M, Hinokio Y, Lindner TH, Mashima H, Schwarz PE, del Bosque-Plata L, Horikawa Y, Oda Y, Yoshiuchi I, Colilla S, Polonsky KS, Wei S, Concannon P, Iwasaki N, Schulze J, Baier LJ, Bogardus C, Groop L, Boerwinkle E, Hanis CL, Bell GI: Genetic variation in the gene encoding calpain-10 is associated with type 2 diabetes mellitus. Nat Genet 26:163–175, 2000
Love-Gregory LD, Wasson J, Ma J, Jin CH, Glaser B, Suarez BK, Permutt MA: A common polymorphism in the upstream promoter region of the hepatocyte nuclear factor-4 gene on chromosome 20q is associated with type 2 diabetes and appears to contribute to the evidence for linkage in an Ashkenazi Jewish population. Diabetes 53:1134–1140, 2004
Hattersley AT, McCarthy MI: What makes a good genetic association study Lancet 366:1315–1323, 2005
Duggirala R, Blangero J, Almasy L, Dyer TD, Williams KL, Leach RJ, O’Connell P, Stern MP: Linkage of type 2 diabetes mellitus and of age at onset to a genetic location on chromosome 10q in Mexican Americans. Am J Hum Genet 64:1127–1140, 1999
Vionnet N, Hani El H, Dupont S, Gallina S, Francke S, Dotte S, De Matos F, Durand E, Lepretre F, Lecoeur C, Gallina P, Zekiri L, Dina C, Froguel P: Genomewide search for type 2 diabetes-susceptibility genes in French whites: evidence for a novel susceptibility locus for early-onset diabetes on chromosome 3q27-qter and independent replication of a type 2-diabetes locus on chromosome 1q21–q24. Am J Hum Genet 67:1470–1480, 2000
Wiltshire S, Hattersley AT, Hitman GA, Walker M, Levy JC, Sampson M, O’Rahilly S, Frayling TM, Bell JI, Lathrop GM, Bennett A, Dhillon R, Fletcher C, Groves CJ, Jones E, Prestwich P, Simecek N, Rao PV, Wishart M, Bottazzo GF, Foxon R, Howell S, Smedley D, Cardon LR, Menzel S, McCarthy MI: A genomewide scan for loci predisposing to type 2 diabetes in a U.K. population (the Diabetes UK Warren 2 Repository): analysis of 573 pedigrees provides independent replication of a susceptibility locus on chromosome 1q. Am J Hum Genet 69:553–569, 2001
Reynisdottir I, Thorleifsson G, Benediktsson R, Sigurdsson G, Emilsson V, Einarsdottir AS, Hjorleifsdottir EE, Orlygsdottir GT, Bjornsdottir GT, Saemundsdottir J, Halldorsson S, Hrafnkelsdottir S, Sigurjonsdottir SB, Steinsdottir S, Martin M, Kochan JP, Rhees BK, Grant SF, Frigge ML, Kong A, Gudnason V, Stefansson K, Gulcher JR: Localization of a susceptibility gene for type 2 diabetes to chromosome 5q34–q35.2. Am J Hum Genet 73:323–335, 2003
Grant SF, Thorleifsson G, Reynisdottir I, Benediktsson R, Manolescu A, Sainz J, Helgason A, Stefansson H, Emilsson V, Helgadottir A, Styrkarsdottir U, Magnusson KP, Walters GB, Palsdottir E, Jonsdottir T, Gudmundsdottir T, Gylfason A, Saemundsdottir J, Wilensky RL, Reilly MP, Rader DJ, Bagger Y, Christiansen C, Gudnason V, Sigurdsson G, Thorsteinsdottir U, Gulcher JR, Kong A, Stefansson K: Variant of transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2 diabetes. Nat Genet 38:320–323, 2006
Yi F, Brubaker PL, Jin T: TCF-4 mediates cell type-specific regulation of proglucagon gene expression by beta-catenin and glycogen synthase kinase-3beta. J Biol Chem 280:1457–1464, 2005
Wilson ME, Scheel D, German MS: Gene expression cascades in pancreatic development. Mech Dev 120:65–80, 2003
Papadopoulou S, Edlund H: Attenuated Wnt signaling perturbs pancreatic growth but not pancreatic function. Diabetes 54:2844–2851, 2005
Benetti R, Copetti T, Dell’Orso S, Melloni E, Brancolini C, Monte M, Schneider C: The calpain system is involved in the constitutive regulation of beta-catenin signaling functions. J Biol Chem 280:22070–22080, 2005
Gallois Y, Vol S, Caces E, Balkau B: Distribution of fasting serum insulin measured by enzyme immunoassay in an unselected population of 4,032 individuals: reference values according to age and sex: D.E.S.I.R. Study Group: Donnees Epidemiologiques sur le Syndrome d’Insulino-Resistance. Diabetes Metab 22:427–431, 1996
Duvillie B, Attali M, Bounacer A, Ravassard P, Basmaciogullari A, Scharfmann R: The mesenchyme controls the timing of pancreatic -cell differentiation. Diabetes 55:582–589, 2006(Stephane Cauchi, David Meyre, Christian )
2 INSERM U780-IFR69, Villejuif, France
3 Corbeil Hospital, Corbeil, France
4 INSERM ERIT-M 0106, Lille, France
5 Necker Hospital, INSERM E363, Paris, France
6 Atherosclerosis Department, Pasteur Institute, Lille, France
7 INSERM U545, Lille, France
8 University of Lille 2, Lille, France
9 Department of Metabolic Research, Clinica Universitaria de Navarra, Pamplona, Spain
10 Section of Genomic Medicine, Hammersmith Hospital, Imperial College, London, U.K
FBAT, family-based association test; GLP-1, glucagon-like peptide 1; PAR, population-attributable risk; SNP, single nucleotide polymorphism; TDT, transmission disequilibrium test
ABSTRACT
Recently, the transcription factor 7-like 2 (TCF7L2) gene has been associated with type 2 diabetes in subjects of European origin in the DeCode study. We genotyped the two most associated variants (rs7903146 and rs12255372) in 2,367 French type 2 diabetic subjects and in 2,499 control subjects. Both the T-allele of rs7903146 and the T-allele of rs12255372 significantly increase type 2 diabetes risk with an allelic odds ratio (OR) of 1.69 (95% CI 1.55–1.83) (P = 6.0 x 10–35) and 1.60 (1.47–1.74) (P = 7.6 x 10–28), respectively. In nonobese type 2 diabetic subjects (BMI <30 kg/m2, n = 1,346), the ORs increased to 1.89 (1.72–2.09) (P = 2.1 x 10–38) and 1.79 (1.62–1.97) (P = 5.7 x 10–31), respectively. The rs7903146 T at-risk allele associates with decreased BMI and earlier age at diagnosis in the type 2 diabetic subjects (P = 8.0 x 10–3 and P = 3.8 x 10–4, respectively), which is supported by quantitative family-based association tests. TCF7L2 is expressed in most human tissues, including mature pancreatic -cells, with the exception of the skeletal muscle. In the subcutaneous and omental fat from obese type 2 diabetic subjects, TCF7L2 expression significantly decreased compared with obese normoglycemic individuals. During rat fetal -cell differentiation, TCF7L2 expression pattern mimics the key marker NGN3 (neurogenin 3), suggesting a role in islet development. These data provide evidence that TCF7L2 is a major determinant of type 2 diabetes risk in European populations and suggests that this transcription factor plays a key role in glucose homeostasis.
To date, positional cloning for type 2 diabetes genes has not been very successful in detecting type 2 diabetes putative susceptibility genes. Although several positive reports have emerged (1,2), few have been consistently replicated (3). Several studies have shown linkage between type 2 diabetes and chromosome 10q in Mexican-American, French, English, and Icelandic populations (4–7) with a strong association between type 2 diabetes and variation in the transcription factor 7-like 2 (TCF7L2) gene found in Icelandic, Danish, and American populations (8). Little is known about the physiological implication of this transcription factor in glucose homeostasis. It has been suggested that intestinal proglucagon gene expression may be regulated by the Wnt/TCF7L2 pathway in enteroendocrine cells (9). Thus, TCF7L2 variants may modify type 2 diabetes susceptibility through modulation of glucagon-like peptide-1 (GLP-1) secretion. Because French type 2 diabetic families exhibit evidence for linkage in a chromosome 10q region encompassing the TCF7L2 locus, we assessed the contribution of TCF7L2 genetic variation to type 2 diabetes genetic risk. Tissue profiling analyses were also performed in rodents and humans to help understand the regulation of TCF7L2 and its involvement in the physiology of type 2 diabetes.
Among the previously reported TCF7L2 variants, the association between the T-allele of rs7903146 and the T- allele of rs12255372 have been replicated (8). These two variants were genotyped in 2,367 French type 2 diabetic subjects and in 2,499 normoglycemic nonobese control middle-aged individuals (fasting glucose <5.6 mmol/l and BMI <30 kg/m2) (Table 1). We found a strong association between type 2 diabetes and the two variants in our case-control study (Table 2). The genotypic distributions for both variants were in Hardy-Weinberg equilibrium, and the T-alleles of rs7903146 and rs12255372 were both found to significantly increase type 2 diabetes risk with an allelic odds ratio (OR) of 1.69 (95% CI 1.55–1.83) (P = 6.0 x 10–35) and 1.60 (1.47–1.74) (P = 7.6 x 10–28), respectively. Testing for the two-variant haplotype did not produce a better result than the two variants analyzed individually. Analyzing the 1,346 nonobese type 2 diabetic subjects (BMI <30 kg/m2) by themselves increases the allelic OR to 1.89 (1.72–2.09) (P = 2.1 x 10–38) and 1.79 (1.62–1.97) (P = 5.7 x 10–31), respectively.
As the two variants exhibit strong linkage disequilibrium (r2 = 0.77, D' = 0.89), further analyses focused on rs7903146 (the single nucleotide polymorphism [SNP] with the strongest association to type 2 diabetes). The population-attributable risk (PAR) for rs7903146 was estimated at 31% using our case-control sample. When the allelic OR is estimated in the nonobese subjects versus the control subjects (BMI <30 kg/m2), the attributable risk reaches 37.7%. The T-allele frequency was 30.5% in control individuals and 42.5% in type 2 diabetic subjects. We tested departure from the additive model in a logistic regression framework, where an index variable identifies heterozygous individuals. No significant deviation was observed (P = 0.74), supporting the hypothesis of a risk increase proportional to the number of variant alleles (genotypic OR = 1.66 for C/C vs. C/T and genotypic OR = 2.86 for C/C vs. T/T). In this study, there is a discrepancy in sex between type 2 diabetic and control subjects, but no statistical heterogeneity was found between male and female ORs (P = 0.32). The adjusted allelic OR, using the Mantel-Haenszel approach, was 1.70 (1.56–1.85) (P = 3.2 x 10–35), very close to the allelic OR observed with the pooled data. There is evidence of an association of the at-risk allele with lower BMI in the type 2 diabetic group (30.40 ± 6.30 kg/m2 for C/C, 30.01 ± 5.71 kg/m2 for C/T, and 29.29 ± 5.75 kg/m2 for T/T; P = 8.0 x 10–3), and T- allele carriers were 2.4 years younger at type 2 diabetes diagnosis (48.78 ± 9.79 years of age for C/C, 47.62 ± 10.29 years of age for C/T, and 46.42 ± 10.14 years of age for T/T; P = 3.8 x 10–4). There is no evidence of an association with the variant and type 2 diabetes complications such as history of coronary heart disease or severe retinopathy or nephropathy (data not shown). In the individuals without diabetes, there was no evidence of an association between the T at-risk allele and homeostasis model assessment–derived -cell function (fasting insulinemia [μU/ml] x 20)/(fasting glycemia [mmol/l] – 3.5), homeostasis model assessment for insulin resistance {[fasting insulinemia (μU/ml) x fasting glycemia (mmol/l)]/22.5}, fasting insulinemia, fasting glycemia, total cholesterol, LDL cholesterol, HDL cholesterol, glycated hemoglobin, BMI, or triglycerides at baseline and during the 9-year follow-up (online appendix Table 1 [available at http://diabetes.diabetesjournals.org]). There is a trend for a decrease in fasting insulinemia between CC and TT individuals (39.17 ± 19.01 pmol/l for C/C, 38.08 ± 25.79 pmol/l for C/T, and 35.56 ± 19.64 pmol/l for T/T; P = 0.08).
We investigated the familial association with type 2 diabetes–associated traits in the 148 type 2 diabetes pedigrees, which showed evidence of linkage on chromosome 10q in 633 individuals, 432 with type 2 diabetes. We observed an overtransmission of the T at-risk allele in affected offspring (76.9% in 60 informative pedigrees, P = 1.8 x 10–4). The association with BMI and the age at type 2 diabetes diagnosis was confirmed by family-based association tests (FBATs) (P = 0.01 and P = 9.9 x 10–4, respectively). We determined that the TCF7L2 gene variant does not explain the linkage signal by testing the correlation between familial at-risk allele frequency and familial linkage score (nonparametric linkage) at different locations. This correlation was first tested at the marker position where the maximal logarithm of odds score is found (5) (D10S1655) and then at the marker closest to the TCF7L2 gene (D10S597). Lack of evidence of a correlation showed that the SNP does not explain the linkage.
Figure 1 shows that TCF7L2 is highly expressed in most human tissues, including mature pancreatic -cells, with the exception of the skeletal muscle. In human subcutaneous and omental adipose tissues, TCF7L2 expression is significantly impaired in obese subjects developing type 2 diabetes (P = 0.03 for omental fat and P = 0.02 for subcutaneous fat) (Fig. 2A). During rat fetal -cell differentiation, TCF7L2 expression pattern mimics neurogenin 3 (NGN3) (early-step key marker) (P < 0.02) but not paired box gene 6 (Pax6) (late-step key marker), suggesting a role in the gene expression cascade previously described in islet development (10) (Figs. 2A and B). Pancreatic TCF7L2 expression does not vary in NGN3 knock-out mice, possibly because it is expressed in other tissues, as well as the endocrine pancreas (data not shown).
DISCUSSION
Our data provides strong evidence that the transcription factor TCF7L2 is a major susceptibility gene for type 2 diabetes in the French population. This is supported by both case-control analysis and familial transmission disequilibrium test (TDT) performed on 4,866 subjects, as well as FBAT analyses that exclude obvious stratification biases. However, the FBAT analysis was not completely independent from the case-control study, since 126 unrelated type 2 diabetic subjects from the case-control study were from the 148 pedigrees. In addition to a significant association for the rs7903146 T at-risk allele (allelic OR = 1.7, P = 6.0 x 10–35), 31% of diabetes cases can be attributable to the presence of the T at-risk allele of the rs7903146 variant, which suggests that TCF7L2 is one of the strongest type 2 diabetes polygenes found to date. The contribution of TCF7L2 is even higher in nonobese (PAR = 37.7%). However, to calculate the PAR, the population genotypic frequencies were estimated by the control genotypic frequencies, and the PAR was probably overestimated due to clinically selected type 2 diabetic subjects. A more appropriate PAR estimation requires a larger general population analysis, which is currently underway. We found that the rs7903146 T at-risk allele modulates BMI in type 2 diabetic subjects (–1.1 kg/m2) and decreases the age at diagnosis by 2.4 years (further supported by quantitative FBAT). The strong association of rs7903146 T at-risk allele with type 2 diabetes in nonobese subjects makes it unlikely that the TCF7L2 diabetogenic effect directly involves insulin sensitivity or fat deposition.
The lack of association with the linkage peak suggests that other variants in the same region are also associated with type 2 diabetes, but as the linkage peak is weak (5), further studies are required in other pedigrees in regions of strong linkage in chromosome 10q (4,6,7). There was no evidence of an association with any metabolic parameter, but there was a trend for a decrease in fasting insulinemia in T at-risk allele carriers within the 2,499 control subjects (online appendix Table 1). This does not provide evidence of the gene effect on insulin secretion/sensitivity; therefore, further investigation in a sample where oral glucose tolerance is measured is required.
Recently, Yi et al. (9) found that TCF7L2 was not expressed in murine pancreatic islets but suggested that it may activate the proglucagon gene in enteroendocrine cells. In contrast, we report significant expression in human fluorescence-activated cell sorter–purified pancreatic -cells, suggesting that this transcription factor may be involved in -cell development and/or function. A putative role in -cell differentiation is suggested by the strong correlation between NGN3 and TCF7L2 expression during the early steps of rat pancreas development. In this regard, attenuated Wnt signaling might restrict pancreatic growth (11).
We found that the TCF7L2 gene is significantly expressed in human tissues that are vital for glucose homeostasis, including visceral and subcutaneous fat. Its expression is downregulated in obese subjects developing type 2 diabetes, but its relevance in physiology still requires evaluation. TCF7L2 expression needs to be measured in other tissues to assess the tissue specificity of such differential expression. Recent work suggests that the Calpain system is involved in the constitutive regulation of -catenin signaling functions (12), which raises the hypothesis of a potential interaction between TCF7L2 and Calpain-10.
In conclusion, our results not only replicate the Icelandic data but also provide strong evidence that TCF7L2 is a major determinant of diabetes risk in the European population. As previously suggested, an additive model of inheritance fits well with our data (8), which triples the risk of developing type 2 diabetes for homozygous T/T carriers (8.9% of type 2 diabetic subjects). Further work remains to be completed to evaluate the contribution of TCF7L2 in other ethnic groups and to elucidate the pathways controlled by this transcription factor. Its -cell expression and its downregulation with type 2 diabetes occurrence suggest that it plays an important role in glucose homeostasis.
RESEARCH DESIGN AND METHODS
The clinical characteristics of all French type 2 diabetic subjects are described in Table 1. Type 2 diabetic subjects were diagnosed with diabetes according to World Health Organization criteria or were currently being treated for diabetes. Altogether, 355 probands from families with type 2 diabetes were recruited by the Centre National de la Recherche Scientifique–Institut Pasteur Unit in Lille, and 2,012 type 2 diabetic subjects were patients of the Endocrinology-Diabetology Department at Corbeil-Essonnes Hospital. Two groups of healthy control subjects of French origin were analyzed (fasting glucose <5.6 mmol/l, age at exam >40 years, BMI <30 kg/m2), including 228 normoglycemic subjects from the family collection recruited in Lille and 2,271 normoglycemic nonobese control subjects from the DESIR (Angers) prospective population-based cohort (13) were used for replication. The nonobese subgroup of the type 2 diabetic subjects (n = 1,346, BMI <30 kg/m2) was analyzed to estimate the association of the variants with the nonobese type 2 diabetic subjects. Genotypic homogeneity was tested by 2 test, and no significant difference was found between control subjects (Angers vs. Lille) or between type 2 diabetic subjects (Corbeil-Essones vs. Lille). In this study, 8.8% of type 2 diabetic subjects have a Maghrebian origin. Analyses have been repeated with Maghrebian or European subjects only (online appendix Tables 2 and 3). Genotypic homogeneity was tested by 2 test, and no significant difference was found between type 2 diabetic subjects (Maghrebian vs. European). In addition, familial association tests were performed in 148 French pedigrees recruited by UMR8090 that had been previously scanned for linkage with type 2 diabetes (5). These families consisted of a total of 633 individuals, of whom 432 presented with diabetes (sex ratio male:female 198:234, mean age at diagnosis 49.5 ± 10.6 years, and mean BMI 27.9 ± 4.5 kg/m2), 72 were glucose intolerant (sex ratio 36:36, mean age at diagnosis 59.1 ± 9.6 years, and mean BMI 27.4 ± 4.7 kg/m2), and 129 were normoglycemic subjects (sex ratio 40:89 and mean BMI 25.1 ± 4.1 kg/m2). A total of 126 unrelated type 2 diabetic individuals from the case-control study were from the 148 pedigrees with type 2 diabetes. The genetic study was approved by the ethical committee of Hotel-Dieu in Corbeil-Essonnes and CHRU in Lille, and informed consent was obtained from all participants.
Genotyping methods.
High-throughput genotyping of SNPs rs12255372 and rs7903146 was performed using TaqMan SNP Genotyping Assays (Applied Biosystems, Foster City, CA). The PCR primers and TaqMan probes were designed by Primer Express and optimized according to the manufacturer’s protocol. There was a 98% genotyping success rate, and the genotyping error rate was assessed by sequencing 384 control and 384 type 2 diabetic individuals and by regenotyping a random 10% of subjects. No difference was found with the first genotyping results; therefore, the genotyping error rate was 0%.
TCF7L2 mRNA expression.
Human cDNAs from the human multiple tissue cDNA panel (BD Biosciences Clontech), fluorescence-activated cell sorter–purified pancreatic -cells (provided by the Human Pancreatic Cell Core Facility, University Hospital, Lille, France), and subcutaneous and omental adipocytes were used for mRNA expression analysis (Fig. 1). -Cell purity was confirmed by immunochemistry (98% insulin-positive cells) and PCR (absence of amplification with chymotrypsin primers, specific for exocrine cells, and presence of amplification with Pdx1 primers, specific for -cells). The forward primer used for the multiple tissue cDNA panel was 5'-ACCATGTCCACCCCCTCACGCCTCT-3', and the reverse was 5'-TGGGGCTTCTTCTTTTCTTCTTCCTTTTTG-3'. These two primers were high-performance liquid chromatography purified. PCR was performed in a 25-μl mixture containing 10 mmol/l Tris-HCl, 1.5 mmol/l MgCl2, 10 mmol/l of each dNTPs, 2.5 units Taq polymerase (Promega), 30 mmol/l of both forward and reverse primer, and 3 μl single-strand cDNA, using the hot-start PCR method of 95°C for 2 min, 45 cycles of 95°C for 30 s, 68°C for 2 min, 72°C for 2 min, and then 72°C for 10 min. PCR products were separated on 2% (wt/vol) agarose gel and visualized using ethidium bromide and ultraviolet transillumination, and -actin amplification was used as internal control. Omental and subcutaneous adipose tissue from normoglycemic (n = 3, all women, aged 44 ± 7 years, BMI 39.2 ± 3.8 kg/m2), glucose intolerant (n = 3, all women, aged 45 ± 7 years, BMI 45.3 ± 7.0 kg/m2), and type 2 diabetic (n = 3, two women and one man, aged 55.7 ± 5.8 years, BMI 49.0 ± 7.4 kg/m2) individuals were analyzed for TCF7L2 expression in HG-U133_Plus2 (Affymetrix microarray) (Fig. 2A). Affymetrix chip HG-U133_Plus2 contains 15 probes for TCF7L2 genes. For six probes, the signal is too low to detect any expression level. The probe "212761_at" was selected to represent TCF7L2 gene, since it has the higher signal and was located in the 3' untranslated region of the gene. Expressions were computed using the GeneChip Robust Multi-array Analysis method, and normalization was performed for each condition using the quantile method. Both methods are implemented using R language in the bioconductor package. TCF7L2, Pax6, and NGN3 gene expressions were assessed in a rat pancreas development model in RAE230 (Affymetrix microarray). Rat pancreas embryonic samples were cultured in triplicate (n = 3 different embryos from the same mother) during different time lapses (0, 1, 3, 5, and 7 days), mimicking the endocrine cell differentiation and proliferation (Fig. 2B). Pancreatic TCF7L2 expression was also measured in NGN3 knockout mice by TaqMan (14).
Statistical methods.
Tests for deviation from Hardy-Weinberg equilibrium and for association were performed with the De Finetti program (http://ihg.gsf.de/cgi-bin/hw/hwa1.pl). The PAR was based on a model-free hypothesis, with two level of exposition, heterozygous and homozygous variants. To calculate the PAR, the population genotypic frequencies were estimated by the control genotypic frequencies. A multivariate linear regression model taking into account age and sex was performed for BMI parameters. The same analysis was done with BMI adjustment for diabetes-onset parameters. Analysis on binary and quantitative traits was done by the TDT and quantitative TDT methods implemented in FBAT software. Evidence for association between identity-by-descent sharing and possession of the T at-risk allele was also investigated using the Genotype-IBD Sharing Test, which assigns family-specific weights based on the genotype of the affected family members and the model of interest (dominant, recessive, and additive) and tests for a positive correlation between this weight and the family-based identity-by-descent sharing as represented by the nonparametric linkage score. All reported P values are two tailed. SPSS 10.1 software was used for general statistical analyses.
ACKNOWLEDGMENTS
This work was partly supported by the French Governmental Agence Nationale de la Recherche and the charity Association Franaise des Diabetiques.
We thank Marianne Deweider and Frederic Allegaert for DNA bank management, Cecile Lecoeur and Stefan Gaget for help with phenotype databases, and Kirsten Ward for grammar and spelling improvement of the manuscript. We are indebted to all the subjects who participated in the study.
FOOTNOTES
Additional information for this article can be found in an online appendix at http://diabetes.diabetesjournals.org.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
REFERENCES
Horikawa Y, Oda N, Cox NJ, Li X, Orho-Melander M, Hara M, Hinokio Y, Lindner TH, Mashima H, Schwarz PE, del Bosque-Plata L, Horikawa Y, Oda Y, Yoshiuchi I, Colilla S, Polonsky KS, Wei S, Concannon P, Iwasaki N, Schulze J, Baier LJ, Bogardus C, Groop L, Boerwinkle E, Hanis CL, Bell GI: Genetic variation in the gene encoding calpain-10 is associated with type 2 diabetes mellitus. Nat Genet 26:163–175, 2000
Love-Gregory LD, Wasson J, Ma J, Jin CH, Glaser B, Suarez BK, Permutt MA: A common polymorphism in the upstream promoter region of the hepatocyte nuclear factor-4 gene on chromosome 20q is associated with type 2 diabetes and appears to contribute to the evidence for linkage in an Ashkenazi Jewish population. Diabetes 53:1134–1140, 2004
Hattersley AT, McCarthy MI: What makes a good genetic association study Lancet 366:1315–1323, 2005
Duggirala R, Blangero J, Almasy L, Dyer TD, Williams KL, Leach RJ, O’Connell P, Stern MP: Linkage of type 2 diabetes mellitus and of age at onset to a genetic location on chromosome 10q in Mexican Americans. Am J Hum Genet 64:1127–1140, 1999
Vionnet N, Hani El H, Dupont S, Gallina S, Francke S, Dotte S, De Matos F, Durand E, Lepretre F, Lecoeur C, Gallina P, Zekiri L, Dina C, Froguel P: Genomewide search for type 2 diabetes-susceptibility genes in French whites: evidence for a novel susceptibility locus for early-onset diabetes on chromosome 3q27-qter and independent replication of a type 2-diabetes locus on chromosome 1q21–q24. Am J Hum Genet 67:1470–1480, 2000
Wiltshire S, Hattersley AT, Hitman GA, Walker M, Levy JC, Sampson M, O’Rahilly S, Frayling TM, Bell JI, Lathrop GM, Bennett A, Dhillon R, Fletcher C, Groves CJ, Jones E, Prestwich P, Simecek N, Rao PV, Wishart M, Bottazzo GF, Foxon R, Howell S, Smedley D, Cardon LR, Menzel S, McCarthy MI: A genomewide scan for loci predisposing to type 2 diabetes in a U.K. population (the Diabetes UK Warren 2 Repository): analysis of 573 pedigrees provides independent replication of a susceptibility locus on chromosome 1q. Am J Hum Genet 69:553–569, 2001
Reynisdottir I, Thorleifsson G, Benediktsson R, Sigurdsson G, Emilsson V, Einarsdottir AS, Hjorleifsdottir EE, Orlygsdottir GT, Bjornsdottir GT, Saemundsdottir J, Halldorsson S, Hrafnkelsdottir S, Sigurjonsdottir SB, Steinsdottir S, Martin M, Kochan JP, Rhees BK, Grant SF, Frigge ML, Kong A, Gudnason V, Stefansson K, Gulcher JR: Localization of a susceptibility gene for type 2 diabetes to chromosome 5q34–q35.2. Am J Hum Genet 73:323–335, 2003
Grant SF, Thorleifsson G, Reynisdottir I, Benediktsson R, Manolescu A, Sainz J, Helgason A, Stefansson H, Emilsson V, Helgadottir A, Styrkarsdottir U, Magnusson KP, Walters GB, Palsdottir E, Jonsdottir T, Gudmundsdottir T, Gylfason A, Saemundsdottir J, Wilensky RL, Reilly MP, Rader DJ, Bagger Y, Christiansen C, Gudnason V, Sigurdsson G, Thorsteinsdottir U, Gulcher JR, Kong A, Stefansson K: Variant of transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2 diabetes. Nat Genet 38:320–323, 2006
Yi F, Brubaker PL, Jin T: TCF-4 mediates cell type-specific regulation of proglucagon gene expression by beta-catenin and glycogen synthase kinase-3beta. J Biol Chem 280:1457–1464, 2005
Wilson ME, Scheel D, German MS: Gene expression cascades in pancreatic development. Mech Dev 120:65–80, 2003
Papadopoulou S, Edlund H: Attenuated Wnt signaling perturbs pancreatic growth but not pancreatic function. Diabetes 54:2844–2851, 2005
Benetti R, Copetti T, Dell’Orso S, Melloni E, Brancolini C, Monte M, Schneider C: The calpain system is involved in the constitutive regulation of beta-catenin signaling functions. J Biol Chem 280:22070–22080, 2005
Gallois Y, Vol S, Caces E, Balkau B: Distribution of fasting serum insulin measured by enzyme immunoassay in an unselected population of 4,032 individuals: reference values according to age and sex: D.E.S.I.R. Study Group: Donnees Epidemiologiques sur le Syndrome d’Insulino-Resistance. Diabetes Metab 22:427–431, 1996
Duvillie B, Attali M, Bounacer A, Ravassard P, Basmaciogullari A, Scharfmann R: The mesenchyme controls the timing of pancreatic -cell differentiation. Diabetes 55:582–589, 2006(Stephane Cauchi, David Meyre, Christian )