Introduction

In terms of aetiology, genetics and pathogenesis, Type 1 and Type 2 diabetes mellitus are considered to be two distinct diseases. There is, however, accumulating evidence that inflammatory mediators are important in the pathogenesis of both diseases. Beta cell loss in Type 1 and Type 2 diabetes mellitus may be a consequence of apoptosis and necrosis induced by inflammatory mediators. In Type 1 diabetes mellitus, beta cell destruction results from an immune-mediated reaction towards the beta cell. Proinflammatory cytokines are released during inflammation of the pancreatic islets and in synergy they lead to apoptosis and necrosis of beta cells [1]. In Type 2 diabetes mellitus, beta cell deficit and increased beta cell apoptosis have also been reported [2], as well as elevated circulating levels of proinflammatory cytokines in patients with recent onset of the disease [3, 4, 5]. In addition, hyperglycaemia in Type 2 diabetic patients is accompanied by increased levels of proinflammatory cytokines [6]. Finally, numerous studies have shown that several proinflammatory cytokines induce insulin resistance [7, 8, 9]. Thus, cytokines that are produced locally as part of the autoimmune infiltrate in Type 1 diabetes mellitus or are expressed in beta cells and/or at high levels in circulation in response to hyperglycaemia in Type 2 diabetes may be a common pathogenetic denominator in beta cell failure in the two diseases.

A number of natural inhibitors of cytokine signalling have recently been characterised. The suppressor of cytokine signalling (SOCS) family is a group of proteins originally identified as negative feedback regulators of IFN- γ signalling [10]. We recently reported that SOCS3 also down-regulates IL-1 signalling [11]. Interestingly, SOCS3 up-regulation in response to IL-1 in native rat islets is delayed and may therefore be insufficient to down-regulate IL-1-mediated signalling and thereby increase beta cell sensitivity to cytokines [11]. TNF-α induces insulin resistance both indirectly by stimulating stress hormone production, and directly by sustained induction of SOCS3, which decreases insulin-induced IRS1 tyrosine phosphorylation and its association with the p85 regulatory subunit of phosphatidylinositol-3 kinase [7]. SOCS3 was shown to modulate insulin signalling by targeting IRS1 and IRS2, the two key signalling proteins in insulin action [12]. It has also been proposed that the adipocyte hormone leptin, which induces SOCS3, impedes insulin signalling. Growth hormone (GH) also induces SOCS3, thereby leading to insulin resistance [13].

Insulin resistance, which often precedes the onset of Type 2 diabetes mellitus by several years, affects a large segment of the general population. There is increasing evidence that genetic components code for insulin resistance [14, 15, 16], which is a feature of the offspring of parents with Type 2 diabetes mellitus and, as shown by longitudinal studies of families affected by the disease, a major risk factor for developing Type 2 diabetes mellitus [17].

For these reasons, we consider SOCS3 to be a candidate gene in the pathogenesis of Type 1 diabetes and insulin resistance. SOCS3 maps to chromosome 17q25 and consists of only one exon spanning 850 nucleotides. The upstream sequence is a G/C rich region and the promoter area is by definition a GpC island [18]. In the present study we performed a mutation scanning of the promoter region, the exon and the 3′UTR of the human SOCS3 gene. We also examined the identified variants for an association with Type 1 diabetes mellitus and insulin sensitivity.

Subjects, materials and methods

Subjects

Mutation scanning was performed in seven control subjects and 21 Type 1 diabetic patients by single-strand conformational polymorphism (SSCP) and direct sequencing. To estimate the frequencies of minor alleles, the single nucleotide polymorphisms (SNPs) identified were tested further in a panel of 100 Type 1 diabetic patients. An identified promoter variant was examined in three separate groups of subjects involving:

  1. 1.

    A collection of families affected by Type 1 diabetes mellitus. This group comprised 250 Danish families (1097 individuals), consisting of 99 simplex and 151 multiplex families. The median age (range) at onset for probands was 11 years (0–29 years) [19, 20]. HLA stratification was performed for all probands and/or affected offspring, who were subsequently divided into high-risk DR3/4 heterozygous and non-high-risk (non-DR3/4 heterozygous) HLA-types.

  2. 2.

    A group of 212 glucose-tolerant Danish first-degree relatives of Type 2 diabetes mellitus patients. This group was examined by a 75-g OGTT and an IVGTT in combination with an intravenous tolbutamide injection .The detailed phenotypic characteristics of this population sample have been reported previously [21].

  3. 3.

    A random population-based sample of young, healthy, unrelated Danish subjects (n=370) from Copenhagen. Aged 18 to 32 years, these subjects underwent an IVGTT in combination with an intravenous tolbutamide injection [22].

Because the penetrance of a single allele on a physiological trait like insulin sensitivity levels is likely to be higher in glucose-tolerant subjects than in patients with a complex metabolic disorder like Type 2 diabetes mellitus, we chose to investigate offspring of Type 2 diabetes mellitus subjects and healthy unrelated subjects. The subjects of the study gave informed consent and the investigations were approved by the responsible ethics committee and carried out in accordance with the Declaration of Helsinki as revised in 2000.

Mutation scanning and genotyping

SSCP and sequencing were performed as described previously [23, 24]. Sequence data were analysed using SeqScape version 2.0 (Applied Biosystems, Foster City, Calif., USA). Genotypings of polymorphisms were done by primer extension using the SnapShot Multiplex Kit (Applied Biosystems). In the PCR reaction the following primers were used 5′CTCCGCGCTCAGCCTTTCTCTGC 3′ and 5′TACCTGGTCCCGAATCGAAGTCT 3′. As a template we used 100 ng DNA in a total volume of 25 µl in a thermal cycler (Perkin Elmer, Boston, Mass., USA). In addition the PCR mixture contained 1.75 units Herculase-enhanced DNA polymerase (Stratagene, La Jolla, Calif., USA), 1× Herculase PCR buffer (Stratagene), 1 µmol/l of each primer, 25 µmol/l deoxynucleoside (5′-)triphosphate, DMSO 8% and H2O. Denaturing, annealing and extension temperatures were 95 °C, 64 °C and 72 °C respectively, for 30 seconds each for a total of 35 cycles. In the last 25 cycles 10 seconds were added to each cycle, followed by a final extension for 10 min at 72 °C. Primer extensions were performed according to the manufacturer’s instructions, using the following primer 5′ CGGGAGCTGGGCCGGGCGGGCGGC 3′, terminating one nucleotide before the polymorphism. The assay was performed on an ABI Prism 3100 (Applied Biosystems) and analysed using Genemapper version 2.0 (Applied Biosystems). Evaluations of the SNP database (NCBI [25]) revealed three rare variations of the SOCS gene region. One of these (rs 1061489), which causes a tyrosine to histidine substitution (Y125H), was also evaluated by RFLP-PCR assay in the current study.

Phenotypic characterisation

In Groups 2 and 3, comprising glucose-tolerant first-degree relatives of Danish Type 2 diabetic patients and healthy, unrelated, Danish subjects respectively, the insulin sensitivity index (SI) was measured applying Bergman’s minimal model computing data, obtained during an intravenous glucose and tolbutamide tolerance test. The insulinogenic index for insulin was calculated as (insulin t=30 min−insulin t=0 min/glucose t=30 min). The insulinogenic index for C-peptide was calculated as (C-peptide t=30 min−C-peptide t=0 min/glucose t=30 min).

Statistical analysis

The families with Type 1 diabetes mellitus were evaluated by the Sib Transmission Disequilibrium Test (Sib-TDT) [26, 27]. In relatives of Type 2 diabetic patients and healthy glucose-tolerant subjects, differences in continuous variables between carriers of the polymorphism were tested using a general linear model or a mixed model for analysis of variance with age, BMI and sex as covariates. All residuals were tested for normal distribution and proper transformation of the variables was made if necessary (log-transformation). SPSS for Windows version 11.0 (SPSS, Chicago, Ill., USA) or the SAS System for Windows version 8.2 (SAS Institute, Cary, N.C., USA) were used for statistical analysis. We considered p values of less than 0.05 statistically significant. All genotype distributions were tested for Hardy Weinberg equilibrium using a likelihood ratio test.

Results

Mutation scanning

Three novel mutations were identified in the promoter region, but none in the coding or 3′UTR region. Two mutations (C −1044→A and G −202→A) had allele frequencies below 1% in the group of 100 Type 1 diabetic individuals and were not tested further. The third, a C −920→A substitution in the putative promoter had an allele frequency of 8%. Two different software analyses [28, 29] demonstrated that the mutation causes deletion of an activator protein 2 (AP2) transcription factor binding site within the sequence encompassing the polymorphism. AP2 is involved, for example, in angiogenesis, tumour invasion/metastasis, and chronic inflammation [30]. Sib-TDT analysis showed no significant linkage to Type 1 diabetes mellitus in the 250 families with Type 1 diabetes (Table 1). Stratification by high- and low-risk HLA-type also failed to reveal significant differences (data not shown). In addition the dbSNP (rs1061489) in the coding region (Y125H) was genotyped by RFLP in a group of 100 Danish Type 1 diabetic patients. None carried this mutation. Two other dbSNPs (rs 12059, rs 2280148) were not identified in the group screened, preventing construction of haplotypes covering the region.

Table 1 Sib Transmission Disequilibrium Test (of the C −920→A polymorphism of SOCS3) in the 250 families with Type 1 diabetes mellitus (1097 individuals)
Full size table

For offspring of Type 2 diabetic patients and the control subjects, phenotypic data stratified according to genotype are shown in Tables 2 and 3.

Table 2 Clinical and biochemical data of 212 glucose-tolerant first-degree relatives of Type 2 diabetic patients, stratified according to the C −920→A polymorphism of SOCS3
Full size table
Table 3 Clinical and biochemical data of 370 young healthy Danish Caucasians stratified according to the C −920→A polymorphism of SOCS3
Full size table

In the 370 young healthy Danish Caucasians (Table 3) the SI-values using IVGTT and Bergman minimal modelling were significantly higher among homozygous carriers of the A-allele than among heterozygous and wild-type subjects (AC and CC), despite higher BMI. In the offspring of Type 2 diabetic patients the same trend was found (Table 2). However, due to the low number of homozygous subjects, we were unable to perform any valid statistical analyses in this group of subjects. In the phenotype variables describing beta cell function there was a tendency towards lower acute insulin response (AIR) in the A/A groups with higher SI-values.

Discussion

We performed a mutation scanning of approximately 1000 bp upstream of the 5′UTR region, and of the coding region and approximately 200 bp of the 3′UTR region of the human SOCS3 gene. Three mutations were found in the promoter region of subjects of Danish Caucasoid origin. The most frequent polymorphism was tested in three different groups of individuals. In a population-based family collection no significant linkage to Type 1 diabetes mellitus was observed. Because of the known effects on IRS1 and IRS2 and the fact that prolonged cytokine stimulation leads to insulin resistance [31], this promoter polymorphism was also tested in a group of glucose-tolerant offspring of Type 2 diabetic patients (Group 2) and a group of young healthy subjects (Group 3). Comparing insulin sensitivity values from wild-type, heterozygous and homozygous carriers of the polymorphism in Group 3, the allele dose response suggests a recessive mode of inheritance of this relatively rare insulin-sensitive phenotype. Although the number of homozygous subjects was small, the observation was statistically significant (p=0.027), and a trend towards higher insulin sensitivity in subjects homozygous for the A allele was also found in Group 2. In the phenotype variables describing pancreatic beta cell function, the tendency towards lower AIR in the AA groups with higher SI-values is consistent with the known correlation between these variables. Because the phenotype data are not independent of each other, a conservative Bonferroni correction was not performed.

The presence of the AA genotype is compatible with a protective mechanism against the development of diabetes and even though this group had a higher BMI, insulin sensitivity was above normal. Whether the mutation in the promoter region influences the level of SOCS3 expression in beta cells when exposed to cytokines is still under investigation, but a possible molecular mechanism of the A-allele could be decreased SOCS3 expression leading to less degradation of IRS1 and IRS2 in insulin-sensitive tissues. Functional studies are being carried out to address this possibility. De novo methylation of the GpC island and the following silencing of neighbouring genes could be part of the reason for the effect of SOCS3 overexpression in beta cell lines and this also needs to be evaluated experimentally. Taking the data together, it could be hypothesised that homozygosity for the A-allele is a marker for insulin sensitivity, independent of genetic risk of getting Type 2 diabetes mellitus.

In conclusion, a mutation scanning of the human SOCS3 gene identified no variations associated with Type 1 diabetes mellitus. Homozygosity for a C −920→A promoter polymorphism with a minor allele frequency of 8% was shown to be associated with increased whole-body insulin sensitivity in a population of young healthy unrelated individuals. This finding should be interpreted with caution, and large-scale genetic epidemiology and functional studies are needed to clarify the potential role of this variant in the regulation of whole-body insulin sensitivity.