Variable number of tandem repeats of the insulin gene determines susceptibility to latent autoimmune diabetes in adults.

Abstract

Background: The different clinical presentations of latent autoimmune diabetes in adults (LADA) and type 1 diabetes mellitus may be the result of susceptibility genes in determining the mode of onset. We analyzed the 5' polymorphisms of the insulin mini-satellite region (INS), a variable number of tandem repeats (VNTR) [repeat units; RU]. We evaluated the association of the different INS-VNTR alleles in patient susceptibility to LADA autoimmune diabetes. To our knowledge, this constitutes the first study of this kind performed in a Caucasian population.

Methods: From an group of 160 Argentinean patients previously characterized as having LADA, we selected 44 patients who presented with humoral autoimmunity for genotyping and compared them to 88 patients with type 1 diabetes and 138 healthy individuals. The INS-VNTR allele classes were determined by Southern blotting (class I: 21-44RU; class III: 138-159RU). Subjects with class I alleles were further studied using PCR amplification to determine the exact length of the alleles (short 1S: 22-37RU; medium 1M: 38-41RU; large 1L: 42-43RU). Allelic and genotype frequencies were estimated by chi(2) tests for independence with 2 x 2 contingency tables and the relative risks (RR) were determined using GraphPad InStat software.

Results: We observed differential associations among the class I alleles when comparing patients with LADA (80.6%) and type 1 diabetes (81.3%) with the controls (70%; p < 0.005). This increase was largely due to the high frequency of the 1S/S genotype (63.6% LADA vs 37% controls, with a p-value of 0.0019 [p1]; 53.4% type 1 diabetes vs 37% controls, with a p-value of 0.0149 [p2]). Remarkably, all LADA patients genotyped as class I homozygous had the shorter (S) class I allele (100%). Differences in the overall 1S distribution were observed: in LADA the 94.4% of the alleles were equal to or smaller than 35RU, while in patients with type 1 diabetes it was 78.3% and in controls 74.1%. Moreover, the relative risks associated with the 1S/S genotype for patients with LADA showed a substantial increase with respect to those with type 1 diabetes (52%) when we compare them to the controls (1S/S LADA/control, 2.282 [RR1] vs type 1 diabetes/control, 1.497 [RR2]).

Conclusion: The presence of the 1S allele could be considered a risk factor in LADA patients, as previously reported for type 1 diabetes. The class I INS-VNTR allele in LADA increases genetic susceptibility to disease development.