Diagnosis of monogenic diabetes: 10-Year experience in a large multi-ethnic diabetes center.

AIMS/INTRODUCTION: Monogenic diabetes accounts for approximately 1-2% of all diabetes, and is difficult to distinguish from type 1 and type 2 diabetes. Molecular diagnosis is important, as the molecular subtype directs appropriate treatment. Patients are selected for testing according to clinical criteria, but up to 80% of monogenic diabetes in the UK has not been correctly diagnosed. We investigated outcomes of genetic testing in our center to compare methods of selecting patients, and consider avenues to increase diagnostic efficiency. MATERIALS AND METHODS: We reviewed 36 probands tested for monogenic diabetes in the last 10 years in a large adult diabetes outpatient clinic, serving an ethnically diverse urban population. We compared published clinical criteria and an online maturity onset diabetes of the young calculator applied to these 36 patients, and presented the predictions together with the molecular results. RESULTS: The overall mutation detection rate was 42%, reflecting the strict clinical selection process applied before genetic testing. Both methods had high sensitivity for identifying patients with mutations: 88 and 89% for the clinical criteria and online calculator, respectively. Cascade testing in a total of 16 relatives led to diagnosis of a further 13 cases. CONCLUSIONS: Existing patient selection criteria were effective in identifying patients with monogenic forms of diabetes, but the number of patients missed using these strict criteria is unknown. Because of the potential savings resulting from correct molecular diagnosis, it is possible that testing a larger pool of patients using less stringent selection criteria would be cost-effective. Further evidence is required to inform this assessment.

Resequencing the susceptibility gene, ITGAM, identifies two functionally deleterious rare variants in systemic lupus erythematosus cases.

INTRODUCTION: The majority of the genetic variance of systemic lupus erythematosus (SLE) remains unexplained by the common disease-common variant hypothesis. Rare variants, which are not detectable by genome-wide association studies because of their low frequencies, are predicted to explain part of this "missing heritability." However, recent studies identifying rare variants within known disease-susceptibility loci have failed to show genetic associations because of their extremely low frequencies, leading to the questioning of the contribution of rare variants to disease susceptibility. A common (minor allele frequency = 17.4% in cases) nonsynonymous coding variant rs1143679 (R77H) in ITGAM (CD11b), which forms half of the heterodimeric integrin receptor, complement receptor 3 (CR3), is robustly associated with SLE and has been shown to impair CR3-mediated phagocytosis. METHODS: We resequenced ITGAM in 73 SLE cases and identified two previously unidentified, case-specific nonsynonymous variants, F941V and G1145S. Both variants were genotyped in 2,107 and 949 additional SLE cases, respectively, to estimate their frequencies in a disease population. An in vitro model was used to assess the impact of F941V and G1145S, together with two nonsynonymous ITGAM polymorphisms, A858V (rs1143683) and M441T (rs11861251), on CR3-mediated phagocytosis. A paired two-tailed t test was used to compare the phagocytic capabilities of each variant with that of wild-type CR3. RESULTS: Both rare variants, F941V and G1145S, significantly impair CR3-mediated phagocytosis in an in vitro model (61% reduction, P = 0.006; 26% reduction, P = 0.0232). However, neither of the common variants, M441T and A858V, had an effect on phagocytosis. Neither rare variant was observed again in the genotyping of additional SLE cases, suggesting that their frequencies are extremely low. CONCLUSIONS: Our results add further evidence to the functional importance of ITGAM in SLE pathogenesis through impaired phagocytosis. Additionally, this study provides a new example of the identification of rare variants in common-allele-associated loci, which, because of their extremely low frequencies, are not statistically associated. However, the demonstration of their functional effects adds support to their contribution to disease risk, and questions the current notion of dismissing the contribution of very rare variants on purely statistical analyses.