Analysis of Viral Infections' Exposition Preceding the Type 1 Diabetes (T1D) Diagnostic in Children of the Isis-Diab Cohort. Search for Explanations of of the Disease's Early Onset

Overview

The purpose of this study is to investigate viral factors determining the early onset of T1D. Thanks to the quantification of viral exposures of T1D patients before the disease onset with questionnaires and environmental databases analyses, and through whole genome association studies of these patients, investigators could attempt to identify gene-virus interactions determining the age of T1D onset.

The "hygiene hypothesis", which has been proposed to explain the observed increase of the incidence of T1D, relies on experimental evidence acquired in mouse models. However, epidemiological data are still lacking to validate this hypothesis in man. This is critical, because -in opposition with the hygiene hypothesis- there are many reasons to believe that, at contrary, certain virus can trigger the disease. Genetic predisposition to a severe infection form (particularly primary infection) was demonstrated for several infectious diseases (Casanova JL, Science 317:617-619, 2007; Casanova JL, EMBO J 26:915-922, 2007). It usually corresponds to deficiencies in genes involved in the host's immune response, the transmission of which being Mendelian. Genetic factors affect the ability of enteroviruses and other viruses to damage beta cells and to induce diabetes. Recently, Nejentsev et al have demonstrated a link between enteroviruses and diabetes genes: they have indeed identified 4 rare IFIH1 polymorphisms that reduce the T1D risk. However, this gene encodes an enzyme recognizing the DNA enterovirus, causing immunity activation; mutations inhibit gene activation. Except one study on HIV-1, there is to our knowledge no genome wide association studies (GWAS) in humans on the role of host's genetic polymorphisms in the risk of infection, clinical expression, duration of viral shedding, or response to therapy or to anti-viral vaccines. We relied on our cohort of T1D children (Isis-Diab) to investigate the possible relation between viral exposures, genetic polymorphisms, and subsequent T1D. The search for viral factors responsible for the increased T1D prevalence in youth children is difficult to implement. The absence or scarcity of infections is difficult to assess robustly at the individual level. The analysis of digestive, ENT or blood samples in the search for viruses themselves can only be done at T1D diagnosis and is therefore unlikely to be positive several years after the causal infection. It is not possible to reconstitute retrospectively viral events, which an individual has been exposed between birth and date of diabetes diagnosis. That is why our project proposes to use a proxy of viral infections crossed by a child, quantifying viral exposures to which he was submitted before the T1D diagnosis. We focused on early childhood's viral infections that may interfere with early forms of T1D. We combine 2 data sources: * The geolocation of the child's address will locate places where he lives. Spatio-temporal data from Sentinel Network, collected since 1984, will provide access on the following viral exposures: seasonal influenza, viral diarrhea (mostly enteroviral), mumps, measles, chickenpox. If the defect of our approach is that it does not see real viral infections experienced by children, but only the level of exposure to which they were exposed according to their address, the advantage is an objective spatio-temporal description of virus epidemics around children, making it more or less likely infection of these. Only France has such Sentinel data. * Vaccination (including MMR) and information on infectious past of children will be collected from data recorded in his book health. We will focus on the mother's pregnancy and the child between birth and 2 years of age. Environmental data from these 2 sources will be crossed with already available genetic data from GWAS to identify gene-virus associations potentially determining age of T1D onset. This "high dimensionality" analysis will be addressed with "machine learning" programs. If avoidable risks are identified, it would be possible to think to design clinical trials for prevention of the identified forms of T1D.