The primary aim of this thesis was to test whether gestational exposure to environmental factors may induce islet autoimmunity. The second aim was to determine to what extent the exposure to viruses or other environmental factors is a risk factor for type 1 diabetes in the offspring. As children with type 1 diabetes have a higher risk of developing celiac disease (CD), an additional aim was to determine whether markers of possible infections during early pregnancy was associated with development of tissue transglutaminase (tTG) autoantibodies or CD in the offspring. These aims have been summarized in four independent studies.
Study population and Methods
The Diabetes Prediction in Skåne (DiPiS) is a population based study where blood samples were obtained at the time of delivery between September 2000 and August 2004 from all mothers in this region (Skåne). The blood samples were analyzed for HLA-DQ alleles and islet autoantibodies.
The prospective cohort study Celiac Disease Prediction in Skåne (CiPiS) is part of the DiPiS study, which aims to determine the etiology indicators of celiac disease (CD) in newborn children. Children with HLA-risk alleles associated with CD were screened for tissue transglutaminase autoantibodies (tTGA). CD was established by intestinal biopsy in children with confirmed tTGA.
Early pregnancy serum samples (gestational week 10-16) were collected from the Southern Sweden Microbiological biobank (SSM-Biobank). This biobank contains more than 120,000 stored plasma samples that have been obtained between 1986 and 2009 from all pregnant mothers at their first visit to the Maternity Care Center.
By combining the SSM-Biobank with DiPiS and CiPiS, a total of 24,000 mothers were identified. All mothers included in this thesis (n=1,748) were analyzed for HLA genotype, islet autoantibodies (GADA, IA-2A and IAA in both early pregnancy samples and at delivery), nine different cytokines, one chemokine (early pregnancy samples), IgM class enterovirus (EV) antibodies and EV-RNA (early pregnancy samples).
Study I: The objective of our first study was to determine seroconversion to islet autoantibodies in non-diabetic mothers during pregnancy. This was achieved by analyzing end point titers of GADA, IA-2A and IAA in both early pregnancy samples and samples at delivery. Mother’s positive for GADA (92%), IA-2A (84%) or IAA (65%) at delivery had increased titers already in early pregnancy. Titers declined for GADA (p<0.0001), IA-2A (p<0.0001) and IAA (p<0.0001). Seroconversion during pregnancy was observed for GADA in 10 (8%), IA-2A in 3 (16%) and IAA in 37 (35%) mothers.
Study II: In this study, we collected early pregnancy serum samples from mothers who later gave birth to children who developed high titers of tTGA or confirmed CD. We then measured an array of Th1/Th2 cytokines (nine cytokines and one chemokine) with the aim of investigating whether CD is triggered already in utero as denoted by quantitative changes in the mother’s cytokine profile. We observed that levels of seven out of ten cytokines were significantly increased in mothers who gave birth to children with CD when compared to controls.
Study III: As studies have shown that unbalanced gestational cytokine profiles have been associated with maternal autoimmune disease, preeclampsia, and recurrent spontaneous abortions we analyzed in the present study cytokines in serum samples collected during early pregnancy from mothers who gave birth to children developing multiple, persistent islet autoantibodies, type 1 diabetes, or both, before seven years of age. We found that IFN-γ (p=0.02) and IL-1β (p=0.04) were elevated in the index mothers. All cytokines except IL-4 were highly correlated (p<0.0001).
Study IV: Gestational EV infections have been associated with risk HLA-DR as well as with type 1 diabetes in the child. We analyzed enterovirus RNA (EV-RNA) and IgM (EV-IgM) in relation to type 1 diabetes HLA-DQ risk genotypes and to islet autoantibodies in non-diabetic mothers studied both in early pregnancy and at delivery. EV-IgM, but not EV-RNA was detected during early pregnancy in 12% (44/365) islet autoantibody positive mothers compared to 11% (156/1457) of the controls (p=n.s.). In early pregnancy, mothers with HLA-DQ 2/2 or 2/X genotypes showed, in adjusted logistic regression, an increased risk for islet autoantibodies (OR 1.85, 95% CI 1.34-2.54; p=0.001). EV-IgM was not associated with HLA-DQ in early pregnancy. However, after adjusting for parity, maternal age, year of birth and season of early pregnancy, early pregnancy EV-IgM combined with DQ2/2 or 2/X increased the risk for the mother to be positive for islet autoantibodies at delivery (OR 3.10, 95% CI 1.35-7.15; p=0.008).
From these studies we conclude that pregnant non-diabetic mothers with islet autoantibodies at delivery had significantly higher titers of autoantibodies during early pregnancy than at delivery. As the statistical power in the seroconverting mothers was insufficient, further studies are needed to determine if the risk for type 1 diabetes in the offspring differs between mothers who already had increased titers of islet autoantibodies during early pregnancy or acquired them during pregnancy. Moreover, we conclude that a shift in the Th1/Th2 mediated cytokine pattern during early pregnancy, possibly caused by viral infection may be associated with CD in the offspring. Furthermore, the results from study III suggest that increased levels of IFN-γ and possibly IL-2 during early pregnancy was associated with an increased risk for multiple, persistent islet autoantibodies, type 1 diabetes, or both, before seven years of age in the offspring. Our final study concludes that EV-IgM in early pregnancy increased the risk for islet autoantibodies at delivery in non-diabetic mothers who carry HLA-DQ 2/2 or 2/X type 1 diabetes risk genotypes.