Transmission of Maternal Islet Antibodies and Risk of Autoimmune Diabetes in Offspring of Mothers With Type 1 Diabetes

BABYDIAB prospectively follows offspring of mothers and/or fathers with type 1 diabetes from birth. Cord blood is obtained at birth, and venous blood samples are obtained at age 9 months, and at ages 2, 5, 8, and 11 years (7). Recruitment began in July 1989 and ended in November 2000, and follow-up has continued thereafter. A total of 1,610 children of a parent with type 1 diabetes were recruited at birth and participated in the 9-month follow-up. These included 1,030 children of mothers with type 1 diabetes. Cord blood for determination of all three antibodies (insulin antibody, GADA, and IA-2A) at birth was available in 720 of 1,030 offspring of mothers with type 1 diabetes and in 285 of 580 offspring of fathers with type 1 diabetes and nondiabetic mothers. The cord blood volume was insufficient for islet antibody determination in the remaining 290 offspring of mothers with type 1 diabetes and 295 offspring of fathers with type 1 diabetes. After birth, insulin autoantibodies (IAAs), GADAs, and IA-2As were measured in samples from all scheduled follow-up visits. The median follow-up time from birth to last sample was 6.5 years (range 0.75–12.5 years), for a total of 9,773 person-years. HLA DR and DQ genotypes were determined in 602 of the 720 offspring with birth islet autoantibody measurements. Type 1 diabetes diagnosis was defined according to World Health Organization (WHO) criteria. All parents gave written informed consent to participate in BABYDIAB. The study was approved by the ethical committee of Bavaria, Germany (Bayerische Landesärztekammer no. 95357).

Insulin antibody(A), GADA, and IA-2A were determined by radiobinding assays as previously described (7,8). The 99th percentile of control children were used to define the upper limits of normal and corresponded to 8.5 local units/ml or 25 WHO units/ml for GADA, 2.5 local units/ml or 4 WHO units/ml for IA-2A, and 1.5 local units/ml for IAA. The assays had sensitivities and specificities of 80 and 94% (GADA), 58 and 100% (IA-2A), and 30 and 98% (IAA), respectively, in the First Diabetes Autoantibodies Standardization Program Proficiency Workshop (9). All measurements were performed on coded samples that were operator blinded.

HLA-DRB1, HLA-DQA1, and HLA-DQB1 alleles were typed using PCR-amplified DNA and nonradioactive sequence-specific oligonucleotide probes as previously described (10).

Time-to-event methods (life table analysis and Cox proportional hazards model) were used to analyze islet autoantibody and diabetes development in the offspring. Positive autoantibody outcome in children was defined as islet autoantibodies to at least two autoantigens (insulin, GAD, or IA-2) in at least two consecutive samples (persistent multiple islet autoantibody positive). Islet autoantibodies against a single autoantigen have not been demonstrated to be associated with diabetes development, and children with single islet autoantibodies were therefore considered negative. Children with autoantibodies in one sample and no subsequent sample for confirmation were also considered negative. The age of onset of islet autoantibody positivity was defined as the age at the first positive sample. Analyses considered censoring in losses to follow-up and in autoantibody-negative children at their age during the follow-up visit of their last antibody-negative sample. Offspring of mothers with type 1 diabetes were categorized according to their antibody status at birth as insulin antibody positive and negative, GADA positive and negative, IA-2A positive and negative, and autoantibody (GADA or IA-2A) positive or negative. Hazards ratios (HRs) were calculated using Cox’s proportional hazards model and were adjusted for the duration of type 1 diabetes in the mother (<10 years or >10 years), low birth weight (<2,700 g), and short gestation age (<36 weeks). All P values were two-tailed. All statistical analyses were performed using SPSS (Chicago, IL).

At birth, insulin antibodies were detected in 620 of 720 (86%) offspring of type 1 diabetic mothers, GADAs were in 401 of 720 (55.7%) offspring, and IA-2As in 267 of 720 (37%) offspring (Table 1). GADAs or IA-2As were detected at birth in 478 (66%) offspring. Of the 720 offspring, 45 (6.3%) had no islet antibodies in their cord blood, 54 (7.5%) had GADAs (44%) and/or IA-2As (25%) without insulin antibodies, 196 (27.2%) had insulin antibodies only, and 424 (58.9%) had GADAs and/or IA-2As together with insulin antibodies. One of the 285 offspring of fathers with type 1 diabetes had IAAs at birth, none had IA-2As, and none had GADAs.

Thirty-one offspring of type 1 diabetic mothers developed multiple islet autoantibodies during early childhood, and 16 of these developed diabetes. Autoantibody status at birth significantly affected diabetes risk in offspring of mothers with type 1 diabetes. Offspring who were islet autoantibody (GADA or IA-2A) positive at birth had a significantly lower risk for the subsequent development of multiple islet autoantibodies (1.3% by age 5 years) and diabetes (1.1% by age 8 years) than offspring who were islet autoantibody negative at birth (5.3%, P = 0.008 for multiple islet autoantibodies; 3.0%, P = 0.04 for diabetes) (Table 1) (Fig. 1). Risks were also reduced in the offspring with GADAs or IA-2As at birth after adjustment for maternal diabetes duration, gestational age, and birth weight (adjusted HR for developing multiple islet autoantibodies 0.25, P = 0.007; adjusted HR for developing diabetes 0.25, P = 0.02). The effect of birth autoantibody status on diabetes risk was most striking in children who did not have the high-risk HLA DRB1*03/DRB1*04-DQB1*0302 genotype (0.5 vs. 3.5% multiple islet autoantibodies by age 5 years, P = 0.007; 0.3 vs. 2.4% diabetes by age 8 years, P = 0.05) and was negligible in HLA DRB1*03/DRB1*04-DQB1*0302-positive children (17.8 vs. 26.3% multiple islet autoantibodies, P = 0.8; 18.4 vs. 15.6% diabetes, P = 0.7) (Fig. 2). No significant difference in the risk for developing multiple islet autoantibodies or diabetes was found between offspring with or without cord blood insulin antibodies (Table 1). The risk of developing multiple islet autoantibodies was 4 ± 0.9% (mean ± SE) in the 284 offspring of fathers with type 1 diabetes who were islet autoantibody negative at birth.

Antibodies bind antigens and can act as cell-surface receptors on B-cells or monocytes and dendritic cells to facilitate or enhance presentation of antigen to T-cells (11–14). As a result, antibodies may play a role in the pathogenesis of autoimmune disease (15). In autoimmune diabetes, however, it has generally been considered that autoantibodies are purely markers of disease because blocking B-cell function in experimental models does not affect the development of disease (16), and removal of immunoglobulin from type 1 diabetic patients has only a minimal effect on disease severity (17,18). While these observations show that autoantibodies alone are insufficient for diabetes development, more recent studies (15,19–21) have demonstrated that B-cell-deficient NOD mice have little insulitis and have a markedly reduced diabetes incidence, indicating that B-cells and perhaps antibodies play a role in the initiation of diabetes. This hypothesis was recently reinforced by the finding (5) that the removal of immunoglobulin during gestation markedly reduced the incidence of diabetes. It remains to be determined whether these observations were due specifically to the presence or absence of autoantibodies to islet antigens in both the NOD mouse and humans.

In humans, the BABYDIAB study provided an opportunity to examine whether fetal and neonatal exposure to autoantibodies and to endogenous autoantigen (GAD or IA-2) or to antibodies against exogenously administered autoantigen (insulin) modified diabetes risk. The BABYDIAB study is the largest study of offspring of parents with type 1 diabetes and includes >1,000 offspring of mothers with type 1 diabetes. We and others have previously shown that islet antibodies in cord blood of offspring from type 1 diabetic mothers correlate highly with the level of antibodies found in maternal blood at delivery, suggesting the transmission of these antibodies through the placenta (1,2) and that these maternally transmitted antibodies usually become undetectable within the first year of life (3,4). Here we determined the prevalence of islet antibodies at birth and followed offspring for the development of persistent diabetes-associated multiple islet autoantibodies and for the development of diabetes. Antibodies to exogenously administered insulin were detected at birth in 86% of offspring of type 1 diabetic mothers, and 66% of offspring had autoantibodies to GAD and/or IA-2 at birth. We found that the presence or absence of insulin antibodies did not affect the risk of developing diabetes or diabetes-associated autoantibodies. Remarkably, however, offspring with autoantibodies at birth had a significantly lower diabetes risk than offspring who were autoantibody negative at birth. The decreased risk did not appear secondary to potential confounders, such as maternal diabetes duration, birth weight, and gestational age.

Assuming that antibodies detected at birth are representative of antibody titers during pregnancy, the findings in the BABYDIAB offspring suggest that fetal exposure to GAD and/or IA-2 autoantibodies may protect against subsequent diabetes. Consistent with this observation is the overall decreased diabetes risk in offspring of type 1 diabetic mothers compared with that of offspring of type 1 diabetic fathers and nondiabetic mothers (22,23) and the previous report (24) of reduced development of islet autoimmunity in offspring of mothers with type 1 diabetes compared with children of fathers with type 1 diabetes. Although many factors are likely to contribute to this difference, it is possible that the fetal autoantibody exposure that occurs in the majority of offspring of type 1 diabetic mothers plays a role in reducing their diabetes risk. Such protection could be due to a more efficient elimination of autoreactive T-cell clones by antibody-mediated presentation of autoantigen during fetal and neonatal life or to a state of immune ignorance against autoantigen by antibody-mediated blocking or masking the presentation of relevant autoantigen peptides (25). Interestingly, the protection conferred by fetal exposure to autoantibodies was most striking in children who did not have the typical high diabetes risk HLA DRB1*03/DRB1*04-DQB1*0302 genotype, suggesting that the mechanism is unlikely to require HLA restriction of antigen presentation.

Overall, our findings in humans do not support the hypothesis that fetal exposure to islet autoantibodies increases diabetes risk and, if these observations are confirmed, suggest that fetal autoantibody exposure may protect from future endogenous islet autoimmunity and diabetes.

This study was supported by grants from the Juvenile Diabetes Research Foundation (JDRF nos. 1-2000-619 and 1-2003-646) and the Charlotte-Fiévet-Stiftung. This forms part of the dissertation of K.K. at the Technical University of Munich.

The authors thank Annette Knopff, Karolina von Dalwigk, Ulrike Mollenhauer, Heike Naserke, Michael Hummel, Markus Walter, and Doris Huber for their expert help in data collection and antibody measurements. The authors also thank all of the obstetric departments, pediatricians, and family doctors in Germany who participated in the BABYDIAB study.