Aim: To determine islet autoantibody profiles at time of seroconversion and track changes in antibody profiles on follow-up (1–2 years after seroconversion) with respect to age at first antibody appearance, genetic predisposition and rate of progression to type 1 diabetes (T1D).
Methods: Islet autoantibodies (IAA, GADA, IA-2A, IA-2βA, ZnT8A) were measured in follow-up serum samples of 142 confirmed autoantibody-positive children from the BABYDIAB study (median follow-up from birth 11.7 years). Antibody binding to epitope regions on GAD65 (NH2, MID, COOH) and IA-2 (JM, PTP), polymorphic variants of ZnT8 (R325, W325), and affinity of IAA, GADA and IA-2A was determined. Children were genotyped at HLA-DR/DQ genes.
Results: Sixty-two (44%) children developed autoantibodies before age 4 years (EARLY immunization), and 80 (56%) thereafter (LATE immunization) (median autoantibody-positive follow-up 7.3 and 5.4 years; P=0.1). At seroconversion, 27 (44%) of the EARLY children had already developed multiple autoantibodies (vs. 17 [21%] LATE children; P=0.006), IAA was the most frequent EARLY antibody (87% vs. 49% in LATE group; P<0.0001), and GADA the most frequent LATE antibody (59% vs. 48% in EARLY group; P=0.2). Within 2 years from seroconversion, another 14 EARLY children (total 66%) progressed to multiple autoantibodies (vs. another 5 LATE children [total 28%]; P<0.0001), and the prevalence of IA-2A and ZnT8A in the EARLY group increased to 48% and 55%, respectively (vs. 13% and 25% in LATE group; P<0.001 for both antibodies). EARLY children also developed higher titers of IAA (P=0.0002) and GADA (P=0.01) compared to LATE children. In both groups, progression to multiple autoantibodies was associated with antibodies of high affinity (≥109 L/mol) and presence of HLA DR4, DQ8, DR4-DQ8, or DR3/DR4-DQ8 (all P<0.0001 vs. single antibodies). In contrast, the majority of single autoantibodies was of low affinity (<109 L/mol) and/or transient, and had relatively low antibody titer. Furthermore, 27% of single GADA showed restricted binding against NH2-terminal GAD epitopes (P<0.0001), and 21 (34%) single antibody-positive children carried protective HLA alleles (vs. 3 [4%] multiple antibody-positive children; P<0.0001). T1D developed in 36 (58%) EARLY and 8 (10%) LATE children (P<0.0001). There was no difference in the progression rate from multiple autoantibodies to T1D between both groups.
Conclusions: Children who progress in islet autoimmunity and eventually develop T1D show similar antibody characteristics regardless of EARLY or LATE immunization. Likewise, non-progression is associated with characteristics that are similar in both groups. However, the frequency of predictive antibody characteristics differs between EARLY and LATE children, and corresponds to the frequency of predisposing or protective HLA genotypes.
