Immune responses to glutamic acid decarboxylase and insulin in patients with gestational diabetes

Abstract

Pregnancy is a natural state of immunoprotection and tolerance. We studied subjects with gestational diabetes (GDM) to evaluate the influence of pregnancy on the humoral immune response to the autoantigen GAD and to injected insulin. Antibodies against glutamic acid decarboxylase (GADA) subclasses and epitope reactivity were determined in 34 GADA-positive pregnant patients with GDM, in 20 GADA-positive relatives of people with TID and in 25 GADA-positive patients with newly diagnosed TID. Partum levels of insulin antibodies (IA), IgG1- and IgG4-IA were measured in 131 women with GDM treated with human insulin from the time of diabetes diagnosis (including 22 with GADA) and were compared to 19 patients with TID after 3 months of insulin treatment. GADA titre and subclasses were similar among all groups. GADA in GDM patients bound fewer epitopes than GADA in relatives of patients with TID (all epitopes being present in 23%versus 65%, P < 0.01). In particular, antibodies to the minor GADA epitopes GAD6596-249, GAD651-100 and GAD67 were less frequent in patients with GDM compared to relatives (P < 0.01). Antibodies to insulin (IA) were found in 17% of patients with GDM. They were more frequent in GDM patients with GADA compared to GADA-negative patients (41%versus 12%, P < 0.005). IgG1 was the dominant insulin antibody subclass response in both patients with GDM and TID but levels of IgG1-IA and IgG4-IA were significantly lower in patients with GDM compared to patients with TID (P < 0.004). Antibody responses in women with gestational diabetes appear to be dampened and restricted, but without change in subclass usage.

Fig. 1

Levels of GAD antibody subclasses IgG1 and IgG4 (s.d.s.) in GADA-positive patients with GDM (black circles), TID (black triangles) and first-degree relatives of patients wit TID (FDR) (open circles). The through line represents the upper limit of normal values, small lines the median values. Levels of GADA subclasses were expressed as standard devation scores (s.d.s.).

Fig. 2

Prevalence (upper panel) and levels (lower panel) of antibodies to epitopes of the COOH terminal region GAD65_436−585, the middle epitopes GAD65_235−442 (mid a) and GAD65₉₆₋₂₄₉ (mid b) to the NH₂-terminal region GAD65₁₋₁₀₀, and to GAD67 in GADA-positive patients with GDM (black bars and circles) and GADA-positive relatives of patients with TID (open bars and circles). P_corr levels < 0·01 are marked with an asterisk (*). Small lines represent the median values.

Fig. 3

Levels of insulin antibodies (units) at delivery in 25 GADA-positive patients with GDM (black circles) and 106 GADA-negative patients with GDM (open circles) 3·1 and 2·9 months after commencement of insulin treatment (insulin exposure). The through line represents the upper limit of normal values of insulin antibodies, small lines the median values.

Fig. 4

Levels of insulin antibody subclasses IgG1 and IgG4 (mU/ml) in patients with GDM and GADA (black circles), GDM without GADA (open circles) and in patients with TID (open triangles) 3 months after commencement of insulin treatment. Small lines represent median values.