Maternal immunity to insulin does not affect diabetes risk in progeny of non obese diabetic mice

Abstract

It has been suggested that maternal environment, in particular maternal autoantibodies, modify the risk of developing autoimmune diabetes in offspring. The aim of this study was to determine whether modification of maternal environment and maternal diabetes risk through immunization affects autoimmune diabetes risk in the progeny. The risk of developing insulin antibodies and of developing diabetes was determined in 113 female progeny of non obese diabetic (NOD) dams that were immunized with insulin, control antigen or vehicle before or during pregnancy. Although NOD dams immunized with insulin were rendered diabetes resistant (40% diabetes by age 36 weeks versus 100% in control dams), diabetes development in their female offspring (72%, 26/36) was similar to that in female offspring of dams immunized with glucagon (82%, 22/27) or vehicle (76%, 19/25). Furthermore, no significant differences in diabetes development or insulin autoantibody titres were observed between female progeny of insulin autoantibody positive NOD dams (82% diabetes by age 36 weeks, 18/22), insulin autoantibody negative NOD dams (75%, 41/55), and NOD dams that had antibodies against exogneous insulin (71%, 22/31). The findings suggest that modification of the maternal risk for autoimmune diabetes via antigen-specific immunization is not transferred to progeny and that fetal exposure to insulin autoantibodies does not increase the risk for diabetes development.

Introduction

The transfer of humoral factors from the mother to offspring during gestation and/or the first weeks of life represents an important mechanism of immunological protection in offspring until their own immune system has matured [1]. Protection is conferred through the transfer of antibodies which neutralize infections and which become vaccine-like to promote neonatal immunity against potentially lethal pathogens [1]. Female hormones are known to enhance antibody responses suggesting that humoral factors other than antibodies can be transferred and modify protection [2,3].

While maternal transfer is considered to be important in the context of immune protection, other immunity such as antibodies to nonpathogens and immune mediators such as cytokines and chemokines are also transferred. In specific circumstances such as autoimmunity and in the presence of a genetic predisposition to disease the transmission of such humoral factors may alter the natural course of disease development in the offspring. In rodent autoimmune diabetes, maternal transmission of antibody has been suggested to be protective when against pathogenic autoreactive T cells [4] and a risk factor when against autoantigen [5]. In particular, maternal autoantibodies against islet antigens including insulin were suggested to contribute to the development of diabetes in the non obese diabetic (NOD) mouse. The concept that risk for autoimmune diabetes may be modified by changing the composition of maternal humoral factors is tantalizing with respect to disease prevention. In order to examine this possibility, we used an experimental model and studied diabetes development in offspring of NOD dams that had been rendered diabetes resistant through immunization with insulin or left diabetes prone through immunization with control antigen or vehicle. The findings suggest that maternal protection, even when administered during pregnancy is not transferred to NOD female offspring, and that diabetes development in NOD female mice is not related to the presence or absence of maternal insulin antibodies or autoantibodies.

Materials and Methods

Offspring of immunized and control female NOD mice

NOD mice were originally obtained from Taconics (Germantown, NY, USA) and the colony established in animal facilities at the Diabetes Research Institute Munich. The incidence of diabetes within unmanipulated female NOD mice at the time of the study was 87% by age 36 weeks. For all animal experiments, the principles of animal care (NIH publication no. 85–23, revised 1985) and the national laws on Protection of Animals were followed. The study was approved by the ethics committee of Bavaria (license no. 209·1/211-2531-51/02).

For the present study, 46 female NOD mice before the age of 4 weeks were litter matched and randomized into 7 treatment groups (group A–G, see Fig. 1). Mice in groups A and D were immunized with human insulin (Aventis, Frankfurt, Germany) in incomplete Freund's adjuvant (IFA) (Sigma-Aldrich Corp., St. Louis, MO, USA), mice in groups B and E with human glucagon (Sigma) in IFA, and mice in groups C and F with saline in IFA. Mice in groups A, B and C were immunized before pregnancy at 4, 5, 6, and 7 weeks of age; and mice in groups D, E, and F were immunized during pregnancy at the 1st, 2nd, and 3rd week of gestation. Mice in group G served as untreated controls. Venous blood samples from orbital vein were obtained from all dams at age 4 weeks (before immunization and pregnancy) and at 11 weeks (shortly before delivery) for insulin antibody determination. Twenty-eight of the 46 randomized dams became pregnant and delivered 235 mice of which 113 were female and 122 male (Fig. 1).

Fig. 1.

Schematic representation of study design.

All 113 female newborn mice were followed to diabetes development or the age of 36 weeks. Urine glucose levels were measured 3 times weekly by using urine glucose sticks (Glucetur Test, Boehringer Mannheim, Mannheim, Germany) beginning at 10 weeks of age. Mice were considered to be diabetic after two consecutive urine glucose values > 5·5 mmol/l and blood glucose levels above 13·9 mmol/l (Glucometer Elite, Bayer Diagnostics GmbH, Munich, Germany). Venous blood was obtained at age 4, 10, and 36 weeks for the measurement of insulin antibodies.

Insulin antibody measurement

Antibodies to insulin were measured by a radiobinding assay as described previously [6]. The upper limit of normal was determined from the 99th centile values obtained in sera from BALB/c and C57/B6 female mice and corresponded to 4 local units. The assay is represented as laboratory B in the animal models of diabetes workshop [7]. All measurements were performed on coded samples that were operator blinded.

Statistical analysis

Life table analysis was used to analyse diabetes development in NOD mice. The Mann–Whitney U-test was used to compare insulin antibody levels in NOD mice progeny. All P-values are two-tailed. All statistical analyses were performed using the Statistical Package for Social Science (SPSS, Chicago, IL, USA).

Results

Relationship of birth islet antibody status and subsequent development of insulin autoantibodies and diabetes in NOD mice progeny

IA or IAA of IgG isotype were detected before pregnancy and/or just before delivery in 12 of 28 NOD dams. These included 8 of 10 dams that were immunized with insulin plus IFA, and 4 of 18 of the dams that were not immunized with insulin. Of the 113 female progeny, 31 were from IA positive dams that had been immunized with insulin, 22 were from IAA positive dams that had not received exogenous insulin, and 55 were from IAA negative dams. 28 of 31 (90%) female progeny of IA positive dams were IA/IAA positive at age 4 weeks indicative of residual maternally acquired IA (Fig. 2a). Fourteen (45%) were positive at age 10 weeks. IAA prevalence and titre at age 10 weeks did not differ between progeny of IA positive mothers (45%), IAA positive mothers (46%) and IAA negative mothers (51%), suggesting that insulin antibody status at birth did not affect the subsequent development of endogenous IAA.

Fig. 2.

Insulin antibody titres (a) and diabetes development (b) in female progeny of NOD dams. Progeny are classified as insulin antibody positive at birth if they were from insulin antibody positive dams that were immunized with insulin, insulin autoantibody positive at birth if they were from insulin antibody positive dams that were not immunized with insulin, and insulin autoantibody negative at birth if they were from insulin autoantibody negative dams. Insulin antibody titres in the progeny are shown at age 4 weeks and 10 weeks (a). In b, diabetes development in progeny that are insulin autoantibody positive at birth (n = 22) is represented by the thick solid line, insulin antibody positive at birth (n = 31) is represented by the broken line, and insulin autoantibody negative at birth (n = 55) is represented by the thin solid line.

Diabetes development was slightly, but not significantly faster in progeny from IAA positive dams as compared to progeny of IAA negative dams (P = 0·14, Fig. 2b). Diabetes development in progeny of IA positive dams did not differ from that in progeny of IAA negative dams (P = 0·91).

Relationship of maternal immunization and subsequent development of diabetes in NOD mice progeny

Diabetes development in NOD dams immunized with insulin before or during pregnancy was significantly less than in nonimmunized mice (40%versus 100% by age 36 weeks; P = 0·01). Diabetes development in female progeny from insulin immunized dams was similar to that in the progeny of control dams and of NOD dams immunized with glucagon plus IFA, or vehicle (Fig. 3). No differences were observed between groups immunized before pregnancy compared to groups immunized during pregnancy.

Fig. 3.

Diabetes development in NOD dams (a) and female progeny of NOD dams (b, c) according to immunization protocol in dams. For dams, those immunized with insulin plus IFA are represented by the thick solid line and nonimmunized dams are represented by the broken line. For the progeny, diabetes development is shown separately for progeny of dams (b) immunized before pregnancy and (c) during pregnancy. Progeny that are from dams that were immunized with insulin plus IFA (n = 36) is represented by the thick solid line, from dams immunized with glucagon plus IFA (n = 27) is represented by the thick broken line, from dams immunized with saline plus IFA (n = 25) is represented by the thin solid line. The thin broken line in both b and c represents diabetes development in the progeny from unimmunized dams (n = 25).

Discussion

Antibodies bind antigen 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 [8–12]. As a result, antibodies may play a role in the pathogenesis of autoimmune disease [13]. Although autoantibodies found in autoimmune diabetes are considered to be purely markers of disease and unnecessary for diabetes development [14], recent studies 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 [13,15–18]. This hypothesis was recently reinforced by the finding that the removal of immunoglobulin during gestation markedly reduced the incidence of diabetes [5]. It remains to be determined whether these observations were due specifically to the presence or absence of autoantibodies to islet antigens both in the NOD mouse and in man.

We examined autoantibody and diabetes development with respect to birth antibody status in NOD mice. Antibodies to insulin were examined since these are the only validated humoral marker in this model [7], and they are of IgG isotype and therefore could cross the placenta [6]. The presence of autoantibodies to insulin in NOD dams did not increase the likelihood of progeny becoming IAA positive at age 10 weeks, and diabetes development was not significantly faster in progeny of IAA positive dams as compared to the progeny of IAA negative dams. Diabetes incidence in the progeny of IAA negative dams was almost 80% by age 36 weeks indicating that the presence or absence of IAA during fetal life was unlikely to be responsible for the markedly reduced diabetes incidence previously observed after removal of maternal immunoglobulin [5]. We also examined progeny of NOD dams that were immunized with insulin and had high titres of IgG-IA. Diabetes and IAA development in these progeny were very similar to those in progeny of insulin antibody negative dams indicating that fetal exposure to high titre antibodies against exogenous insulin does not influence sensitization to autoantigen. Interestingly, although immunization with insulin prevented diabetes in the dams, diabetes development in their progeny was similar to that in the progeny of control dams, and dams immunized with control antigen or vehicle, indicating that protective factors either do not cross the placenta or do not influence diabetes risk in progeny, even if immunization is performed during fetal life.

Altogether our findings in NOD mice indicate that fetal exposure to insulin antibodies does not affect diabetes development and therefore suggest that previous observations of reduced diabetes development in progeny of IgG deficient dams is likely to be due to the removal of as yet unidentified autoantibodies or IgG per se.