Reduced IL-4 associated antibody responses to vaccine in early pre-diabetes.

AIMS/HYPOTHESIS: The aim of this study was to determine whether beta-cell autoimmunity is associated with immune response bias to exogenous antigens. METHODS: IgG subclass responses against tetanus toxoid and rubella were measured after vaccination in children with (n=36) and without (n=73) islet autoantibodies participating in the BABYDIAB prospective study of offspring of parents with Type I (insulin-dependent) diabetes mellitus. All children had been vaccinated against tetanus toxoid antigen before 6 months of age and at 18 months of age, and against live attenuated rubella virus at 18 months of age and again before 5 years of age. Tetanus toxoid specific IgG subclasses and cytokine responses were compared in a second cohort of subjects. RESULTS: Responses to tetanus toxoid in islet-autoantibody-negative children were characterized by early IgG1 antibodies at 9 months of age followed by the appearance of IgG4 and lesser IgG2 antibodies at 2 years of age. Children who had developed islet autoimmunity before one year of age (n=15) did not have the shift to IgG4 and IgG2 anti-TT after booster vaccination (p<0.01), and had undetectable or IgG1 restricted responses. This defect was independent of HLA class II genotype, was restricted to children who had islet autoimmunity before 1 year of age, and was most evident in children who already had multiple islet autoantibodies by 9 months of age. IgG4 and IgG2 anti-TT correlated with IL-4 (p<0.005), but not IFNgamma responses. Antibody responses to the IFNgamma-inducing rubella vaccination were strongly IgG1 dominated and no differences were observed between islet autoantibody positive and negative children. CONCLUSIONS/INTERPRETATION: These data are consistent with a reduced capacity to make IL-4 promoted antibody responses to exogenous antigen in early pre-diabetes.

Prevalence, characteristics and diabetes risk associated with transient maternally acquired islet antibodies and persistent islet antibodies in offspring of parents with type 1 diabetes.

Accurate assessment of type 1 diabetes risk in young children requires discrimination between antibodies that are produced by the child and antibodies acquired through the placenta from an islet antibody-positive mother. We studied 682 offspring from mothers with type 1 diabetes and 329 offspring from fathers with type 1 diabetes and nondiabetic mothers for insulin (auto)antibodies, glutamic acid decarboxylase antibodies, and tyrosine phosphatase IA-2 antibodies before age 1 yr and again at age 2 yr to ascertain transience or persistence. Antibodies were detected at age 9 months in 5 (1.5%) offspring from fathers with type 1 diabetes; all were insulin (auto)antibodies only, all persisted and developed multiple antibodies, and 1 developed type 1 diabetes. In contrast, 31 (4.5%) offspring from mothers with type 1 diabetes had antibodies at 9 months; 12 (1.8%) persisted at age 2 yr, and 19 (2.8%) did not persist, suggestive of transient residual maternal antibodies. Multiple antibodies at 9 months were usually persistent (3 of 4 offspring), as were single insulin (auto)antibodies in offspring from mothers with type 1 diabetes (8 of 13 offspring), whereas persistent glutamic acid decarboxylase antibodies (1 of 12) and tyrosine phosphatase IA-2 antibodies (0 of 2) were rare. Offspring with persistent antibodies at age 9 months had a high type 1 diabetes risk (100% by age 5 yr for those with multiple antibodies and 27% for single antibodies at 9 months), whereas offspring with transient antibodies had 0% type 1 diabetes risk (P < 0.01). Transience was associated with very high antibody levels at birth. For insulin (auto)antibodies, the measurement of subclass was also informative. Residual maternal antibody was indicated by similar insulin (auto)antibodies subclasses at 9 months and at birth, whereas different subclasses were indicative of nonmaternal antibody. Moreover, the presence of IgG1-insulin (auto)antibodies was associated with antibody persistence and type 1 diabetes risk. These strategies are helpful in discriminating high and low risk antibodies before age 1 yr and should be important for prognosis and reducing unnecessary parent anxiety.

Immunoglobulin G insulin autoantibodies in BABYDIAB offspring appear postnatally: sensitive early detection using a protein A/G-based radiobinding assay.

Insulin autoantibodies (IAA) are early sensitive markers of pre-diabetes in the young. The aim of this study was to assess whether, using IgG-specific measurement with a protein A/G assay, IAA are already present at birth, and whether this assay is suitable for early autoantibody screening. Cord blood and follow-up samples from offspring of parents with type 1 diabetes included in the BABYDIAB study were analyzed. Although insulin antibodies in cord blood from children of mothers with type 1 diabetes were readily detected and correlated well with levels in the maternal circulation, no insulin binding was detected in 247 cord blood samples from children of father probands. IgG IAA were detected at 2 yr in all 21 children who had multiple islet autoantibodies or who later developed type 1 diabetes, but were confirmed in only 6 of 58 with IAA by the conventional IAA assay in the absence of other islet autoantibodies. False positive IAAs in the conventional assay were often attributable to hemolysis. Hemolysis did not affect protein A/G IAA measurement, and results in whole capillary blood samples were comparable to those in corresponding serum samples (r2 = 0.99). These data show that IgG IAA appear early and after birth, and that the protein A/G IAA assay is sufficiently sensitive for early screening. The specificity of this assay requires further evaluation.

Early development and spreading of autoantibodies to epitopes of IA-2 and their association with progression to type 1 diabetes.

Autoimmunity precedes clinical type 1 diabetes, and indicators of maturing autoimmune responses may be useful markers for disease prediction. To study this, epitope maturation of autoantibodies to the related protein tyrosine phosphatase (PTP)-like autoantigens IA-2 and IA-2beta was examined in sequential samples from birth in a cohort of 21 offspring developing multiple islet autoantibodies and a similar cohort of 48 relatives of patients with type 1 diabetes recruited at an older age. Initial reactivity in offspring was heterogeneous against the IA-2 juxtamembrane region (10/21) and PTP domains (13/21), and both specificity and extent of initial IA-2/IA-2beta autoantibodies were associated with HLA class II genotype. Intra-IA-2 epitope spreading and/or intermolecular spreading to IA-2beta epitopes were observed in seven offspring. In contrast, in older relatives, IA-2/IA-2beta Ab reactivity was stable and spreading rare. Development of diabetes in children was associated with the presence of Abs to the IA-2 juxtamembrane region (risk by age 5 yr, 52% vs 0% in those with PTP domain Abs only; p < 0.02), and 5 of 26 relatives who developed diabetes had IA-2 Abs only against the juxtamembrane region. The findings show that autoantibody reactivity to IA-2/IA-2beta is dynamic in the young, show that the juxtamembrane region of IA-2 is an early IA-2 autoantibody target, and suggest that these Abs are a risk factor for development of type 1 diabetes in infancy.

Comparison of a novel micro-assay for insulin autoantibodies with the conventional radiobinding assay.

Measurement of insulin autoantibodies (IAA) with a novel micro radiobinding assay which requires only 20 microl of serum was compared with that in a conventional radiobinding assay which uses 600 microl of serum. IAA were measured with both assays in samples from 94 new onset insulin-dependent diabetes mellitus (IDDM) patients, 97 control subjects. and 48 first degree relatives of IDDM patients selected for having IAA in the conventional radiobinding assay. Overall, 227 (95%) of 239 samples tested were concordant. and IAA levels correlated well (r2 = 0.7) between the two assays. Discordant results were obtained in 7 new onset patients, 4 control subjects, and 1 first degree relative, and these had low IAA levels in the respective assays. Sensitivity and specificity in the new onset IDDM patients and control subjects were 69% and 98% for the micro radiobinding assay and 72% and 98% for the conventional radiobinding assay. The use of the micro radiobinding assay should greatly facilitate islet related antibody screening, particularly in children.

The T-cell receptor beta chain CDR3 region of BV8S1/BJ1S5 transcripts in type 1 diabetes.

We recently described the T-cell receptor (TCR) beta chain CDR3 motif S-SDRLG-NQPQH (BV8S1-BJ1S5) in an islet-specific T-cell clone (K2.12) from a type 1 diabetic patient (AS). A similar motif (RLGNQ) was also reported in a T-cell clone of non-obese diabetic (NOD) mice by others. In order to determine the frequency of our motif in selected and unselected T-cell populations, we cloned and sequenced the CDR3 region of BV8S1-BJ1S5 transcripts. These transcripts were derived from unstimulated peripheral blood T lymphocytes from two type 1 diabetic patients (AS and FS) and their non-diabetic sibling (WS), as well as from an islet-specific T-cell line of one of the patients. In addition, we compared the structure and composition of the CDR3 region in BV8S1-BJ1S5 transcripts from peripheral blood T cells between the patients and their non-diabetic sibling (>50 sequences each). We found that 30% of the islet-specific T-cell line cDNA clones expressed the entire sequence-motif, whereas it was absent in the clones of unstimulated peripheral blood T cells from both patients and their non-diabetic sibling. The average length of the CDR3 region was shorter in the patients (mean AS 9.9, FS 9.9, versus WS 10.7, p = 0.0037) and the number of inserted nucleotides in N nucleotide addition at the DJ-junction lower (mean AS 3.5, FS 3. 2, versus WS 5.2, P = <10(-4)) as compared with their non-diabetic sibling. Moreover, the pattern of amino acid usage in the CDR3 region was dissimilar at positions 5 and 6, where polar amino acids predominated in both diabetic siblings. In contrast, basic amino acids are preferentially used at position 5 in the clones of the non-diabetic sibling. These data provide information on the general structure of the TCR(BV8S1-BJ1S5) CDR3 region in type 1 diabetes and may indicate differences in the amino and nucleic acid composition of the TCR beta chain CDR3 region between two type 1 diabetic patients and their non-diabetic sibling.