Determination of anti GAD65 autoantibodies with an ELISA before and after standardization with the new international reference serum

Presence of GAD65 autoantibodies in the serum of children with autism or ADHD

Antibodies against glutamic acid decarboxylase 65 (GAD65) have been detected in the serum of patients with several neurological disorders. The presence of antibodies against GAD65 has not yet been examined in the serum of patients with neurodevelopmental disorders such as autism or attention-deficit/hyperactivity disorder (ADHD). In this study, GAD65 antibodies and total IgG were assayed in the serum of normal subjects and patients diagnosed with autism or ADHD. GAD65 antibodies were detected in the serum of 15% of children with autism (N = 20), 27% of children with ADHD (N = 15) and of none of the controls (N = 14). The serum of 60% of autistic and 53% of ADHD patients reacted with Purkinje neurons in mouse cerebellum. Serum from 20% of ADHD patients reacted also with the cells in the molecular and granule cell layers and cells in the vicinity of the Purkinje neurons. No association was found between the titer of GAD65 antibodies and total IgG levels, and presence of seizures or mental retardation. None of the ADHD patients were diagnosed with mental retardation. Serum anti-GAD65 antibodies may be a common marker of subgroups of patients with autism and ADHD. Reactions of serum antibodies with the cells in the cerebellum in these patients suggest direct effects on brain function. The subgroup of children with autism and ADHD that tests positive for GAD65 antibodies needs further characterization in a larger study.

Development of 2 alternative enzyme-linked immunosorbent assays for routine screening of glutamic acid decarboxylase autoantibodies

BACKGROUND

Antibodies to GAD65 (GADA) are considered highly predictive humoral markers of the type 1 diabetes mellitus and also of the insulin requirement in adult-onset patients presumptively classified as type 2 diabetics or LADA.

METHODS

We present 2 methods for GADA assessment. The first one (fluid phase, ELISA Protocol A) is carried out in a 2-step procedure in which serum GADA are first allowed to react with a fixed dose of GAD65-biotin in solution and the residual free antigen is later assayed by a conventional ELISA. In the second test (solid phase, ELISA Protocol B) GADA are measured in an ELISA that depends on the ability of divalent autoantibodies to form a bridge between immobilized TrxGAD65 and liquid-phase biotinylated TrxGAD65.

RESULTS

All normal control samples scored negative in both variants of ELISA and RBA, hence specificity was 100% for all methods; the relative sensitivity of ELISA Protocol A respect of the RBA was 94% and that of ELISA Protocol B was 76%.

CONCLUSIONS

Although ELISA Protocol A exhibited a better performance in terms of relative sensitivity than ELISA Protocol B, the simplicity of execution and the intended use of the assay must also be taken in consideration for the final choice.

Autoimmune thyroiditis in non-obese subjects with initial diagnosis of Type 2 diabetes mellitus

Autoimmune thyroiditis is often associated with Type 1 diabetes mellitus (T1DM). In non-obese adult-onset diabetes diagnosed initially as Type 2 diabetes mellitus (T2DM), there is a proportion of cases with so far undiagnosed T1DM. The objective of this study was to estimate the frequency of autoimmune thyroiditis (AT) among non-obese (BMI <30.0 kg/m2) patients with T2DM and to compare the frequency of AT in subgroups of patients according to the presence of glutamic acid decarboxylase antibodies (GADA), insulin requirement, and post-breakfast C-peptide levels. The study included 118 adult patients (55 men and 63 women) with the initial diagnosis of T2DM and age at the onset of diabetes > 35 yr. Median of age was 66 yr (range 39-82), and median duration of diabetes was 9 (range 1-27) yr. AT was diagnosed using thyroid peroxidase antibodies, TG-antibodies, US and TSH levels. Nineteen per cent of the subjects were found to have AT, and the frequency of AT did not significantly differ between the groups of GADA+ and GADA- subjects. There was no difference in the frequency of AT between the group treated with hypoglycemic agents and/or diet and the group requiring insulin. The frequency of AT was higher in the group with post-breakfast C-peptide levels < or = 0.8 nmol/l compared to the group with post-breakfast C-peptide levels > 0.8 nmol/l (37% vs 16%), however the group with post-breakfast C-peptide levels < or = 0.8 nmol/l had longer duration of diabetes.

Islet autoantibodies and insulin dependent diabetes mellitus in cystic fibrosis

Cystic fibrosis-related diabetes mellitus (CF-DM) is thought to be secondary to beta-cell destruction by fibrous tissue replacing the exocrine pancreas. The aim of this study was to investigate the hypothesis that other factors may also be responsible. Glutamic acid decarboxylase (GAD) and islet cell (IA-2) antibodies were measured by quantitative ELISA in a group of patients with CF (n=30) in comparison to a group of newly diagnosed DM type 1 (IDDM) patients (n=30) and normal subjects (n=30). GAD antibodies were positive (>32 ng/ml) in 50% of the CF, 93% of the IDDM and 0% of the control group. IA-2 antibodies were detected (>0.9 U/ml) in 40% of the CF, 93% of the IDDM and 0% of the control group. Among the fifteen CF patients with positive GAD and IA-2 antibodies, four already had IDDM and another five abnormally low (<45 mU/l) first phase insulin response (FPIR) indicating a prediabetic state. We conclude that factors other than mechanical may be involved in the development of CFDM. The presence of autoantibodies predicting IDDM supports the hypothesis that CF-DM may have a multifactorial pathogenesis.

Highly-sensitive and specific enzyme-linked immunosorbent assays for GAD65 autoantibodies using a thioredoxin-GAD65 fusion antigen

Autoantibodies against glutamic acid decarboxylase (GAD65) are present in the sera of most patients with recently diagnosed insulin-dependent diabetes mellitus (IDDM). These antibodies appear years before the clinical symptoms, and they are considered to be early markers of the disease. To detect GAD65 autoantibodies (GADA), we developed new enzyme-linked immunosorbent assays (ELISA) with a fusion protein thioredoxin-GAD65 (Trx-GAD65) produced in E. coli as the antigen. These assays were compared with the reference radiobinding assay (RBA). Since most GADA are directed against native epitopes, and adsorption of GAD65 to plastic may cause disruption of its native conformation, the new assays rely on the following immobilization procedures: (a) capture ELISA (c-ELISA) with Trx-GAD65 (protocol A) or biotin-Trx-GAD (protocol B) indirectly immobilized by a non-adsorptive process; (b) ELISA with antigen-antibody preincubation in solution (p-ELISA) in which GADA were reacted first with Trx-GAD65 (protocol C) or biotin-Trx-GAD (protocol D) and the free antigen was determined by conventional ELISA. The results obtained with 42 newly diagnosed IDDM patients and 30 normal individuals were as follows: RBA had 79% sensitivity (percentage of IDDM patients detected) and 97% specificity (100% minus the percentage of false positives). c-ELISA showed low sensitivity (36 and 50%, respectively for protocols A and B), and high specificity (100 and 97%, respectively). p-ELISA were highly-sensitive (74 and 79%, respectively) and specific (97 and 93% for protocols C and D, respectively). Thus, protocols C and D had a performance similar to the reference method. The results reported here provide the basis for simple, highly-sensitive, specific, and widely-applicable tests for GADA that eliminate many of the drawbacks of the radioactive methods.