Yukako Ozawa

Radioimmunoassay Detects the Frequent Occurrence of Autoantibodies to the Mr 65,000 Isoform of Glutamic Acid Decarboxylase in Japanese Insulin-Dependent Diabetes

Glutamic acid decarboxylase antibodies (GAD65Ab) are common in new onset Caucasian insulin-dependent diabetic (IDDM) patients but it is unclear if this marker is also prevalent in patients of other ethnic backgrounds. We determined antibodies against human recombinant GAD in Japanese diabetic patients using a radioimmunoassay with competition between in vitro translated 35S-GAD65 and non-labelled recombinant human GAD65 (rhGAD65). GAD67 antibodies (GAD67Ab) were similarly analyzed but without antigen competition. In 73 Japanese diabetic patients, GAD65Ab were found in 11/16 (69%) of patients with short-duration (less than 5 yrs) IDDM, 6/23 (26%) with long-duration (5 or more yrs) IDDM and 10/20 (50%) with slowly progressive diabetes. High GAD65Ab levels were associated with concomitant autoimmune diseases (p = 0.021). GAD67Ab were found in 4/16 (25%) of patients with short-duration IDDM, 3/23 (13%) with long-duration IDDM and 2/20 (10%) with slowly progressive diabetes. In 14 non-insulin dependent diabetic (NIDDM) patients, GAD65Ab and GAD67Ab were not found (0/14) and 1/50 (2%) healthy controls were positive in either assay. Among the GAD67Ab-positive samples, 8/9 (88%) were also high level GAD65Ab positive, 7/9 (77%) were displaced by an excess of rhGAD65 and the antibody levels correlated (r2 = 0.573; p = 0.003). Our data are consistent with a strong association of GAD65Ab also in Japanese IDDM, and suggest that, when present, GAD67Ab are frequently directed to epitope(s) common to GAD65 and GAD67.

Lack of association between CTLA-4 gene polymorphism and IDDM in Japanese subjects

Susceptibility to insulin-dependent diabetes mellitus (IDDM) is determined by both environmental and genetic factors. The main gene associated with predisposition to IDDM is HLA. Recent studies have described linkage and association of IDDM to the CTLA-4 gene (IDDM12) in Caucasians. CTLA-4 is a candidate gene for T-cell-mediated autoimmune diseases because it is a negative regulator of T-cell proliferation. We investigated the distribution of a CTLA-4 gene polymorphism in 110 Japanese patients with IDDM and 200 control subjects. In 84 patients, we also investigated associations between this CTLA-4 gene polymorphism and GAD65 antibody positivity. An A/G transition at position 49 of exon 1 was analyzed by the polymerase chain reaction-restriction fragment length polymorphism method. GAD65 antibody was detected using a radioligand binding assay. There was no significant difference in the distribution of CTLA-4 alleles in patients and controls and no difference was observed in prevalence of CTLA-4 alleles when GAD65 antibody-positive and -negative individuals in the IDDM groups were compared. The present study did not support an association between the CTLA-4 gene and IDDM in the Japanese population.

Analysis of MHC Class II antigens in Japanese IDDM by a novel HLA-typing method, hybridization protection assay

We examined HLA Class II antigens in 116 Japanese IDDM patients [84 typical IDDM (T-IDD); 32 slowly progressive IDDM (S-IDD)] by the hybridization protection assay (HPA) which is a novel HLA typing method based on hybridization of acridinium-ester-labeled DNA probes to amplified DNA. We detected HLA-DRB1, -DQA1 and -DQB1 genes by this method which is capable of analyzing over 50 samples within 4 h with high sensitivity. Positive associations were found in DRB1*0405, DRB1*0802, DRB1*0901, DQA1*0301, DQB1*0303 and DQB1*0401, negative correlations in DRB1*0403, DR2, DR12, DRB1*0801 or 03, DQA1*0101 or 02, DQA1*0501, DQB1*0301 and DQB1*0602 alleles. The absence of aspartic acid (Asp) at position 57 of the DRB1 chain and the presence of arginine (Arg) at position 52 of the DQA1 chain correlated positively with both types of IDDM. There were no significant differences in HLA between T-IDD and S-IDD. These results suggest that the absence of Asp at position 57 of the DRB1 chain and the presence of Arg at position 52 of the DQA1 chain are significant Japanese IDDM patients and that DRB1*0802, in which the amino acid at position 57 is aspartic acid, may play a role in the pathogenesis of IDDM. Also, T-IDD and S-IDD have common bases in the HLA gene.

Autoantibody Against ICA512 Did Not Improve Test Sensitivity for Slowly Progressive IDDM in Adults

only 1) if symptoms suggestive of hypoglycemia were of sufficient severity to seek medical advice; 2) if whole venous blood glucose levels were <2.5 mmol/1 at the time of symptoms; 3) if symptoms were relieved by eating carbohydrate. Of the 162 nondiabetic twins, 5 sought medical advice with hypoglycemic symptoms relieved by carbohydrate; hypoglycemia was confirmed on extensive testing in only 3 of them. All 3 subjects were twins of IDDM patients, 2 of these 3 twins subsequently developed diabetes. However, in one (S.C.), hypoglycemia was probably due to self-injection of insulin since she was admitted unconscious to the hospital at age 15 years with a blood glucose of 0.9 mmol/1 and normal serum Cpeptide level (0.049 nmol/1), despite an extremely high serum insulin level (1,380 pmol/1) indicating an exogenous source of insulin. This twin already had impaired glucose tolerance, islet cell antibodies (ICA), and glutamic acid decarboxylase 65 (GAD65) antibodies but had no insulin autoantibodies; IDDM developed at age 19 years. Thus, spontaneous clinical hypoglycemia was detected in just 1 of the 55 twins who developed diabetes under observation. This twin (VS.), started having symptomatic hypoglycemia after large meals (blood glucose on three occasions of 2.0 mmol/1) at age 34 years. She developed diabetes 4 years later with GAD65 antibodies and decreased first-phase insulin response. Her diabetes is controlled by diet alone at present. The third twin (C.A.) with documented hypoglycemia (blood glucose on three occasions <2.5 mmol/1) remains nondiabetic without antibodies and is, we estimate, unlikely to develop diabetes. Symptoms of hypoglycemia before the onset of NIDDM have been reported to be common (5) but alone are not sufficient to establish the diagnosis. Furthermore, low blood glucose levels may occur without symptoms (6). By strict diagnostic criteria, spontaneous hypoglycemia occurred in only 2 of 162 twins (1.2%) and then only in twins of IDDM patients. Extrapolating from our observations in twins, we conclude that clinical hypoglycemia, whether fictitious or real, is not a feature of the prediabetic period.

Age of onset, not type of onset, affects the positivity and evanescence of IA-2 antibody

Autoantibody against IA-2 (IA-2A) was found to be discordant with autoantibody against glutamic acid decarboxylase (GADA) with respect to both positivity and titer in Japanese, the same as in Caucasians. In this study, 247 type 1 diabetic patients were tested in order to clarify how the type of onset, age of onset, and duration of diabetes affect the frequency and evanescence of IA-2A. Among the young onset patients, the frequency of IA-2A was higher (52.2%), but evanescent (54.5, 66.7 and 36.7% in the insulin therapy duration < or =1, 2-5 years, and > or =6 years groups, respectively), whereas among adult onset patients, the frequency was lower (19.3%) but persistent (19.6, 13.3 and 23.5%, respectively). In addition, in the follow-up study, two of three IA-2A-positive young onset patients converted to negative in only three years, while all five adult onset patients remained positive for over 5 years. Among the adult onset patients, IA-2A frequency was similar in the slowly progressive type and the abrupt onset type. In view of the above findings, IA-2A positivity and evanescence in type 1 diabetic patients appear to be affected by age of onset, not type of onset.

Immune Response to Heat-Shock Protein Correlates With Induction of Insulitis in I-Eα d Transgenic NOD Mice

To evaluate the correlation between heat-shock protein (HSP) and insulitis, we compared lymphocyte proliferative response to Mycobacterium leprae HSP65 of NOD mice with that of I-Eα d transgenic NOD (I-E+NOD) mice, which show no insulitis. We found that splenocytes from 15-week-old NOD mice showed a more marked proliferative response to HSP than did those from age-matched I-E+NOD mice (P < 0.05). We then transferred splenocytes from 12-week-old NOD mice into I-E+NOD mice to induce insulitis in the recipients and examined antibody levels against HSP. By 6 weeks posttransfer, insulitis was successfully transferred to four out of five recipients of NOD splenocytes and antibody levels against HSP were significantly higher in the NOD splenocyte-transferred group than in controls, which showed no insulitis (P < 0.01). These results suggest that immune response to HSP correlates with insulitis in NOD mice. Our results support the assertion that HSP is a useful antigen for investigating the etiology of IDDM.

HLA class II is associated with the frequency of glutamic acid decarboxylase M r 65 000 autoantibodies in Japanese patients with insulin-dependent diabetes mellitus

Autoantibodies to glutamic acid decarboxylase (GAD65Ab) are common in both caucasian and Japanese patients with insulin-dependent diabetes mellitus (type 1), while the type 1-associated HLA haplotypes differ. In the present study, we analyzed GAD65Ab in relation to HLA-DQ and-DR alleles in Japanese type 1 patients. GAD65Ab were found in 58% short-duration (less than 5 years) type 1,23% long-duration type 1,56% slowly progressive type 1,3% type 2 patients, and 1.7% healthy individuals. In 75 HLA-typed type 1 patients, the GAD65Ab frequency was higher in short-duration patients with DRB1*08 allele (100%,Pc<0.05). GAD65Ab frequencies in DQB1*0302, DQB1*0303, and DRB1*09-positive, long-duration type 1 patients were lower than those in short-duration type 1 patients (14%, 19%, and 20%,Pc<0.02 compared with short-duration type 1, 90%, 75%, and 71%, respectively), while the frequency varied less in DQB1*04 individuals (44% and 30% in short- and long-duration type 1 patients, respecitively). These findings were also observed among patients with DRB1*04, i.e., the haplotype DRB1*0405-DQB1*0401 showed less variation in frequency of GAD65Ab (44% and 35% in short- and long-duration type 1 patients, respectively), while DRB1*04xx-DQB1*0302 showed lower frequency in long-duration type 1 than short-duration (13% and 100%, respectively). Thus, HLA class II is associated with frequency GAD65Ab, and this association might be affected by disease duration in Japanese type 1 patients.

Absence of Antibodies to ICA512/IA-2in NIDDM Patients With the Mitochondrial DNA bp 3243 Mutation

Diabetes with the mitochondrial DNA bp 3243 mutation is a new subtype of diabetes characterized by maternal inheritance and hearing disturbance (1). Although phenotype of diabetes is usually noninsulin-dependent, Oka et al. (2) detected the bp 3243 mutation in three islet cell antibody (ICA) patients with slowly progressive IDDM; a finding that suggests the immune response to islet cells may be triggered by cell damage arising from mitochondrial dysfunction. Recently, the major part of antigens for ICAs has been revealed to be the protein tyrosine phosphatase-like IA-2 (ICA512/IA2) (3). We examined the antibodies to both ICA512/IA-2 and human GAD65 by sensitive radioligand-binding assays (4) in 14 diabetic patients with the bp 3243 mutation, including two patients with the syndrome of myoclonus, epilepsy, lactic acidosis, and stroke-like episode (MELAS). Although eight patients were taking insulin, fasting plasma C-peptide concentrations and urinary excretion of C-peptide revealed that these patients except one were noninsulin-dependent (Table 1). All 14 patients with the bp 3243 mutation were negative for both anti-ICA512/IA-2 and anti-GAD. This finding is consistent with previous reports in which neither of the ICAs nor the antibodies to bovine GAD were demonstrated in limited numbers of NIDDM patients (5,6) or in MELAS patients and their family members (7) with the bp 3243 mutation, but the finding contrasts with results of the recent report by Kobayashi et al. (8) in which 13 of 31 diabetic patients with the bp 3243 mutation were ICA detected by a sensitive indirect immunofluorescence method. The difference may have been an artifact of patient selection. ICAs might be detected in NIDDM patients with mitochondrial mutation who also have a background of IDDM. Kobayashi et al. speculated that those who have IDDM-related human leukocyte antigen (HLA) react to islet cell antigen and produce antibodies to it when mitochondrial dysfunction occurs because ICA patients tended to have IDDMrelated HLA. In this study, eight patients had IDDM-related HLA (Table 1), indicating that HLA alone is not sufficient for the development of antibodies to ICA512/IA-2 or GAD in mitochondrial diabetes. Another possibility is low specificity of antigens for ICAs assayed by indirect immunofluorescence. Although the principal islet antigens that react with ICAs are ICA512/IA-2 and GAD65, several other antigens have been proposed. In fact, among 11 patients with the bp 3243 mutation in this study in whom ICAs were sought, very low titers (10 Juvenile Diabetes Foundation QDF] units in both) were detected in two (9). It is conceivable that antibodies to antigens other than ICA512/IA-2 or GAD65 exist in the patients with bp 3243 mutation. In the patients observed by Kobayashi et al., antibodies to GAD were detected only in two patients. In typical IDDM patients with recent onset, prevalence of ICA and antibodies to GAD is usually similar (10). Furthermore, ICA disappears shortly after the complete destruction of islet cells, while antibodies to GAD remain positive after absolute insulin dependency. The fact that most of the patients with mitochondrial bp 3243 mutation included in this study and in previous studies do not have antibodies to GAD suggests immunological background is somewhat different between ICA mitochondrial diabetes and IDDM. In conclusion, the bp 3243 mutation in mitochondrial DNA is unlikely to be a trigger for production of antibodies to ICA512/LA-2 and GAD, and immune response to islet cell antigens in diabetes with mitochondrial mutation, if any, seems to be different from that in IDDM.