Kimiyoshi Tsuji

HLA class II molecules and autoimmune hepatitis susceptibility in Japanese patients

To investigate the association between autoimmune hepatitis and HLA alleles in Japanese patients, serological typing and class II genotyping were performed using the polymerase chain reaction-restriction fragment length polymorphisms (PCR-RFLP) method. Serological typing showed that HLA-B54, -DR4, -DR53, and -DQ4 were significantly more frequent in patients with autoimmune hepatitis than in controls. HLA-DR4 was most frequently associated with autoimmune hepatitis (88.7%). In PCR-RFLP typing, the frequency of DRB1*0405 was significantly higher in autoimmune hepatitis than in controls. However, there was no significant difference in the frequency of Dw between the patients and the controls who were DR4-positive. The significant increase observed in DQA1*0301 and DQB1*0401 was explained by a linkage disequilibrium with DR4. Six DR4-negative patients had DR2, but there was no significant difference in the frequency of the DR2-associated Dw-alleles compared with the DR2-positive controls. No DPB1 allele was significantly associated with autoimmune hepatitis. These findings suggest that the basic amino acid at position 13, which is present only on the DR2 and DR4 B1 molecules (Arg on DR2 and His on DR4), contributes to the susceptibility to autoimmune hepatitis among the Japanese.

Time Course of Islet Cell Antibodies and β-Cell Function in Non-Insulin-Dependent Stage of Type I Diabetes

The time course of islet cell antibodies (ICA) and serum C-peptides responses (CPRs) to oral glucose tolerance tests (OGTTs) were studied prospectively up to 60 (mean 35) mo in 32 ICA-positive subjects [28 with non-insulin-dependent diabetes (NIDDM) and 4 subjects with impaired glucose tolerance (IGT); mean age 45 yr], 96 matched subjects [56 with NIDDM, 8 with IGT, and 32 normal first-degree relatives of patients with insulin-dependent diabetes (IDDM); mean age 45 yr] who were negative for ICA at the beginning of the study. In addition, the effects of human leukocyte antigens(HLA) on the time course of ICA and (β-cell function were evaluated. In 10 subjects (8 with NIDDM and 2 with IGT) who were ICA positive, ICA became undetectable, even by sensitive ICA assay, 15 ± 2 mo (mean ± SE) after initiation of this study. In these subjects, integrated serum CPR values (σCPR) and 2-h blood glucose values in response to OGTTs improved significantly (P < .05-.01). In contrast, the remaining 22 subjects who were ICA positive were persistently positive for ICA. CPR and blood glucose responses deteriorated progressively in these 22 subjects, and 7 subjects in this group progressed to the insulin-dependent state. Serum CPR and blood glucose responses to OGTTs showed no remarkable changes in 64 patients (56 with NIDDM and 8 with IGT) and 32 normal first-degree relatives of patients with IDDM who remained negative for ICA throughout the study. The frequencies of HLABW54 and HLA-DR4 in 10 subjects who became ICA negative during the follow-up period were lower than those in 22 subjects with ICA throughout the study (uncorrected P < .02). The frequencies of these two antigens were higher in the 22 subjects with persistently positive ICA than in normal controls (uncorrected P < .02), whereas these differences were not observed in 10 subjects who became ICA negative during the study. The reversed β-cell function after negative conversion of ICA observed in our study yields a new insight into the natural history of type I diabetes, especially in late-onset cases. It suggests that HLA-related genetic predisposition is a prerequisite to the slowly progressive β-cell destruction through pancreatic autoimmunity.

Association of HLA-A24 With Complete β-Cell Destruction in IDDM

A sensitive C-peptide immunoreactivity radioimmunoassay demonstrated the presence of subtle, but definite residual β-cell function in patients with IDDM of long duration. Although HLA antigens are known to influence susceptibility to IDDM, their contribution to the extent of pancreatic β-cell destruction has not yet been examined extensively. We studied the relationship between residual β-cell function and HLA class I and class II antigens in 111 unrelated Japanese IDDM patients. Using the sensitive C-peptide immunoreactivity radioimmunoassay, the presence or absence of residual β-cell function was evaluated by the C-peptide immunoreactivity response to a 100-g oral glucose load. DNA typing for HLA-DQA1 and HLA-DQB1 antigens was performed in addition to serological typing of HLA-A, HLA-B, HLA-C, and HLA-DR antigens. A C-peptide immunoreactivity response > 0.033 nM was regarded as an indication of the presence of residual β-cell function, not the assay error. Surprisingly, 35 of 37 (94.6%) patients without residual β-cell function had HLA-A24, whereas only 39 of 74 (52.7%) patients with residual β-cell function had this antigen (corrected P = 9.795 × 10(–6). Any other HLA antigens, including the DR and DQ loci, showed no difference in the frequency with regard to residual β-cell function. The duration of diabetes was similar between the groups with and without residual β-cell function. The duration of diabetes was similar between the groups with and without residual β-cell function. Even when the patients were stratified according to the duration of diabetes, HLA-A24 was more common in those with early complete loss of β-cell function (duration of diabetes <1 yr) (P = 0.035) and was less common in those with residual β-cell function despite a long duration of diabetes (>10 yr) (P = 0.001). The correlation between HLA-A24 positivity and complete β-cell loss also was confirmed in younger-onset (<30 yr old) and elder-onset (≥30 yr old) groups. The C-peptide immunoreactivity response in patients with HLA-A24 (0.09 ± 0.02 nM, mean ± SE, n = 74) was significantly lower than that in patients without HLA-A24 (0.19 ± 0.03 nM, n = 37, P < 5.0 × 10−5). Further typing of HLA-A24 by one-dimensional isoelectric focusing gel electrophoresis revealed that the isoelectric point of HLA-A24 was identical in charge in 17 of 18 patients and 7 normal control subjects (isoeletric point 6.32, HLA-A24.1). We conclude that a specific HLA class I antigen, HLA-A24, promotes pancreatic p-cell destruction in IDDM patients with other disease-susceptible HLA antigens.

Southern hybridization analysis of DNA polymorphism in the HLA-D region

Restriction fragment-length polymorphisms (RFLP) were systematically analyzed by Southern hybridization with restriction endonuclease-digested genomic DNA from 28 HLA-homozygous B cell lines with Dw1-Dw19 specificity using the DR beta and DQ beta chain cDNAs as probes. These probes detected polymorphic fragments unique to each HLA-DR specificity. Furthermore, the DQ beta chain probes permitted us to distinguish between different Dw specificities with an identical DR type much more efficiently than with the DR beta chain probe. Distribution analysis of restriction fragments hybridizing to DR beta in relation to the DR and DQ specificities showed several sets of them forming ten clusters, some of which correlate with DRw53, DQw1, and DR alleles. This DNA typing technique allows the direct definition of HLA types at the gene level and provides a powerful tool for isolating genes controlling HLA-associated diseases.

A simple and rapid method for HLA-DP genotyping by digestion of PCR-amplified DNA with allele-specific restriction endonucleases

We previously reported a simple and rapid method for HLA-DQA genotyping by digestion of polymerase chain reaction-amplified DQA genes with allele-specific restriction endonucleases. Here we report the application of this method to DP genotyping. The second exon of the HLA-DPB genes was selectively amplified from genomic DNAs of 72 HLA-D homozygous B-cell lines by the polymerase chain reaction method. Amplified DNAs were digested with ApaI, SacI, BstUI, FokI, and RsaI, which can recognize allelic sequence variations in the polymorphic segments of the DPB second exon and then subjected to electrophoresis in polyacrylamide gels. Sixteen different polymorphic patterns of the restriction fragments were found, and twelve were identical to patterns predicted from the known DNA sequences correlating with each HLA-DPw specificity defined by cellular typing. The other four patterns were distinct from those of the known DPw specificities, suggesting the presence of novel DP alleles. This polymerase chain reaction-restriction fragment length polymorphism method provides a simple and rapid technique for accurate definition of HLA-DP types at the nucleotide level, replacing the technically demanding method of primed lymphocyte typing.

Microsatellite polymorphism between the tumor necrosis factor and HLA-B genes in Behçet's disease

Behçets disease is associated with the HLA-B51 antigen. However, it has not yet been clarified if the HLA-B51 gene itself is the susceptibility gene related to this disease or if it is some other non-HLA gene in linkage disequilibrium with HLA-B51. Therefore, we screened one of the HSP70 genes, HUM70t (HSP70-Hom), around the class III region and the microsatellite sequence located between the HLA-B and TNF genes for genetic polymorphism in BD. A comparison between patients with BD and healthy controls revealed no significant difference in the frequency of the HUM70t polymorphism. In the microsatellite sequence, Tau-a, in the region between the HLA-B and TNF genes, the frequency of 14 repetitions of GT was increased significantly and that of 11 repetitions was decreased significantly in the patient group. Further, the allelic distributions of the B51 antigen-associated microsatellite polymorphism differed significantly between patients and healthy controls, and in the B51 antigen-negative subjects, analysis of the microsatellite polymorphism also revealed a significant difference in the haplotype frequency between the patient and control groups. These results suggest that the HLA-B51 gene may not be the primary locus responsible for BD, and implicate some other gene(s) located between the TNF and HLA-B genes.

Analysis of genes within the HLA region affecting susceptibility to ulcerative colitis

To clarify the molecular relationship between HLA loci and ulcerative colitis (UC) in Japanese patients, we performed HLA-DP genotyping by the PCR-RFLP method and studied tumor necrosis factor beta-chain genetic polymorphism by Southern hybridization, in addition to conventional serologic typing. Significant increase was observed in Bw52, DPw9 (DPB1*0901), and DR2 (DRB1*1502) in Japanese patients with UC. Linkage analysis indicated that A24-Bw52-DR2-DPw9 alleles constitute a common haplotype in Japanese UC patients. Among the patients not carrying Bw52, B13 was significantly increased and B44 was relatively increased. These Bw52, B13, and B44 alleles share the unique amino acids, serine and aspartic acid at positions 67 and 77, respectively. These positions are in the second hypervariable region of the alpha 1-domain of the HLA-B13, B44, Bw52, and B49 antigens (B49 is quite rare in the Japanese population). The inflammatory region in UC patients was found to vary depending on their HLA-B alleles. These results suggest that the HLA-B locus itself plays an important role in the susceptibility to Japanese UC.

Gm ALLOTYPES IN MYASTHENIA GRAVIS

Gm typing and acetylcholine receptor antibody assay were performed on serum samples from 74 patients with myasthenia gravis (31 male, 43 female) and from 236 unrelated normal blood-donors. The haplotype Gm1,2,21 was significantly more common in patients with myasthenia gravis (relative risk = 3.24), especially those with thymoma (relative risk = 6.99). The frequency of haplotype Gm1,2,21 was further increased in patients with severe generalised myasthenia gravis (relative risk = 10.52). The frequency of Gm1,2,21 was also increased in patients with high acetylcholine receptor antibody titres (greater than 5 pmol/ml). The results indicate the presence of a pathogenic gene close to the IgG heavy-chain gene complex in the 6th chromosome in the patients with myasthenia gravis, especially those with thymoma.

The Prevalence of Islet Cell Antibodies in Japanese Insulin-dependent and Non-insulin-dependent Diabetic Patients Studied by Indirect Immunofluorescence and by a New Method

Islet cell antibodies (ICA) were measured in Japanese patients with insulin-dependent diabetes mellitus (IDDM) and non-insulin-dependent diabetes mellitus (NIDDM) by a standard, indirect immunofluorescence method (IF method) and by a newly established, threelayer immunofluorescence method applying a biotin-avidin system (BAS method). In addition, the relationship between ICA and HLA was studied in IDDM patients. ICA titers detected by the BAS method correlated well with those determined by the standard IF method (rs = 0.987, P < 0.01). The BAS method had about an eightfold higher sensitivity for ICA than the IF method. The overall prevalence of ICA detected by the BAS method (ICA-BAS) versus that by the IF method (ICAIF) was 41% (82/198) versus 28% (56/198) in IDDM patients and 3% (19/593) versus 2% (14/593) in patients with NIDDM. In IDDM patients, ICA-BAS was all positive <1 mo after the onset of diabetes, while the prevalence of ICA-IF was 83% (20/24) during the same period. The prevalence of ICA-IF decreased rapidly with the duration of disease, reaching a value of 6% (3/55) in the patients with a disease duration of 10 yr or more. The incidence of ICA-BAS also decreased with the duration of disease, although to a lesser degree than ICA-IF. No association was found between HLA types and persistence of ICA-BAS or -IF. These results suggest that pancreatic autoimmune processes occur in almost all Japanese IDDM patients. Although IDDM is less common in Japan than among Caucasians, the prevalence of ICA seems to be the same. The higher sensitivity of the BAS method may be of significant diagnostic value, especially in patients with a long duration of disease.

IgG Heavy-Chain (GM) Allotypes and Immune Response to Insulin in Insulin-Requiring Diabetes Mellitus

IN guinea pigs1 , 2 and mice, 3 4 5 the ability to develop humoral and cell-mediated immune responses to heterologous insulins is in part controlled by immune-response genes linked to the major his...