To relate genetic variation in Graves' disease (GD) susceptibility to polymorphism at MHC loci, clinical and family studies were undertaken in eastern Hungary. Among 1980 relatives of 534 index patients, 2.9% of siblings, 2.7% of offspring, and 3.0% of parents had GD. HLA haplotype combinations in affected sibling pairs were determined in the present data and combined with data in the literature (12 sibling pairs from Farid 1981, 12 from Chan et al. 1980, and 15 from Sasazuki et al. 1983); 43, 23, and 1 affected sibling pairs shared, respectively, 2, 1, and 0 HLA haplotypes. This distribution is inconsistent with simple dominant inheritance, but is consistent with simple recessive inheritance of HLA-related susceptibility over a range of gene frequencies (0.2-0.4). A frequency of 0.3 gives the best fit and is consistent with penetrance of 7.1% for the recessive susceptibility genotype; the data, however, can accommodate penetrance values up to 16%. The distribution of HLA haplotypes in 33 families related disease susceptibility more strongly to DR than to other loci. The distribution of HLA-B8 genotypes in 256 patients was in close agreement with Hardy-Weinberg equilibrium proportions, also favoring recessive inheritance of MHC-related susceptibility. The probability that an individual will be affected with GD can be predicted, based on sex, HLA genotype, and family history. For example, 14.9% of DR3-positive women with an affected first degree relative are likely to be affected. These predictions can be tested as family data accumulate.

The authors have studied in detail human leukocyte antigen (HLA) association in 87 Hungarian patients with thyroid epithelial carcinoma. The authors also examined in a small group of patients, five parameters of cell-mediated immunity and related them to HLA as well as to lymphocytic infiltration of the tumor/normal tissue interface. HLA-DR1 was significantly associated with thyroid carcinoma; the strongest association was in patients with follicular histologic features and DR1 homozygotes were not at greater risk for thyroid cancer. The HLA-DR3 was nonsignificantly increased in patients with papillary or mixed histologic features. The HLA-DR1, 3 heterozygotes were highly associated with follicular carcinoma, carried no risk for papillary carcinoma, and an intermediate risk for tumors with mixed histologic features. Because of the small proportion of DR1, 3 heterozygotes in the follicular and mixed histologic group, its predictive value at the population level was low. Better predictive potential was shown for the phenotype DR1 and/or DR3. Neither metastatic disease nor age at diagnosis (less than 45 years) could be related to HLA phenotypes. Patients in all histologic variants showed some measure of cell-mediated immunity compared to controls. Patients with papillary carcinoma showed an overall better response than those with tumors with follicular or mixed histology. The HLA-DR could not be related to cell-mediated immune response. Patients with papillary carcinoma with a good cell-mediated immune response occurred with much lower infiltration of the tumor boundary with lymphocyte whereas the follicular carcinoma less cell-mediated immunity was associated with dense lymphocytic infiltration, suggesting the biological relevance of lymphocytic infiltration may be different for the two histologic variants.

We investigated the distribution of HLA and immunoglobulin G heavy chain markers (Gm) in 117 patients with Graves' disease, 62 with ophthalmopathy and 55 without. With Graves' disease per se, there is a closer association with HLA-DR3 than with B8. The opposite was true for Graves' patients with ophthalmopathy (odds ratio for ophthalmopathy associated with B8 was 12.4 and with DR3 was 7.7, both with P less than 0.0005). HLA-DR7 interacts with B8 in modifying the risk for eye disease; using the phenotype B8- DR7- as reference, the odds ratios were 16.7 for B8+ DR7+, 8.7 for B8+ DR7- and 0.26 for B8- DR7+. Thus, DR7 enhanced the risk for ophthalmopathy in the presence of B8+ but had a protective influence in its absence. Although Gm showed no association with eye disease, it modified the risk for ophthalmopathy associated with HLA-B8; the odds ratios were 20.9 for B8+ Gmfb homozygozity (fb+), 15.3 for B8+ fb- and 1.7 for B8- fb+ (B8- fb- = 1.00). We conclude that the genetic factors contributing to Graves' ophthalmopathy are different from those related to liability for Graves' hyperthyroidism.

Fifty-two patients with thyroid epithelial cell cancer were studied for evidence of association with human leukocyte antigens (HLA). Twenty-eight patients (53.8%) and 19.4% of 160 controls were HLA-DRI-positive, conferring a relative risk of 4.85 (x2 = 21.3, P < 0.0001). HLA-DRl was increased in all histologic types of thyroid cancer. Interestingly 10 of 12 patients with metastatic disease were DR1-positive compared to 18 of 41 patients without metastases (relative risk = 6.1, xz = 4.7, P < 0.05). This study suggests that major histocompatibility complex-linked gene@) determine susceptibility to and the biologic behavior of thyroid cancer. Cancer 58:52-54, 1986. ECENT EVIDENCE suggests that the course of epithelial

BACKGROUND. Using serologic human lymphocyte antigen (HLA) typing, the authors previously described a strong association between well differentiated thyroid carcinoma and HLA D-related 1 (HLA-DR1) in a population of unselected patients from Eastern Hungary. METHODS. In the current study, the authors used polymerase chain reaction-single strand conformational polymorphism to determine the HLA-DR type in 75 patients with well differentiated thyroid carcinoma from the same area as their previous population, and they compared the current results with the results from a group of 170 healthy controls. RESULTS. A significant increase in HLA-DR11, rather than HLA-DR1, was observed in patients with well differentiated thyroid carcinoma among a population of patients from the same area that was studied previously. After excluding technical reasons to account for differences in disease association, they postulated that interim environmental factors, possibly radiation fallout , may have resulted in differences in genetic susceptibility to thyroid carcinoma. Consideration of the potential antigenic peptides that may be restricted by the two HLA-DR alleles may have allowed for the binding of similar peptides to initiate an immune response, likely leading to progressive immunomodulation of the tumor. Discriminat function analysis indicated a significant relation between tumor size and metastases and less lymphocytic infiltration of the tumor, but this was not related to HLA-DR phenotypes. CONCLUSIONS. The authors found that the study of major histocompatability complex alleles holds promise for understanding the events that initiate and maintain tumor immunomodulation.