We performed HLA typing in order to clarify the association. between the cancer risk, the patients' outcome, and the phenomenon of over-dominant selection (heterozygote advantage) at the major histocompatibility complex (MHC) in Japanese cancer patients. The subjects consisted of 3219 Japanese individuals (2275 male and 944 female), 2776 of whom had several types of cancers that were confirmed pathologically by means of resected or biopsied specimens, 318 who had benign diseases, and 125 normal control subjects. HLA antigens were serologically tested using the NIH standard microlymphocytotoxicity method for HLA-A, B, C, DR and DQ antigens. Eight HLA-A antigens, 20 HLA-B antigens, 5 HLA-C antigens, 12 HLA-DR antigens, and 4 HLA-DQ antigens were examined.
Cancer subjects showed a significant lower frequency of human leukocyte antigen-A33 (11.5% vs. 15.6%; P=0.0152; relative risk (RR)=0.7062), B44 (10.8% vs. 14.2%; P=0.0328; RR=0.7281), and DR9 (26.9% vs. 31.8%; P=0.0302; RR=0.7871) than noncancer subjects, respectively. They also showed a significant lower frequency in heterozygote patients who had all detectable and different forms of antigens at HLA-A, -B, -C, -DR, and -DQ loci than in non-cancer subjects, having 8.4%, and 15.8%, respectively (p<0.00001, RR=0.4882). But patients who had heterozygote advantage at MHC did not show good outcomes. Our data strongly suggest that individuals with heterozygous or homozygous alleles at HLA class I and II are more resistant to cancer, have more susceptibility to cancer, respectively, but with no effect on their outcomes.
View full abstract