Autoantibodies to LKM1 are present in 1 to 10% of patients with hepatitis C virus (HCV)-related chronic hepatitis (3, 4, 11, 21). The presence of these autoantibodies is associated with an increased risk of developing hepatitic flares and thyroid disorders in HCV-infected patients (17, 18). The antigenic target of anti-LKM1 antibodies is cytochrome P450IID6 (CYPIID6), a 50-kDa microsomal enzyme involved in the metabolism of xenobiotics (12, 19, 41). Molecular mimicry between HCV proteins and CYPIID6 has been suggested as a possible mechanism for the origin of these autoantibodies (2, 13). Certain HLA and CYPIID6 alleles are associated with the prevalence of LKM1 autoantibodies in some populations, suggesting the involvement of host genetic factors in the induction of these antibodies (2, 9). Immunoglobulin (Ig) GM and KM allotypes-hereditary antigenic determinants of IgG heavy chains and κ-type light chains, respectively—are associated with susceptibility to several autoimmune and infectious diseases (7, 8, 24, 27, 30, 31, 34). They also influence immune responsiveness to infectious epitopes as well as to certain autoantigens (10, 25, 26, 28, 29). Of particular relevance here, certain GM and KM determinants interact to influence the outcome of HCV infection (22). These observations led us to hypothesize that GM and KM allotypes may contribute to the generation of anti-LKM1 autoantibodies in HCV-infected subjects. Between 2002 and 2004, 129 HCV-infected patients were consecutively enrolled at the Department of Internal Medicine, Cardioangiology, and Hepatology, Alma Mater Studiorum-University of Bologna. Criteria for inclusion in the study were the following: serum anti-HCV and HCV RNA positivity, abnormal alanine transaminase levels at least twice in the past 6 months, and chronic inflammation on liver histology. Other causes of liver disease were excluded by appropriate tests. The study population also included 90 ethnically matched blood donors who were negative for anti-HCV antibodies. The study was approved by the appropriate ethics committees for human research. Anti-HCV antibodies were tested by third-generation enzyme immunoassay (Ortho HCV version 3.0 ELISA; Ortho-Clinical Diagnostics, Inc., Raritan, NJ) according to the manufacturer's instructions, and HCV RNA was tested by nested PCR using primers derived from the highly conserved 5′ noncoding region of the viral genome. Anti-LKM1 antibodies were measured by indirect immunofluorescence on cryostat sections of rat liver, kidney, and stomach specimens at a serum dilution of 1:40, and titers were determined to extinction, as previously reported (20). Serum samples were typed for G1M (1/a, 2/x, 3/f, 17/z), G2M (23/n), G3M (5/b1, 6/c3, 13/b3, 21/g), and KM 1 and 3 allotypes by a standard hemagglutination-inhibition method (38). The notation follows the international system for human gene nomenclature (35). Logistic regression and Fisher's exact test were used to determine the significance of the association between GM and KM phenotypes and the prevalence of anti-LKM1 antibodies. Odds ratios (ORs) were calculated to measure the strengths of the associations observed. ORs are not presented for comparisons where the cell counts were less than or equal to 5, as the use of large-sample theory to calculate confidence intervals for the ORs can only be justified when all of the expected cell counts are greater than 5. Statistical significance was defined as P < 0.05. Because of almost absolute linkage disequilibrium between particular GM alleles in a given race, data were analyzed as a group (phenotypes) rather than according to the presence or absence of individual markers (36). Subjects with very unusual GM phenotypes were classified as “other” for statistical analyses. All analyses were conducted using SAS version 8.1 software. The distribution of GM and KM phenotypes in relation to the presence or absence of autoantibodies to LKM1 is given in Table Table1.1. The GM 1,3,17 23 5,13,21 phenotype was significantly associated with the prevalence of anti-LKM1 antibodies. Its frequency was significantly higher (45%) in subjects with anti-LKM1 antibodies than in those lacking these antibodies (14%) or in random blood donors (20%; data not shown). Among HCV-infected subjects, those with GM 1,3,17 23 5,13,21 were over five times as likely to possess anti-LKM1 antibodies as those lacking this phenotype (OR = 5.13). TABLE 1. Distribution of GM and KM phenotypes in HCV-infected subjects in relation to the presence or absence of autoantibodies to LKM1 In addition to its main effect, this phenotype also interacted with the KM 1,3 phenotype (Table (Table2).2). Subjects with GM 1,3,17 23 5,13,21 but lacking KM 1,3, as well as those positive for both heterozygous phenotypes, were more likely to possess anti-LKM1 antibodies than those lacking both these phenotypes (P = 0.007 and 0.037, respectively). No other significant associations were found. TABLE 2. Distribution of combined GM 1,3,17 23 5,13,21 and KM 1,3 phenotypes in HCV-infected subjects in relation to the presence or absence of autoantibodies to LKM1 The results presented here show a distinct association between the presence of the heterozygous GM 1,3,17 23 5,13,21 phenotype and the prevalence of autoantibodies to LKM1. The most probable haplotypes responsible for this phenotype are GM 1,17 21 and GM 3 23 5,13. One explanation of these findings could be that the GM locus directly affects the autoreactivity to LKM1 antigens. Perhaps the B cells carrying the heterozygous phenotype GM 1,3,17 23 5,13,21 on their Ig receptors are more efficient in the uptake, processing, and subsequent presentation of LKM1 antigenic peptides to the collaborating T cells, resulting in autoantibody production. Although GM and KM markers are located in the constant region, there is a growing body of evidence for the involvement of these regions in antibody specificity usually associated with the variable (V) region of the Ig molecule. Possible mechanisms include a direct contribution to the formation of idiotypic determinants, an effect on V-region protein conformation, modulation of antibody binding affinity, and linkage disequilibrium with alleles coding for the V-region epitopes (5, 15, 16, 33, 37). The significant interaction between the GM 1,3,17 23 5,13,21 and KM 1,3 phenotypes may be a reflection of a preferential association of heavy and light chains of particular genotypes in the synthesis of LKM1 antibodies. Such nonrandom pairing of heavy and light chains has been reported for experimental animals (6, 32). It may be relevant to note that in addition to the involvement of GM and KM allotypes in the persistence and/or clearance of HCV infection mentioned earlier, we have reported interactive effects of these determinants in humoral immune responses to group B streptococcus antigens and the Epstein-Barr virus (1, 23) as well as cellular immune responses to the streptococcal cell wall antigens (39). Although statistically significant, the results of the interaction analyses should be viewed with caution. The sample size for these comparisons was small, and the findings need to be replicated in a larger study population. In addition to the direct involvement of GM and KM genes in autoreactivity to LKM1, the associations reported here could also result from linkage disequilibrium between GM and KM alleles and alleles of an as-yet-unidentified immune response gene(s) for the induction of LKM1 antibodies. Evidence for epistatic contributions of unlinked genes to the risk of human diseases is accumulating rapidly (14). In future studies, for further dissection of the associations reported here, it will be important to determine the possible interactive effects of Ig allotypes and other candidate genes on humoral and cellular immunity to the cross-reactive epitopes of HCV and CYPIID6. Among the candidate genes, HLA seems to be the most relevant, as it has been shown to be associated with the prevalence of anti-LKM1 antibodies (2). Furthermore, the simultaneous presence of particular GM and HLA phenotypes has been shown to dramatically increase (over 39 times) the risk of acquiring autoimmune chronic active hepatitis (40). KM allotypes also interact with particular HLA alleles and contribute to the occurrence of certain autoantibodies (10). To our knowledge, this is the first report of an association of GM and KM allotypes with a prevalence of anti-LKM1 antibodies.