A Case-Control Study of Candidate Immunoregulatory Genes Reveals Haplotypes That Influence Inhibitor Risk in Severe Hemophilia A.

Inhibitor antibodies to FVIII develop in ∼20% of patients with severe hemophilia A, and are the most important adverse events associated with FVIII replacement therapy. Mutations in the FVIII gene are the major determinant of inhibitor risk, but variations in immune response genes, such as IL10 or TNFα, (Astermark, et al, 2006a, 2006b) may also confer risk of inhibitor development and variations in the CTLA4 gene may protect against inhibitors (Astermark et al, 2006c).

Using a case-control study design we sought to confirm and extend previous observations of inhibitor risk modifying genes seen in family studies. Candidate genes were selected based on previous clinical or animal studies and included cytokines involved in the TH1/TH2 immune responses, and genes thought to decrease demand for FVIII therapy (e.g., FV Leiden and prothrombin 20210 polymorphisms).

We used the CEU population data in the HapMap database to select haplotype tagging SNPs in IL1α/β, IL2, IL4, IL6, IL10, IL12A/B, IL13, IL17A, IL22, TNFα, CTLA4, interferon-γ, TGFβ, zinc α-2 glycoprotein I, IL1RN, and the FVIII gene, as well as the FV Leiden and prothrombin 20210 gene polymorphisms. A total of 366 tagging SNPs were selected with a goal of covering the entire coding and regulatory regions of the genes (spanning from 20kb 5' to 10kb 3' of the gene's coding sequence), resulting in 100% coverage of potential haplotypes (r² = 1.0). DNA was purified from 915 Caucasian, severe hemophilia A patients (282 inhibitor cases and 633 non-inhibitor controls) who participated in the Multicenter Hemophilia Cohort Studies I & II. Subjects were classified as having an inhibitor to FVIII on the basis of at least one inhibitor titer ≥1.0 Bethesda unit. SNP genotypes were determined by Sequenom MALDI-TOF spectroscopy. Haplotype frequencies were estimated using expectation maximization (EM) algorithm, and generalized linear models (GLM) were used to assess the marginal effect of SNPs and haplotypes on FVIII inhibitor development risk after adjusting for HIV infection and HCV persistence (prevalence in controls, 46% and 77%, respectively).

Of the 366 SNPs, 298 (81%) had unambiguous genotypes in >80% of the subjects and consequently qualified to the association analyses. Significant associations were seen between loci in the IL10, IL12A, IL1, IL2, TNFα, & IL17A genes and inhibitor development. Particularly, individuals carrying an 8 SNP haplotype located ∼10 kb 5' to the IL10 initiation start site were at higher risk to develop inhibitors than individuals with the most common haplotype (OR:1.54, 95% CI:1.15–2.06, p-value = 0.004). This haplotype covers a region that includes the IL10G microsatellite previously studied by Astermark et al (2006a), as well as the nearby IL10R microsatellite. Interestingly, the effect of this haplotype on FVIII inhibitor development was larger among HIV+ subjects in our study (OR: 1.81, 95% CI: 1.18–2.77 vs. OR: 1.28, 95% CI: 0.85–1.93 for HIV+ and HIV- subjects, respectively). A similar phenomenon was seen with haplotypes in the IL12A and IL2 genes. Additional haplotypes in genes for IL1α, IL1β, TNFα, and IL17A significantly increased or decreased risk for inhibitor development. No association was seen between FVIII haplotypes and inhibitors in this group of Caucasian subjects with hemophilia A, in contrast to the findings in African-Americans by Viel et al (2009). Also, no significant effect on inhibitor risk was seen for the factor V Leiden or prothrombin 20210 polymorphisms.

Our large case-control study confirms the findings in family studies that IL10 and TNFα genotypes confer risk for inhibitor development in Caucasians with severe hemophilia A, which differed by HIV status and were of lower magnitude than seen previously in family studies. We identified four other genes in which certain haplotypes alter inhibitor risk. More work is needed to identify inhibitor associations in non-Caucasian populations and to corroborate and more specifically define our novel associations in Caucasians. Moreover, whole genome association and other studies should be considered to identify additional modifier genes and potential targets for intervention to prevent inhibitors.

No relevant conflicts of interest to declare.