Abstract
Graft-versus-host disease (GVHD) is a major cause of morbidity and mortality following allogeneic hematopoietic stem cell transplantation (HSCT). Advances in molecular typing of major histocompatibility complex (MHC) alleles have decreased but not eliminated the occurrence of acute GVHD. Minor histocompatibility antigen (mHA) disparities between patients and their molecularly HLA-identical donors are likely to represent the residual targets of donor immunity. The majority of mHAs are encoded by SNPs which are disparate between patient and donor. Identification of these immunogenic SNPs would provide both a clinical benefit as well as insight into the biology of GVHD. We used Affymetrix high-density arrays to assess 10,043 SNPs spanning the entire genome. We have examined genomic DNA from 36 patient/donor pairs (21 with acute GVHD grades II-IV and 15 without GVHD). All patients and donors were HLA and sex-matched siblings and all patients received T-cell depleted transplants. T cell depletion is important in this context as it removes the potential pharmacogenomic confounding factor of differential sensitivity to immunosuppressive agents. Because the donors and recipients are related, large areas of the genome are identical by descent. Similarly sized blocks of chromosome are disparately inherited. We hypothesized that genomic regions important to the immunopathogenesis of GVHD would be more likely to be disparately inherited in pairs with GVHD than in pairs without GVHD. As a control of our method we compared the area of chromosome 6 surrounding the MHC complex and demonstrated that there was minimal SNP disparity in either GVHD pairs or asymptomatic pairs. We also focused our attention on the region of chromosome 6 outside the MHC area, hypothesizing that disparities in this area would have to result from recombination events. Interestingly the only major area of disparity on Chr 6 was in the gene GMDS, which has been proposed to play a role in the extravasation of activated lymphocytes. We then examined other chromosomes for evidence of genomic regions that that were selectively disparate in GVHD pairs. We were able to identify several genomic regions that appeared to be associated with this outcome. The Lander-Green algorithm was used to estimate the allele sharing between the siblings for each SNP marker and then the disparity score was defined as the average allele sharing of the GVHD group - average allele sharing of the asymptomatic group. The five genomic regions with the highest disparity scores (ranked in order of score) and the gene closest to the region are: 11q14.3 (gene cysteine and histidine-rich domain (CHORD)-containing, zinc binding protein 1), 9p22.2 (gene SH3GL2), 5p15 (gene KIAA0947), 10p15.1 (gene aldo-keto reductase family 1, member C4), and 9p21.2 (Chr 9 ORF 72). The highly disparate regions ranged between 70 and 2200 kb in size. Interestingly the total number of disparate SNPs was not different between pairs with GVHD and pairs without GVHD, supporting the hypothesis that a limited number of SNPs are important immunologic targets. This technique of genome-wide disparity analysis is a promising addition to our ability to define important mHA.
Author notes
Corresponding author