GVHD meets GWAS

In this issue of Blood, Sato-Otsubo et al perform a genome-wide association study (GWAS) to identify minor histocompatibility antigens associated with clinically relevant graft-versus-host disease (GVHD).¹ 

The outcome of allogeneic hematopoietic cell transplantation (HCT) is influenced by several individual factors and their interaction. Among these factors, the number of HLA mismatches between donors and recipients is the most relevant factor associated with the development of GVHD.²  GVHD, a frequent posttransplant complication, is still the leading cause of mortality after HCT, and therefore HLA matching is of the utmost importance for a successful HCT. Previous reports have shown serious clinical complications even after HLA-identical sibling HCT.^3,4  In fact, polymorphisms in the CTLA-4 gene in HLA-identical sibling HCT has been associated with an increased risk of developing GVHD.⁵  Additional genetic disparities between donors and recipients are involved in clinical complications after fully HLA-matched HCT and when possible they should be identified for an improved donor selection. Almost 20 years ago, the first minor histocompatibility (H) antigens were characterized at the molecular level. Since then, a still growing list of >50 minor H antigens have been identified, and their mismatch consequences in transplantation have been described. Single nucleotide polymorphisms (SNPs) are the most common form of minor H antigens, and several approaches have been proposed to identify them frequently by using immunologic assays as a starting point.

As a result of the Human Genome and the International HapMap Projects in the early 2000s, GWASs become available to be applied in different settings, expanding dramatically our capacity to understand genetic variability. GWASs have been used to identify disease pathways, therapeutic targets, and biomarkers that can be used for monitoring response to treatment, just to mention only a few applications.

The Ogawa group has previously reported 6 associations between SNPs located around the HLA-DPB1 locus and HCT outcome in a large cohort involving ∼1800 patients and donors.⁶  To further improve sensitivity and mapping resolution, the same group combined GWAS with cytotoxicity assays.⁷ 

In this study, Sato-Otsubo et al identify SNP mismatches, located in autosomal chromosomes, associated with increased risk of severe acute GVHD (see figure). As mentioned by the authors, the potential clinical relevance of the study is the risk evaluation of severe GVHD that could not be predicted by standard high-resolution DNA typing. The study was performed in a large patient cohort completely matched for HLA-A, -C, -B, -DRB1, and -DQB1 loci. The Sato-Otsubo et al data based on the identified SNP disparities and the consequent risk of severe GVHD raise the question of whether transplantation should be at least reconsidered as the therapeutic option in patients that are otherwise HLA fully matched. Although improved donor selection based on SNP mismatches between donors and recipients will likely result in a better HCT outcome, this might not be a realistic option in patients without a HLA-matched donor. As stated by the authors, this strategy may be used instead to reduce the risk of severe GVHD or lead to a more effective GVHD prophylaxis. It should be noted also that an enhanced allo-HCT antileukemic effect may be achieved by specific antileukemia immunity through vaccination or genetically modified donor T cells rather than trying to control donor alloreactivity against the recipient.

Regarding GWASs, to achieve an adequate statistical power, this kind of approach requires a large cohort of patients and dedicated computational tools among other requirements. Therefore, GWASs, although a powerful method, are time-consuming, and costs might be prohibitive for some institutions. As suggested by the authors, an alternative to circumvent these obstacles is performing a large GWAS whose results can then be used by different transplantation centers.

The study by Sato-Otsubo et al cannot precisely identify the putative minor H antigens that correspond to the identified loci and therefore the detected SNPs responsible for causing severe GVHD. The mere association of SNPs disparities between donors and recipients can only predict severe GVHD with a certain probability and is tempting to speculate on the mechanism and pathways by which the identified loci affects the risk of GVHD. Several post-GWAS analyses have been proposed to biologically validate the real functional variants and establish whether the functional consequences from these variants take place at the transcriptional or the protein coding level. Additional confounding factors such as conditioning regimen, GVHD prophylaxis, and genetic variants that affects immune responses will further influence allo-reactions even after performing extensive post-GWAS analysis. Identifying and quantifying the individual contribution of each factor on the development of GVHD is challenging and, as stated by the authors, probably unrealistic. Despite inherent limitations of GWASs, reports such as this one by Sato-Otsubo et al are clinically useful and provide clear evidence of the association between minor H antigens and the risk of severe acute GVHD.