Occurrence and Impact of Minor Histocompatibility Antigens' Disparities on Outcomes of Hematopoietic Stem Cell Transplantation from HLA-Matched Sibling Donors

We have examined the alleles of eleven minor histocompatibility antigens (MiHAs) and investigated the occurrence of immunogenic MiHA disparities in 62 recipients of allogeneic hematopoietic cell transplantation (allo-HCT) with myeloablative conditioning performed between 2000 and 2008 and in their HLA-matched sibling donors. Immunogenic MiHA mismatches were detected in 42 donor-recipient pairs: in 29% MiHA was mismatched in HVG direction, in another 29% in GVH direction; bidirectional MiHA disparity was detected in 10% and no MiHA mismatches in 32%. Patients with GVH-directed HY mismatches had lower both overall survival and disease-free survival at 3 years than patients with compatible HY; also higher incidence of both severe acute GvHD and extensive chronic GVHD was observed in patients with GVH-directed HY mismatch. On contrary, GVH-directed mismatches of autosomally encoded MiHAs had no negative effect on overall survival. Results of our study help to understand why posttransplant courses of allo-HCT from siblings may vary despite the complete high-resolution HLA matching of a donor and a recipient.


Introduction
The allogeneic hematopoietic cell transplantation (allo-HCT) still remains a curative treatment of many severe diseases, especially hematooncological malignancies. The successful donor search is one of the most important factors deciding about the feasibility of transplantation. It starts with search among the patient's siblings as the HLAmatched sibling donor is regarded as the optimal one. The odds ratio for HLA compatibility in siblings is 1 : 4. The probability of having a matched sibling donor by a particular patient is determined by the formula 1 − (0.75) n , where n equals the number of siblings. Despite the improved matching of donor-recipient pairs that was possible after the implementation of high-resolution methods of molecular HLA typing, the better outcomes of transplantations are still limited by high number of complications: graft versus host disease (GVHD), engraftment problems (lack or loss of engraftment), and relapse [1]. The long-term survival after allo-HCT is being estimated in the range of 40-70%. Failures are mainly due to infectious complications and GVHD (30-40% each), organ toxicity of chemotherapy (20%), and relapse (20-30%) [2].
HLA matching remains the most important factor influencing both donor selection and transplantation outcomes. However, research of the human genome revealed that polymorphism of nucleotides in genes that are non-HLA related (e.g., NOD2/CARD15 or genes encoding cytokines: TNF-alpha, IL-10, IL-6, interferon gamma, IL-1, and TGFbeta) may also determine the individual immunological phenotype of donor-recipient pairs, thus influencing GVHD, infections, and overall survival [3]. Minor histocompatibility antigens (MiHAs) belong to immunogenetic non-HLA related factors encoded by polymorphic genes, which may differ between the recipient and the donor and thus they may have impact on transplant outcomes.
The impact of antigens independent from Major Histocompatibility complex on transplantation results was first observed by Counce et al. in 1950s [4]. They explored graft rejection in inbred mice, which had undergone the transplantation of skin cells and neoplasmatic cells. Genes which were not associated with MHC responsible for slower course of rejection were called weak histocompatibility genes [4,5]. The first hypothesis concerning potential impact of MiHA on the outcome of BMT (bone marrow transplantation) was based on a case of a female recipient (with severe aplastic anemia) who received a transplant from her brother. Graft rejection after BMT was diagnosed and reactivity of cytotoxic T cells isolated from peripheral blood of recipient was directed to antigens present on donor's cells which were not associated with HLA [6].
Minor histocompatibility antigens are polymorphic peptides consisting of 9-12 amino acids. After binding to the antigen recognition site of either class I or class II HLA molecules present on a cell surface MiHAs can be recognized by T-lymphocytes. Thus the occurrence of MiHA depends on the presence of specific HLA antigens, which is called the MHC restriction. MiHAs are encoded by either autosomal chromosomes or by Y-chromosome [7][8][9]. Disparities of MiHA may result from polymorphism of amino acids, gene deletions [10], or from several intracellular mechanisms [11]. MiHA disparity may originate from a single or several amino-acid substitution in the part of MiHA peptide recognized by TCR (T-cell receptor), like in the case of HY and HA-1. Amino-acid polymorphism may be present in the region of MiHA that binds to HLA molecule, causing different expressions of peptide-HLA complex in the donor than in the recipient. Polymorphism may also pertain proteins responsible for intracellular processing of peptides, what leads to the presence or absence of peptides (e.g., HA-2 or HA-8) on cell's surface [12], or phosphoproteins (e.g., SP-110, MiHA discovered in 2006 by Warren et al.) [13].
Most MiHA possess only one immunogenic allele, which is sufficient to induce MiHA immunogenicity [12]. Up to date 18 autosomal and 10 Y-chromosome encoded MiHAs have been identified; those tested in our study are presented in Tables 1 and 2. There are two patterns of MiHAs' tissue distribution: restricted and broad. Autosomal HA-3, HA-8, and most of MiHAs encoded by Y-chromosome are present in most tissues, including those crucial for GVHD: skin, intestines, and liver [11,12]. Most of autosomal and 2 MiHAs encoded by Y-chromosome (B8/HY and B52/HY) appear only in hematologic cells including leukemic cells, dendritic cells, NK, and multiple myeloma cells [40]. Thanks to their restricted distribution all of them may be potentially exploited in immunotherapy. The other type of MiHAs' tissue distribution is their appearance on epithelial neoplasmatic cells, for example, HA-1 and ACC-1/ACC-2 [41,42], although in normal conditions they are restricted only to hematopoietic cells and are not present on epithelial cells.
Detection of MiHA bases most often on genomic typing with PCR-SSP method. The assessment of detected immunogenic disparities is simplified by the online availability of Leiden University Medical Center's dbMinor database [43]. Disparities of immunogenic MiHA alleles between the donor and the recipient may trigger GVHD and HVG reactions, which may lead to graft rejection or to GVH/GVL reaction [44][45][46]. T-lymphocytes directed against recipient specific MiHAs were detected in patients with GVHD [47]. In the group of 92 recipients of allo-HCT from unrelated donors, a higher incidence of chronic GVHD was observed in those with HY disparity [48]. Many clinical trials confirm that disparities of autosomally encoded MiHAs (like HA-1, HA-2, and HA-8) may increase the incidence of GVHD [15,17,22], while others did not confirm such dependence [49]. Female recipients after transplantation from male donors may experience graft failure due to HVG reaction against HY antigens resulting in a worse survival [3]. MiHA present on recipient's neoplasmatic cells (HA-1, HA-2, HA-8, HB-1, and HY) may constitute the target of cytotoxic CD8+ Tlymphocytes crucial for GVL reaction [12,50], leading to the decrease of relapse rate [51]. Use of cytotoxic T-lymphocytes recognizing selectively only MiHA present on neoplasmatic cells enables the separation of GVL effect from GVHD [52]. Such MiHAs can be used both in vivo for the production of vaccines enhancing GVL reaction and in vitro as a load to antigen presenting cells stimulating reactivity of cytotoxic Tcells [53]. HA-1 and HA-2 are the most intensively explored MiHAs in immunotherapy [12,[52][53][54].
The aim of this study was to determine MiHA alleles and genotypes enabling to detect their immunogenic disparities in sibling donor-recipient pairs and to explore their influence on the results of allo-HCT.

Transplantation Procedure.
Conditioning treatment was myeloablative (CyTBI: cyclophosphamide + total body irradiation in 12 pts, BuCy: busulfan + cyclophosphamide in  of respective donor-recipient pairs. The study has been approved by the responsible Ethical Committee of Medical University of Silesia.

Statistical Methods.
Median, minimal, and maximum values were used to show numeric parameters of donorrecipient groups. Statistical analysis of MiHA mismatches' impact on transplantation outcomes was conducted in accordance to recommendation of EBMT [56]. MiHA mismatches were grouped according to mismatch direction (GVH or HVG), tissue distribution (restricted or broad), and the way of coding (autosomal or by Y-chromosome) in search for their influence on transplant results. Analysed endpoints included overall survival (OS), disease-free survival (DFS), aGVHD, and limited and extensive cGVHD. Kaplan-Meier method was used to estimate the probability of impact of MiHA mismatches on overall survival and disease-free survival. Results were presented as percent ±95% confidence interval (CI). The cumulative incidence method was used to evaluate the probability of relapse and GVHD (acute or chronic) in order to account events which may influence the outcome as a competing risk. Results were presented also in percent ±95% CI. Results with significance level P < 0.05 were considered statistically significant.     Table 4 and the distribution of 11 MiHA alleles and genotypes in 62 related donor-recipient pairs is presented in Tables 5 and 6.

Impact of Immunogenic MiHA Mismatches on Allo-HCT
Outcomes. Analysis of overall survival showed unfavorable impact of GVH-directed Y-chromosome encoded MiHA mismatches (P = 0.011), as presented in Figure 1 and Table 7, and favorable trend in case of GVH-directed autosomal MiHA disparities (P = 0.045), as presented in Figure 2 and Table 7. GVH-directed mismatches of Y-chromosome encoded MiHA influenced unfavorable the disease free-survival (P = 0.05), as shown in Figure 3 and Table 7.
Serious (grade III or IV) acute GVHD was observed in 24 patients and it was influenced by Y-chromosome encoded GVH-directed MiHA mismatches (P = 0.037), which is presented in Figure 4 and Table 7.

Discussion
Minor histocompatibility antigens belong to genetic factors which may vary between the donor and the recipient despite identical HLA and thus they may influence allo-HCT results. Knowledge of MiHA alleles and genotypes   enables to detect their disparities, which could be helpful not only in optimal matching of a donor/recipient pair and in understanding transplant results, but also it may create a chance to the use of MiHA in immunotherapy aiming to improve patients' survival [52][53][54]. The largest meta-analysis of MiHA distribution was performed by Spierings et al. who described the results of a multicenter trial of 10 MiHA  distribution in 5 different ethnic groups worldwide. The study revealed significant differences in the frequency of MiHA alleles in dependence of geographical location, with UGT2B17 being the most variable MiHA [57]. Two MiHA trials have been performed in Polish population till now: in the first one HA-1 was analyzed in a group of 30 sibling pairs [58], another trial concerned the group of 92 unrelated pairs  Table 5. [12]. In our current study alleles and genotypes of 11 MiHAs have been estimated in 62 sibling donor-recipient pairs. Basing on our results and several other studies estimating the occurrence of specific MiHA mismatches in allo-HCT [59,60], HA-1 can be regarded as a candidate target for immunotherapeutic applications.
We have observed the unfavorable impact of GVHdirected mismatches of Y-chromosome encoded MiHAs on OS (P = 0.011) and DFS (P = 0.05). Y-chromosome encoded MiHA represents MiHA with "broad" tissue distribution. Attack of donor's T-lymphocytes on recipients' tissues precipitated by HY mismatch could explain the increased occurrence of severe forms of acute and chronic GVHD, leading to earlier deaths of recipients. In our study recipients of allo-HCT from siblings did not receive antithymocyte globulin, what probably influenced the worse course, including the fatal course of their GVHD. We have shown that GVH-directed mismatches of HY correlate significantly with serious (III or IV) aGVHD and extensive cGVHD. These results correspond to the reported influence of sex difference on transplant outcomes, especially in the case of female donor to male recipient (FDMR) transplants [61,62]. Oppositely, Markiewicz et al. in a study of 92 unrelated donor-recipient pairs found that HY mismatches in GVH direction influenced more favorable GVL effect than unfavorable GVHD, what resulted in the increased DFS (P = 0.05) [12,63]. The probable explanation of this difference in MiHAs impact on OS and DFS between related and unrelated allo-HCT may be the use of stronger standard immunosuppressive prophylaxis including pretransplant antithymocyte globulin in unrelated allo-HCT setting. Increased incidence of serious acute and extensive chronic GVHD associated with mismatches of Ychromosome encoded MiHAs, leading to a worse overall survival, may justify the administration of anti-thymocyte globuline before allo-HCT from sibling female donor to male recipient. Such approach could probably reduce the risk of GVHD originating from GVH-directed HY mismatch.
The analysis of GVH-directed mismatches of autosomal MiHAs, oppositely to HY, showed favorable trend to increase the OS, which was 76% in a mismatched versus 53% in a compatible groups at a 4-year posttransplant. Unlike GVHdirected HY disparities, those of autosomal MiHAs did not increase the occurrence of serious GVHD in our study, which contributed to the better survival. There are reports  describing the role of autosomal MiHAs in GVHD: for example, higher risk of aGVHD in the case of autosomal HA-1 incompatibility was reported in Tunisian group of 60 sibling donor-recipient pairs [64]. Others described increased incidence of cGVHD in the case of mismatched autosomal MiHAs localized on hematopoietic cells: HA-1, HA-2, and HA-8 [15,16,23,65]. There are also reports that report no impact of autosomal MiHAs on GVHD [49,66]. One could expect that disparities of MiHAs with broad tissue distribution present in the host should precipitate the posttransplant reaction of donor's lymphocytes and induce the GVHD. Unexpectedly, the tissue distribution of neither GVH-nor HVG-directed MiHA mismatches did not influence the incidence of GVHD.
Much lower probability of relapse following allo-HCT was observed by us in patients with HVG-, but not with GVH-directed MiHA mismatches. This finding, although intriguing, needs further confirmation as we do not find a reasonable explanation for this result. Japanese group found that GVH-directed HA-1 mismatches were associated with lower risk of relapse [51]. Similarly, experience of Polish group studying MiHAs in unrelated allo-HCT showed seldom episodes of relapse occurring when GVH-directed HY mismatches were present [63].
Results of our study help to explain why posttransplant courses of allo-HCT from siblings may vary despite complete high-resolution HLA-match and why cells interactions between the donor and the recipient may lead to serious complications.

Conclusions
GVH-directed HY mismatch significantly increased the occurrence of serious acute GVHD and extensive chronic GVHD and finally caused decreased overall survival. GVHdirected mismatches of autosomally encoded MiHAs had no negative effect on overall survival, which in fact was even longer. Findings of our study help to explain why the occurrence of immunological complications and in consequence final results of allo-HCT from high-resolution HLA-matched sibling donors are variable.