Syed Ali Abutalib, MD
Katharina Fleischhauer, MD
Ephraim Joseph Fuchs, MD, MBA
Besides relapse of the malignant disease, graft-vs-host disease is still one of the greatest concerns, in terms of adverse effects, following a hematopoietic cell transplantation (HCT) in a patient with a hematologic malignancy. To mitigate these concerns, investigators are continually analyzing donor protocols and delving into the etiology of the disease to find better HCT parameters to prevent the development of graft-vs-host disease in these patients.
HLA-Mismatch Guidelines
Dr. Abutalib: What are the acceptable HLA-mismatch numbers outside of clinical trials (eg, 7/8 vs 6/8 vs 5/8)?
Dr. Fleischhauer (University Hospital Essen): I am not sure there are any rules about that yet. At our center, we try to avoid < 7/8 but accept them in exceptional cases. In terms of human leukocyte antigen (HLA) matching, I don’t see why haploidentical family donors should be any better than haploidentical unrelated donors. At least from the immunogenetic point of view, there is no reason for this, and it is also supported by the encouraging outcome data of the ACCESS trial.1 Many years ago, there was discussion about whether matched unrelated donors could be as good as matched sibling donors; I think, nowadays, nobody argues this (unrelated donors could be even better due to the younger donor age). I anticipate this might be similar in a few years for haploidentical family/unrelated donors. Time will tell.
Dr. Fuchs (Johns Hopkins Medicine): At Johns Hopkins, our upper limit on mismatches is 6/12 HLA alleles, and we will accept some double mismatches, for instance at HLA-DPB1. When we were doing lots of HLA-haploidentical related donor transplants, one of the first analyses we conducted was to review outcomes according to the number of mismatches. The numbers were heavily weighted to 6/10 and 5/10 matched donors, but we did not see any detrimental impact of an increasing number of mismatches between donor and recipient in haploidentical transplants. There is no reason to think it would be any different using unrelated donors, provided posttransplant cyclophosphamide is being used.
The fundamental question is whether your patient could benefit from an allogeneic transplant or not. If the answer is “yes,” you try to find an HLA-matched sibling or unrelated donor and then resort to HLA-haploidentical related or partially mismatched unrelated donors. You don’t say, “well, I would have done the transplant if we had a 9/10 matched unrelated donor, but I won’t do it if the best we have is an 8/10 matched unrelated donor.”
What Is PBM-GVH?
Dr. Abutalib: In a recent Journal of Clinical Oncology paper, the term “PBM-GVH” (peptide-binding motif–graft-vs-host) was used.2 Does this imply a conventional graft-vs-host disease vector direction based on donor and recipient HLA typing, or is additional testing required to determine PBM–graft-vs-host?
Dr. Fleischhauer: The term PBM-GVH was first described in the Journal of Clinical Oncology article by Crivello et al,3 in which PBM–graft-vs-host mismatches are those between HLA molecules presenting different peptide repertoires and hence higher immunogenicity. So, in contrast to conventional graft-vs-host disease mismatches, it’s not just the number but the quality of the mismatch.
In the setting of conventional graft-vs-host disease prophylaxis (not posttransplant cyclophosphamide), PBM–graft-vs-host mismatches were associated with inferior survival in two independent retrospective cohorts from the Center for International Blood and Marrow Transplant Research (CIBMTR)3 and the European Society for Blood and Marrow Transplantation (EBMT).2 However, as pointed out previously, this could be different and even potentially opposite under posttransplant cyclophosphamide. So, for the moment, I would not recommend avoiding these mismatches if you use posttransplant cyclophosphamide (but would try to avoid them if you use conventional graft-vs-host disease prophylaxis). No additional testing is needed to determine them; together with CIBMTR, we developed a Web tool that allows for the determination of the PBM mismatch status based on the high resolution HLA types of patient and donor.4
Dr. Fuchs: I will say that if posttransplant cyclophosphamide can nullify the detrimental effect of HLA mismatches, it should also be able to nullify the effect of differences in minor histocompatibility antigens, which is what I suspect PBMs are.
Dr. Fleischhauer: There is a fundamental difference between PBMs and minor histocompatibility antigens, as the latter are immunogenic because of amino acid variability of the peptide that is presented by a matched HLA molecule. In contrast, PBMs are immunogenic because of amino acid variability of the mismatched HLA molecule, presenting any (also invariable) peptide that is unique to the HLA mismatch. This means that the number of potentially immunogenic peptides is significantly higher for PBMs than for minor histocompatibility antigens. It should also be noted that PBM mismatches have more unique peptides and hence are more immunogenic than PBM matches, although both are HLAmismatches. It is possible that posttransplant cyclophosphamide might abrogate the effect of PBM matches but not those of PBM mismatches. More studies are needed to prove or disprove this hypothesis.
Association of Mismatch Status With Survival
Dr. Abutalib: My understanding is that a better donor would typically be one with a DR mismatch rather than a class I mismatch. Is this correct, or is it too early to draw such a conclusion?
Dr. Fleischhauer: Under conventional (non–posttransplant cyclophosphamide) graft-vs-host disease prophylaxis, a DR mismatch was not associated with inferior survival, whereas a class I mismatch was.2 This makes sense if one thinks of HLA-DR as a potential graft-vs-leukemia target expressed less on graft-vs-host disease target organs than on hematopoietic cells (which after transplant are of donor origin except for relapsing leukemia). In the context of posttransplant cyclophosphamide, Fuchs et al5 showed that HLA-DR mismatches in graft-vs-host direction were associated with better disease-free survival and less relapse. So, I would agree with you that a DR mismatch (especially in graft-vs-host direction) might be preferable to a class I mismatch.
“In terms of HLA matching, I don’t see why haploidentical family donors should be any better than haploidentical unrelated donors.”— KATHARINA FLEISCHHAUER, MD
Tweet this quote
Dr. Fuchs: The study by Arrieta-Bolaños and colleagues seems to indicate that class II mismatches are better tolerated than class I mismatches.2 There are some papers asserting that class II mismatches are actually beneficial,6,7 whereas others find no benefit to class II mismatches, all in the context of posttransplant cyclophosphamide with graft-vs-host disease prophylaxis. However, I have not seen anything to suggest that class II mismatches are detrimental when posttransplant cyclophosphamide is used; however, the 2021 article in Blood did show a detrimental impact of HLA-C mismatching.6 In a larger context, I hypothesize that CD8-positive T cells, which recognize class I mismatches, are the mediators of classical graft-vs-host disease (skin rash, diarrhea, liver damage), whereas class II mismatches are associated more with phenomena like cytokine-release syndrome (seen early after mismatched transplants) and engraftment syndrome (which we don’t see much in the era of posttransplant cyclophosphamide).
Dr. Abutalib: Does allele mismatch become insignificant if the antigen is matched in class I?
Dr. Fleischhauer: Under conventional (non–posttransplant cyclophosphamide) graft-vs-host disease prophylaxis, antigen class I mismatch was worse than allele mismatch, but allele mismatches also still had a negative impact on outcome.2 There was no significant interaction with posttransplant cyclophosphamide in this study, suggesting the association might be similar under posttransplant cyclophosphamide. But as mentioned, the numbers of patients were small, and other studies (with similarly small patient numbers) suggested otherwise; so with posttransplant cyclophosphamide, we still don’t know.
Dr. Fuchs: I agree that antigen mismatches are more detrimental than allele mismatches, but I still believe allele mismatching is significant, even with posttransplant cyclophosphamide.
Dr. Abutalib: How important is DQ matching if DR is mismatched?
Dr. Fleischhauer: When DR is mismatched, often DQ is also mismatched, so it is difficult to dissect. But we did not see an association with outcome in the Journal of Clinical Oncology study by Arrieta-Bolaños et al,2 in line with the U.S. standard of considering 8 rather than 10 alleles. However, DQ mismatching might be exploited for graft-vs-leukemia, as suggested by the recent study by Petersdorf et al,8 showing lower relapse with certain DQ mismatches in a mostly posttransplant cyclophosphamide–based cohort of haploidentical transplants.
Dr. Fuchs: In the study we did,5 the best outcomes were obtained when DRB1 was mismatched, but DQB1 was matched. These patients did better than patients who were mismatched at both loci, but I don’t have a good explanation for this finding. I should also note that when some effect is found when analyzing data from one large database (eg, the CIBMTR database) but not from the analysis of another database (eg, EBMT), maybe the effect size is not all that large. My guess is that the effect of the HLA-DQB1 phenomena is small, because HLA-DQB1 is a low-expression class II molecule.
Dr. Abutalib: Does DP mismatch matter in this setting?
Dr. Fleischhauer: As mentioned, Fuchs et al5 showed better overall and disease-free survival with nonpermissive DP mismatch in graft-vs-host direction after haploidentical hematopoietic cell transplantation under posttransplant cyclophosphamide. But these data need to be validated, ideally prospectively, in independent cohorts.
Dr. Fuchs: Initially, we found there was a benefit to HLA-DPB1 mismatching in the haploidentical setting with posttransplant cyclophosphamide, as previously mentioned. We have also looked at the effect of nonpermissive HLA-DPB1 mismatching in the unrelated donor setting with either conventional graft-vs-host disease prophylaxis (eg, tacrolimus and methotrexate) or with posttransplant cyclophosphamide. With conventional prophylaxis, nonpermissive mismatching of HLA-DPB1 in the graft-vs-host direction was detrimental to outcome. With posttransplant cyclophosphamide, there was no detrimental effect of DPB1 nonpermissive mismatching, but we did not find a beneficial impact of such mismatches. Still, posttransplant cyclophosphamide eliminated the detrimental impact of HLA-DPB1 nonpermissive graft-vs-host mismatches on transplant outcome, which means that such mismatches do not need to be considered when posttransplant cyclophosphamide is used. These data have been submitted for publication.
Finally, I would like to point out two things: (1) I have never seen a paper in which mismatched transplants did as well as matched transplants when all patients received the same graft-vs-host disease prophylaxis regimens. (2) Posttransplant cyclophosphamide has done such a good job of mitigating the detrimental impact of mismatching that other criteria such as donor age and cytomegalovirus status may be as important or more important in donor selection.
DISCLOSURE: Dr. Abutalib reported a financial relationship with AstraZeneca. Dr. Fleischhauer and Dr. Fuchs reported no conflicts of interest.
REFERENCES
1. Al Malki MM, Bo-Subait S, Logan B, et al: Post-transplant cyclophosphamide-based graft-versus-host disease prophylaxis following mismatched unrelated donor peripheral blood stem cell transplantation. 2024 ASCO Annual Meeting. Abstract 6503.
2. Arrieta-Bolaños E, Bonneville EF, Crivello P, et al: Human leukocyte antigen mismatching and survival in contemporary hematopoietic cell transplantation for hematologic malignancies. J Clin Oncol 42:3287-3299, 2024.
3. Crivello P, Arrieta-Bolaños E, He M, et al: Impact of the HLA immunopeptidome on survival of leukemia patients after unrelated donor transplantation. J Clin Oncol 41:2416-2427, 2023.
4. CIBMTR: Class I HLA Peptide Binding Motif (PBM) Matching Tool. Available at https://pbm-matching-tool.b12x.org/. Accessed May 28 2025.
5. Fuchs EJ, McCurdy SR, Solomon SR, et al: HLA informs risk predictions after haploidentical stem cell transplantation with posttransplantation cyclophosphamide. Blood 139:1452-1468, 2022.
6. Fuchs EJ, O’Donnell PV, Eapen M, et al: Double unrelated umbilical cord blood vs HLA-haploidentical bone marrow transplantation: The BMT CTN 1101 trial. Blood 137:420-428, 2021.
7. Solomon SR, Aubrey MT, Zhang X, et al: Class II HLA mismatch improves outcomes following haploidentical transplantation with posttransplant cyclophosphamide. Blood Adv 4:5311-5321, 2020.
8. Petersdorf EW, McKallor C, Malkki M, et al: HLA haplotypes and relapse after hematopoietic cell transplantation. J Clin Oncol 42:886-897, 2024.
Dr. Abutalib is Director of the Malignant Hematology and Transplantation & Cellular Therapy Programs at the Advocate/Aurora St. Luke’s Medical Center, Milwaukee, and Associate Professor at Rosalind Franklin University of Medicine and Science, Chicago. Dr. Fleischhauer is Director of the Institute for Experimental Cellular Therapy, University Hospital Essen, Essen, Germany. Dr. Fuchs isProfessor in the Department of Oncology at Johns Hopkins Medicine, Baltimore.
