AUTHORS
Michelle Y. Y. Lee, PhD, Christoph A. Thaiss, PhD2, Sally Arai, MD3, Sushma Bharadwaj, MD4, Saurabh Dahiya, MD5, Hany Elmariah, MD,MS6, Mathew J. Frank, MD, PhD7, Laura Johnston, MD8, Robert Lowsky, MD9, Lekha Mikkilineni, MD, MA10, David Miklos, MD, PhD11, Lori Muffly, MD12, Robert Negrin, MD13, Andrew Rezvani, MD14, Parveen Shiraz, MD15, Judith Anne Shizuru, MD, PhD16, Surbhi Sidana, MD17, Melody Smith, MD,MS18, Hyma Vempaty19, Wen-Kai Weng, MD, PhD20, Bing Zhang, MD21, Vanessa E Kennedy, MD22
AFFILIATIONS
1 Stanford University, Stanford, California, United States, The Arc Institute, Palo Alto, California, United States
2 Stanford University, Stanford, California, United States, The Arc Institute, Palo Alto, California, United States
3 Stanford University, Stanford, California, United States
4 Stanford University, Stanford, California, United States
5 Stanford University, Stanford, California, United States
6 Stanford University, Stanford, California, United States
7 Stanford University, Stanford, California, United States
8 Stanford University, Stanford, California, United States
9 Stanford University, Stanford, California, United States
10 Stanford University, Stanford, California, United States
11 Stanford University, Stanford, California, United States
12 Stanford University, Stanford, California, United States
13 Stanford University, Stanford, California, United States
14 Stanford University, Stanford, California, United States
15 Stanford University, Stanford, California, United States
16 Stanford University, Stanford, California, United States
17 Stanford University, Stanford, California, United States
18 Stanford University, Stanford, California, United States
19 Stanford University, Stanford, California, United States
20 Stanford University, Stanford, California, United States
21 Stanford University, Stanford, California, United States
22 Stanford University, Stanford, California, United States
ABSTRACT
Introduction
In allogeneic hematopoietic cell transplantation, donor–recipient HLA matching is associated with improved survival and reduced graft-versus-host disease (GVHD). Match grade is conventionally assessed at HLA-A, -B, -C, and -DRB1 (8/8 match), with or without HLA-DQ (10/10). Even matched pairs, however, may still differ at higher resolution, including 3rd field (synonymous coding variation) and 4th field (noncoding variation). These “hidden” mismatches are not apparent on conventional typing. Their clinical and biological significance is unknown.
Methods
We studied 546 unrelated donor–recipient pairs transplanted at Stanford University from 2014–2024 who were 10/10 matched at high resolution, indicating identical antigen-binding-site protein sequences. Pairs were reevaluated at extended resolution (up to 4th field) at HLA-A, -B, -C, and -DRB1 to identify hidden mismatches. Outcomes included acute GVHD (aGVHD; primary), chronic GVHD (cGVHD), overall survival (OS), relapse-free survival (RFS), and graft failure. Cox regression was used for time-to-event outcomes and Fine–Gray regression for competing risks. Models adjusted for diagnosis, disease risk, donor age, sex match, ABO compatibility, conditioning intensity, and GVHD prophylaxis; fully adjusted models also included HLA-DRB3/4/5 and -DP mismatch status.
Results
Among all pairs, the median recipient age was 61 years (IQR, 48–68); 70.3% had donors aged ≤35 years, 44.6% received myeloablative conditioning, and 15.6% received post-transplant cyclophosphamide (PTCy) GVHD prophylaxis. Overall, 289/546 pairs (52.9%) carried ≥1 hidden mismatch, with a mean of 0.78 mismatches per pair overall and 1.48 among pairs with ≥1 mismatch.
Presence of any hidden mismatch was associated with aGVHD (Grade [Gr] II–IV: sHR 1.46, P=.039; Gr III–IV: sHR 1.71, P=.054). Day-100 cumulative incidence of Gr II–IV and III–IV aGVHD was 26.3% vs 17.9% and 12.5% vs 6.6%, respectively, for pairs with vs without a hidden mismatch. Effect sizes remained similar after full adjustment for HLA-DRB3/4/5 and -DP matching (Gr II–IV: sHR 1.41, P=.061; Gr III–IV: sHR 1.68, P=.066). Each additional hidden mismatch increased the risk of Gr II–IV (sHR 1.19, P=.056) and Gr III–IV aGVHD (sHR 1.23, P=.115). Hidden mismatch did not impact cGVHD, OS, RFS, or graft failure.
Hidden mismatches were overwhelmingly a 4th field phenomenon: 391/428 (91.4%) allele-level mismatches occurred at the 4th field and 37/428 (8.6%) at the 3rd. A graded decrease in aGVHD was observed across pairs with a 3rd-field mismatch, a 4th-field-only mismatch, and no mismatch (Day-100 Gr II–IV: 35.3% vs 25.1% vs 17.9%, Gray P=.060; Gr III–IV: 20.6% vs 11.4% vs 6.6%, Gray P=.042), suggesting that even sequence differences not altering the antigen-binding-site protein may still have functional consequences.
PTCy prophylaxis was associated with significantly lower aGVHD: day-100 cumulative incidence of Gr II–IV aGVHD was 14.3% with PTCy versus 24.1% without PTCy (adjusted sHR 0.45, P=.008), and Gr III–IV aGVHD was 2.4% versus 11.3% (adjusted sHR 0.15, P=.009). This reduction was more pronounced among pairs carrying a hidden mismatch (Gr II–IV: 14.0% vs 29.1%; Gr III–IV: 2.0% vs 14.8%), although the small number of events among PTCy recipients (13 Gr II–IV events) limited the power to formally evaluate effect modification.
Because hidden mismatches are predominantly noncoding and do not alter the antigen-binding-site protein, we investigated potential influence on HLA expression. Using Saluki, a deep-learning model of transcript stability, we quantified donor–recipient divergence in predicted HLA mRNA half-life (dHL) in 390 pairs with complete sequence information. Greater dHL was associated with Gr III–IV aGVHD (sHR 1.32, P=.037; Gr II–IV: sHR 1.15, P=.16). After adjustment for hidden-mismatch status, estimates attenuated and were nonsignificant (Gr III–IV: sHR 1.25, P=.15; Gr II–IV: sHR 1.13, P=.31), suggesting that dHL may capture part of the biological signal of hidden mismatch without independent prognostic effect.
Conclusions
Hidden HLA mismatches at HLA-A, -B, -C, and -DRB1 are common among conventionally 10/10-matched unrelated donor–recipient pairs and are associated with increased acute GVHD. Divergence in predicted HLA mRNA stability supports a potential expression-related mechanism. These findings support external validation of higher-resolution HLA variation as a consideration in donor selection.
DISCLOSURES
Michelle Y. Y. Lee, PhD: No relevant conflicts of interest to declare.