HLA-DQ8 Haplotype Increases the Risk of Relapse after Allogeneic Stem Cell Transplantation Possibly Affecting Graft-Versus-Leukemia: A Single-Center Study on 400 Patients

Background: HLA DQ8 is a HLA DQ variant associated to celiac disease. Recently, HLA DQ has been studied in the setting of allogeneic hematopoietic stem cell transplantation (HSCT) and found to have increased cumulative incidence of relapse (CIR) and reduced disease-free survival (DFS) when presenting with specific genotype. However, the precise role of different DQ variants in donor HLA typing has not been determined yet. Our study aims at exploring the implication of HLA DQ8 haplotype on transplantation outcomes.

Methods: We conducted a monocentric, retrospective study including adult patients who received HSCT for hematological diseases at the Fondazione IRCCS Policlinico San Matteo of Pavia from January 2013 to February 2024. We evaluated the expression of HLA DQ8 haplotype by donors’ cells.

The primary endpoint was the association of HLA DQ8 with cumulative incidence of relapse (CIR). Secondary end-points included acute and chronic graft-versus-host disease (aGVHD and cGHVD), non-relapse mortality (NRM), and disease-free survival (DFS), overall survival (OS) and graft-free relapse-free survival (GRFS). Probabilities of OS, DFS and GRFS were calculated using Kaplan-Meier estimator method. GVHD, NRM, and CIR as cumulative incidences (CI) with death without the event of interest as competing factor. Multivariate analysis was conducted after adjustment for well-known factors associated with transplant outcomes including: age at HSCT, DRSS category (high/very high vs others), donor sex (female donor-male patient vs other combinations), type of donor (MRD vs others), stem cell source (peripheral blood vs bone marrow), total conditioning intensity (TCI) category (high vs others) and use of PTCy (yes vs no). All p-values <0.05 were considered statistically significant.

Results: A total of 400 patients have been recruited in the present study. The median follow-up from the date of transplant was 23.3 months (range 0.2-131.5). Thirty-eight out of 400 patients (9.5%) have been transplanted from a donor expressing HLA DQ8 haplotype (cohort DQ8+) while 362 patients (90.5%) have been not (cohort DQ8-). According to HLA DQ8 haplotype, DQ8+ patients had similar features in terms of median age at transplant (51 vs 54 years, p=0.217), sex (52.6% vs 55.9%, p=0.611), HCT-CI category (high HCT-CI 42.1% vs 38.4%, p=0.803) and disease-risk index (high/very high DRSS: 78.4% vs 69.3%, p=0.407), respectively. Also, primary indication for transplant was similar among the two cohorts (acute leukemias were 47.7% vs 62.7% in the DQ8+ and DQ8- populations, respectively, p=0.08).

Regarding transplantation platform features: TCI was high in 84.2% vs 83.6% (p=0.899), post-transplant cyclophosphamide (PTCy) was used in 31.6% vs 25.7% (p=0.441), stem cell source was peripheral blood in 81.6% vs 87.3% (p=0.377), donor was MRD in 28.9% vs 21.5% (p=0.473) in the DQ8+ and DQ8- groups, respectively.

Patients transplanted from donors expressing HLA DQ8 variants had higher incidence of relapse compared to patients transplanted from other DQ haplotypes. Indeed, 1-, 2- and 5-year CIR were 21.6% (CI 95%: 10.2-35.8) vs 10.3% (CI 95%: 7.4-13.8), 29.7% (CI 95%: 16.1-44.6) vs 13.8% (CI 95%: 10.4-17.7) and 33.2% (CI 95%: 18.6-48.6) vs 16.2% (CI 95%: 12.4-20.4) (p=0.002), respectively.

According to secondary outcomes, DQ8+ group have shown a slightly lower incidence of aGHVD (100d CI 26.7% vs 43.1%, p=0.056), and a significantly lower cumulative incidence of cGVHD (1y 23.0% vs 38.3%, p=0.023).

No statistical significant differences have been observed in terms of NRM (1y CI 16.2% vs 21.3%, p=0.545), DFS (1y 62.2% vs 68.4%, p=0.590), OS (1y 61.7% vs 71.7%, p=0.228) and GRFS (1y 48.8% vs 55.0%, p=0.197) in the DQ8+ and DQ8- cohorts, respectively.

By multivariable analysis, after adjustment for the above mentioned factors, DQ8 haplotype confirmed its detrimental role on the risk of relapse (HR 3.10, CI 95%: 1.65-5.82, p<0.001); conversely, its presence was protective against cGVHD (HR 0.52, CI 95%: 0.27-0.99, p=0.047).

Conclusion: The presence of the HLA DQ8 haplotype appears to confer a higher risk of relapse and a lower incidence of GVHD after transplantation. Our results suggest that HLA DQ8 expression affects the development of alloreactivity by promoting a tolerogenic response, which may reduce the GVL effect. Whether this information could guide donor selection should be prospectively confirmed.