HLA-DP in unrelated hematopoietic cell transplantation revisited: challenges and opportunities

Katharina Fleischhauer and Bronwen E. Shaw

Article Figures & Data


  • Figure 1.

    Molecular targets of GVHD and GVL after HCT from HLA-matched siblings or UD. HLA molecules and mHAg peptides are depicted as Y or bars, and APCs and T cells as large or small oval shapes, respectively. T cells that undergo positive or negative selection in the thymus are labeled with a T or an X, respectively. Self-peptides presented by HLA molecules on the surface of an APC are indicated with gray dots; mHAg peptides are indicated with orange or pink dots. (A) Genotypically HLA-matched siblings. In the left top box, the 2 parental HLA haplotypes are schematically represented, along with 2 examples of mHAgs of which 1 is shared (gray bars) and the other is not shared (orange and pink bars) by the 2 siblings. In the bottom box, the T-cell repertoire of the sibling donor after thymic education is shown. Self-reactive T cells against HLA molecules have been clonally deleted by negative selection, whereas T cells alloreactive to the unshared mHAg (pink T cells) have undergone positive selection and mediate GVHD and/or GVL after HLA-matched sibling HCT. (B) HLA-matched UD. The top box shows the HLA component of a typical 10 of 10 UD-recipient pair matched for both HLA-A, -B, -C-DRB1 alleles (in black and gray) as well as for the LEL-DRB3/4/5 and -DQ (in dark and light green), but mismatched for both LEL-DPB1 alleles (in dark and light blue and red, respectively). Moreover, in this example, patient and donor are mismatched for the same mHAg as in panel A. The bottom box shows the donor T-cell repertoire after positive and negative selection, which contains T cells alloreactive not only to mismatched mHAgs but also to both mismatched LEL-DPB1 alleles. In this example, 1 of the 2 mismatched HLA-DPB1 alleles is nonpermissive (dark red) whereas the other one is permissive (light red), giving rise to direct T-cell responses of different magnitude but generally stronger than the response to mHAg mismatches. Together, these alloreactive T cells mediate GVHD and/or GVL after HLA-matched UD-HCT.


  • Table 1.

    Major milestone publications impacting practice in the clinical consideration of HLA-DPB1

    ReferenceNo.Patient and HCT characteristicsEraHLA matching* and clinical outcomesMajor findings/impact
    461 CML200111/12 in GVH direction; case report, acute rejectionFirst evidence that mismatched HLA-DPB1 can be the target of CD4+ T-cell–mediated allograft rejection
     Case report
    47143 Various diseases1996-200110/10; HLA-DPB1 allele mismatches; increased aGVHD, decreased relapseFirst evidence of a GVL effect targeted to HLA-DPB1 allele mismatches after UD-HCT
     Myeloablative conditioning
     T-cell depletion
    62118 Various diseases1995-200210/10; HLA-DPB1 nonpermissive TCE mismatches; increased aGVHD and NRMFirst description of a T-cell epitope matching score for HLA-DPB1
     Myeloablative conditioning
     T-cell depletion
    50423 Various diseases1996-200310/10; HLA-DPB1 allele mismatches; increased aGVHD, decreased relapse, better OS in ALLConfirmation of Shaw et al47 and possibly improved survival of ALL patients after HLA-DPB1 mismatched HCT
     Myeloablative conditioning
     T-cell depletion
    5572 Thalassemia1992-200410/10; HLA-DPB1 nonpermissive HVG mismatches; increased graft rejectionFirst association of nonpermissive TCE mismatches with outcome of HCT for nonmalignant disease
    79N/A N/AN/AN/AFirst protocol of targeted epitope-specific HLA-DPB1 typing for UD-HCT
    385929 Various diseases1984-20055-8/8; HLA-DPB1 allele mismatches; increased aGVHD, decreased relapseFirst definitive confirmation of GVL targeted to HLA-DPB1 allele mismatches after UD-HCT in a T-cell–replete cohort
     Registry study (14th IHIW)
    531 Lymphoblastic lymphoma200610/10; 1 permissive and 1 nonpermissive HlA-DPB1 TCE mismatchFirst isolation of CD4+ T cells alloreactive to mismatched HLA-DPB1 from a patient clearing residual malignant disease after donor lymphocyte infusion from an UD
     Case report
    63621 Various diseases1999-200610/10; HLA-DPB1 nonpermissive TCE mismatches; increased aGVHD and NRM, worse OSFirst demonstration of a negative impact of nonpermissive HLA-DPB1 mismatches in a large multicenter registry cohort
     Myeloablative and reduced-intensity conditioning
     Registry study (IBMDR)
    51488 AML, ALL, CML1996-200610/10 and <10/10; HLA-DPB1 allele mismatches; decreased relapse, increased NRM, worse OS in early disease in 10/10; improved OS in late disease in <10/10Differential impact of HLA-DPB1 allele mismatches in 10/10 or <10/10 according to disease status
     Myeloablative and reduced-intensity conditioning
     T-cell depletion
    494 Healthy donorsN/AIn vitro generation of 16 T-cell clones alloreactive to a variety of HLA-DP specificities by stimulation with transfected HELA cellsConfirmation that CD4+ T-cell alloreactivity to HLA-DPB1 can be elicited by both permissive and nonpermissive mismatch combinations
    6524 Healthy donorsN/AIn vitro mixed lymphocyte reactions between 10/10 matched, HLA-DPB1 permissive or nonpermissive UDFirst experimental evidence for more potent T-cell alloreactivity to nonpermissive than to permissive HLA-DPB1 TCE mismatches
    56115 AML, ALL, CML, MDS, UD-HCT1990-20026/12 to 12/12; retrospective matched control study graft failure vs engraftmentGraft failure is associated with donor-specific HLA antibodies, over half of them against HLA-DP
    57592 AML, ALL, MDS and others, UD-HCT2005-20099/12 to 12/12; prospective HLA antibody testingDonor-specific antibodies in 1.4% of patients, all against mismatched HLA-DP; significantly associated with graft failure
    398539 AML, ALL, CML, MDS1993-200710/10 and 9/10; HLA-DPB1 nonpermissive TCE mismatches; increased aGVHD and NRM, worse OSFirst definitive demonstration of a negative impact of nonpermissive HLA-DPB1 mismatches on OS in 10/10 and 9/10 UD-HCT
     Registry study (15th IHIW)
    5424 Various diseases2000-200810/10; CD4+ DLI for mixed chimerism or relapse; ex vivo analysis of T cells alloreactive to HLA-DPB1 in responding or nonresponding patientsFirst demonstration of reliable emergence of CD4+ T cells alloreactive to HLA-DP in patients responding to DLI after UD-HCT
     T-cell–depleted UD-HCT
    68N/A N/AN/AN/ALaunch of the first free online tool for HLA-DPB1 nonpermissive/permissive TCE match assignment
    362 AMLN/A10/10; HLA-DPB1 nonpermissive GVH mismatch in both patients; severe aGVHD of the gut after CMV reactivationHost-derived CMV-specific T cells can trigger CD4+ T-cell alloreactivity against mismatched HLA-DPB1 after UD-HCT, leading to severe GVHD
     2 case reports
    661281 AML, ALL, CML, MDS1988-200310/10; HLA-DPB1 nonpermissive GVH mismatches significantly lower relapse than permissive mismatches; no impact of concomitant mismatching for DPA1GVL effect by nonpermissive GVH mismatches; no impact of DPA1 on nonpermissive TCE mismatches
     Registry study (CIBMTR)
    48003 AML, ALL, CML, MDS1999-201110/10 and 9/10; HLA-DPB1 nonpermissive TCE mismatches; increased NRM, worse OS in 10/10 but not in 9/10First confirmation of a negative impact of nonpermissive HLA-DPB1 mismatches in 10/10 on OS in an independent validation cohort from a registry
     Registry study (CIBMTR)
    714 Healthy donorsN/AGeneration of 12 site-directed mutants of HLA-DPB1*09:01 at 10 key amino acid positions; investigation of their allorecognition by 17 T-cell effectors alloreactive to wild-type HLA-DPB1*09:01; determination of the median impact of each substitution on T-cell alloreactivity (FD)Development of a FD score system for individual amino acid polymorphism in HLA-DPB1, and for HLA-DPB1 alleles. The FD score of any current or future HLA-DPB1 allele can be used to predict its assignment to TCE groups
    527898 Various diseases1993-201010/10; role of HLA-locus mismatches including HLA-DPB1 allelic disparity; increased aGVHD, decreased relapseConfirmation of lower relapse risk associated with HLA-DPB1 allele mismatches; no such association was observed with HLA-DRB1 allele mismatches
     Registry study (JMDP)
    642029 AML, ALL, CML, MDS1988-200810/10; truly unidirectional GVH HLA-DPB1 expression mismatches; increased aGVHD when patient carries high-expression mismatchFirst demonstration that expression levels determine nonpermissive GVH HLA-DPB1 mismatches; conclusive association between 3′ UTR expression SNPs and exon HLA-DPB1 types
     Registry Study (CIBMTR)
    801342 Various diseases2000-200810/10; increased grade III-IV aGVHD with 2 allelic HLA-DPB1 mismatches, increased relapse with TCE nonpermissive HVG disparitiesHLA-DPB1 impacts aGVHD and relapse by number of allelic mismatches and TCE HVG disparities, respectively
     Registry study (RFGM)
    69N/A N/A2013N/ADescription of the Optimatch UD search tool from the German registry (ZKRD) including the possibility to select for HLA-DPB1 permissive donors
    70N/A N/A2016N/ADescription of the HapLogic UD search tool from the US registry (NMDP) including the possibility to select for HLA-DPB1 permissive donors
    71416 AL, MDS2005-201410/10; HLA-DPB1 nonpermissive TCE or ΔFD between patient and donor; ΔFD > 2.665: worse OS and EFS due to less relapse and less NRM; better predictive than nonpermissive TCE in this cohortRefinement of nonpermissive HLA-DPB1 TCE mismatches by ΔFD scores might further improve the predictive value of the algorithm
     Myeloablative conditioning
     T-cell depletion and T-cell replete UD-HCT
    75N/A AMLN/ADevelopment of an in vitro protocol to generate CD4+ T cells alloreactive to mismatched HLA-DPB1 on AML blasts or from RNA-transfected dendritic cells; preclinical humanized mouse model of GVL targeted to HLA-DPB1First protocol for the isolation and expansion of CD4+ T cells from the naive repertoire using primary AML blasts or transfected autologous dendritic cells for cellular immunotherapy; first demonstration of the in vivo efficacy of CD4+ T cells alloreactive to HLA-DPB1 in mediating GVL against AML
    67595 Patients searching for an UDN/A10/10; probability to find an HLA-DPB1 allele-matched or permissive UD70% of patients have a 10/10, young, HLA-DPB1 TCE permissive UD
    • aGVHD, acute GVHD; AL, acute leukemia; ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CIBMTR, Center for International Blood and Marrow Transplant Research; CML, chronic myeloid leukemia; CMV, cytomegalovirus; DLI, donor lymphocyte infusion; EFS, event-free survival; IBMDR, Italian Bone Marrow Donor Registry; IHIW, International Histocompatibility Workshop; JMDP, Japanese Marrow Donor Program; MDS, myelodysplastic syndrome; N/A, not applicable; NMDP, National Marrow Donor Program; NRM, nonrelapse mortality; OS, overall survival; RFGM, Registre France Greffe de Moelle; UTR, untranslated region; ZKRD, Zentrales Knochenmarkregister Deutschland.

    • * HLA matching is reported as the number of matched HLA-A, -B, -C, -DRB1, -DQB1 alleles (10 of 10 or other), followed by the algorithm used to consider HLA-DPB1 mismatches (alleles, nonpermissive TCE, or other).