Complete Remission of Immunocytoma without Graft Versus Host Disease Caused by Allo-HLA-DP Specific T Cells.

Mismatching for HLA-DPB1 in unrelated donor hematopoietic stem cell transplantation (URD-SCT) has been associated with a significant decreased risk of disease relapse, indicating that HLA-DP might be a target for a graft versus leukemia (GVL) effect in HLA-class II expressing hematological malignancies. To determine whether a specific GVL effect could be caused by allo-HLA-DP specific T cells, we analyzed the immune response in a patient with a refractory immunocytoma responding to donor lymphocyte infusion (DLI) after HLA-DP mismatched URD-SCT. Patient and donor were fully matched for HLA-A, -B, -C, -DR and -DQ, but differed for both HLA-DP alleles (donor HLA-DPB1*0402/0501; patient HLA-DPB*020102/0301). The patient received a T cell depleted URD-SCT after a non-myeloablative conditioning regimen, resulting in mixed chimerism (75% donor) without GVHD. Because of a hematological relapse, a single DLI was given 6 months after SCT, resulting in a profound anti-leukemic effect with only grade I GVHD, treated with topical corticosteroids. 6 weeks after DLI, malignant cells in peripheral blood (PB) had dropped from 72% to 47%. 7 weeks later, only 3% malignant cells were present, and after 4 months, complete remission and conversion to full donor chimerism in the absence of GVHD was observed. To determine whether allo-HLA-DP specific T cells were involved in the immune response, leukemia-reactive donor T cell clones were isolated from PB or bone marrow at different time points during the response to DLI. Patient derived T cells were overnight stimulated with irradiated leukemic cells harvested before transplantation, and clonal IFNγ producing T cells were sorted and expanded. 21 CD4+ T cell clones, 19 CD8+ T cell clones and 6 NK cell clones were tested for recognition of patient or donor derived cells as measured by IFNγ production and cytotoxic activity. The CD8+ or NK clones did not recognize patient leukemic cells. However, all 21 CD4+ clones produced INFγ in response to patient leukemic cells but not to donor cells. To determine whether these CD4+ T cell clones were capable of killing the leukemic cells, a CFSE based cytotoxicity assay was performed. 8 clones showed 30–90% lysis of the leukemic cell population. To further analyze the specificity of these CD4+ clones, blocking and panel studies were performed. Blocking with the HLA-DP specific mAb B7.21 abrogated IFNγ production by all clones, confirming HLA-DP restricted recognition. A panel study using 12 unrelated EBV-LCL expressing different HLA-DP alleles identified 18 clones specific for HLA-DPB1*0301, and 3 clones specific for HLA-DPB1*0201. To analyze the polyclonality of the immune response, the distribution of TCR Vβ chains was characterized by RT-PCR and sequence reactions. 7 different Vβs were found within the HLA-DPB1*0301 specific clones and 3 different Vβs within the HLA-DPB1*0201 specific clones. T cells using the same Vβ could be isolated at different time points during the clinical response, demonstrating the significance of this anti-HLA-DP response. In conclusion, we observed in a patient with an HLA-class II positive B cell malignancy a profound GVL effect without GVHD, caused by a polyclonal immune response comprising both T helper and cytotoxic CD4+ HLA-DP specific T cell clones directed against both HLA-DP alleles. These data indicate that in HLA-class II expressing hematological malignancies HLA-DP mismatched SCT may be preferable over a fully matched SCT making use of HLA-DP as a specific target for immunotherapy.