Selective H2-O Tissue Expression Reduces Risk for Graft-versus-Host Disease in an In Vivo Transplantation Model

Allogeneic stem cell transplantation (aSCT) is frequently used to treat patients with hematologic malignancies. The therapeutic effect relies mainly on the graft-versus-leukemia (GVL) effect, in which donor T cells eliminate residual malignant cells. Unfortunately, the beneficial GVL effect is often accompanied by detrimental graft-versus-host disease (GVHD). A successful separation of both effects could not yet be achieved. In previous work, we identified 2 groups of HLA-class II restricted antigens depending on their behavior towards HLA-DM. DM-resistant antigens are presented in the presence of HLA-DM, whereas presentation of DM-sensitive antigens relies on the inhibitory molecule HLA-DO. Due to the unique expression pattern of HLA-DO, DM-sensitive antigens cannot be presented efficiently by non-hematopoietic cells even under inflammatory conditions. This suggests that CD4⁺ T cells directed against DM-sensitive antigens may be able to separate GVL from GVHD.

In this study, we wanted to demonstrate convincingly that HLA-DO expression strongly influences the severity of GVHD in allogeneic stem cell transplantation.

We generated a modified CD4 donor lymphocyte infusion (DLI) depleted of CD4⁺ T cells directed against DM-resistant antigens to address its potential to induce GVHD in an in vivo major histocompatibility complex (MHC) mismatch transplantation model dependent on selective tissue expression of H2-O using H2-O wildtype, knockout, and transgenic recipients.

Intriguingly, we could demonstrate that our modified CD4 DLI targeting DM-sensitive antigens induced only mild GVHD in wildtype recipients with endogenous selective H2-O expression and none in H2-O knockouts while assessing the immunogenic potential of DM-sensitive antigens in H2-O transgenic recipients.

The results of the present work provide evidence that DM-resistant antigens are main targets of GVHD and addressing DM-sensitive antigens might be a promising tool to improve outcome after aSCT by separating GVL from GVHD.