Allogeneic hematopoietic cell transplantation (allo-HCT) has the potential to cure patients with Acute Myeloid Leukemia (AML) [1]. More than 90% of patients with newly diagnosed AML fall into an intermediate or poor risk category per European Leukemia Network (ELN) criteria [2], and in this patient population, allo-HCT in first complete remission (CR1) improves overall survival (OS) and leukemia-free survival (LFS) compared to non-transplant approaches [3]. Treatment-related mortality (TRM) was previously the primary cause of treatment failure following allo-HCT, however, with improvements in supportive care along with the increased use of reduced-intensity conditioning (RIC) regimens, relapse has supplanted TRM as the primary cause of treatment failure and occurs after more than 30% of allogeneic transplants [4,5]. The prognosis for patients relapsing after allo-HCT remains poor. The majority of patients are unable to achieve a subsequent remission, and 2-year-OS is consistently <15% [[6], [7], [8], [9]].

Understanding factors that lead to relapse and impact survival following transplant can help influence the choice of subsequent therapies, including intensive induction therapy, low-intensity therapy, targeted therapy, novel immunotherapy agents, and additional cellular therapy [10]. Cellular therapy is of particular interest as the efficacy of allo-HCT in AML primarily revolves around the graft-versus-leukemia (GVL) effect and is mediated by T lymphocytes, often correlating with graft-versus-host disease (GVHD). Horowitz et al. demonstrated the antileukemia effect of GHVD [11], suggesting that the post-transplant alloimmune effects which prevent relapse are also responsible for their side effects. This is supported by evidence that patients who relapse in the setting of GVHD have poorer survival, likely due to the failure to attain an adequate GVL effect [5,9]. In contrast, particularly in patients with no evidence of GVHD, the GVL effect can be enhanced with donor lymphocyte infusions (DLI). DLI alone can lead to a remission in up to 29% of AML patients who relapse post-transplant [12], and remissions can be durable if DLI is administered after achieving remission with chemotherapy [13].

AML relapse post-transplant appears to occur by two primary mechanisms, downregulation of HLA class II on leukemic blasts and T-Cell exhaustion. Downregulation of HLA class II molecules occurs in approximately 50% of patients relapsing after transplant, and this is not present in AML patients relapsing prior to transplant [14]. MHC genes are involved in antigen presentation and stimulation of antitumor response. Therefore, downregulation of MHC class II genes serve as a mechanism by which leukemic blasts evade the GVL effect [14]. The second mechanism is T-Cell exhaustion. Exhaustion of donor bone marrow memory T cells correlates with relapse. These memory T cells express an increased number of inhibitory receptors, including PD-1, TIM-3 CTLA-4, and 2B4. This results in a loss of cytokine production, proliferative capacity, and lytic activity, all of which lead to a decrease in donor-mediated immunity [15]. Understanding the immunologic mechanisms responsible for post-transplant relapses can have implications for determining the most effective therapeutic options.

DLI effect is primarily mediated by the immunologic antitumor activity of donor T cells, including antigen-presenting cells, CD4+ T cells, CD8+ T cells, regulatory T cells (T regs), and natural killer (NK) cells [11,16,17]. The GVL effect is mediated by the reversal of T cell exhaustion in CD8+ T cells [18]. DLI can also reconstitute the T cell receptor repertoire and normalize the clonal expansion of CD8+ T cells, further enhancing the GVL response [19].

DLI can be used for HLA-matched or mismatched allo-HCTs, from matched related, unrelated, or haploidentical related donors [8]. With the introduction of post-transplant cyclophosphamide (PTCy), T-cell replete haploidentical stem cell transplantation has rapidly expanded [20]. Due to a greater HLA disparity than in HLA-matched donors, haploidentical DLI may exert a more substantial GVL effect [21]. DLI can be used in the treatment or in the prevention of AML relapse. DLI alone has not been shown to control overt relapse, particularly with a high tumor burden. Achieving remission through salvage therapy before DLI has consistently demonstrated superiority to DLI or chemotherapy alone in inducing remission post-allo-HCT for AML relapse [[22], [23], [24]]. Pre-emptive and prophylactic DLIs have also been evaluated in post-transplant settings. Administering DLI before overt hematological relapse may improve outcomes and limit toxicity compared to therapeutic DLI [8], however further exploration of its precise role and effectiveness in this setting is needed. Unfortunately, complete responses and durable remissions related to DLI remain unsatisfying, with 2-yr overall survivals ranging from 15 to 25% [13,25], leaving much to be desired and the need for further work to improve DLI therapy in patients with relapsed AML post-allo-HCT.

In this review, data for DLI use in AML relapse post-transplant will be presented along with adverse effects, pre-DLI therapeutic options, practical concepts, and newer strategies using cellular engineering.