IFN-γ and CD38 in AML: a T-cell engagement made in heaven?

In this issue of Blood, Murtadha et al demonstrate the efficacy of a new CD38-targeting immunotherapy for acute myeloid leukemia (AML).¹ 

Survival rates in newly diagnosed AML have plateaued over the last several decades. Survival in the setting of relapsed disease remains dismal, making AML a disease that would benefit from novel therapies. Numerous immunotherapies have been trialed in patients with AML but with only limited success,^2,3 due to a variety of barriers, including on-target/off-tumor toxicity, an immunosuppressive microenvironment, and antigen heterogeneity.^4,5 AML is both genotypically and phenotypically heterogeneous compared with other hematologic malignancies, making it challenging to find a universal target. Furthermore, the condition has significant intrapatient heterogeneity, including multiple subpopulations existing within a single patient, which limits the potential of targeting a single antigen.

Experience with the monoclonal CD38-targeting antibody daratumumab has shown that the CD38 antigen has a favorable on-target/off-tumor toxicity profile. Targeting of CD38 has demonstrated encouraging clinical efficacy, such as approval for use in multiple myeloma and exploration in numerous other hematologic malignancies, including T-cell acute lymphoblastic leukemia.⁶ Although the majority of bulk AML cells express CD38, there is limited expression on leukemic stem cells (LSCs), bringing the viability of targeting CD38 into question, as LSCs are implicated in the propagation and relapse of AML.⁷ 

Modulation of CD38 has been established but has primarily been limited to preclinical modeling.⁸ In their study, Murtadha et al hypothesize that modulation of CD38 would make it a more compelling target in AML if its expression on LSCs could be successfully induced. Through a series of elegant experiments, they demonstrate that CD38 surface expression in AML blasts is increased in the presence of interferon gamma (IFN-γ), primarily through the upregulation of the transcriptional regulator interferon regulatory factor-1. Because T cells release IFN-γ with activation, there was sufficient mechanistic rationale to explore a novel CD38-directed T-cell engager in AML.

This led to the creation of the single-chain bispecific antibody CD38-CD3 BIONIC (biologics nested inside chains; BN-CD38) by inserting a CD38 nanobody between the light and heavy chains of an anti-CD3 antibody. This compact approach was favored over traditional double-chain approaches because it was hypothesized it would enhance function by mimicking the normal immunologic synapse by decreasing the distance between the T-cell receptor and the major histocompatibility complex. The authors show that coculture of BN-CD38 with T cells and AML cell lines led to activation of T cells (via surface expression of CD69 and CD25), secretion of proinflammatory cytokines (including IFN-γ), and clearance of CD38⁺ cells. These changes were not demonstrated with their mutated CD38 control and were largely abrogated in the presence of antihuman IFN-γ–neutralizing antibody.

Most importantly, the authors show the successful incorporation of BN-CD38 in multiple patient-derived xenograft models. This includes one model from a patient with relapsed AML with a CD34⁺/CD38⁻ fraction of >20% and treated with weekly BN-CD38 and human T-cell infusions. In this model, the majority of BN-CD38–treated mice remained disease-free compared with the control mice, which had a median survival of 12 weeks (P = .0007). Beyond inducing the qualitative expression of CD38 on LSCs, it was also noted that BN-CD38 increased CD38 expression on lower expressing bulk cells. This is potentially of equal clinical importance because level of antigen expression has been linked to response with antibody–drug conjugates in AML and other hematologic malignancies.^9,10 

An intriguing possibility is that the phenomenon of IFN-γ–induced upregulation of CD38 could be extrapolated to other cell-based therapies, such as chimeric antigen receptor (CAR) T cells, which have shown limited success in the AML space but are also associated with robust IFN-γ expression. Murtadha and colleagues provide further support for the use of CD38-targeting CAR T cells in AML while also raising the possibility of combinatorial strategies, such as adding daratumumab to CAR T cells targeting alternative antigens.

Interestingly, there was no modulation of CD38 on normal immune cells and, paradoxically, there was limited on-target/off-tumor toxicity based on the lack of clearance of normal CD38-expressing cells, including normal hematopoietic stem cells. The lack of targeting of normal hematopoietic cells suggests that the hematologic toxicities of BN-CD38 should be relatively nominal, potentially overcoming a common limitation of other immunotherapeutic AML approaches. However, although reassuring, this finding is equally perplexing as it is distinct from other immunotherapies for hematologic malignancies where on-target/off-tumor toxicity is sometimes used as a surrogate for functionality (eg, CD19-targeting blinatumomab and CAR T-cell therapies eliminate both malignant and normal CD19-expressing B cells). Thus, future studies should aim to confirm this finding in other humanized model systems and seek to identify possible underlying protective mechanisms.

With a rigorous series of in vitro, in vivo, and ex vivo experiments, Murtadha et al demonstrate the viability of targeting CD38 in AML with a T-cell engager, particularly due to the IFN-γ–induced CD38 upregulation and clearance of LSCs (see figure). Their work provides mechanistic clarity into immunotherapeutic treatment barriers and contributes to the growing body of evidence supporting the successful incorporation of immunotherapy for AML.

Conflict-of-interest disclosure: The authors declare no competing financial interests.