Novel CAR T cells to combat antigen escape in AML

In this issue of Blood, Silva et al¹ report on an innovative, dual-targeted, chimeric antigen receptor T-cell (CAR-T) therapy that addresses the significant challenge of antigen heterogeneity and escape in acute myeloid leukemia (AML). This dual-targeting approach uses a bispecific T-cell engager molecule (TEAM) to activate both CAR T cells and bystander T cells in mouse models.

AML, the most prevalent leukemia in adults, is marked by its aggressive course and high relapse rates. Despite recent therapeutic advancements, outcomes remain dismal, particularly in relapsed or refractory cases. Although CAR-T therapy has transformed the treatment of B-cell malignancies, its application in AML is hindered by on-target off-tumor toxicity and the high propensity for antigen escape, due to the antigenic heterogeneity of AML.^2,3 

Silva et al sought to address these challenging conditions through a novel dual-targeting CAR-T strategy that engages 2 key antigens, CD70 and CD33, to enhance the breadth and persistence of immune response against AML. The study’s core innovation lies in the integration of TEAM, a bispecific antibody secreted by CD70-targeted CAR T cells.⁴ CD70, a molecule associated with T-cell memory formation, is overexpressed in AML cells. By directing CAR T cells to target CD70, the therapy attacks a critical component of AML immune evasion mechanisms. What sets this approach apart is the simultaneous targeting of CD33, a common myeloid marker, via the TEAM bispecific molecule. This molecule not only engages CAR T cells, but also redirects unmodified bystander T cells to attack CD33-expressing AML cells. This bispecific mechanism offers a unique advantage: even if AML cells downregulate CD70, which could otherwise result in immune evasion, the therapy maintains pressure on the AML cells by engaging CD33-targeted bystander T cells. By recruiting exhausted or unmodified T cells, this strategy amplifies the immune response and compensates for the antigen escape that typically hampers single-antigen CAR-T therapies in AML.

Through a series of eloquently designed in vitro and in vivo experiments, the authors provide strong evidence that their dual-targeting strategy is more effective than single-target CAR-T therapies. In vitro studies demonstrated superior cytotoxicity of the dual-targeted CAR T cells, referred to as “7033 CAR T” cells, against AML cells, even when CD70 expression was downregulated. In mouse models, the 7033 CAR T cells showed improved tumor control, persistence, and expansion compared with single-antigen CAR T cells. Furthermore, transcriptional analyses revealed that the dual-targeted CAR-T cells had upregulated natural killer–like features (NCR1), which may contribute to their enhanced longevity and functionality in the immunosuppressive AML microenvironment.

Although the dual-targeting approach represents a substantial advance, it also raises concerns about immune-related toxicities, particularly given the expression of CD33 on healthy myeloid cells. CD33-targeted therapies, such as gemtuzumab ozogamicin, have previously been associated with serious off-target toxicities, including sinusoidal obstruction syndrome and liver damage.⁵ Anti-CD33 CAR T cells have shown similar toxicities in early trials, raising concerns about their safety in AML treatment.⁶ Silva et al sought to mitigate this risk by localizing the secretion of the bispecific TEAM molecule within the leukemia microenvironment, reducing systemic exposure. However, although this localized secretion could potentially limit off-target effects, the risk of immune-related toxicities remains a major consideration for clinical translation.

Moreover, other CAR-T therapies have commonly resulted in cytokine release syndrome and immune effector cell–associated neurotoxicity syndrome.^7,8 These complications arise due to the robust immune activation that CAR T cells induce, especially in heavily pretreated patients with AML, who have compromised immune systems. Silva et al report increased persistence and cytotoxic potential of their dual-targeted CAR T cells, including elevated transcriptional signatures associated with T-cell proliferation (eg, JAK2 and MYC). Although this enhanced persistence may translate to more durable remissions, it also increases the risk of immune overactivation, which could exacerbate cytokine release syndrome or immune effector cell–associated neurotoxicity syndrome. Thus, early-phase clinical trials will need to focus on optimizing dosing and closely monitoring for these immune toxicities. Despite these promising results, several additional challenges must be addressed before this approach can be widely adopted in the clinic. The complexity of manufacturing dual-targeted CAR T cells that also secrete bispecific antibodies presents significant logistical hurdles. Ensuring the consistency, quality, and scalability of this therapy for broader clinical use is nontrivial and will require substantial refinement of current CAR T–cell production processes. Moreover, further studies will be necessary to determine whether the enhanced persistence of these dual-targeted CAR T cells translates into long-term tumor control without inducing severe toxicities, such as autoimmunity or graft-versus-host disease.

Another key concern is the potential for AML cells to develop resistance to dual-targeting CAR-T therapy. Although the simultaneous targeting of CD70 and CD33 reduces the likelihood of antigen escape, there remains the possibility that AML cells could downregulate both antigens and/or develop alternative escape mechanisms.³ To prevent or delay relapse, combination therapies should be explored. For example, combining this CAR-T approach with hypomethylating agents could help sustain antigen expression, whereas immune checkpoint inhibitors could be used to prevent T-cell exhaustion and prolong the efficacy of the therapy.

Silva et al’s dual-targeting CAR-T strategy has the potential to become a significant advancement in the treatment of AML, addressing the challenge of antigen escape and heterogeneity. By leveraging the cytotoxic potential of both CAR T cells and bystander T cells through the bispecific TEAM molecule, this approach has the potential to improve outcomes for patients with relapsed and refractory AML.

Conflict-of-interest disclosure: N.E.K. declares no competing financial interests.