Two to tango: engineered T cells against AML

In this issue of Blood, Dao et al¹ present a new dual-antigen targeted engineered T-cell platform for safe and efficient T-cell therapy of acute myeloid leukemia (AML).

Development and clinical implementation of targeted T-cell therapy for AML has so far been impeded by the phenotypic similarities between malignant and normal myelopoiesis, the paucity of tumor-specific antigens, the complex bone marrow microenvironment, and AML disease heterogeneity. Now, Dao et al redirected T cells against Wilms tumor 1 (WT1) with a novel antibody T-cell receptor (AbTCR) construct and exploited CD33 as input to activate a chimeric costimulatory signaling receptor (CSR) to enhance specificity, safety and efficacy (see figure). From their phage-display library, the authors identified new antibodies (ESK2) recognizing the WT1 RMF (RMFPNAPYL) peptide/HLA-A∗02:01 complex with enhanced specificity compared with their previously identified ESK1 antibodies. The fragment antigen binding regions of 2 lead candidates were linked to the constant chains of the γδ T-cell receptor (TCR) and expressed in a viral vector along with a CSR recognizing CD33 through a single-chain variable fragment (scFv) linked to CD28. When human T cells transduced with AbTCR + CSR encountered AML cells expressing WT1 (RMF)/HLA-A∗02:01 and CD33, the T cells recognized both target antigens and received T-cell activation signals from the γδ TCR/CD3 complex (signal 1, by AbTCR) and CD28 costimulation (signal 2, by CSR) (see figure). Investigation of engineered T-cell function showed that AbTCR⁺ T cells specifically recognized and killed WT1 (RMF)/HLA-A∗02:01⁺ AML targets and that both killing and interferon-γ production were increased on delivery of CD28 costimulation through the CSR (see figure). The CSR acted in cis, requiring both the WT1 (RMF)/HLA-A∗02:01 complex and CD33 expressed on the same target cell. More important, no toxicity against normal peripheral blood mononuclear cells, granulocytes, or cord blood CD34⁺ cells was detected in vitro. In vivo in mouse xenograft models, strongest antileukemic activity was observed when mice were treated with T cells expressing both AbTCR + CSR. Thus, the proposed AbTCR + CSR combination resulted in increased AML specificity and enhanced antileukemic activity while sparing normal hematopoiesis and reducing on-target off-tumor activity of the engineered T cells (see figure).

Being highly overexpressed in hematologic and solid tumors, with restricted low expression in some adult normal tissues (reproductive organs, kidney podocytes, mesothelial lining, and CD34⁺ cells), and exerting critical biological functions in tumors, WT1 is an intensively investigated highly prioritized oncofetal tumor-associated antigen. WT1 has several immunogenic epitopes presented in the context of human leukocyte antigen (HLA) class I. The WT1 (RMF)/HLA-A∗02:01 epitope targeted in the study by Dao et al is historically the most investigated WT1 epitope in vaccine or TCR-based adoptive cell transfer studies. Safety and therapeutic benefit have been reported in 12 patients with AML infused prophylactically with WT1 TCR-transgenic T cells after allogeneic hematopoietic stem cell transplant.² However, the RMF epitope requires the immunoproteasome for natural processing,³ introducing some uncertainties about cell surface presentation in malignant cells. Some recent analyses are pointing toward a potential immune escape mechanism after WT1 TCR–T-cell therapy, where reduced levels of specific immunoproteasome subunits were reported in 2 resistant patients.⁴ Immunoproteasome-independent WT1 epitopes and corresponding class I restricted αβ TCRs have also been identified, opening new interesting therapeutic prospects for targeting WT1.^4-6 Eventually, an AbTCR against a WT1 epitope processed by standard proteasomes could be developed.

The concept of splitting the T-cell activation and costimulation signals to enhance target specificity and engineered T-cell function has been explored in the field for over a decade,⁷ but has recently gained momentum in various combinations with conventional chimeric antigen receptors (CARs) or native or transgenic αβ TCRs.^8,9 The possibility of simultaneously targeting both intracellular and cell surface antigens in a cooperative manner, with αβ TCRs or with AbTCRs, significantly broadens the number of potentially targetable antigens and allows the exploitation of their different signaling sensitivity thresholds. A novel dual-antigen targeting strategy for AML was recently presented, targeting 2 cell surface receptors with scFvs, combining a CAR with a chimeric costimulatory receptor.¹⁰ Target antigens were chosen on the basis of differential antigen density between normal and malignant hematopoiesis identified on a set of samples from patients with AML and healthy donors.¹⁰ This T-cell engineering strategy allowed tuning of engineered T-cell sensitivity by targeting adhesion G protein-coupled receptor E2 (ADGRE2) and C-type lectin domain family 12 member A (CLEC12A), and improved antileukemic activity with reduced toxicity to normal hematopoiesis, compared with single-antigen targeting. The study by Dao et al combined WT1 with CD33 targeting, WT1 by itself already having an excellent safety and specificity profile. When combined with a CSR exploiting the highly expressed pan-myeloid antigen CD33 for the delivery of costimulation, enhanced leukemia specificity and antileukemic activity were demonstrated. Although CD33 antigen densities were characterized recently,¹⁰ we have no information available regarding the densities of WT1 (RMF)/HLA-A∗02:01 on AML and normal hematopoietic stem and progenitors. It would be interesting to evaluate whether the combination of AbTCR + CSR shifts the activation threshold of the AbTCR toward recognition of lower-density target cells.

In conclusion, Dao et al present a compelling preclinical proof of concept for a novel approach of dual-antigen targeting for AML with engineered T cells conferring enhanced specificity, safety, and antileukemic activity. Clinical translation is eagerly awaited.

Conflict-of-interest disclosure: C.A. receives licensing fees and royalties from Immatics (through previous institution, Baylor College of Medicine); participated in advisory boards for Kite/Gilead, Janssen, and Celgene/Bristol Myers Squibb; received sponsored travel from Gilead; and has several patents and pending patent applications in the field of engineered T-cell therapies.