A bispecific approach to improving CAR T cells in AML

In this issue of Blood, He et al describe a new technique, so-called STAR (sequential target antigen and antibody retrieval), to rapidly identify tumor-specific nanobodies (heavy chain–only antibodies with a small-size single-variable domain) to bind to various epitopes, and their corresponding antigens.¹  This approach allowed for the targeting of previously unappreciated cell surface antigens, thereby potentially expanding the ability to generate functional chimeric antigen receptor (CAR) T-cell development in acute myeloid leukemia (AML) and other cancers.

In recent years, a number of immunotherapy approaches have been investigated in patients with AML.²  Redirecting the patient’s own immune system to target cancer cells using bispecific T-cell engaging antibodies, CAR T or CAR NK cells, or T-cell or macrophage checkpoint inhibitors, is a highly attractive approach that aims to simultaneously enhance antitumor activity while reducing the burden of systemic toxicities seen with traditional cytotoxic chemotherapies.

Emerging data support T-cell immune suppression in the AML microenvironment, that progressively worsens (due to both T-cell depletion and T-cell dysfunction) with disease progression and advanced salvage status. Increased regulatory T-cell infiltration, coinhibitory checkpoint expression on effector T cells, myeloid-derived suppressor cells and macrophages in the tumor environment, reduced MHC expression, and lower mutational burdens all contribute to the immune dampening in AML. The complicated interplay between these immune dampening forces, the hitherto limited (albeit growing) knowledge of the immune system in AML, the lack of highly specific and tumor-restricted antigens or checkpoint pathways in AML, the success of BCL2 inhibition, targeted therapies, monoclonal antibodies, and hedgehog inhibitors, and the lack of stellar successes thus far with T-cell–directing therapies have resulted in a slower pace of clinical development of immune approaches in AML compared with solid tumors, lymphoma, and acute lymphoblastic leukemia (ALL). This may now be changing. A number of trials combining PD1 and/or CTLA4 antibodies are showing early promising results in frontline and relapsed AML/myelodysplastic syndrome (MDS), especially when combined with hypomethylating agents.^3,4  Furthermore, inhibition of macrophage checkpoint CD47 appears to be highly effective when CD47 antibody is combined with hypomethylating agents in frontline MDS and high-risk AML.⁵ 

CAR T cells are tumor-reactive T cells genetically engineered to express the binding site of specific antibodies and are capable of targeted tumor killing.⁶  Anti-CD19 CAR T cells have already shown remarkable success in the treatment of B-cell malignancies.⁷  Clinical trials with CAR T cells in patients with AML are in a much earlier phase of development. Ritchie et al reported the safety, feasibility, and persistence of CAR T cells for up to 10 months postinfusion in 5 first salvage AML patients.⁸  Two patients achieved stable disease (duration 23 months in 1 patient), and an additional 2 patients had transient response (blasts reduction/cytogenetic remission). No severe adverse events (AEs) or cytokine release syndrome was observed. Budde et al reported 6 patients (5 AML and 1 blastic plasmacytoid dendritic cell neoplasm) who had relapsed/refractory disease following allogeneic stem cell transplant (ASCT) with a median of 4 prior lines of therapy.⁹  The patients received 1 to 2 doses of CD123 CAR T-cell construct, and 2 of them achieved complete remission (CR) with successful bridge to second ASCT. Two additional patients achieved blasts reduction not classified as CR. All toxicities were reversible and manageable with only 1 grade 3 AE (rash), and no treatment-limiting AE. This trial is ongoing (NCT02159495), and updated data are awaited. Key issues impeding the development of CART cell therapy in AML, compared with ALL and lymphoma, have been the difficulty in identifying target antigens, which are relatively specific to AML bulk and stem cells, while sparing normal monocyte/macrophage populations and healthy human stem cells, and the availability of monoclonal antibodies suitable for developing CAR T-cell therapy and able to endow T cells with cytotoxicity.

He et al identified multiple nanobodies to AML using a nanobody phage library and then selected the nanobodies able to activate and redirect CAR T cells to the leukemia microenvironment and induce leukemia cell death. CD13 targeting Nb 157 induced potent T-cell proliferation, T helper 1 cytokine production, degranulation, and eradicated AML from the bone marrow, blood, and spleen in vitro (THP-1 and HL60 lines), in NSG THP-1 mice, and in patient-derived xenograft models. CD13 is overexpressed in AML cells and leukemia stem cells (LSCs), but is also expressed on human hematopoietic stem cells (HSCs), myeloid progenitors, and monocytes as well as non–leukemia cells, including colon and kidney epithelium. CD13 CAR T cells, although demonstrating potent anti-AML efficacy, almost completely eliminated HSCs, myeloid progenitors, and peripheral monocytes.

To develop a more tolerable and safer CAR T cell, the authors then developed a bispecific and split CAR (BissCAR) targeting CD13 and TIM3. TIM3 is a more specific marker with expression predominantly on exhausted T cells and LSCs, and limited expression on normal HSCs and myeloid progenitors. The BissCAR demonstrated elimination of leukemia as potently as seen with the anti-CD13 alone CAR T cell but with limited and reversible toxicity to normal HSCs, myeloid progenitors, circulating monocytes, and healthy organ systems. This suggests that dual targeting of optimally selected antibodies and tumor-associated antigens may be a new approach to develop a safer and more tolerable CAR therapy with maintained efficacy and is especially important in AML wherein preservation of normal HSCs and myeloid progenitors is likely to be critical to successful clinical CAR therapy development. This approach warrants clinical evaluation.