Potential of CAR T-cell therapy in acute myeloid leukemia

The success of anti-CD19 CAR T-cell therapy in relapsed pediatric B-lineage acute lymphoblastic leukemia (B-ALL) has so far not been translated into treatment for acute myeloid leukemia (AML). This might be due to a lack of unique antigens on the surface of AML cells that would allow selective targeting of these cells while sparing healthy hematopoietic stem/progenitor cells (HSPC). However, recent years have seen a lot of research activity in the field as researchers are aiming to find a way to introduce chimeric antigen receptor (CAR) T-cell therapy into the treatment of AML.

Katherine Cummins and Saar Gill from the University of Pennsylvania, US, wrote an article on CAR T-cell therapy for AML, which appeared in the June issue of Haematologica.¹ The article discusses CAR T-cell therapies in clinical development, novel target antigens, other immunotherapy options, and potential issues associated with these therapies.

CAR T-cell based clinical trials

A lot of effort is currently being put into assessing the efficacy and safety of CAR T-cells developed against a range of antigens in different subsets of patients with AML. The 13 clinical trials currently underway are summarised in table 1.

Table 1. Chimeric antigen receptor (CAR) T-cell therapies in clinical development.

Abundantly present on AML blasts, CD33 and CD123, are one of the promising targets for CAR T-cell therapy. In particular, CD123 has been identified as a marker of leukemia-initiating cells² also expressed by other hematologic malignancies.³ However, both antigens are also present on normal HSPC, raising safety questions regarding potential myeloablation.⁴ Therefore, patients on these therapies will most likely require a rescue allogeneic stem cell transplant (allo-HSCT) donor. Additional toxicity is associated with CD123 expression on endothelial vessels⁵ that has resulted in a fatal cytokine release syndrome (CRS) and capillary leak syndrome (CLS).⁶ One approach to minimize those risks of vascular toxicity is the use of CAR-T cells with a limited capacity to expand in vivo⁷.

Novel approaches

Other strategies to use adaptive immune responses in AML include targeting natural killer group 2D (NKG2D) ligands, using dual-specific CART, conducting bone marrow transplant with gene-edited allograft followed by CAR T-cells as well as trying other immunotherapy options.

NKG2D-CAR T-cells

NKG2D ligands, evaluated as antigens for CAR T-cells, are predominantly expressed on cancer cells.⁸ However, cellular stress, including CART induced CRS can upregulate their expression also in normal cells⁹. Initial results of phase I clinical trial of autologous NKG2D-CAR T-cells in seven patients with AML showed no objective clinical responses and a limited life-span of the CART population without dose-limiting toxicities.¹⁰ Encouragingly, in the follow-up study using higher doses one of the two patients achieved an objective clinical response.¹¹ Additionally, a combination of anti-NKG2D-CAR T-cells with azacytidine, which enhances the expression of NKG2D ligands on AML blasts, is in development to assess whether this could improve efficacy.¹²

Dual specificity CARs and new target antigens

Researchers are also employing CAR T-cells expressing a dual-specific CAR directed against two different co-expressed target antigens, to enhance the killing of malignant cells. So far, ADGRE2, CCR1, CD70 and LILRB2, CLL1 have been identified as potentially useful targets in the dual CART approach.^13,14 Moreover, there is a potential for the CLL1 to be used as a stand-alone antigen.

Bone marrow transplant with a gene-edited allograft followed by CAR T-cells

In order to increase the persistence and to protect normal hematopoiesis, an antigen can be edited out from the allo-HSPCs before transplantation. In one study, the CD33 was knocked-out from HSPC and their progeny using CRISPR/Cas9.¹⁵ Although this novel approach minimizes hematopoietic toxicity, it also reduces the in vivo persistence of CAR T-cells. However, the feasibility and clinical implications of this approach are not yet clear, and there are some concerns about the severity of CRS response due to the role of CD33 in immunomodulation.¹⁶

Other immunotherapies for AML

Lack of antigen specificity in AML has led to exploring immunotherapy options beyond CAR T-cell therapy. These include engineered T cell receptor (TCR) cells against tumor-associated antigens (TAA) and neoantigens. TCRs have potentially increased anti-tumor specificity compared to CAR T-cells as they recognize intracellular antigens presented on MHC of malignant cells. The TCR chains are cloned from patients or normal donors with an immune response to the TAA and may be further affinity-enhanced to amplify reactivity to the target. The use of Wilms’ tumor 1 (WT1), as a TAA, has demonstrated safety and some efficacy as a single TCR therapy^17,18 and is assessed in combination with IL-2 (NCT02550535).

Despite encouraging results in solid tumors using PD1/PD-L1 immuno-checkpoint inhibitor, their use in patients with AML has been disappointing so far.¹⁹ Also, an attempt to use vaccination against leukemia-specific peptides has  failed to show a clinical impact²⁰ (NCT00433745). In contrast, the expansion of T-cells targeting ex vivo generated autologous dendritic cells and AML fusion cells has resulted in promising results. This approach has been assessed as consolidation therapy after allo-HSCT, alone and in combination with decitabine (NCT03679650). In a recently reported clinical trial in a small group of patients after standard chemotherapy and a median follow-up of 57 months, it resulted in remission in 71% of patients and a good safety profile.²¹

Potential issues with CAR T-cell therapies

As some patients with AML achieved long-term remission after donor lymphocyte infusions and allo-HSCT, CAR T-cell therapy was anticipated to have a similar positive effect on disease control. Unfortunately, various mechanisms of immuno-evasion have been described, including loss of HLA,²² downregulation of antigen expression, upregulation of proteins protecting from apoptosis, and alterations in the composition of T-cell populations such as an expansion of regulatory T-cells and exhaustion of T-cells.²³ It is therefore likely that AML will also generate resistance to CART. However, even if the satisfactory efficacy is achieved, the therapy must also have an acceptable toxicity, which will be assessed against other new targeted therapies such as those against BCL2, FLT3, and IDH1/2. Moreover, the manufacturing of T-cells possesses its own challenges due to the cost, complexity, and length of the process, with one of the studies reporting the median time of 45 days from enrolment to infusion.²⁴

Conclusions

In the future, CAR T-cell therapy for AML might be a possible way to cure patients. It is unlikely the therapy will benefit all patients and might be considered unsuitable for older and more fragile patients. But hopefully, in the next few years, we will be able to identify subgroups of AML patients that would be most likely to respond to the therapy. The authors are hoping that the combination of established allo-HSCT and the CAR-T cells will improve the outcome of the R/R patients with AML, and the CART field will continue to develop into safe, feasible and effective therapies.

Once the ongoing trials start reporting their findings, we should have a better understanding of the role of CAR T-cell therapies can play in the treatment of AML patients, and how they compare to the recent developments in the therapeutic area.