A Novel Target Antigen for the Treatment of Acute Myeloid Leukemia by CAR T Cells

Main Text

There is strong evidence that the presence of immune infiltrate in tumors is correlated with a good prognosis following therapy.1, 2 This has led to the development of T cell-based immunotherapies for cancer. One such therapy has involved isolating and expanding tumor reactive lymphocytes from the tumor, called tumor-infiltrating lymphocyte (TIL) therapy. This form of adoptive immunotherapy utilizing TILs has shown significant success in patients with advanced melanoma, with greater than 50% of patients responding to treatment.³ However, the broad use of TIL therapy has been precluded by the inability to isolate sufficient number of tumor reactive T cells from most cancers. To overcome this problem, the genetic modification of T cells with αβ T cell receptor (TCR) transgenes⁴ or recombinant receptors, called chimeric antigen receptors (CARs),⁵ have been developed that arm the T cell with anti-tumor activity regardless of major histocompatibility I (MHCI) expression. The promise of this approach has been recently highlighted in certain B cell malignancies, such as acute lymphoid leukemia (ALL) and chronic lymphoid leukemia utilizing CAR T cells specific for CD19, and has resulted in objective responses in up to 90% of ALL patients.⁶ These striking results have raised the prospect of utilizing CAR T cells targeting novel antigens for other malignancies, including acute myeloid leukemia (AML). Previous CAR T cell therapy trials in AML did not result in long term responses, and CAR T cells against AML antigens, such as CD123 and CD33, exhibited unwanted toxicity toward normal hematopoietic progenitor cells (HPCs). Therefore, there is interest in generating CAR T cells with higher specificity toward an antigen expressed exclusively on AML cells that spare normal myeloid progenitor cells.

In this issue of Molecular Therapy, Tashiro et al.⁷ investigate the effectiveness of CAR T cells targeting the C-type lectin-like molecule 1 (CLL-1). This membrane glycoprotein serves as a potential target antigen given its high expression on a majority of AML cells (expressed in 18/19 patients tested) and, importantly, its absence from normal hematopoietic stem cells. CLL-1 has previously been validated as a target in AML through the use of monoclonal antibodies.⁸ In the new study, the authors first demonstrate antigen-specific function of T cells that had been gene-modified with a second-generation CAR receptor (anti-CLL-1BBζ). The data included cytotoxicity, proliferation, and cytokine release against both CLL-1⁺ AML lines and primary AML patient samples. Furthermore, anti-leukemic activity of adoptively transferred anti-CLL-1 T cells against disseminated AML in a xenograft mouse model was shown. An important finding was that anti-CLL-1BBζ T cells were found to be cytotoxic against mature monocytes, but not to normal myeloid progenitor cells, in vitro. This feature should allow full myeloid recovery following elective ablation of anti-CLL-1 T cells, thereby protecting the patient from any dangerous and long term toxic side effects.

There have been several reports of tumor relapse in ALL patients following CAR 19 T cell therapy as a result of the emergence of CD19⁻ variant clones.⁹ One potential approach to circumvent this problem for the treatment of AML and other cancers may be to redirect T cells to multiple antigens. For example, genetic modification of T cells with CARs recognizing both CLL-1 and Lewis Y for AML may prove to be more effective long term and protect against potential editing of tumor antigen and the emergence of antigen-negative clones. Given that Tashiro et al.⁷ found CLL-1 expression to be fairly heterogenous (33%–99% expression), targeting multiple tumor antigens would also likely enhance efficacy in patients with variable expression of CLL-1.

Another issue for effective CAR T cell therapy is the immunosuppressive tumor microenvironment. Tumors can evade immune recognition by activating various checkpoint pathways, including PD-1 and CTLA-4. Tumors often show increased expression of ligands for these receptors, which results in a significant reduction of effector cell function. In the clinic, immunotherapy mediated by antibodies such as pembrolizumab, which blocks interaction of PD-1 with PD-L1 or PD-L2, or ipilimumab, which blocks the interaction between CTLA-4 and CD80/CD86, has resulted in striking effects in cancers such as melanoma, lung and renal cancers, and, in some cases, is durable. There have been several reports in preclinical mouse models demonstrating that checkpoint blockade using anti-PD-1 can significantly augment CAR T cell responses.10, 11, 12 Based on these reports, several clinical trials are currently underway examining adoptively transferred T cells in combination with checkpoint blockade (clinicaltrials.gov: NCT02652455). A recent case study report described a patient with diffuse large B cell lymphoma who had progressive disease following CAR 19 T cell therapy but responded after administration of pembrolizmumab. This result correlated with expansion of CAR 19 T cells and an increase in transgene copy number.¹³ This observation clearly warrants testing of CAR T cell combination therapy with checkpoint blockade in other blood cancers, such as AML, and also in solid tumors. The relative effect of targeting other immunosuppressive pathways could further improve CAR T cell activity. A recent study reported that targeting the CD73/adenosine pathway using antagonist drugs blocking the A_2A adenosine receptor could augment CAR T cell activity against several syngeneic Her-2⁺ tumors in mice.¹¹ Thus, combination approaches that target the immunosuppressive effects of adenosine and other checkpoint pathways may result in a greater proportion of patients responding to CAR T cell adoptive immunotherapy.

A serious concern with immunotherapies such as CAR T cells is the potential for on-target, off-tumor activity leading to fatal toxicity. This has been reported in several trials including patients with CAR19 T cells¹⁴ (CAR19 trial; clinicaltrials.gov: NCT02535364). Thus, targeting an antigen on tumor cells that spare normal hematopoietic cells would be most favorable. In the current study by Tashiro et al.,⁷ targeting the CLL-1 antigen appears to be ideal, given it is expressed predominately on AML cells, and, although it is present on mature monocytes, it is not expressed on immature myeloid precursor cells. This should allow reconstitution of the myeloid population using methods that enable elective ablation of the CAR T cells. One potential approach, as shown in the current study, utilizes CAR T cells expressing an inducible caspase-9 suicide gene that can be effectively eliminated by administration of the activating dimerizer drug (chemical inducer of dimerization [CID]). Importantly, this approach has been validated in the clinic.¹⁵ The use of CAR T cells engineered with a suicide gene therefore provides an inbuilt safety mechanism for patients. It will be important for future studies to investigate the toxicity associated with anti-CLL-1 CAR T cells in an immunocompetent setting, potentially in a self- antigen model, and whether deletion of CLL-1-specific CAR T cells using this system increases the risk of relapse.

In summary, the era of immunotherapy has entered an exciting new phase with recent results in patients using checkpoint inhibitors for various cancers or CAR T cells directed against CD19 for B cell malignancies. In the future, targeting of T cells using CARs directed against other antigens, such as CLL-1, that are predominately expressed on tumor cells and not on normal hematopoietic stem cells may significantly broaden both the effectiveness and safety of adoptive T cell immunotherapy for many cancers, including AML.