Use of CD70 Targeted Chimeric Antigen Receptor (CAR) T Cells for the Treatment of Acute Myeloid Leukemia (AML)

Background:

CAR-T cells have led to a revolution in the treatment of advanced hematologic malignancies. Since these cells target antigens that are expressed on the cellular surface, it is imperative that there is near ubiquitous tumor expression with minimal expression vital human tissues. Finding targets with these characteristics in myeloid malignancies has been challenging. Typical markers expressed on the surface of AML are also expressed on essential innate immune effector cells (e.g. neutrophils) which, if targeted, could lead to prolonged absence of this immune arm, which is not survivable or replaceable. Current approaches rely on the use of CAR-T cells against common myeloid targets (e.g. CD123, CD33) as an ablative strategy with a planned allogeneic stem cell transplant rescue to eradicate the CAR-T cells afterwards. These solutions have resulted in significant toxicity with several deaths resulting from CD123-targeted CAR-T cells. Another approach has involved gene editing donor progenitor cells to delete CD33, repopulation of the marrow with these CD33 negative cells, and then treatment with CD33-targeted CAR-T cells. (Kim, Cell 2018). However, this approach is challenging, costly, and genomic editing of stem cells remains a concern.

CD70 is an immune checkpoint found on antigen presenting cells and activated T cells. Multiple studies have shown a strong degree of expression on AML blasts and leukemic stem cells, with minimal normal tissue expression (Perna, Cell 2017, Riether J Exp Med 2017). A Phase 1 study of a CD70 targeted antibody drug conjugate in combination with azacitidine (which has been shown to increase CD70 expression on leukemic stem cells) for untreated AML patients has shown impressive results (Blood 2018 132:2680, Blood 2017 130:2652). Based on these findings, we explored CD70-targeting CARs for the treatment of AML.

Methods:

Based on our success with a trimeric ligand-based CAR of another TNFα family member, APRIL, for multiple myeloma (Schmidt Blood 2018 132:2059), we generated monomeric and trimeric second-generation ligand-based CAR constructs to target CD70 on AML. In vitro effector function was compared by cytotoxic potency and cytokine production. In vivo anti-tumor efficiency was assessed in a xenograft mouse model of AML. Effect of surface CD70 expression on AML cell lines after co-culture with azacitidine was assessed.

Results:

CAR T cell manufacturing of both constructs was accomplished successfully (transduction efficiency 70-93%) from three different healthy donors with no apparent fratricide. CD70 CARs were efficacious in in vitro cytotoxicity assays targeting an AML cell line Molm13. Unexpectedly, monomeric CD70 targeted CAR-T cells were superior to trimeric in cytotoxicity assays and, thus, were carried forward for in vivo assays. Next, we treated NSG mice that had been engrafted with Molm13 and demonstrated a substantial dose-dependent therapeutic effect with prolonged survival of CAR treated mice compared to those treated with untransduced T-cells (UTD). Treated mice demonstrated a CAR-T robust expansion in the peripheral blood assessed by flow cytometry that was commensurate with individual animal treatment responses. Bone marrow from these mice revealed substantially reduced CD70 in all groups. Preliminary in vitro co-culture of AML cells with azacitidine showed increased CD70 expression.

Conclusion:

CD70 based CAR-T targeting of AML is effective in vitro and in vivo. Combination treatment with azacitidine may increase target antigen expression and lead to synergistic activity and represents a viable therapeutic strategy that warrants further investigation. Treatment of AML engrafted NSG mice with CD70 CAR-T cells in conjunction with azacitidine is ongoing.