Affinity-Tuned T-Cell Engager for Dual Targeting of B-Myeloid Mixed-Phenotype Acute Leukemia (B-MPAL)

Introduction. B-myeloid mixed-phenotype acute leukemia (B-MPAL) is a high-risk leukemia subtype presenting with both lymphoid (e.g. CD19) and myeloid (e.g. CD33) surface antigens. Cure rates in children and adolescents remain lower than for almost all other acute lymphoblastic leukemia (ALL) subtypes, with overall survival ranging from 75-80%. Outcomes for adults are notably poorer, with overall survival ranging from 20-50% for patients over 40 years of age. Induction therapies range from those used for ALL versus those used for acute myeloid leukemia (AML) with a paucity of information from randomized clinical trials due to the uncommon nature of this subtype.

B-Myeloid MPAL cells frequently express both CD19 and CD33 while non-neoplastic human hematopoietic cells don't express both targets, creating an opportunity to selectively target MPAL cells with minimal off tumor toxicity. We hypothesize that a multispecific T-cell engager antibody (MTE) engineered to preferentially bind cells expressing CD19 AND CD33 will selectively kill MPAL cells compared to a bi-specific antibody that requires CD19 OR CD33 expression. We thus engineer MTEs that bind both CD19 and CD33 on MPAL cells and CD3 on T cells. We further fine-tune the affinity of the CD19 and CD33 binders so that the MTE binds poorly to single positive (SP) normal cells that express either CD19 or CD33 but binds well to double positive (DP) leukemia cells. For this we take advantage of both affinity and avidity, the latter of which occurs only when the MTE interacts with both CD19 and CD33 expressed on the same cell. Such therapies could be used to induce deep remissions, improve outcomes when integrated with chemotherapy, and/or serve as a bridge to hematopoietic cell transplantation. Blinatumomab, a bispecific antibody recognizing only CD19, has been a successful addition to treating both childhood and adult ALL. However, patients require months of intravenous immunoglobulin replacement therapy due to the consequential depletion of healthy B-cells, which share CD19 as a critical surface antigen with most B-cell ALL. The new therapeutic strategy we are developing will improve safety compared to currently available immunotherapies.

Methods. We first identified that CD19 and CD33 are appropriate targets for B-Myeloid MPAL after analyzing aggregated transcriptome data for 31 MPAL patients. Next, we successfully designed and produced 30 tri-specific antibodies targeting CD19/CD33/CD3. All the binders were selected because they are either components of FDA approved therapeutics, have gone through human clinical trials, or have been extensively studied to de-risk safety and immunogenicity concerns. Our MTE framework is derived from the asymmetric “knob into holes” human IgG1 scaffold incorporating effector silencing mutations. Our in vitro pipeline to test those antibodies included binding assays as well as T-cell cytotoxic assays using cell lines expressing both targets, only one, or none. Top candidates were further tested for efficacy in in vivo mouse models. We use a luciferase labeled JIH-5 cell line that innately expresses CD19 and CD33, xenografted into NSG-SGM3 immunocompromised mice. We dosed engrafted mice with both MTE and human T-cells from a healthy donor for repeated weekly cycles. We followed disease progression every week using IVIS imaging.

Results. Our preliminary data demonstrate our initial in vitro success in targeting CD19/CD33/CD3 antigens as well as dramatic improvement of specificity by generating lower affinity variants of known CD19 and CD33 binders. For multiple candidates we observed a much lower cytotoxicity for SP cells, with up to a 7 log difference in IC₅₀. This difference in IC₅₀ potentially expands the therapeutic window because DP cells are specifically targeted for lysis while single positive cells are spared. In addition, we present initial pre-clinical in vivo data for our top candidates in xenograft mouse models, showing successful control of disease in engrafted mice at treatment doses that are well tolerated.

Conclusions. Taking advantage of the unique features of MPAL we developed a novel immunotherapy with improved half-life, potency, selectivity, and safety. This work serves as a proof of concept that the combinatorial use of binders with the proper affinity for their respective targets allow for the generation of highly specific immunotherapies with minimal cytotoxicity for normal cells.

No relevant conflicts of interest to declare.