A Novel Full Length Anti-FLT3 CD3 Bispecific Antibody for the Treatment of Acute Myeloid Leukemia

FLT3 is a receptor tyrosine kinase expressed on the surface of acute myeloid leukemia (AML) patient blasts. FLT3 is the most frequently mutated gene in AML patients, and these mutations are associated with poor prognosis. Despite the development of small molecule inhibitors of FLT3 function and neutralizing FLT3 antibodies, there remains a need for antibodies that target the broad AML patient population with improved efficacy and safety.

We chose to use Pfizer's proprietary full length humanized CD3 bispecific IgG molecule platform. The first step in the development was finding a suitable targeting epitope on FLT3 as not all epitopes result in optimal T cell activation in the context of CD3 binding. Through a combination of in vitro and in vivo studies, FLT3 antibodies targeting extracellular domain 4 of FLT3 were found to be more effective at AML cell depletion than other domains in the full-length bispecific IgG format, significantly outperforming the antibodies targeting the most membrane proximal region of domain 5. The final candidate antibody was engineered to have picomolar affinity for recombinant human FLT3 (<50 pM) to ensure that low levels of FLT3 present on AML (<5,000 per cell) could efficiently trigger T cell cytotoxicity.

Anti-tumor activity of the final molecule was further characterized in vitro and in vivo. Three AML cell lines, Eol-1, Molm-13 and MV-411 express high, medium and low levels of FLT3, respectively, and were targeted by activated healthy donor T cells in the presence of the FLT3 bispecific with an efficiency that correlated with FLT3 surface density (EC₅₀ range was 0.5 pM to 40 pM at E:T ratios of 1:1). Ability to target primary AML cells was confirmed in assays with primary AML blasts and autologous patient T cells (EC₅₀s 7 - 15 nM, for E:T ratios of up to 1:50). In orthotopic xenograft models with the same set of cell lines, all three cell lines were efficiently eliminated by activated T cells in the presence of a single dose of FLT3 bispecific (dose range 10 to 100 μg/kg), confirming the anti-tumor activity of the bispecific antibody in vivo.

The safety of Pfizer's FLT3 bispecific antibody was also studied in preclinical studies. We first analyzed gene expression of FLT3 and three other commonly considered AML targets (CD33, CD123, and CLL-1) in publicly available gene expression databases. FLT3 had the highest differential expression between AML and healthy tissues with blood and brain showing expression of low levels of FLT3 RNA. In tissue cross-reactivity studies, we detected little to no binding of the candidate FLT3 antibody to human brain tissue sections. In blood, the highest expression of FLT3 was in hematopoietic stem cells and progenitors, dendritic cells and monocytes consistent with previous publications and known roles of FLT3 in hematopoiesis and dendritic cell homeostasis. Of note, expression of FLT3 in monocytes and whole blood was the lowest compared to three other AML targets.

To address any potential toxicities of FLT3 bispecific, we performed exploratory studies in cynomolgus monkeys. Affinity of FLT3 bispecific for cynomolgus FLT3 was ~20x lower than for human FLT3 (~1 nM), as determined in binding assays with cells expressing human and cynomolgus FLT3. No major clinical signs or toxicological findings were observed up to the highest dose level tested (3 mg/kg) following two administrations at weekly intervals. On target activity was demonstrated by nearly complete elimination of FLT3+ dendritic cells in the blood of treated monkeys two days after administration. In addition, we detected elimination of FLT3+ CD34+ stem cells in the bone marrow of treated monkeys two days after the second dose. Remarkably, both dendritic cells and FLT3+ CD34+ stem cells rebounded to baseline levels observed in control animals when analyzed at study termination, two weeks following the second dose. Consistent with minimal expression of FLT3 on healthy tissues and the presence of FLT3+ on rare blood subsets, cytokine increases typically associated with CD3 bispecific administration were minimal (only 6 to 17 and 46 to 89 fold over baseline for IFN-γ and IL-6, respectively, at the highest dose tested).

In conclusion, the robust anti-tumor activity of Pfizer's FLT3 bispecific antibody combined with good tolerability in cynomolgus monkeys, reversible hematological toxicity and absence of non-hematological toxicity support its further clinical development in AML.