Anti-IL1RAP/CD3 Bispecific Antibody (BsAb) Is a Promising Novel and Effective Therapy for Acute Myeloid Leukemia (AML)

Despite increasing insight of the mechanisms of leukemogenesis, implementation of risk-stratification treatment strategies and use of molecular targeting therapeutics, only a minority of AML patients are cured with chemotherapy-based regimens. Most patients are either primary refractory to induction therapy or subsequently relapse following a brief remission likely due to persistence of chemo-resistant leukemia stem cell (LSC)-enriched subpopulations. Therefore there are urgent unmet therapeutic needs. The principle of CD3 based BsAb therapy is to recruit and activate T cell effector function upon engaging target cells expressing a tumor associated antigen (TAA). Several TAA (e.g.CD33, CD123) have been found to be differentially overexpressed on the surface of AML blasts. Recently, the expression of IL-1 receptor accessory protein (IL1RAP) was reported to be higher on LSCs and more differentiated leukemic progenitor cells (LPCs) than on normal hematopoietic stem cells (HSC), making it an attractive TAA target for immunotherapy. Thus, we hypothesize that an anti-IL1RAP/CD3 BsAb could preferentially target LSCs and LPCs while sparing normal HSCs, thereby offering a window of therapeutic opportunity for AML patients.

To test this hypothesis we first determined the expression pattern of IL-1RAP. IL1RAP was differentially expressed on AML cell lines: HL60 and MV4-11 (high), THP-1 (intermediate), and KG-1a (low). Its expression levels were variable on primary blasts from AML patients (n=10). CD34⁺ peripheral blood mononuclear cells (PBMC) from healthy donors showed little or no IL1RAP expression. Next, we screened 12 anti-IL1RAP monoclonal antibodies from our human phage display antibody library. Based on the results of the BIAcore and antibody dependent cellular cytotoxity (ADCC) assays performed on MV4-11 cells, we selected one lead candidate with binding affinity of ~20nM and EC50 4.9nM for further development. Using knobs-into-holes strategy, we paired this IL1RAP antibody with an anti-CD3 antibody. The dual binding ability of this BsAb was confirmed by flow cytometric analysis. We subsequently proved T cell dependent cytotoxicity (TDCC) of this BsAb using both AML cell lines and AML primary blasts as targets. The anti-IL1RAP/CD3 BsAb induced over 95% of specific killing on all tested AML samples at E: T (Effector : Target) ratio of 1:1 compared with no cytotoxicity effect on the IL-1RAP-negative Raji cells in the 48-hour long term killing (LTK) assay. In line with the cytotoxic data, there was also substantial release of cytokines including IFN-γ, TNF-α and Granzyme B from the AML cells but not Raji cells. TDCC reported over 90% of IL1-RAP specific killing in all but one (only 60%) patient samples (n=10), which had the lowest IL1RAP expression level. The killing of leukemia cells was also observed by time-lapse live cell microscope imaging using luciferase-expressing MV4-11^luci+ cells with addition of 100ng/ml BsAb and T cells at E:T ratio of 5:1, which documented effective T cell recruitment and cytolytic activities against MV4-11^luci+ cells. We also tested the anti-leukemic activity of this BsAb in vivo using a NSG xenograft mouse model. NSG mice transplanted with 1x10⁶ MV4-11^luci+ cells were treated with 200µg anti-IL1RAP/CD3 BsAb plus 3x10⁶ unstimulated T cells, vehicle, T cells alone or anti-CD19/CD3 BsAb (control Ab) plus T cells twice a week for 30 days. In vivo imaging demonstrated effective elimination of the leukemia signal (P<0.001) in MV4-11^luci+-engrafted NSG mice treated with T cells and anti-IL1RAP/CD3 BsAb (photon flux (PF) =5.5x10⁵) compared with intense leukemia infiltration in controls treated with either vehicle (PF=3.5x10⁹), T cells alone (PF=4x10⁹), or control anti-CD19/CD3 BsAb plus T cells (PF=2x10⁹). In a separate experiment, treatment with anti-IL1RAP/CD3 BsAb plus T cells of NSG mice engrafted with primary AML blasts led to significant decrease in disease burden than treatment with isotype antibody plusT cells (p=0.004). Further characterization and determination of disease response in these animals is ongoing.

In conclusion, we have successfully engineered and characterized an anti-IL1RAP/CD3 BsAb derived from our human naïve phage display library. Our data demonstrated that this BsAb can effectively eradicate AML cells in vitro and in vivo, supporting the clinical translation of this novel immunotherapeutic agent for treatment of AML patients.