Abstract ND03: AMX-818: Next generation conditionally active, masked T-cell engager for the treatment of HER2-expressing solid tumors

Abstract

Abstract T Cell Engagers (TCEs) are a highly potent, cancer killing modality that redirect T cell cytotoxicity against tumors that express a selected tumor-associated antigen, or TAA, bypassing the requirements for T cell recognition of tumor specific MHC/peptides. In addition to their induction of direct cytotoxicity, their potency is amplified by cytokine driven actions downstream of T cell activation that enhance the anti-tumor immune response. Because of this two-fold mechanism of action, TCEs have demonstrated clinical activity in patients with “cold” tumors that are not responsive to other immune therapies such as checkpoint inhibitors. While a razor-thin therapeutic index (TI) has been observed in hematologic malignancies for TCEs like blinatumomab, and similarly with CAR-Ts, developing solid tumor T-cell-directed therapies has been hampered by on-target off-tumor toxicities including cytokine release syndrome (CRS), other toxicities, and other challenges related to immunogenicity. Attempts to circumvent CRS include step-up dosing, prophylactic steroids, complex molecular designs, have had limited success. Unmasked TCEs targeting HER2 have been unable to achieve a therapeutic index in the clinic due to dose limiting CRS toxicities at doses below 1 ug/kg. To address this challenge, we have engineered conditionally-activated T cell engagers or XPATs (XTENylated Protease Activated T Cell Engagers), that exploit the intrinsically high protease activity in tumors relative to healthy tissues. XPATs utilize our universal protease-releasable masking technology (Pro-XTEN) to provide preferential unmasking and activation in tumors, thereby mitigating off-tumor toxicities. Several key features of the technology have clinical validation including half-life extension, protease-driven unmasking and low immunogenicity, as over 200 patients have been treated with XTENylated products. AMX-818 is a conditionally active, masked T cell engager targeting HER2. AMX-818 is comprised of a tandem scFv core with a HER2-binding domain and a proprietary low affinity CD3 binding domain with each domain masked by a protease-releasable XTEN polymer that works by spatial/steric shielding of the active domains. AMX-818 is designed to overcome the on-target, off-tumor toxicities associated with existing TCEs by mitigating toxicity against healthy tissues that express the target antigen. Extensive in vitro and in vivo preclinical testing demonstrate the potential for AMX-818 to drive anti-tumor activity similar to an unmasked TCE, yet significantly expand the safety margin in NHPs. In vitro, proteolytically-unmasked AMX-818 demonstrates potent cytotoxicity against tumor lines with EC50s in the single-digit pM range. AMX-818 reduces target-directed T cell cytotoxicity and T cell activation by up to 4 orders of magnitude, while singly-masked variants of AMX-818 show intermediate activity relative to unmasked HER2 T cell engager. Only minimal complete unmasking of AMX-818 is required to generate potent cytotoxicity. In the established HER2-high (BT-474) and HER2-low (HT-55) xenograft models, AMX-818 induced protease-dependent tumor regressions comparable to the unmasked (active) TCE while remaining stable in circulation. In vivo, preferential cleavage of XPATs were demonstrated in patient-derived xenograft tumors relative to healthy organs (average of 24% unmasked TCE was observed in tumors). In NHPs, AMX-818 demonstrated a wide safety margin (MTD of AMX-818 = 42mg/kg vs MTD of fully unmasked HER2 TCE = 0.2mg/kg/day), supported by its protease stability in circulation and a maximum tolerated Cmax that was >400-fold higher than that of its active form (unmasked HER2 TCE). No CRS or systemic T cell activation was observed even at 50 mg/kg, supportive of minimal CRS risk for XPATs, whereas the unmasked form induced lethal CRS at > 0.3 mg/kg/day. Only 1-3% of the singly-cleaved XPAT metabolites were detected in plasma from NHPs treated with high doses of AMX-818 (25 & 42 mg/kg), while the fully unmasked TCE form was undetectable. In vitro studies in which AMX-818 was incubated for 7-days at 37°C in plasma from healthy and diseased NHPs and humans, (conditions which over-represent the accumulation of metabolites in vivo) demonstrate that AMX-818 is stable in plasma with comparable amounts of fully unmasked TCE generated between species. The similar rates of cleavage between species and the strong safety margin observed with AMX-818 in NHPs predict limited systemic accumulation of active TCE in human patients. In a repeat-dose GLP toxicology study, the highest dose evaluated (6 mg/kg) was the NOAEL, suggesting that HER2-XPAT has a wide TI based on predicted human PK and predicted efficacious exposures. The collective data demonstrate a significant decrease in toxicity with AMX-818 compared to its potent, fully unmasked TCE counterpart. Together with the data supporting stability of AMX-818 in plasma and healthy tissue and its preferential cleavage in tumors, the nonclinical data package supports further evaluation in a planned clinical trial. AMX-818 represents a promising solution to widen the TI for TCE activity in HER2-high and HER2-low expressing tumors. Citation Format: Bryan Irving. AMX-818: Next generation conditionally active, masked T-cell engager for the treatment of HER2-expressing solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr ND03.