GNUV201, a novel human and mouse cross-reactive PD-1 monoclonal antibody for cancer immunotherapy.

Abstract

e14509 Background: Several PD-1 antibodies have been recently approved as anti-cancer therapies which work by blocking the interaction of PD-1 with its ligand PD-L1, thus restoring anti-cancer T cell activities. However, cross-reactive anti-PD-1 antibodies to both human and pre-clinically relevant animal models are still required to predict more precisely the efficacy and toxicity in humans, especially in various combination settings of anti-PD-1 with other treatments. Methods: Therefore, we have developed an antibody named GNUV201, a highly selective and interspecies cross-reactive humanized monoclonal antibody, with conventional hybridoma technology. Results: GNUV201 equally binds to mouse and human PD-1 (EC50= 32 pM and 28 pM, respectively), while marketed Keytruda and Opdivo bind only to human PD-1. According to the results of co-crystal structure and alanine-scanning mutagenesis to examine the interaction mode between GNUV201 and human PD-1, the epitope recognized by GNUV201 is on the “FG loop” region of human PD-1, which is well conserved in both human and mouse PD-1, supporting GNUV201’s inter-species cross-reactivity, and this position is distinct from those of Keytruda (“C’D loop”) and Opdivo (N-term). Notably, the structural feature in which the protruding epitope loop fits into GNUV201’s binding pocket supports its enhanced binding affinity due to slower dissociation (8.5 times slower than Keytruda in SPR). Furthermore, GNUV201 shows stronger binding affinity at pH 6.0 (KD = 0.9 nM at pH 6.0, 5.8 times smaller KD than at pH 7.4) which mimics more closely the hypoxic and acidic tumor micro-environments (TME), especially due to further improved Koff. This phenomenon is not observed with other anti-PD-1 competitors, implying that GNUV201 achieves better target occupancy in tumors compared to its marketed competitors. GNUV201 was also shown to successfully block the interaction of hPD-1 with hPD-L1 at similar single digit nM levels compared with Keytruda and Opdivo in vitro and strongly suppress tumor growth in syngeneic models. In particular, tumor growth regressed in B16F11 (mouse melanoma) and was inhibited in MC38 (mouse colon cancer) syngeneic models and MC38/hPD-1 knock-in mouse models during treatment, when compared with the original tumor size. Conclusions: In summary, we expect GNUV201 to be an excellent antibody candidate whose efficacy and toxicity in the human system can be predicted from preclinical study results due to its human-mouse cross-reactivity and whose efficacy can be further improved with better target occupancy due to its slower dissociation and preferential binding in low pH tumor microenvironments. We are currently engineering GNUV201 to further improve antigen binding kinetics, which could lead to enhanced efficacy and pharmacokinetics in humans, and to develop the best-in-class PD-1 antibody based on superior comparative data from on-going efficacy studies.