Document resource
Background PD-1 blockade therapy revolutionized treatment of solid tumors, but response and eligibility rates remain low. Bispecific antibodies combining PD-1 blockade with other immune checkpoint inhibitors, such as PD-1xCTLA-4, present a promising strategy to augment monotherapies. We previously identified a novel immune checkpoint pathway comprised of HHLA2 (B7-H7), expressed on solid tumors, and its inhibitory receptor KIR3DL3, expressed on T and NK cells. KIR3DL3 blockade showed increased anti-tumor effects and is currently being evaluated in phase I clinical trials. To supplement our KIR3DL3 blocking antibody, we generated a panel of bispecific antibodies (bsAbs) that block both KIR3DL3 and PD-1 to further improve upon each monotherapy.Methods A panel of fully human bsAbs were generated in three conformations to determine the best structure for maintaining binding and blocking both KIR3DL3 and PD-1. KIR3DL3xPD-1 bsAbs were tested in various binding assays (biolayer interferometry, on-cell binding) and receptor blocking assays (immune synapse, NFAT activation, and tumor cell lysis). Based on these results, one candidate bsAb was selected for ongoing further characterization.Results Of the three tested bsAb conformations, two maintained binding and blocking to both target receptors with the anti-PD-1 arms being in the native IgG conformation and anti-KIR3DL3 as the scFv arms. Of the two remaining bsAbs, both could block binding of KIR3DL3 and PD-1 to their ligands in vitro and one, PK-E1, was selected for further characterization. PK-E1 retained the ability to block KIR3DL3 on primary NK and NK92 cells and increase lysis of HHLA2+ target cells HCC827 and K562, identical to that of the parental anti-KIR3DL3 mAb. Additionally, PK-E1 was internalized by PD-1 expressing cells but not KIR3DL3 expressing cells, indicating an additional mechanism of action for PK-E1 like that of pembrolizumab (Keytruda) and nivolumab (Opdivo).Conclusions As novel bispecific constructs, we set out to determine which bsAb design was optimal to maintain PD-1 and KIR3DL3 blocking among three constructs. We found that, like other reports, anti-PD-1 binding arms required the native IgG orientation to retain binding and blocking, while the anti-KIR3DL3 binding arms functioned in both native and scFv formats. Of the three constructs one was chosen for further characterization, PK-E1, a PD-1 IgG4(H)-scFv(KIR3DL3) due to its similar performance to PK-E3, PD-1 IgG4(L)-scFv(KIR3DL3), but more robust expression. We are continuing characterization of PK-E1 with additional assays of immune cell activation, and in vivo for anti-tumor effects.