Cadonilimab, a tetravalent PD-1/CTLA-4 bispecific antibody with trans-binding and enhanced target binding avidity

Abstract

Clinical studies have shown that combination therapy of antibodies targeting cytotoxic T-lymphocyte antigen-4 (CTLA-4) and programmed cell death-1 (PD-1) significantly improves clinical benefit over PD-1 antibody alone. However, broad application of this combination has been limited by toxicities. Cadonilimab (AK104) is a symmetric tetravalent bispecific antibody with a crystallizable fragment (Fc)-null design. In addition to demonstrating biological activity similar to that of the combination of CTLA-4 and PD-1 antibodies, cadonilimab possess higher binding avidity in a high-density PD-1 and CTLA-4 setting than in a low-density PD-1 setting, while a mono-specific anti-PD-1 antibody does not demonstrate this differential activity. With no binding to Fc receptors, cadonilimab shows minimal antibody-dependent cellular cytotoxicity, antibody-dependent cellular phagocytosis, and interleukin-6 (IL-6)/IL-8 release. These features all likely contribute to significantly lower toxicities of cadonilimab observed in the clinic. Higher binding avidity of cadonilimab in a tumor-like setting and Fc-null design may lead to better drug retention in tumors and contribute to better safety while achieving anti-tumor efficacy.

Keywords: PD-1/CTLA-4 bispecific antibody; drug retention; tumor microenvironment.

Figure 1.

Tetravalent design of anti-PD1/CTLA-4 bispesific antibody cadonilimab based on co-expression of PD-1 and CTLA-4 in tumor tissue. (a) Correlation analysis of PD-1 and CTLA-4 mRNA expression levels in various tumor types from TCGA dataset using cBioportal. The X- and Y-axis represent the mRNA expression level transformed by Log(value+1). TCGA, The Cancer Genome Atlas. (b) Schematic diagram of cadonilimab tetravalent structure.

Figure 2.

Cadonilimab induce IL-2 and IFN-γ production in mixed lymphocyte reaction assays. (a, b) Cadonilimab promoted activation of human peripheral blood mononuclear cells (hPBMCs), which induced a more robust secretion of IL-2 and IFN-γ in mixed culture of hPBMCs and human DCs; Compared with nivolumab plus ipilimumab, cadonilimab did not significantly improve secretion of IL-2 and IFN-γ. (c, d) Cadonilimab enhanced secretion of IL-2 and IFN-γ in mixed culture of hPBMCs and Raji-PDL1 cells. Data are shown as mean ±SEM for n = 2 and analyzed using one-way ANOVA. *P < .05, **P < .01 and ***P < .001 vs. isotype control.

Figure 3.

Cadonilimab demonstrates preferential higher avidity binding to higher density of PD-1 and CTLA-4. In a Fortebio assay, cadonilimab showed higher binding avidity with a (a) high density of PD-1 and CTLA-4, where PD-1 (50 nM) and CTLA-4 (50 nM) were loaded onto the sensor, compared with (b) that of a relative low density of PD-1, where PD-1 (10 nM) was loaded onto the sensor. Penpulimab (parental PD-1 antibody of cadonilimab) showed similar binding avidity under these antigen conditions (a) and (b).

Figure 4.

Tetravalence binding of Cadonilimab to PD-1 and CTLA-4. (a) Trans-binding of Cadonilimab to CHO-K1-CTLA-4 cells (Orange cells) which transfected with human CTLA-4, and Jurkat-PD1 cells (light blue cells) labeled with Hoechst 33342. Jurkat-PD1 cell which is a suspension lymphoblasts cell line became adherent to CHO-K1-CTLA-4 cells when co-cultured with cadonilimab, but not with nivolumab plus ipilimumab. (b) Crosslinking of cadonilimab with cells expressing PD-1 and CTLA-4 in FACS assay. AK104 or control antibodies were added to a 1:1 mix of Far red-labeled CTLA-4-expressing CHO-K1 cells and CSFE-labeled PD-1-expressing CHO-K1 cells. (c) Cadonilimab could bind to human PD-1 and CTLA-4 with similar binding activity regardless of binding PD-1 or CTLA-4 firstly. Cadonilimab was fixed onto the sensor in Fortebio assay, and sequential binding to human PD-1 (PD1-hFc) and CTLA-4 (CTLA-4-hFc) or in reverse order was performed and measured.

Figure 5.

Cadonilimab with Fc null-mutations for both safety and efficacy concerns. (a) Antibody-dependent cell-mediated cytotoxicity (ADCC) activities were determined by measuring lactase dehydrogenase (LDH) release from 293 T-CTLA4-PD1 cells. (b) Complement-dependent cytotoxicity (CDC) activities of cadonilimab, penpulimab in hG1WT format, and ipilimumab and nivolumab were determined by measuring LDH release from CHO-K1-PD1-CTLA4 cells. (c) Antibody-dependent cellular phagocytosis (ADCP) activities of cadonilimab, penpulimab in hG1WT format plus anti-CTLA-4 mAb in hG1WT format, and nivolumab plus ipilimumab were studied by examining phagocytosis of CHO-K1-PD1-CTLA4 cells by murine bone marrow derived macrophages (MBMMs). (d, e) Cadonilimab decreased the release of inflammatory cytokines IL-8 and IL-6 from human peripheral monocyte-derived macrophages (HPMMs). Data are expressed as mean or mean ±SEM and analyzed using one-way ANOVA. *P < .05 and ***P < .001 vs. isotype control; ###P < .001 vs. negative control.