The anti-tumor efficacy of 3C23K, a glyco-engineered humanized anti-MISRII antibody, in an ovarian cancer model is mainly mediated by engagement of immune effector cells

Ovarian cancer is the leading cause of death in women with gynecological cancers and despite recent advances, new and more efficient therapies are crucially needed. Müllerian Inhibiting Substance type II Receptor (MISRII, also named AMHRII) is expressed in most ovarian cancer subtypes and is a novel potential target for ovarian cancer immunotherapy. We previously developed and tested 12G4, the first murine monoclonal antibody (MAb) against human MISRII. Here, we report the humanization, affinity maturation and glyco-engineering steps of 12G4 to generate the Fc-optimized 3C23K MAb, and the evaluation of its in vivo anti-tumor activity. The epitopes of 3C23K and 12G4 were strictly identical and 3C23K affinity for MISRII was enhanced by a factor of about 14 (KD = 5.5 × 10-11 M vs 7.9 × 10-10 M), while the use of the EMABling® platform allowed the production of a low-fucosylated 3C23K antibody with a 30-fold KD improvement of its affinity to FcγRIIIa. In COV434-MISRII tumor-bearing mice, 3C23K reduced tumor growth more efficiently than 12G4 and its combination with carboplatin was more efficient than each monotherapy with a mean tumor size of 500, 1100 and 100 mm3 at the end of treatment with 3C23K (10 mg/kg, Q3-4D12), carboplatin (60 mg/kg, Q7D4) and 3C23K+carboplatin, respectively. Conversely, 3C23K-FcKO, a 3C23K form without affinity for the FcγRIIIa receptor, did not display any anti-tumor effect in vivo. These results strongly suggested that 3C23K mechanisms of action are mainly Fc-related. In vitro, antibody-dependent cytotoxicity (ADCC) and antibody-dependent cell phagocytosis (ADCP) were induced by 3C23K, as demonstrated with human effector cells. Using human NK cells, 50% of the maximal lysis was obtained with a 46-fold lower concentration of low-fucosylated 3C23K (2.9 ng/ml) than of 3C23K expressed in CHO cells (133.35 ng/ml). As 3C23K induced strong ADCC with human PBMC but almost none with murine PBMC, antibody-dependent cell phagocytosis (ADCP) was then investigated. 3C23K-dependent (100 ng/ml) ADCP was more active with murine than human macrophages (only 10% of living target cells vs. about 25%). These in vitro results suggest that the reduced ADCC with murine effectors could be partially balanced by ADCP activity in in vivo experiments. Taken together, these preclinical data indicate that 3C23K is a new promising therapeutic candidate for ovarian cancer immunotherapy and justify its recent introduction in a phase I clinical trial.