Evaluation of anti-human complement component 5 antibody in humanized C5 mice

Abstract

Purpose : Complement has been implicated in ocular inflammatory and retinal degenerative diseases. Blocking C5 cleavage by a monoclonal antibody (mAb) is a possible therapeutic strategy for these disorders. The present study was undertaken to evaluate pharmacokinetic/pharmacodynamic (PK/PD) properties of a therapeutic mAb in humanized C5 mice.

Methods : Lead anti-human C5 mAb (REGN3918) was selected by classical (CP) and alternative pathway (AP) hemolysis bioassays in vitro. Humanized C5 mice (C5^hu/hu, Regeneron) were used to assess PK/PD in vivo/ex vivo. C5^hu/hu mice were generated and characterized using VelociGene and VelociMouse technologies. Mice were stratified according to their serum C5 level. Each mouse received a single subcutaneous injection of REGN3918 (15 mg/kg) or isotype control mAb. Mouse serum was collected at different time points: Pre-dose, day 1, terminal bleed time points (days 10-60; n≥4 each). PK was measured by ELISA or Gyros immunoassays. PD was analyzed by ex vivo CP assay. Excess human C3 (80ug/ml) was added ex vivo prior to hemolysis assay. Human C5 and mAb concentrations in serum were evaluated using a LC-MRM-MS method.

Results : REGN3918 blocked CP hemolysis (5% human serum) in a dose-dependent manner, with IC50 of 14.1 nM in a 10 mins lysis assay or 18.3 nM with 60 mins lysis. Maximal inhibition of hemolysis was 91-95%. In this assay REGN3918 also blocked C5a generation in a dose-dependent manner with IC50 of 8.5 nM as measured by C5a ELISA. In AP hemolysis (10% human serum) REGN3918 showed maximal inhibition of 88%, with IC50 of 27.4 nM. The isotype control antibody did not block hemolysis under identical assay conditions. Average serum level of human C5 in C5^hu/hu mice was 40.9 μg/ml. There is a difference between male (55.4 ± 1.7 μg/ml, n=47) and female (24.7 ± 0.6 μg/ml, n=49) mice. Following a single SC dosing of REGN3918, compared with the pre-dose, CP activity induced by the mouse serum ex vivo was significantly reduced at all time points up to and including day 35. This was correlated with the C5/REGN3918 molar ratio measured by LC-MRM-MS method.

Conclusions : Our findings demonstrate that C5^hu/hu mice can be used to evaluate PK and PD profiles of anti-human C5 mAb. LC-MRM-MS is a valuable method to confirm the bivalent binding of REGN3918 to C5 and its PD effect in vivo.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.