R. Shi, H. Wang, S. Suchard, S. Nadler, M. Honczarenko, S. Zhang, T. Salcedo, K. Price, C. Fleener, B. Ganguly, J. Mora, J. Haulenbeek, R. Liu, B. Murthy, Z. Yang; Bristol-Meyers Squibb, Princeton, NJ

BACKGROUND: Therapeutic proteins continue to be investigated in autoimmune diseases. For high risk biologics, the first-in-human starting dose should be based on animal in vitro / in vivo data and human in vitro data to establish minimum anticipated biological effect level (MABEL). This presentation highlights starting dose calculation for three antagonist T-cell co-stimulation domain antibodies (dAb) targeting CD28, CD40L, and CD40 using the MABEL approach.
METHODS: For all three compounds, Pharmacokinetics/Pharmacodynamics (PK/PD) analysis was conducted to correlate exposure with receptor occupancy (RO) via Emax model, which is further correlated with the T-cell-dependent antibody response (TDAR). Human PK exposures were predicted assuming the same PK parameter estimates in monkeys for CD28, using the observed human PK parameters of Abatacept for CD40-L (showed similar monkey PK), and allometrically scaling PK parameters for CD40 assuming target mediated drug disposition. A ≤ 10% RO was targeted to achieve minimal pharmacologic activities for these dAbs based on the observed or derived human RO EC50.
RESULTS: MABEL starting doses are 0.01 mg (CD28), 0.6 mg (CD40L), and 0.5 mg (CD40). All three dAbs showed negligible pharmacological effects and were well tolerated after the MABEL starting doses. There was good agreement between the model predictions and observed RO. The safety margins for these doses were up to ~ 2,000,000-fold below the NOAELs.
CONCLUSION: A PK/PD model approach successfully assisted the MABEL dose calculation for three dAbs, which were administered safely in healthy subjects.