Pharmacokinetic, receptor occupancy and pharmacodynamic (PK/RO/PD) modeling of ALX148, a CD47 blocker
Elena Vasileva (1), Oleg Demin Jr (1)
(1) InSysBio, Moscow, Russia
Introduction: ALX148 is a promising anti-CD47 blocker currently undergoing clinical trials, which is designed to enhance the activity of anti-cancer target antibodies. Optimal doses selection is increasingly important in clinical setup and can be guided by assessment of target receptor occupancy (RO) and pharmacodynamics (PD) effect in the site of action. While direct measurement of actual RO in tumor tissues is challenging, mechanistic pharmacokinetic and RO modelling can provide valuable information that can be extrapolated to the clinic.
Objectives:
- To develop a semi-mechanistic PK/RO/PD model of ALX148, a CD47 blocker;
- To describe clinical ALX148 data on PK and CD47 RO;
- To predict CD47 RO in the tumor tissues of Non-Hodgkin lymphoma;
- To estimate ALX148 effect on antibody-dependent cellular phagocytosis (ADCP) on the basis of CD47 occupancy in the tumor tissues of Non-Hodgkin lymphoma patients.
Methods: The model structure was developed on the basis of a two-compartmental PK model with an additional mechanistic description of ALX148 and CD47 interaction in the central compartment. Two-step binding on the surface of red blood cells (RBCs) and T cells (tumor) was implemented into the model. Clearance is described as a combination of linear and nonlinear components, with the latter being reflected by ALX148 binding to the CD47 receptor on RBCs and further complex internalization. Parameters for the model were either taken/calculated from in vitro and in vivo data or identified via fitting. The model was fitted and validated against ALX148 clinical data on PK in plasma and CD47 RO on RBCs after administration of 3 different dosages. Distribution to the tumor tissues of Non-Hodgkin lymphoma (spleen, lymph nodes, bone marrow) was described using antibody biodistribution coefficients [1]. RO predictions were made for RBCs and tumor cells taking into account their amount in the body and CD47 expression (molecules per cell). ADCP enhancement by ALX148 was described on the basis of RO using the direct PD effect model. Parameters for the PD effect model were fitted against in vitro data on phagocytosis using a separate model, which was developed to reproduce experimental phagocytosis assay for 4 cancer cell lines (Daudi, DLD-1, MM1.R, OE19).
Results: Developed PK/RO/PD model successfully described ALX148 clinical data on PK and RO in blood plasma. Predicted concentrations in lymph nodes, spleen, and bone marrow were similar and approximately 3 times lower than in blood plasma. Such high concentration in tumor tissues is caused by the relatively small size of ALX148 (molecular weight is 78 kDa). High doses (10 mg/kg QW and 30 mg/kg Q2W) of ALX148 resulted in more than 98% of mean trough steady state CD47 RO in the tumor tissues, whereas a lower 95% confidence band was above 90%. Corresponding values after administration of 3 mg/kg QW were lower and ranged between 85-90%. ADCP of cancer cells was predicted to be increased by ~1.8 times during the treatment with both regimens of ALX148: 10 mg/kg QW and 30 mg/kg Q2W. Slightly lower ADCP enhancement was observed in the case of 3 mg/kg QW ALX148. The difference in ADCP induction after administration of 3 mg/kg QW and 10 mg/kg QW of ALX148 was statistically significant: p-value < 0.001.
Conclusions: Developed semi-mechanistic PK/RO/PD model predicted that at least 10 mg/kg of ALX148 should be administered weekly to achieve more than 90% of CD47 occupancy in tumor tissues with an average 1.8-fold increase in ADCP. Similar doses were chosen for further phase 2 clinical trials of ALX148. The introduced approach can be also applied for an optimal dose selection of other anti-CD47 antibodies taking into account their specific features as binding properties, size, etc.
References:
[1] Shah and Betts. MAbs. 2013 Mar 1; 5(2): 297–305.
