Quantitative characterization of in vitro bystander effect of antibody-drug conjugates.

Abstract

Antibody-drug conjugates (ADCs) are designed to target antigen expressing (Ag+) cells in a tumor. Once processed by the Ag+cells, ADCs can release cytotoxic drug molecules that can diffuse out of Ag+cells into the neighboring antigen-negative (Ag-) cells to induce their cytotoxicity. This additional efficacy of ADCs on Ag- cells in the presence of Ag+cells is known as the 'bystander effect'. Although the importance of this phenomena is widely acknowledged for effective killing of a heterogeneous tumor, the rate and extent of the bystander killing in a heterogeneous system is not quantitatively understood yet. Thus, the objectives of this manuscript were to: (1) synthesize and characterize a tool ADC Trastuzumab-vc-MMAE that is capable of exhibiting bystander effect, (2) quantify the time course of the bystander effect for the tool ADC using in vitro co-culture systems created using mixture of various HER2-expressing cell lines, and (3) develop a pharmacodynamic (PD) model that is capable of characterizing the bystander effect of ADCs. Co-culture studies conducted using GFP labelled MCF7 cells as Ag- cells and N87, BT474, and SKBR3 as Ag+cells revealed that the bystander effect of ADC increases with increasing fraction of Ag+cells in a co-culture system, and with increased expression level of target on Ag+cells. A notable lag time after ADC incubation was also observed prior to significant bystander killing of Ag- cells. Based on our results we hypothesize that there may be other determinants apart from the antigen expression level that can also influence the ability of Ag+cells to demonstrate the bystander effect in a co-culture system. The co-culture analysis also suggested that the bystander effect of the ADC can dissipate over the period of time as the population of Ag+cells declines. A novel PD model was developed to mathematically characterize the bystander effect of ADCs by combining two different cell distribution models to represent the population of Ag+and Ag- cells in a co-culture system. This PD model can be integrated with the systems PK model for ADCs in the future to generate a quantitative framework that is capable of supporting the discovery and development of novel ADCs with optimal bystander killing capabilities.