(62e) Measurement and Modeling of Heterogeneous Single-Cell PI3K Inhibition Dose Responses

Precision health care in cancer increasingly focuses on treating patients with drugs targeted against specific driver mutations present in malignant cells. Targeted drugs typically produce transient responses with subsequent recurrence fueled by drug resistant cancer cells. Drug resistance arises from multiple cell intrinsic and/or environmental causes, including secondary genetic mutations, compensatory signaling pathways, intercellular signaling, and mass transfer limitations for drugs in spatially heterogeneous, disorganized tumor environments. Standard models for drug development and testing in cell-based assays typically rely on population-scale metrics defined by concentration-response curves. These metrics include maximum effect (E_max) and the concentrations of a compound required to reach 50% of the maximum effect (EC₅₀). However, population-scale metrics of efficacy mask heterogeneity among single cells. Drug resistance and tumor recurrence commonly arise from small subsets of cancer cells, motivating new approaches to analyze drug effects in single cells.

In this work, we first simulated concentration-response curves from heterogeneous populations of breast cancer cells. Our simulations reveal that arbitrary distributions of concentration-response parameters, including bimodal distributions or distributions with small subpopulations, can yield identical population level concentration-response curves. Next, we used live, single cell fluorescence microscopy to measure concentration-dependent responses to clinically-approved inhibitors (alpelisib and omipalisib) of the oncogenic phosphatidylinositol-3-kinase (PI3K) - Akt pathway. We quantified effects of these compounds on activation of Akt and ERK, two major kinases driving proliferation and survival in breast cancer and multiple other malignancies, with multiplexed kinase translocation reporters (KTRs). KTRs reversibly move between the nucleus and cytoplasm of cells, providing a dynamic, quantitative readout of kinase activity and inhibition. We developed automated image processing methods to quantify KTRs in thousands of cells, tracking activities of Akt and ERK in single cells over time. We stably expressed KTRs for Akt and ERK in multiple breast cancer cell lines with mutations that activate the Akt pathway. Our studies revealed that cells respond to PI3K inhibitors rapidly and heterogeneously, with different magnitudes and durations of inhibition. We next subjected cells to continuously escalating doses of PI3Ki and extracted dose-dependent kinase activity measurements from individual cells over time. We determined that both breast cancer cell lines displayed heterogeneous EC₅₀ and Emax values for each tested PI3K inhibitor. Approximately 10% of cells displaying EC₅₀ values 10 times higher than the population EC₅₀ and approximately 10% of the cells showing an E_max close to zero, indicating a lack of kinase inhibition.

We next investigated the impact of measured cell-intrinsic heterogeneity on drug resistance relative to variations in pharmacokinetic and mass transport dynamics of small molecule drugs that govern tissue distribution in vivo. We integrated a two-compartment pharmacokinetic model with a Krogh Cylinder model of drug distribution in tissue around a blood vessel. We calibrated this model based on measured PK data for one of our PI3K inhibitors and then calculated different distributions of pharmacokinetic parameters at various locations around a blood vessel. For instance, we compared random spatial distributions of EC₅₀ and E_max values for single cells or distributions that were a function of radius from the blood vessel against the baseline case where all cells had identical concentration-dependent responses. We found that intrinsic variation in cellular dose-response can accentuate tissue heterogeneity, increasing the likelihood of resistance to an administered drug.

In summary, our systems biology approach based on single-cell imaging data provides an innovative framework to analyze susceptibility or resistance to cancer therapies. Predicting how cell-intrinsic and environmental factors regulate efficacy of drugs promises a new pathway to overcome drug resistance and improve outcomes for patients with cancer.