Single-Cell Analysis of Cancer Targeting Circuits: Lessons from Cellular Heterogeneity for Circuit Optimization

Targeting malignant cells in a heterogeneous population is a major challenge. Synthetic circuits integrating gene expression markers can discriminate cancer cells autonomously and at high precision on average. However, expression patterns widely vary from cell to cell and are also influenced by the tumor micro‐environment. In a recent paper (1), we designed circuits integrating two transcriptional activities into a biological output and studied Inputs/Output relationships at the single‐cell resolution. In a model cell population, we characterized our circuits with varying signal processing modules and showed that an AND gate behavior was conserved in all our circuits but with different thresholds and amplification values.

Then, using cancer‐specific promoters we designed cell‐classifiers able to autonomously distinguish between pre‐malignant and malignant populations of WI‐38 lung cancer cells in a mixed environment, and specifically kill malignant ones. We found that a general trade‐off exists between sensitivity and specificity of the targeting, due to the circuit inherent noise but also to the cellular heterogeneity over inputs levels. Besides, killing results varied if cancer cells were cultured separately, homogeneously mixed to the pre‐malignant population or organized as a "tumor" environment. Our study underlines the importance of single‐cell analysis and of reconstructing micro‐environments for the development and optimization of targeting circuits.

(1) Morel, M.; Shtrahman, R.; Rotter, V.; Nissim, L.; Bar-Ziv, R. H. Proc. Natl. Acad. Sci. 2016, 113 (29), 8133–8138.