A network approach to interrogate cellular plasticity and drug resistance in cancer

ABSTRACT We will address a critical problem in clinical oncology, namely how highly heterogeneous, drug resistant tumor cell populations develop, and how they can be targeted. Most tumors develop resistance to almost every type of therapy, including targeted-, radiation-, chemo- or immunotherapy, ultimately leading to cancer deaths. It is

essential to develop novel methods to understand the processes leading to drug resistance under complex in vivo conditions where stromal and immune elements interact with malignant cells. We will study squamous cell carcinomas (SCCs), a major contributor to human cancer burden and one of the most common solid tumor

types that arise in a range of tissues including head and neck, lung, esophagus, bladder, and skin. We will use a well-established multistage, carcinogen-induced, cutaneous SCC mouse model and credential its’ representation of human cancer drug resistance. Mouse cSCCs display many genetic alterations seen in

human SCCs, including mutation of Ras, PI3 Kinase and Notch pathways. The model also incorporates the critical role played by non-mutagenic tumor promoting factors as cancer drivers. This proposal will build on our previous work, accessing our extensive in-house mouse tumor genomics and transcriptomics databases. In

Aim 1, we will use single cell analyses of primary papillomas, carcinomas and metastases induced by chemical carcinogenesis in situ and analyzed before and during chemo- or immunotherapies. We will use our novel biocomputational Metagene approach to identify rewiring of transcriptomic networks within single tumor cells

after therapy. Single cell analyses, namely scRNAseq, CyTOF, MIBI, and FISH, will be combined with our in- house developed analytical tools, to identify high plasticity state tumor cell populations enriched or depleted in response to therapy and their molecular and spatial relationship to other cells and structures within the tumor.

In Aim 2, we will test the fidelity of the chemical carcinogenesis model as a robust representation of human cSCC biology by undertaking longitudinal validation studies of fresh human cSCC tissue collected before and during chemo- or immuno-therapy and analyzed using the same technology. In Aim 3, we will empirically test

the function of candidate genes (Metagene components) expressed in the high plasticity state, for their contribution to drug resistance. CRISPRi/dCas9 and CRISPR/MultiCas12a technology will be used to test gene activities during drug therapy by single or combinatorial gene knockdown in syngeneic tumor models in vivo.

Our strategy will credential the use of the skin chemical carcinogenesis system to model features of human cancer drug resistance. The project is responsive to PAR-23-281, as it undertakes cross-species discovery of the molecular basis for development of drug resistant tumor cells. The knowledge to be gained will contribute

to discovery of biomarkers and therapeutic targets for this drug resistant cell population. Our computational methodology and our extensive transcriptomic data from hundreds of tissue and tumor samples will be shared with the Oncology Models Forum NCIP Hub.