Research Projects

Mutant KRAS has become the focus of drug development in recent years, culminating in the approval of small molecules targeting KRAS G12C. While these and other targeted therapies have extended survival in a subset of patients, most progress within a few months. Using targeted exome sequencing and spatial transcriptomics on patient-matched biopsies, we showed that acquired genetic events at progression are frequently subclonal and coexist with transcriptional adaptive states. In contrast, cancer-cell-autonomous inflammatory programs are induced early and precede overt resistance. We identified TBK1 as a target to suppress this inflammatory adaptive phase and deepen responses to KRAS inhibition.

To resolve how resistance evolves over time, we have assembled cohorts of serial plasma and tissue collected before, during and after treatment with Food and Drug Administration- (FDA-) approved and investigational KRAS inhibitors. The goal is to capture both genetic and non-genetic adaptation in real time across different RAS-targeted therapies and throughout the treatment course.

Our data in patient samples and organoid models show that resistance to targeted and cytotoxic therapies is often driven by non-genetic adaptations, including YAP activation as well as regenerative and fetal reprogramming. We are using barcode lineage tracing and single-cell multiomics to determine whether non-genetic resistance arises from preexisting populations and whether diverse selective pressures reproducibly engage conserved escape mechanisms. Through the ASCEND-CRC program, we are profiling patients’ tumors throughout their treatment course with three primary objectives:

1. To deeply profile the molecular evolution of colorectal cancer (CRC) under treatment
2. To develop predictors of resistance
3. To investigate the impact of novel therapies targeting these acquired vulnerabilities in metastatic CRC

High-copy-number amplifications of RAS pathway genes (KRAS, EGFR, MET, BRAF) encoded on extrachromosomal DNA (ecDNA) are recurrently observed in colorectal cancer (CRC) tumors progressing on targeted therapies. We utilize patient-derived xenographs (PDXs) derived from clinical samples harboring ecDNA amplifications, as well as engineered organoids carrying ecDNAs at loci of interest, to determine the mechanisms underlying resistance and to identify acquired vulnerabilities associated with the presence of high-copy-number ecDNA amplifications.