PROJECT SUMMARY
HNSCC is the sixth leading cause of cancer-related mortality. Most deaths are caused by metastasis after
treatment failure, but unfortunately, a firm understanding of the molecular pathways that drive treatment
resistance remains elusive. For example, it is unclear why the vast majority of tumor cells are successfully
eliminated by treatment, yet a few escape destruction. Are these cells in a privileged cell state that enables
evasion of these drugs? Or does resistance emerge adaptively upon treatment? Given the morbidity and
mortality associated with HNSCC, there is an urgent need to answer these questions, but this has been
prevented by two major obstacles. First, HNSCC tumors are highly heterogeneous, so bulk genomic methods
cannot discern subpopulations of cells that might give rise to resistant clones. Second, nearly all existing genomic
methods are destructive and require specimen lysis at the time of measurement. This “destruction upon
observation” has made it impossible to correlate molecular events that occurred in the past with the final fates of
the cells in which these events took place. To overcome these barriers, we have recently utilized single-cell RNA-
seq (scRNA-seq) to characterize heterogeneity among HNSCC tumors, defining a partial epithelial-to-
mesenchymal (p-EMT) program which predicts HNSCC outcomes (Puram et al., Cell). We have also developed
a novel single-cell ‘Calling Card’ (scCC) technology that can record the genome-wide interactions of any
transcription factor (TF), creating a permanent molecular memory of all binding events that occur at a given
moment or epoch (Moudgil et al., Cell). This allows transient molecular interactions to be captured non-
destructively and read out later (e.g. after drug treatment), allowing us to “go back in time” and determine which
cell states enabled a cell to resist treatment. We accomplish this by fusing any TF to the piggyBac transposase,
which bestows the TF with the ability to direct transposon insertion into the genome near where it binds. We will
use this technology to define the mechanisms by which HNSCC cells persist after cetuximab treatment and then
evolve to produce resistant clones, identifying genes and pathways that can be targeted by adjuvant therapies.
Specifically, we hypothesize there are subpopulations of tumor cells in a pre-existing p-EMT state that
confers immunity to drug treatment. To test this hypothesis, we will first use our molecular memory tool to
determine why some cells acquire p-EMT but others do not (Aim 1). This Aim is critical because EMT plays a
key role in the development of cetuximab resistance, and our extensive experience with p-EMT will allow us to
use this system to mature and benchmark our molecular memory tool. To probe pathways specific to cetuximab
resistance, we will utilize scCC to evaluate HNSCC lines with cetuximab resistance and sensitivity and record
both pre-existing and adaptive changes in cell state. We will validate these findings in vivo, establishing a set of
genes and molecular pathways responsible for therapeutic resistance and thereby revealing new targets to
overcome these mechanisms as well as biomarker predictors of treatment response.