41P The functional GRHL3-FLG axis predicts targeted therapy response in head and neck cancer

Abstract

Head and neck squamous cell carcinoma (HNSCC) is a heterogeneous disease harbouring the most frequent hotspot mutations in the differentiation genes. Disruption of epithelial differentiation acts as a primary driver of HNSCC development and correlates with a poor patient's prognosis. To date, in addition to the non-selective conventional HNSCC treatments, such as chemotherapies and surgeries, the only FDA-approved targeted therapy Cetuximab, an epidermal growth factor receptor inhibitor, has a low response rate with considerable toxicity. Therefore, determining whether functional differentiation can serve as a molecular vulnerability and/or a predictor of targeted therapy in HNSCC is an area of valuable clinical need. A multi-omic approach integrating whole-genome and whole-transcriptome sequencing with drug sensitivity screening was employed in tumours from a spontaneous HNSCC murine model, HNSCC patient’s tumours, and established human cell lines to reveal potential predictive biomarkers for targeted therapies. CRISPR-Cas9-mediated gene knockout and activation validated candidate predictors in HNSCC cell lines treated with inhibitors of the PI3K/mTOR, c-MYC and STAT3 pathways, the key signalling in HNSCC oncogenesis. We identified a novel Grainyhead-like 3 - Filaggrin (GRHL3-FLG) differentiation axis as a predictive biomarker to targeted therapy response in HNSCC. A subset of HNSCC with functional GRHL3-dependent differentiation was the most sensitive to inhibitors of PI3K/mTOR, c-MYC and STAT3 signaling. Furthermore, we identified the GRHL3 transcriptional target gene FLG as a novel tumour differentiation gene and, more importantly, stratified HNSCC subsets as treatment-resistant based on their FLG mutational profile. Moreover, the loss of FLG in sensitive HNSCC cells resulted in a dramatic resistance to targeted therapies while the GRHL3hi-FLGwt signature predicted a favourable patient's prognosis. Functional differentiation (GRHL3hi-FLGwt) provides the first example of differentiation-dependent therapy response and establishes a rationale for clinical investigation of differentiation-paired targeted therapy that may improve outcomes in HNSCC and other heterogeneous cancers.