Illuminating biological pathways for drug targeting in head and neck squamous cell carcinoma

Abstract

Head and neck squamous cell carcinoma (HNSCC) remains a morbid disease with poor prognosis and treatment that typically leaves patients with permanent damage to critical functions such as eating and talking. Currently only three targeted therapies are FDA approved for use in HNSCC, two of which are recently approved immunotherapies. In this work, we identify biological pathways involved with this disease that could potentially be targeted by current FDA approved cancer drugs and thereby expand the pool of potential therapies for use in HNSCC treatment. We analyzed 508 HNSCC patients with sequencing information from the Genomic Data Commons (GDC) database and assessed which biological pathways were significantly enriched for somatic mutations or copy number alterations. We then further classified pathways as either "light" or "dark" to the current reach of FDA-approved cancer drugs using the Cancer Targetome, a compendium of drug-target information. Light pathways are statistically enriched with somatic mutations (or copy number alterations) and contain one or more targets of current FDA-approved cancer drugs, while dark pathways are enriched with somatic mutations (or copy number alterations) but not currently targeted by FDA-approved cancer drugs. Our analyses indicated that approximately 35-38% of disease-specific pathways are in scope for repurposing of current cancer drugs. We further assess light and dark pathways for subgroups of patient tumor samples according to HPV status. The framework of light and dark pathways for HNSCC-enriched biological pathways allows us to better prioritize targeted therapies for further research in HNSCC based on the HNSCC genetic landscape and FDA-approved cancer drug information. We also highlight the importance in the identification of sub-pathways where targeting and cross targeting of other pathways may be most beneficial to predict positive or negative synergy with potential clinical significance. This framework is ideal for precision drug panel development, as well as identification of highly aberrant, untargeted candidates for future drug development.

Fig 1. Identifying Targetable Pathways in the GDC Head and Neck Squamous Cell Carcinoma Cohort.

A) Somatic mutation, copy number alteration, and clinical data for patients with head and neck squamous cell cancer (HNSCC) are selected from the Genomic Data Commons. B) HPV status was annotated as described in the methods for a subset of the cohort. C) The genes with somatic mutations (or copy number alterations) for HNSCC (represented by green, gold or blue in (A) are then mapped onto the pathways in the Reactome database. D) An overrepresentation analysis is done, using the hypergeometric probability distribution, to identify Reactome pathways likely to be aberrant for HNSCC patients (represented by green, gold and blue pathways). E) Proteins and associated FDA-cancer drugs, from the Cancer Targetome database, are then mapped onto the aberrant HNSCC pathways. F) A pathway that contains a cancer drug-associated protein is then considered ‘light’, as designated by the bottom pathway in the box on the right. Pathways with no association to the Cancer Targetome are considered ‘dark’.

Fig 2

A. Top 20 mutated genes in the GDC HNSCC patient cohort (N = 507). We only included variants that had high or moderate impact, which were classified as Missense, Nonsense, Nonstop, Frame Shift Deletion, Frame Shift Insertion, In Frame Deletion, In Frame Insertion, Splice Site and Translation Start Site. We excluded variants with mostly low or modifying impact, which were classified as 3’Flank, 3’UTR, 5’Flank, 5’UTR, IGR, Intron, RNA, Silent, and Splice Regions. B. Top 20 copy number altered genes in the GDC HNSCC patient cohort (N = 296). We only included copy number alterations characterized by “-2” or “+2”, for high confidence deletions and amplifications, respectively.

Fig 3. Light and dark pathway coverage for HNSCC mutation and copy number enriched pathways.

Diagram of the subsets of Reactome pathways in HNSCC that are mutation-enriched or copy-number alteration enriched, with drug-targeted subsets labeled for each. Percentages shown in the figure are out of the total number pathways used for this analysis (1650). Of the HNSCC-specific mutation enriched pathways, 34% are targetable by FDA approved cancer drugs (light pathways) and 66% are open for drug development (dark pathways). Of the HNSCC-specific enriched pathways, 38% are targetable by FDA-approved cancer drugs (light pathways) and 62% are open for drug development (dark pathways.

Fig 4. Overlap of light and dark pathway coverage between data types.

Left Venn diagram shows overlap of mutation-enriched and copy-number enriched light pathways. Right Venn diagram shows overlap of mutation-enriched and copy number enriched dark pathways.

Fig 5. The pathway “Nephrin Interactions” is highly aberrant in the GDC HNSCC cohort and Light to FDA-approved Cancer Drugs.

Nephrin Interactions is an example of a top-ranked light pathway, defined as a pathway highly covered with HNSCC mutations (~100%) and frequently mutated in the patient cohort (36%). Nodes in green are genes mutated in the GDC HNSCC cohort and nodes in yellow are mutated in the cohort as well annotated as targets for cancer drugs in the Cancer Targetome. Drugs are indicated by red diamonds and represent FDA-approved cancer drugs with targets in this pathway. For drug-target interactions shown here, we required supporting binding assay evidence to be <1000nM.