Abstract 2180: Immune, keratinocyte, and fibroblast expression signatures distinguish HNSCC samples by HPV status

Abstract Introduction Head and neck squamous cell carcinoma (HNSCC) is the sixth leading cancer in the world. HPV-positive HNSCC is considered a distinct molecular entity with a better prognosis than HPV-negative. To better understand the mechanisms underlying the clinical and molecular behavior between HPV-positive and HPV-negative HNSCC we applied bioinformatics approaches to transcriptomic data to identify signature patterns associated with 25 different cell types, and characterize cell type heterogeneity by HPV-status. Materials and methods We analyzed 18 HPV (+) and 18 HPV (-) samples collected at University of Michigan hospital from 36 individuals with HNSCC; all were determined to be >70% cellularity by a pathologist. Gene expression was evaluated using Illumina HiSeq RNA-seq. Differentially expressed genes were identified using TopHat, HTSeq, and then edgeR, resulting in 559 and 873 up in HPV(+) or HPV(-), respectively, with FDR < 0.05 and fold change > 3. Cell type-specific expression patterns were identified using > 700 microarray samples from 25 relevant cell types. Using this database and the list of genes differentially expressed by HPV status we identified which cell type(s) contributed the strongest expression pattern by calculating cell type-specific enrichment scores as a weighted Kolmogorov-Smirnov-like statistics against the top ranked 250 genes. Results Significantly higher estimates of cell type composition in HPV(+) samples were observed for T lymphocytes (CD3+ T-cells, CD4+ T-cells, CD8+ T-cells, CD34+ T-cells, γδ T-cells, regulatory T-cells), NK-cells, B cells, and monocytes (p-value < 0.0004). Conversely, endothelial cells, epidermis, dermis, keratinocyte, and fibroblast all had significantly higher estimates of cell type proportion in HPV(-) samples (p-value < 0.0004). For validation, we compared these results to results calculated in an identical way using TCGA HNSC RNA-seq data (39 HPV(+) vs 277 HPV(-) samples) and found that the largest differences in cell types higher in HPV(+) were shown also in T lymphocytes, NK-cells & B cells, and that dermis, fibroblast, epidermis & keratinocyte cell types were again estimated to be higher in HPV(-) samples (p-value < 0.0004). Conclusion Differential expression in HNSCC by HPV status can partly be explained by differences in the tumor cellular composition, and this in turn will help to focus on differential expression within the tumor cells themselves. Our results predict that T lymphocytes and other cell types responsible for the adaptive immune response contribute a larger portion of the cellular composition in HPV(+) tumors compared to HPV(-). Conversely, keratinocytes and fibroblasts are predicted to make up a larger portion of cells in HPV(-) tumors. Surprisingly, the stem cell expression signature was not significantly different by HPV-status, even though HPV(+) tumors tend to be less differentiated than HPV(-) tumors. Citation Format: Lada A. Koneva, Yanxiao Zhang, William R. Swindell, Shama Virani, Alisha Virani, Thomas E. Carey, Laura S. Rozek, Maureen A. Sartor. Immune, keratinocyte, and fibroblast expression signatures distinguish HNSCC samples by HPV status. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2180. doi:10.1158/1538-7445.AM2015-2180