Correlation between intra-tumor microbes with tumor-infiltrating lymphocytes across all cancer samples.
Transcriptome sequencing (RNAseq) is a common method for understanding the gene-level regulation of tissues. The OSU Comprehensive Cancer Center-James (OSUCCC) regularly performs RNAseq on biopsied tumor tissues to support research that seeks to better understand how genes are dysregulated in cancer and how that might relate to patient outcomes. It has recently been discovered that in addition to the human transcripts found in tumor biopsies, there are transcripts that align to “exogenous” sources, including bacteria, fungi, viruses, and others. We sought to explore to the extent to which exogenous sequences could be found in tumor biopsy RNAseq data and how it might relate to other features of the tumor including the inferred immune cell composition. Hence, RNAseq data of tumor biopsies were retrieved from patients that had underwent treatment at OSUCCC. These included a total of 480 tumor samples that consist of 202 samples of lung cancer, 118 samples of sarcoma, 104 samples of bladder cancer, 20 samples of colorectal cancer, 20 samples of renal cell carcinoma and 16 samples of melanoma. Cleaned RNAseq data were aligned to exogenous genomes using Kraken2/Bracken and to the human genome using TopHat2. Human gene counts were deconvolved to relative abundances of immune cells using CIBERSORT. An average of 99.87% of the reads aligned to the human reference genome, across all cancers. T-cells and M1 macrophages were significantly enriched in lung cancer (p-value<0.0001). M2 macrophages were enriched in sarcoma (p-value<0.0001). Within the tumor biopsy RNAseq were reads that aligned to 1328 genomes of bacteria, archaea, viruses and fungi. The most abundant phyla were Proteobacteria, Firmicutes and Actinobacteria, with 39.5%, 20.1% and 5.16% of the total exogenous signal, respectively. Across all cancers, the relative abundance of the CD8+ T-cells positively correlated with Acinetobacter junii (p-value=0.035) and the relative abundance of NK-cells positively correlated with Acidipropionibacterium virtanenii (p-value=0.035). These results suggest a mechanism by which intra-tumoral microbes may affect tumor-infiltrating immune cells and therefore cancer development and/or treatment outcomes.
Keywords:
Cancer, deconvolution, intratumoral microbes, RNAseq, lymphocytes
Poster (ExoTCC) #SURF2020
link to pdf file: surf poster.pdf
Poster Presentation Video (ExoTCC) #SURF2020