Tumor and pan-tumor diversity and heterogeneity of cancer tissue microbiomes: a medical ecology analysis across 32 cancer types

Abstract

AbstractDiversity and heterogeneity are hallmarks of any ecosystems including cancer ecosystems. Tumor heterogeneities have been a hot spot in cancer research because of their critical roles in promoting clonal evolution and metastasis of cancer cells and in influencing cancer progression and therapy efficacy. Cancer tissue microbiome as part of tumor microenvironment can influence tumor heterogeneities both directly and indirectly through their “intimate” intracellular and intercellular interactions with human cells including tumor, immune and normal cells. From an ecological perspective, the relationship between tumor microenvironment and tumor heterogeneity is not unlike that between habitat heterogeneity and community heterogeneity. That is, their heterogeneities should be interwoven with each other, and then the net effects of the microbiomes on cancer development, malignant progression, and therapy responses can be either promotive or suppressive depending on the so-termed immune-oncology-microbiome axis or trio. The objective of this study is to assess and interpret the heterogeneity and often conflated diversity of tumor microbiomes. Our findings, from reanalyzing a big microbiome dataset originally distilled from the TCGA (The Cancer Genome Atlas) database by Pooreet al. (2020, Nature) including 16555 tumor microbiome samples from the primary tumor (PT), solid tissue normal (SN) and blood derived normal (B) of 32 cancer types, include: (i) The tumor microbiome heterogeneity (diversity) cancer relationship HCR (DCR),i.e.,the heterogeneity (diversity) differences between PTvs.SN (B) are only significantly in approximately 10%-40% depending on the cancer types. (ii) The pan-tumor HCR (DCR),i.e.,microbiome heterogeneity/diversity differences of same tissue type (e.g., PT) across cancer types (e.g., lung vs. breast cancers) are approximately twice the range of previous tumor-HCR (DCR) (i.e., 30%-80% for pan-tumorvs. 10%-40% of tumor scale). In both tumor and pan-tumor scales, the heterogeneity differences ranges are wider than the diversity ranges. (iii) The NSR values range between 0.4 and 0.8 and in 75% cases NSR>0.5, suggesting that tumor selection plays a dominant role than stochastic drifts in shaping microbiome diversity/heterogeneity patterns. Furthermore, the NSR values are significantly different between PT and NT (B) in 50%-100% (mostly 70%-80%) cases across 32 cancer types, further confirming that it should be the tumor growth that is largely responsible for the dominance of selection forces. Finally, we postulate that the HCR (DCR) should be dynamic with tumor types, progression, microbial taxa, host genomics and physiology, therapy and diets.