The Human Tumor Microbiome

Introduction:

The tumor microenvironment has a major impact on multiple cancer hallmarks. The human body is host to a huge number of bacteria that are found at various sites of the human body and play major roles in many physiological processes such as host metabolism, function of the immune system and response to anti-cancer therapy. We recently demonstrated that bacteria are present in pancreatic adenocarcinomas and can contribute to chemoresistance through inactivation of the anti-cancer drug, gemcitabine. In this work, we profiled the microbiome of 8 cancer types in over 1800 human samples to fully describe the tumor microbiome landscape. We characterized the number and identity of intra-tumor bacteria and correlated them with the clinical characteristics of the tumors and the patients. In addition, we visualized bacteria to observe their spatial distribution and their relation to the tumor immune profile.

Material and method:

A total of 1824 human samples were acquired through collaboration with 11 medical centers and include tumors (and matched normal tissue) from breast, lung, melanoma, pancreas, ovary, colon, glioblastoma and bone tumors.

Methods for DNA extraction and bacterial 16S sequencing were developed to meet the low biomass nature of the tumor microbiome. Multiple measures were taken to reduce contamination and hundreds of negative samples were used to control the remaining contamination. Bacteria and immune cells were visualized in tumors using both IHC and multiplexed IF methods.

Results and Discussion:

While we detected and visualized bacteria in all tumor types examined, it seems that the presence and frequency of bacteria in tumors varies according to tumor type. We found that some bacterial species are tumor-type specific and that the microbiomes of tumors and their adjacent normal tissue is highly similar. Breast cancer tumors are distinct from other tumor types with a relatively higher number of bacterial strains per tumor sample. We also found statistically significant correlations between the identity of intra-tumor bacteria and clinical metadata such as age, smoking status, tumor stage, tumor genetic markers and response to both targeted and immune therapies.

Visualization of bacteria in the tumor tissues showed that they’re mainly intra-cellular in both cancer and immune cells.

Conclusion:

Extensive profiling of the tumor microbiome showed that bacteria are present in a wide variety of tumor types and that the tumor microbiome is distinct in composition and prevalence for the different tumor types and correlated with various clinical parameters.