A comprehensive analysis of breast cancer microbiota and host gene expression Original paper

What was studied?

This study examined the breast cancer microbiome using a dual-purpose analysis of RNA-seq data from The Cancer Genome Atlas (TCGA). The focus keyphrase breast cancer microbiota appears throughout the workflow because the study uniquely leveraged unmapped RNA-seq reads to characterize microbial signatures within breast tumors and adjacent non-cancerous tissue, while simultaneously relating microbial abundance to host gene expression. By analyzing 668 breast tumors and 72 adjacent tissues, the authors investigated how microbial communities differ by tissue type and cancer subtype, and whether specific microbes correlate with oncogenic pathways. The approach allowed microbial detection and transcriptomic analysis from the same tissue source, reducing confounding and enabling a highly integrated view of host–microbe interactions in breast cancer biology.

Who was studied?

The investigation used human breast tissue samples collected through TCGA. After standard exclusions—male patients, metastatic lesions, neoadjuvant therapy, or missing HER2 status—the final dataset included 668 breast tumors (ER+, HER2+, and triple-negative) and 72 adjacent non-cancerous tissues. A subset of six Mayo Clinic ER+ samples also underwent 16S-rRNA sequencing for validation. The cohort, therefore, represented a large, diverse breast cancer population and allowed subtype-specific microbiome comparisons while providing a healthy-tissue reference.

Most important findings

Microbiome profiling revealed 1,634 initial microbial operational taxonomic units (OTUs), narrowed to 327 consistently present taxa across tumor and adjacent tissues. Proteobacteria dominated tumors, whereas Actinobacteria were more abundant in adjacent tissue—a pattern visible in page-6 bean plots, where tumor distributions skew heavily toward Proteobacteria. Twelve of the twenty-four prevalent species demonstrated significant differential abundance, while nine species showed >2-fold change between tumors and adjacent tissue. These included Mycobacterium fortuitum, Mycobacterium phlei, Acinetobacter baumannii, and Escherichia coli. Microbial–host correlations identified three organisms with strong transcriptomic associations: Haemophilus influenzae, Neisseria subflava, and Listeria fleischmannii. As shown in the enrichment bar graph on page 9, H. influenzae correlated with proliferative pathways (G2M checkpoint, E2F targets, mitotic spindle) even after stringent correction, suggesting a link between microbial presence and tumor cell cycle activation. L. fleischmannii is associated with epithelial–mesenchymal transition genes, indicating a potential contribution to metastatic potential. The hive plot on page 9 visually demonstrates H. influenzae as the most connected species, sharing information with 226 genes.

Key implications

The study establishes that breast cancer tissues harbor distinct microbial signatures and that even low-abundance taxa may influence tumor biology. The dominance of Proteobacteria in tumors and Actinobacteria in adjacent tissues suggests environment-specific niche selection. Associations between H. influenzae and proliferative transcriptional programs imply that certain microbes may contribute to tumor progression. Meanwhile, L. fleischmannii links microbial presence to metastatic pathways. These findings reinforce the importance of incorporating microbiota into breast cancer research, risk assessment, and potentially therapeutic targeting as the field progresses toward microbiome-integrated oncology.

Citation

Thompson KJ, Ingle JN, Tang X, et al. A comprehensive analysis of breast cancer microbiota and host gene expression.PLOS ONE. 2017;12(11):e0188873. doi:10.1371/journal.pone.0188873