Bacteria And Cancer: It's Not Random

 BY DEREK LOWE

I've written before here about the relationship between some types of cancer and associated bacteria in vivo, and here's a new paper with many more details on this phenomenon. There appear to be over thirty types of cancer that have now been associated with bacterial colonization, and it’s a very active field. As is always the case in the biology of disease, the more you look, the more you see, and in this paper the authors have studied two types of tumors (colorectal cancer and oral squamous cell carcinoma) in greater resolution than anyone has before. There are indeed bacteria associated with these, most notably Fusobacterium nucleatum, although the paper notes that the bacterial situation is rather heterogeneous across the patients surveyed. For example, the oral cancer samples showed anywhere from 1 to 42 different bacterial genera in them (!) with a median of 8, while the colorectal samples showed 1 to 31 genera with a median of 2. The oral cancer samples had roughly ten times as many bacterial RNA reads in them for the two dominant genera (Fusobacterium and Bacteroides) as compared to the colorectal ones.

All that is (at least to me) a rather startling variety of bacteria in such tissues, and I think the authors probably felt so, too. Further work on protein markers of inflammation, immune function, and tumor progression strongly suggested that these bacteria were mostly found in highly immunosuppressed microniche environments. It appears, for example, that T cells are largely excluded from regions with high bacterial colonization. Those regions were also notably less vascularized and had significantly lower expression of the p53 tumor-suppression protein. At the same time, though, these regions are significantly lower in cellular markers of tumor proliferation like Ki-67.

But there’s another level to this. There are a number of previous reports of intracellular bacteria in tumor cells, and these authors observed the same phenomenon. This could be reproduced in cell culture using the common HCT116 colorectal tumor cell line and three of the most common invasive bacterial species from the sequencing experiments above. The bacteria could be seen in the cytoplasm of the HCT116 cells by confocal imaging, and single-cell analysis indicated that their presence led to a distinct transcriptional signature in the host cells. Some of these are what you’d expect (markers of infection!) but there were also significant increases in proteins associated with metastatic behavior. Separate experiments backed this up - bacteria-associated cells in culture detached and migrated into collagen gel medium notably more quickly than others. This matches up with other reports in different tumor types.

Looking at fresh biopsy samples from colorectal cancer patients confirmed this - the dominant genera in cell-adherant and cell-penetrant bacteria were Fusobacterium and Treponema, although it has to be noted that (histologically) these were associated with macrophage cell clusters. It’s possible that these were macrophages in the process of engulfing and destroying these bacteria, but it’s also possible that they were successfully being invaded by them. That said, the widest variety of bacterial transcripts (outside of those two genera just mentioned) were in epithelial cells that appeared to have major chromosomal instability, flat-out transformed tumor cells, and this was confirmed by proteomic profiling on them.

So if you haven’t already, you’re going to have to modify your thinking about tumor heterogeneity. None of our earlier assumptions are wrong; they’re just incomplete. Cells in single solid tumors do indeed vary a great deal (as has been clear for some years now), and some (much?) of this is certainly due to their intrinsic genetic instability. Past that, their interactions with other human cells will certainly alter their expression profiles and behavior. And past that, we now see that interactions with bacteria (lots of bacteria) can and do also change things. Fuseobacterium nucleatum has had a lot of attention, but it’s well beyond that, and we now see that the microscopic patterns of bacterial invasion are very distinct and complex. As others have suggested, we may need to stop thinking of tumors as completely human tissues, but rather as mixtures of aberrant human cells and opportunistic bacteria - we’ll see how far that idea takes us.

https://www.science.org/content/blog-post/bacteria-and-cancer-it-s-not-random