re: microbiome + therapeutics, I like this finding: https://www.science.org/doi/10.1126/science.aau6323. Thinking about microbiome effects on drug metabolism for already understood drugs seems more actionable to me, at least in the short term.
It's definitely a cool idea, agree on it being more actionable! I was looking into prodrugs a lot recently, and there are also a few examples there of prodrug -> active drug activation via microbiome interactions (https://pubmed.ncbi.nlm.nih.gov/6147110/)
Regarding the cancer microbiome controversy, I am on the side of the skeptics. The issue of mis-identifying human sequences as bacterial (due to human DNA contamination in bacterial reference genomes) is a big problem, and the authors of the original study didn't properly filter out human-derived sequencing reads. That error, by itself, invalidates the paper.
Also, I don't love the conflation of using fecal microbiota transplant (FMT) for gastrointestinal issues and curing anxiety or depression. If FMT stops you from having diarrhea while commuting, you'll definitely experience lower anxiety, but that doesn't mean the FMT affected your anxiety through the gut-brain axis. That's an issue with both studies you linked to.
As someone who actively works in the tumor microbiome space, I’d love second how hard it is to parse out real signals! Our bioinformatics data are always correlational at best (this microbiome is related to worse response to immunotherapy, etc.) and so many groups have shifted to multidisciplinary methods to: 1) affirm an intratumor microbiome even exists, 2) attempt to quantify an “abundance” of intratumoral microbes, 3) more confidently assert that microbes in tumors can affect tumor biology, and 4) start trying to take a stab at causal relationships between intratumor microbes and cancer growth.
The main ways our group and others have done this is through 1) mouse modeling w/ introduction of microbes of interest, 2) tumor culturomics where we actually grow microbes out of sterile human or mouse tumors, 3) immunostaining of tumor samples to physically visualize microbes inside of tumors, and 4) qPCR of microbial reads in tumors to get at “microbial burden”.
Through these methods and others, we are working to more confidently support that intratumoral microbes are causal actors in the microenvironment!
very much appreciated this piece — i worked in a microbiome lab as part of a public health phd. outwardly optimistic (this was the peak of the hype), but private conversations always ran along the lines: this doesn't replicate, how can we spin this, etc. etc. I left thinking this may be an interesting diagnostic at some point in the future, but not a site of an intervention, with rare exceptions (eg c diff)
Nice post - learned a lot.
re: microbiome + therapeutics, I like this finding: https://www.science.org/doi/10.1126/science.aau6323. Thinking about microbiome effects on drug metabolism for already understood drugs seems more actionable to me, at least in the short term.
It's definitely a cool idea, agree on it being more actionable! I was looking into prodrugs a lot recently, and there are also a few examples there of prodrug -> active drug activation via microbiome interactions (https://pubmed.ncbi.nlm.nih.gov/6147110/)
Regarding the cancer microbiome controversy, I am on the side of the skeptics. The issue of mis-identifying human sequences as bacterial (due to human DNA contamination in bacterial reference genomes) is a big problem, and the authors of the original study didn't properly filter out human-derived sequencing reads. That error, by itself, invalidates the paper.
Haven't gotten the chance to do a full read through yet, but liking it so far!
One thing that I think is important, though, is to specify that most of your generalizations about the microbiome are about the human gut microbiome, with the exception of your aside about the tumor microbiome. Other microbiomes (oral, vaginal, skin) and other species have their own issues. Wrote a bit about the vaginal microbiome myself (https://trevorklee.substack.com/p/bacterial-vaginosis-we-contain-multitudes) as well as ground squirrel microbiomes (https://trevorklee.substack.com/p/ground-squirrel-microbiomes-are-neat).
Also, I don't love the conflation of using fecal microbiota transplant (FMT) for gastrointestinal issues and curing anxiety or depression. If FMT stops you from having diarrhea while commuting, you'll definitely experience lower anxiety, but that doesn't mean the FMT affected your anxiety through the gut-brain axis. That's an issue with both studies you linked to.
As someone who actively works in the tumor microbiome space, I’d love second how hard it is to parse out real signals! Our bioinformatics data are always correlational at best (this microbiome is related to worse response to immunotherapy, etc.) and so many groups have shifted to multidisciplinary methods to: 1) affirm an intratumor microbiome even exists, 2) attempt to quantify an “abundance” of intratumoral microbes, 3) more confidently assert that microbes in tumors can affect tumor biology, and 4) start trying to take a stab at causal relationships between intratumor microbes and cancer growth.
The main ways our group and others have done this is through 1) mouse modeling w/ introduction of microbes of interest, 2) tumor culturomics where we actually grow microbes out of sterile human or mouse tumors, 3) immunostaining of tumor samples to physically visualize microbes inside of tumors, and 4) qPCR of microbial reads in tumors to get at “microbial burden”.
Through these methods and others, we are working to more confidently support that intratumoral microbes are causal actors in the microenvironment!
very much appreciated this piece — i worked in a microbiome lab as part of a public health phd. outwardly optimistic (this was the peak of the hype), but private conversations always ran along the lines: this doesn't replicate, how can we spin this, etc. etc. I left thinking this may be an interesting diagnostic at some point in the future, but not a site of an intervention, with rare exceptions (eg c diff)