Obesity, FMT vs cancer, 'dysbiosis' & microbial predation (reading list)

Comment: Very interesting study on T2D and gut microbiota links in the Swedish SIMPLER cohort of older adults (mean age: 73.9yo) followed up for 5.3 years for diabetes onset: ~383 participants who were free of T2D at baseline developed incident diabetes during follow-up. Interesting points are being made about functional links with risk: higher T2D risk is linked to higher asparagine degradation, lower T2D risk with mannose degradation and non-oxidative 5P pathway.

Comment: How much of the metabolic benefit for obesity/T2D after bariatric surgery is microbiome-mediated? In this study, authors look at 77 participants both before RYGB and SG surgeries and 12 months after and compare microbiome profoiles. Both types of surgeries improve weight loss and metabolic health and T2D remission at 12 months, but Roux-en-Y gastric bypass (RYGB) showed the largest microbiota alterations. Richness increased after both surgery types along with microbial butyrate production. A small caveat was that metformin use (known to affect a lot microbiota richness and structure) covaried strongly with remission status, so couldn’t be controlled well.

Comment: There’s a bunch of efforts being made to study the role of microbiome in cancer treatment, particularly immune checkpoint inhibitor therapy and the possibility that FMT could help deal with side effects. In this work, authors look at whether reproducible gut microbiome features are associated with response to ICI in advanced melanoma, and also whether these microbiome signatures can be generalised across cohorts and whether they differ between ICI-only treatment and ICI+FMT. This is a meta-analysis from 15 melanoma cohorts from 11 primary studies for ~760 samples from 484 individuals. Results are that there aren’t any single universal “good” or “bad” bacteria predicting response; however, ICI-only responders seemed to be enriched for SCFA producers. It all varied according to cohorts, so authors conclude that microbial responses to ICI still seem to be heavily context-dependent.

Comment: This is a very interesting paper if you are interested in what exactly people mean when they talk about “dysbiosis” (including the debate about it). In here, authors argue that gut microbiome network organization, not just microbial abundance, differs across metabolic health states in obesity. Metabolically healthy individuals had more cohesive and robust microbial networks, while metabolically unhealthy individuals had more fragmented and potentially “dysbiotic” networks. Very ecological message in essence, and not that surprising really: microbial community stability and connectivity may be important microbiome features associated with metabolic health, rather than just the presence/absence of single species.

Comment: Interesting 16S-based mechanistic mouse study looking at whether dietary cholesterol and saturated fat independently or synergistically reshape the gut microbiota, and thereby promote progression from MASLD/MASH to hepatic fibrosis. Authors observe a rapid expansion of Parasutterella alongside depletion of classic bile-sensitive or potentially beneficial taxa such as Bifidobacterium and Lactobacillus. The study links these community shifts to altered bile-acid chemistry, especially deoxycholic acid, and then shows that gut-derived material from diseased SPF mice can directly activate human hepatic stellate cells. Cool.

Comment: Here, authors ask whether people with a prior colorectal adenoma resection (~700 women + external validation on CRC datasets on ~2K people) retain long-lasting gut microbiome and fecal metabolome alterations more than a decade later (on average 12y after endoscopy/resection). Some of these alterations were resembling CRC-associated patterns, especially the depletion of F. prausnitzii and the enrichment of taxa such as F. plautii and R. gnavus. The metabolome also showed sphingolipid and BA disruptions, suggesting altered host–microbe metabolism. The study supports (without proving it directly) that a sustained gut microbiome–metabolome state even AFTER adenoma removal may still contribute to elevated CRC risk.

Comment: Very interesting review on how the human gut microbiota changes through life and what are the factors involved, with an original focus on diurnal rhythms of the microbiota across the lifespan (there is an estimated 10% of microbial species/20% of microbial pathways showing diurnal fluctuations). Food for thoughts when we think about experimental design and sampling strategies!

Comment: In the same series as above, a very interesting review on the spatial distribution of the human gut microbiome along the GI tract and how to study it. Very cool.

Comment: A new preprint seen on Bluesky but untested: a new simple statistical test of dysbiosis (or, as highlighted, any microbiome composition data relative to a reference). Apparently outperforms Bray-Curtis and other metrics while simultaneously being very simple to compute. Keen to hear feedback if you’ve used it!

Comment: Authors asked whether gastrointestinal microbial overgrowth is common in patients with confirmed celiac disease (n=256), and whether small intestinal bacterial overgrowth correlates with celiac disease activity as measured by Marsh histology score. The main takeaway is that microbial overgrowth is common in tested celiac disease patients, especially those with refractory disease, but it didn’t find evidence that SIBO is linked to Marsh histologic severity. In other words, SIBO may contribute to persistent symptoms but it shouldn’t be used as a direct marker or driver in treated celiac disease.

Comment: This very interesting trial (involving authors with a commercial interest in selling Akkermansia-based products) suggests that daily pasteurized A. muciniphila MucT supplementation may help reduce weight regain after diet-induced weight loss in adults with overweight or obesity. The main observed signal was a stronger weight-loss maintenance with accompanying preservation of insulin sensitivity. This is especially interesting because the intervention was a pasteurised non-viable mucin-associated bacterium and yet it still produced a measurable host metabolic effect. Also, there is some interesting ecological part on the fact that people with naturally lower levels of Akkermansia in their gut seem to respond better to the treatment, which is a good point for individual/personalised microbiome medicine.

Comment: This came to my attention a few months after its publication but this is a good old “dysbiosis” debate paper, this time quite important and in Science, no less! Here, authors ask whether gut dysbiosis can be characterised not just by which microbes are present or by diversity loss, but by the balance of ecological interactions among microbes. A much more sensible approach in my sense. They built a consumer-resource model of the gut microbiome which included lots of parameters and compared it to real metagenomic data and developped a new potentially interesting metric, the ecological network balance index (ENBI). One of the observations from this was that “healthy” microbiomes appear more competition-dominated, whereas “diseased” microbiomes show a stronger signature of positive cross-feeding interactions.

Comment: Interesting effort to define a bit better the FMT treatment approach for C. difficile and make it more scalable. Authors define a 15-strain live cocktail and compare it to a “same donor” FMT approach in a randomised trial. I must say I am surprised about the result (not in a bad way), which is that the manufactured strain cocktail still engrafted well and appeared clinically comparable to FMT in this small trial. To be followed.

Comment: Great summary of the state of microbiome research and some good future directions for the field, suggested here: define what a healthy gut microbiome means in a clinically useful way; standardise sampling, DNA extraction, sequencing, metadata collection and bioinformatic pipelines; use better-designed, adequately powered, longitudinal and multicentre studies; move from association to causality using mechanistic models; develop evidence-based microbiome diagnostics rather than commercial tests with unclear clinical meaning amd improve microbiome therapeutics through strain-level, functional and ecological understanding.

Comment: Interesting premises for a study here: authors looked at whether the COVID-19 pandemic and the associated changes in hygiene, social contact, lifestyle, and environmental exposure, altered gut microbiome development during infancy. For this, they used a prospective longitudinal birth cohort study (LucKI Gut cohort, Netherlands, n=139 infants). This cohort began before the pandemic and continued through it, allowing comparison of samples collected before and during the pandemic within the same broader study framework. They found that normal infant microbiome maturation was still observed but that the pandemic measurably affected infant gut microbiome composition, slightly (explaining 2.7% of microbiome variation at 11 months old). Pandemic-era samples also somewhat showed altered species abundances (44 species differed significantly). All in all, a few big caveats, and some mild possible effects, but an interesting thing to have looked at with an opportunistic cohort.

Comment: Interesting to see this review in Nature Medicine, for a more clinical audience. It discusses very interestingly how climate change is expected to alter the distribution and burden of infectious diseases. Authors focus on vector-borne pathogens, waterborne/environmentally transmitted pathogens and respiratory pathogens. As expected, temperature, humidity, precipitations etc. can affect vectors and influence pathogen development, and/or amplify outbreaks. The take-home message is that while climate-linked changes are probably very nonlinear and context-dependent, climate change will likely alter where/when ID occur rather than how (changes in vector range, changes in seasonality, following demographic changes, etc.)

Comment: Resistance to predation is one of the most important selective pressure on bacteria, and may help explain why microbes evolve traits often studied in other contexts, such as biofilm formation, surface remodelling, stress resistance, virulence-associated phenotypes, and survival inside host-like phagocytic environments. This review is one of the best I’ve seen summarizing all types of predation that bacteria can encounter (very impressive Figure 1!).

Comment: I liked going through this review which connects abstract ecological theory to the kinds of tools microbiologists actually use. Table 1 is especially good at showing the experimental or computational approaches with the ecological concepts they can test. The paper also makes a strong case that microbial systems are not merely examples to which ecology can be applied, but powerful model systems for testing ecological theory itself, because microbes allow fine control over spatial structure, dispersal, environment, interactions, and time. Interesting considerations if you’re interested in ecological theory and microbiology.

Comment: It was about time that research on endometriosis is picking up the pace. New impressive GWAS on ~1.4 million women (105k caes/1.2m controls) identifying 80 loci linked to endometriosis risk (37 new compared to previous GWAS). Highlighting SYNE1 as significant across all clinical definitions. Authors investigate drug repurposing which highlights potential breast-cancer related agents and progestins to be investigated. PRS was investigated and gives early results. Overall, the study supports endometriosis as a genetically complex, systemic disease and provides some early molecular support for several possible mechanisms. There isn’t any likely single disease pathway but multiple biological processes interacting.

Comment: BMI is problematic for many reasons, but it’s still a way we have to look at population-level trends. But do we have to? In this preprint, authors examine the genetics of clinical obesity, a newer obesity definition intended to capture excess body fat plus dysfunctional adiposity beyond BMI alone; and then they compare genetic signals for clinical obesity with those for BMI. Result: 127 independent loci associated with clinical obesity, of which 63 did not share significant association with BMI, suggesting that clinical obesity captures genetic biology not captured by BMI. Interestingly, one of the most discordant variant was rs15285, located in LPL, with stronger association for clinical obesity than BMI. The variant colocalised with LPL expression QTLs and was associated with higher triglycerides, plus proteomic markers linked to insulin resistance and inflammation. Authors then did a clinical obesity PRS and compared it with a BMI PRS and found stronger associations with cardiovascular risk factors, reclassifying 35% of individuals from their BMI-based polygenic risk.

Comment: If you’ve lived under a rock in the last 5 years, there is an ongoing revolution with the currently available obesity/T2D/weight-loss drugs. This is an interesting review of where obesity drugs could be going after GLP-1/GIP.

Comment: Useful new resource (>24k participants between 2021-2024), and potentially very good to compare with All of Us, UKB and other national-scale genomic medicine efforts for (much needed!) added diversity and representativity. The paper shows what we all know (but very good to see), which is how much clinical interpretation changes when reference databases are built around the population actually being examined.

Comment: Probably worth mentioning for the sample size alone. Very impressive study aiming to identify genetic variants associated with 249 plasma metabolic traits, including common, low-frequency, and rare variants, using datat from >600k individuals (~185k Estonian Biobank + ~435k UKB) (unfortunately a big caveat again, >97% samples are mainly European). A lot of very interesting findings including very promising causal inferrences as you could expect, but overall the fact that even for well-studied circulating metabolic traits, much larger GWAS sample sizes still reveal many new genetic associations and biological insights. Very interesting caution too, that very large GWAS datasets reveal extensive pleiotropy, which means that MR signals should not automatically be interpreted as direct causal effects of metabolites on disease.

Comment: The realisation that microplastics are everywhere and found in every body site imaginable can be quite daunting and discouraging. This study (observational data + randomized intervention trial) shows that carefully controlled low-plastic food and household items exposure can reduce several urinary phthalate and bisphenol markers within only one week, with clearer effects on particular types of plastics. The study does not prove any health benefits, but it is experimental evidence that everyday plastic-associated chemical exposure is modifiable through diet, food packaging, kitchenware and cosmetic practices.

Comment: Not really about the microbiome but potentially interesting. This paper is looking at not just whether obesity increased globally, but how fast obesity prevalence changed over time in different countries. Authors focus on “velocity”: the annual absolute change in obesity prevalence and separately analyse children/adolescents and adults >20yo. It will be interesting to look at the same thing in 10 years given current progresses in obesity treatment.

Comment: Making the rounds in the field of clinical AI lately. Very good perspective with the the core point being simple but important: uncertainty communication should be designed into AI systems, not patched on afterward.

Comment: Too little (or too long) sleep is bad for your health. But how bad? In this great systems biology paper leveraging biobank-scale data, authors treat sleep as a whole-body aging variable rather than just a lifestyle phenotype and examine the scale of changes at various phenotypic levels. The results are converging on a U-shape sleep-aging pattern, with 6-8h/night being a healthy reference. More interestingly, authors connect these biological clocks to disease endpoints, mortality and depression.

Comment: A nice letter to the editor in Nature Medicine on the nomenclature change of PCOS to PMOS and how this will impact (positively) many women lives. Good brief analogy too with NAFLD which became MASLD, highlighting the more complex metabolic origin of the liver diseases.

Comment: Not microbiome/microbiology but putting this here because of obvious diet and metabolic links. This is mirroring my recent reading of the “Ultra-Processed People” book by Chris van Tulleken, which I recommend. In this study, authors follow at ~50k people in the US from 1999-2018. I’ll just copy-paste part of the abstract here, it will be enough:
“Every 10% of energy supplied from UPFs was associated with higher body mass index, HbA1c, diastolic blood pressure, total-to-high-density lipoprotein cholesterol (HDL-C); lower HDL-C and low-density lipoprotein cholesterol (LDL-C); greater prevalence of metabolic syndrome (odds ratio [OR] 5 1.07; 95% confidence interval [CI] 5 1.05, 1.09), diabetes (OR 5 1.03; 95% CI 5 1.00, 1.07), and cancer (OR 5 1.05; 95% CI 5 1.02, 1.08); and higher risk of all-cause mortality (hazard ratio 5 1.04; 95% CI 5 1.02, 1.07). When we adjusted for i.FCS, associations were only partly attenuated, remaining significant. By comparison, adjustment for saturated fat, added sugar, or sodium had little effect. Findings were consistent in population subgroups, except for stronger associations among lower-income adults.”
As a side note, it is quite astounding to see the backlash online against any epidemiology study linking UPFs with poorer health. I don’t exactly know what’s behind this torrent of (badly framed) arguments, or what even is implausible to some people (it seems very plausible that foods containing many additives used for processing only with no nutritional value are not being part of a healthy diet). However, it does seem like a hot topic for the wrong reasons (and actually more of a political rather than scientific one).