Your Gut Is the Translator

May 15, 2026 ·

Thaena Inc.

High protein. Intermittent fasting. GLP-1 drugs. All three have real evidence. All three have changed people's lives. And all three quietly depend on a variable nobody is talking about, which is whether your gut microbiome can do the translating. A new paper in Nature from March, plus a foundational paper from 2004, together suggest that when the microbes are missing, the same dietary signal that should produce adaptive fat burning produces something else entirely. Cortisol.

Lit Review Friday · Learn Something with Thaena · Episode 24 · Published 2026 · Reading time: ~15 minutes

The Paper Everyone on LinkedIn Is Talking About

The week of May 13, 2026, two papers landed in Nature Medicine on the same day, and one of them caught the wellness internet on fire. Sarah Mount and colleagues at Maastricht ran a 32-week randomized controlled trial. People with overweight or obesity went through eight weeks of a low-calorie diet, then continued for twenty-four weeks with either pasteurized Akkermansia muciniphila, a heat-killed gut bacterium, or a placebo. The placebo group regained about three kilograms. The Akkermansia group regained about one. The primary endpoint hit at p=0.012. The dead bacterium did most of the work.

That paper deserves its own conversation, and we will come back to it. But the reason it works, the mechanism that explains why dead bacteria can do anything useful at all, sits in a different paper from two months earlier. A Nature paper from March 2026 by Tanoue and colleagues, plus a foundational paper from 2004 by a Japanese physiologist named Nobuyuki Sudo. When you read those two together, the picture they paint is bigger than weight maintenance.

The story is about how your body decides what to do with food, what to do with the absence of food, and what to do with the pharmacological version of the absence of food. The story is that those decisions are not entirely yours. Microbes are doing most of the translating between what you eat and what your metabolism makes of it. And when those microbes are missing, the same dietary signal that should produce productive adaptation produces something different. It produces stress.

For the broader context on how the microbiome communicates with the host through chemistry, see Chapter II: The Mechanism in the Science Hub.

What Tanoue Actually Did

The Tanoue paper was a collaboration across Keio University in Tokyo, the Broad Institute, Beth Israel Deaconess in Boston, and several other institutions. The author list includes Kenya Honda, Ramnik Xavier, and Shingo Kajimura, three of the heaviest hitters in microbiome biology and adipose tissue research. The question they asked was simple. What happens if you drop the protein in a mouse's diet way down and watch the white fat under the skin?

Standard mouse chow is about twenty percent protein. They took it down to seven percent. Within a few weeks, the inguinal white adipose tissue, the standard storage fat under the skin, began to transform. It started building mitochondria, the cellular furnaces that burn fuel and release heat. It started expressing UCP1, the marker protein of brown fat, the kind of fat that exists specifically to dissipate energy rather than store it. The browning effect was on the same magnitude as you would see if you put the mouse in a cold room or hit it with a beta-adrenergic drug. Diet alone, no cold, no pharmacology, induced a metabolic adaptation that previous biology said required external stress.

Before we go further, two things to be careful about. This was a mouse study, not a human trial. And it is not a paper telling you to eat seven percent protein to lose weight. Protein still matters for muscle, for satiety, for a hundred other things. The point of the experiment was not to design a diet. The point was to ask a mechanistic question. When the body senses dietary scarcity, what physically happens, and who is actually doing the sensing.

The answer, when they repeated the experiment in germ-free mice raised without any bacteria, was that the host was not doing the sensing alone. Same diet, same seven percent protein, zero browning. The white fat just sat there. Mouse cells alone could not translate the low-protein cue into the metabolic response. The bacteria were the translators.

Two Bacterial Messages, Both Required

What the bacteria did, in response to low-protein feeding, was send two parallel chemical signals along two different routes. Both signals had to arrive for the browning response to fire. If either one was missing, the adipose tissue stayed white. The researchers called this a non-redundant requirement. Engineers would call it a two-key launch.

Pathway 1: bile acids talking to fat stem cells

The first signal route went through bile acids. Your liver produces primary bile acids in response to feeding, which then travel through the gut, where bacteria transform them. The transformation matters. In 2020, a landmark Nature paper from Robert Quinn and Pieter Dorrestein's lab at UCSD showed that gut bacteria do something extraordinary to bile acids: they conjugate them with amino acids, producing entirely new compounds that we did not know existed. Phenylalanocholic acid. Tyrosocholic acid. Leucocholic acid. These microbially-conjugated bile acids are found throughout mammalian tissues, including in humans, and they activate the FXR receptor.

FXR, the farnesoid X receptor, sits on cells across the body. In Tanoue's experiment, the relevant FXR receptors were on adipose progenitor cells, the stem cells inside fat tissue that decide what kind of fat cell to differentiate into. When the bacterially-modified bile acids bound to FXR on these stem cells, the signal told them, when you mature, become a furnace, not a warehouse. Mice with the FXR receptor knocked out of their adipose progenitors did not produce the browning response. The receptor was load-bearing.

Pathway 2: bacterial ammonia talking to the liver

The second signal route was stranger and, in some ways, more surprising. When the protein content of the diet drops, the gut faces a nitrogen famine. Bacteria need nitrogen to build their own cellular structures, and amino acids from dietary protein are the usual source. Some gut bacteria carry an ancient gene called nrfA, which encodes an enzyme that lets them scavenge nitrogen by reducing nitrite to ammonia. This is the kind of chemistry that happens at the edge of life on Earth, in soil microbes and deep-sea bacteria. It also happens in your gut, under the right conditions.

When the bacteria produced ammonia in response to the nitrogen famine, that ammonia seeped from the gut into the portal vein, the direct blood highway between the intestine and the liver. The liver sensed the ammonia and responded by secreting a hormone called FGF21. FGF21 is having a moment in metabolic pharmacology right now. Akero Therapeutics, 89bio, and several other companies are racing to develop FGF21 analogs as drugs for metabolic dysfunction-associated steatohepatitis and obesity. The Tanoue paper shows that your gut microbiome already induces FGF21 endogenously, when given the right cue.

FGF21, once secreted, stimulated the growth of sympathetic nerve endings into the white adipose tissue. New nerve wiring grew into the fat. Those nerves released neurotransmitters that activated the UCP1 thermogenic program. The fat cells, now wired and primed, began to burn.

The researchers proved the ammonia step with one of the most elegant experiments in the paper. They added a tiny amount of tungsten to the mice's drinking water. Tungsten is chemically similar to molybdenum, the metal that powers the nrfA enzyme, but it does not actually work in the enzyme's active site. The bacteria took up the tungsten, inserted it into their nitrogen-scavenging machinery, and the machinery jammed. The ammonia production stopped. The liver stopped releasing FGF21. The nerves did not grow. The browning did not happen. One element substitution, eight steps downstream, an entire metabolic adaptation prevented.

For more on bacterially-derived molecules and how the microbiome influences adipose tissue and insulin sensitivity, see the Metabolic Health topic page in the Science Hub.

Why Fasting Needs the Bugs Too

Tanoue's low-protein experiment is one version of a more general question. What happens when the gut senses dietary scarcity. And the same question shows up in the intermittent fasting literature, with a similar answer.

In 2020, Zhigang Liu and colleagues published a paper in Nature Communications on intermittent fasting in diabetic mice. They put the mice on a twenty-eight-day intermittent fasting protocol and watched what happened to cognition. The fasting worked. Hippocampal mitochondrial function improved. Cognitive impairment alleviated. Standard story so far.

Then they tried the same protocol after wiping out the gut microbiota with antibiotics. The benefits of intermittent fasting partially disappeared. The fasting still happened, but the metabolic and cognitive payoff did not arrive. To test the alternative side of the same hypothesis, the researchers administered specific bacterial metabolites to mice without putting them through fasting at all. Three-indolepropionic acid. Serotonin. Short-chain fatty acids. Tauroursodeoxycholic acid. The mice receiving the metabolites alone, without skipping a single meal, showed the cognitive improvements that the fasted mice had shown.

The implication is the same as Tanoue's. The empty stomach is the cue. The bacteria are the messengers. The metabolites are the message. And if the messengers are not there, the cue still happens but the message never gets delivered.

This reframes what we are doing when we fast. We are not really choosing to skip food. We are choosing to send a signal. Whether that signal gets translated into productive metabolic adaptation depends on whether the translators are there to receive it.

When the Microbes Are Missing, What's Left Is Cortisol

Here is the part that has been turning over in my head for weeks. If dietary scarcity becomes productive adaptation when the microbes are there to translate, what happens when the microbes are not there. The dietary signal does not just disappear. It goes somewhere. And the foundational evidence on where it goes comes from a paper published more than twenty years ago, from a quiet corner of Japanese physiology.

In 2004, Nobuyuki Sudo and colleagues at Kyushu University published a paper in the Journal of Physiology that should have been one of the most-cited microbiome papers of the early modern era. They took germ-free mice, mice raised without any bacteria at all in completely sterile isolators, and put them through a standard restraint stress test. They compared the cortisol and ACTH response to the same test in conventional mice with normal microbiomes.

The germ-free mice had a dramatically exaggerated stress response. Their cortisol and ACTH levels spiked far higher than the conventional mice for the same stressor. They also had reduced brain-derived neurotrophic factor expression in the cortex and hippocampus. The hypothalamic-pituitary-adrenal axis, the body's central stress-response system, was uncalibrated. The brake was missing.

Then Sudo did the rescue experiment. He reconstituted the germ-free mice with a single bacterial species, Bifidobacterium infantis. The exaggerated stress response normalized. One bacterium, restoring the brake on cortisol. The microbes were not just helpful for stress regulation. They were necessary to develop a normal stress response in the first place.

The practical implication is uncomfortable. If you fast, or you drop your protein, or you take a GLP-1 drug that mimics fasting, and you have a low-diversity gut, you may not get the metabolic adaptation. You may just get the cortisol response. The same intervention that gives an adaptive payoff to one person may give a stress response to another, depending on what the person's microbiome is bringing to the conversation.

The Sex Differences Nobody Has Measured Properly

If the response to dietary scarcity is microbiome-dependent, and microbiomes vary across people, then the response to popular interventions should vary too. The published data is starting to show this in humans, although most of the popular fasting and protein literature was built on study populations that were either mostly male or did not control for menstrual cycle phase.

In 2024, Dimitrios Kapogiannis and colleagues at the National Institute on Aging published a randomized trial in Cell Metabolism. Forty cognitively intact older adults with insulin resistance went through an eight-week intervention, either 5:2 intermittent fasting or a healthy-living diet. Both diets improved metabolic and brain biomarkers. The intermittent fasting group lost more weight. But buried in the abstract is the sentence that matters for this conversation. "Sex, body mass index, and apolipoprotein E and SLC16A7 genotypes modulated diet effects." Sex modulated the response. The same intermittent fasting protocol did different things to men than to women in the same trial.

And the question that keeps coming up in the women's-physiology discourse is whether the menstrual cycle phase matters. The intuitive story, the one that circulates in the Stacy Sims and adjacent literature, is that the luteal phase is when fasting becomes dangerous because cortisol baseline is already elevated and the body is more sensitive to stress.

Here is the awkward part of the discourse-versus-data question. The popular framing appears to have outrun the primary literature. A 2020 meta-analysis by Ajna Hamidovic and colleagues in Frontiers in Endocrinology, pooling cortisol levels across 778 women, found cortisol slightly higher in follicular than luteal, not the other way around. The effect size was small (Hedges' g = 0.13), which is the honest statement that baseline cortisol is roughly similar across phases, with a small follicular tilt rather than a luteal tilt. And the one small crossover study that has directly tested fasting across cycle phases, Ohara and colleagues in 2015 in BMC Women's Health, found that twelve hours of fasting in the luteal phase actually lowered salivary cortisol and raised parasympathetic activity. n = 7, but it is the only direct test.

The honest summary is that the randomized trial which would settle this, healthy regularly-cycling women randomized to fasting initiated in follicular versus luteal phase with cortisol area-under-curve and metabolic outcomes as primary endpoints, has not been run. What we do have is small, but it points in a different direction than the popular framing. The estrobolome literature on how gut bacteria modulate estrogen circulation, a 2021 review by Xinyu Qi and colleagues in Gut Microbes, sits underneath all of this and is rich enough to make sex-difference biology in the microbiome a serious research priority. The intersection with fasting and cortisol is a layer that nobody has measured yet.

The piece I want to flag for anyone listening is that the data we trust for fasting and protein and most other dietary interventions was mostly built on bodies that are not representative of all bodies. The trial that resolves this is not done. Women have been generalized into men's metabolic data for a long time. Be skeptical of categorical advice in either direction until the actual trial gets run.

GLP-1 and the Maintenance Question

Which brings us back to the paper everyone on LinkedIn is reacting to. The Mount paper on pasteurized Akkermansia muciniphila. Published the same day as a second Nature Medicine paper from Louis Aronne and colleagues, ATTAIN-MAINTAIN, on orforglipron, an oral once-daily GLP-1 receptor agonist used to maintain weight loss after coming off injectable GLP-1 drugs. Two different strategies for the same problem, post-intervention weight maintenance, both published in Nature Medicine on May 13, 2026.

The thread that connects them is the one this whole essay has been about. GLP-1 drugs are pharmacological fasting mimickers. They produce, in the body, the same signal of dietary scarcity that the Tanoue paper studied in mice and the Liu paper studied in fasting protocols. They quiet appetite. They slow gastric emptying. They send the body the message that food is less available than it actually is.

And the question that follows from Tanoue and Sudo and Liu, taken together, is whether the response to that drug-induced scarcity signal depends on whether the microbes are there to translate it. The Mount paper is the closest signal we have. A specific bacterial preparation, pasteurized so the cells are dead, helps people maintain weight loss after dieting. The mechanism implied is not engraftment. The cells are dead. The chemistry on their surface, and the metabolites they had already produced before pasteurization, are what is doing the work.

Which raises the question of what we are doing at Thaena. ThaenaBiotic is a multi-species postbiotic preparation. It is not the same product as pasteurized A. muciniphila. The bacterial composition is different, the manufacturing is different, the molecular profile is different. But the mechanistic class is the same. Dead bacteria delivering chemistry, not trying to engraft. The hypothesis that a multi-species postbiotic could support the metabolic flexibility that GLP-1 drugs are signaling for is sitting in three Nature papers and a 2004 J Physiol paper. The honest version is that we have not run the trial. We have not studied ThaenaBiotic in patients on GLP-1 drugs. We would like to. Someone has to actually run it.

For the broader landscape on what Ozempic mimics, and what your microbiome was already doing, see the GLP-1 & Weight topic page in the Science Hub.

The Honest Limitations

Every paper here has limitations, and the synthesis we have built across them is a hypothesis, not a settled result. Some of what is missing.

The Tanoue paper is a mouse paper. The browning of white fat at seven percent dietary protein has not been demonstrated in humans, and the dietary protein restriction needed to test this in humans would be both ethically and metabolically complicated. The mechanism, the two-key bile acid and ammonia pathway, is supported by elegant germ-free and gene-knockout experiments in mice. The human translation is plausible. It is not proven.

The Sudo paper is also a mouse paper, twenty years old, and the rescue with Bifidobacterium infantis is a model-organism finding that does not directly tell us what would happen in a human with a depleted gut microbiome under fasting stress. The hypothesis we built on top of it, that low-diversity humans under fasting may experience cortisol where high-diversity humans experience adaptation, is testable, and has not been tested.

The Ohara cycle-phase study is n = 7. The Hamidovic meta-analysis effect size is small. The Kapogiannis sex modulation is real but limited to a single intervention type and a small sample. The Mount pasteurized A. muciniphila trial is n = 90 with three authors disclosing affiliation with The Akkermansia Company, the commercial entity that produces the supplement. The Aronne orforglipron trial was sponsored by Eli Lilly, the drug developer. None of these are reasons to dismiss the findings. All of these are reasons to hold the conclusions with appropriate humility about what one study can show.

What none of these papers do, including Mount, is directly test whether a multi-species postbiotic like ThaenaBiotic produces the same metabolic adaptation outcomes that pasteurized A. muciniphila produced in the Mount maintenance trial. That work has not been done. We would like to do it.

One Tool in the Toolbox

None of this is anti-protein, anti-fasting, or anti-GLP-1. All three of those approaches have evidence behind them. All three have helped people. The argument is that the story is more complicated than the wellness internet has packaged it, and the complication is the microbe layer underneath all three.

If your gut microbiome is humming, those interventions can do beautiful things. The dietary scarcity signal becomes adaptive metabolic flexibility. The bile acid pathway runs. The ammonia-FGF21 pathway runs. The cortisol response stays calibrated. Body composition shifts in the direction the intervention was supposed to send it.

If your gut microbiome is depleted, the same interventions may produce a different kind of stress. The signals arrive, but the translators are missing. The metabolic adaptation does not fire. The cortisol response does. And the experience of the intervention is one of fatigue, anxiety, sleep disruption, and rebound, rather than progress.

The variable that nobody is talking about, in the protein craze and the fasting industry and the GLP-1 boom, is the microbe layer. And the variable you can actually move, if you want to bet on better translation, is the microbe layer. Eat the fiber. Feed the diversity. And the chemistry will follow.

ThaenaBiotic is one tool in that toolbox. A sterilized, multi-species postbiotic preparation, designed to deliver the chemistry of a healthy donor microbiome without requiring the bacteria to engraft. It is not the same as fixing the underlying gut. It is not a replacement for fiber, for diversity, for the lifestyle changes that build microbial resilience over time. It is a way to deliver the molecules that the bacteria in a well-functioning ecosystem would have produced, in a form that does not require those bacteria to first establish residency in your gut.

For someone fasting, restricting protein, or taking a GLP-1 drug, the question of whether microbial support could help with metabolic adaptation is real, and the trial that resolves it has not been run. We would like to run it. In the meantime, the hypothesis is sitting in the published literature, asking to be tested.

Frequently Asked Questions

Does this paper mean I should eat low protein?

No. The Tanoue paper is a mouse mechanism study, not a human dietary intervention. Seven percent dietary protein in mice was the experimental tool used to ask a mechanistic question about how the gut microbiome translates dietary scarcity signals. It is not a recommendation. Protein still matters for muscle, for satiety, and for a hundred other physiological processes. The takeaway of the paper is about the role of the microbiome as a translator, not about what humans should eat.

If fasting works through my microbiome, does that mean fasting is bad for people with dysbiosis?

The hypothesis the Tanoue and Sudo papers together support is that the response to dietary scarcity depends on microbial translation. Whether that means fasting is contraindicated for people with severely depleted microbiomes has not been directly tested in a controlled trial. What the data suggests is that the same intervention may produce different outcomes in different people, and the gut microbiome appears to be one of the variables doing the differentiating. The practical move, before considering fasting protocols, is to invest in the diversity and resilience of the underlying ecosystem.

Should women avoid fasting during the luteal phase?

The randomized trial that would answer this has not been run. The popular framing that luteal fasting is risky because cortisol is already elevated is not supported by the published literature. A 2020 meta-analysis of cortisol across cycle phases in 778 women found cortisol slightly higher in follicular than luteal. The one small crossover study that directly tested twelve-hour fasting across cycle phases found that luteal-phase fasting reduced salivary cortisol and increased parasympathetic activity. The honest answer is that the data we have is small, and it points in a different direction than the popular framing.

Is ThaenaBiotic the same as the pasteurized Akkermansia in the Mount study?

No. Pasteurized A. muciniphila is a single-species heat-killed bacterial preparation produced by The Akkermansia Company. ThaenaBiotic is a multi-species postbiotic preparation derived from rigorously screened human donors and sterilized for safety. The mechanistic class is similar, dead bacteria delivering chemistry rather than trying to engraft, but the bacterial composition, manufacturing process, and molecular profile are different. Comparisons between the two products are at the mechanism-class level, not the product level. The trial directly comparing them has not been done.

Would ThaenaBiotic help me if I am on a GLP-1 drug?

The hypothesis that microbial support could matter for people on GLP-1 drugs is supported by the published literature on microbial translation of dietary scarcity signals. It has not been directly tested. We have not studied ThaenaBiotic in patients on GLP-1 drugs in a controlled trial. We would like to. In the meantime, the broader practice of supporting gut diversity through fiber, fermented foods, and a sterilized postbiotic preparation has a defensible mechanistic basis, even if the specific GLP-1 question has not been answered yet.

What is FGF21, and why does it matter that the microbiome induces it?

FGF21 is a hormone produced by the liver in response to metabolic stress signals. It promotes insulin sensitivity, lipid oxidation, and adipose tissue browning. Several pharmaceutical companies are developing FGF21 analogs as drugs for metabolic dysfunction-associated steatohepatitis and obesity. The Tanoue paper shows that the gut microbiome induces FGF21 endogenously, through bacterial ammonia signaling, when dietary protein drops. The implication is that the metabolic effects pharmaceutical companies are trying to recreate may be partly achievable through gut-microbiome-mediated physiology in people whose microbiomes can produce the right signals.

The Bottom Line

The translator is the variable nobody is talking about. The variable doing the work between every dietary intervention and every metabolic outcome. The story is not about whether protein or fasting or GLP-1 drugs are right or wrong. The story is that the answer depends on whether your gut can speak the language the intervention is asking it to speak. And if it cannot, the same signal that should produce adaptation produces stress instead.

Translation, not condemnation. Pro-microbe, not anti-anything. ThaenaBiotic is one tool in that toolbox.

References

1. Tanoue T, Nagayama M, Roochana AJA, et al. Microbiota-mediated induction of beige adipocytes in response to dietary cues. Nature. 2026 Mar 4; 653(8114): 499–509. DOI · FREE FULL TEXT
2. Quinn RA, Melnik AV, Vrbanac A, et al. Global chemical effects of the microbiome include new bile-acid conjugations. Nature. 2020 Feb 26; 579(7797): 123–129. DOI · FREE FULL TEXT
3. Liu Z, Dai X, Zhang H, et al. Gut microbiota mediates intermittent-fasting alleviation of diabetes-induced cognitive impairment. Nat Commun. 2020 Feb 18; 11(1): 855. DOI · FREE FULL TEXT
4. Sudo N, Chida Y, Aiba Y, et al. Postnatal microbial colonization programs the hypothalamic-pituitary-adrenal system for stress response in mice. J Physiol. 2004 Jul 1; 558(Pt 1): 263–275. DOI · FREE FULL TEXT
5. Kapogiannis D, Manolopoulos A, Mullins R, et al. Brain responses to intermittent fasting and the healthy living diet in older adults. Cell Metab. 2024 Aug 6; 36(8): 1668–1678.e5. DOI · FREE FULL TEXT
6. Qi X, Yun C, Pang Y, Qiao J. The impact of the gut microbiota on the reproductive and metabolic endocrine system. Gut Microbes. 2021; 13(1): 1–21. DOI · FREE FULL TEXT
7. Hamidovic A, Karapetyan K, Serdarevic F, et al. Higher circulating cortisol in the follicular vs. luteal phase of the menstrual cycle: a meta-analysis. Front Endocrinol. 2020 Jun 2; 11: 311. DOI · FREE FULL TEXT
8. Ohara K, Okita Y, Kouda K, et al. Cardiovascular response to short-term fasting in menstrual phases in young women: an observational study. BMC Women's Health. 2015 Aug 28; 15: 67. DOI · FREE FULL TEXT
9. Depommier C, Everard A, Druart C, et al. Supplementation with Akkermansia muciniphila in overweight and obese human volunteers: a proof-of-concept exploratory study. Nat Med. 2019 Jul; 25(7): 1096–1103. DOI · FREE FULL TEXT
10. Mount S, Canfora EE, Jocken JW, et al. Pasteurized Akkermansia muciniphila MucT for weight loss maintenance in people with overweight and obesity: a controlled randomized trial. Nat Med. 2026 May 13. DOI
11. Aronne LJ, et al. Orforglipron for maintenance of body weight reduction: the double-blind, randomized phase 3b ATTAIN-MAINTAIN trial. Nat Med. 2026 May 13. DOI
12. Powell CE, McSween AM, Dohnalová L, et al. Gut microbiome-produced bile acid metabolite lengthens circadian period in host intestinal cells. PNAS. 2026; 123(11): e2506313123. DOI

This post accompanies the Lit Review Friday episode of Learn Something with Thaena. It is part of the deep-dive layer beneath The Science of ThaenaBiotic®, an interactive journal on microbiome and postbiotic biology. The science here belongs to the researchers cited above. The hypothesis we built on top of it, that the microbiome is the translator that determines whether dietary scarcity becomes adaptation or stress, is a hypothesis the literature is consistent with. It is not a hypothesis a single trial has proven. We are listening, we are reading, and we are looking for the trial that resolves it.