Multiple Sclerosis and the Microbiome: There Are No Universal Good Bugs or Bad Bugs

For a decade, studies of the gut microbiome in multiple sclerosis kept disagreeing with each other. One paper flagged a bug as the villain; the next found nothing; a third pointed somewhere else entirely. In 2026, a large study from Belgium worked out a big part of why, and the answer is stranger and more useful than another list of good and bad bacteria. It turns out the thing that fooled the field was how fast things move through the gut.

Lit Review Friday · Learn Something with Thaena · Published 2026 · Reading time: ~16 minutes

Why a Decade of MS Microbiome Studies Kept Disagreeing

Start with the disease itself, because the mystery is the whole point. In multiple sclerosis, the immune system attacks myelin, the fatty insulation wrapped around your nerve fibers. Think of your nervous system as wiring and myelin as the coating that lets a signal travel clean and fast. In MS, the immune system strips that coating off, and the signals slow, scatter, or stop. It is a disease of the brain and spinal cord. For roughly a hundred years, the best explanation for why it happens has been a partial one: some genetic susceptibility, some environmental trigger, and a large honest gap in the middle.

So it was genuinely surprising when the clue started pointing at the gut. But for years the gut evidence was a mess. Line up the studies of the microbiome in MS and they mostly contradicted one another: one flagged this bug as increased, the next said it was decreased, a third found something else entirely. It looked like noise. And in 2026, Amber Pauwels and colleagues in Jeroen Raes's lab in Belgium, publishing in Gut Microbes, worked out a large part of why.

The culprit is almost funny. It is how fast things move through your gut, what researchers call transit time. They measured the moisture in people's stool, which is a good stand-in for transit time, and then asked a simple question: of all the things that make one person's gut community different from another's, how much does each one explain? That one number, how wet or dry the stool is, explained more of the variation between people's gut communities than whether the person even had MS. Roughly three times as much.

Here is the trap that sets. People with MS, especially the progressive form, often have slower guts, because MS can damage the nerves that run the bowel. Slow transit changes which bacteria thrive. So if you do not account for it, you go looking for MS bacteria, and what you actually find is the fingerprint of a sluggish gut, and you call it MS. Slow transit is partly a reflection of the dysbiosis itself, and it is the exact thing that has been fooling the field. When this group pooled their cohort with ten earlier studies and did the correction, forty of the bacteria that had been named as MS markers did not replicate in a single other study. Gone.

---

The Villain That Dissolves: Akkermansia and the End of Good-Bug, Bad-Bug

The clearest casualty of this correction is a bacterium called Akkermansia muciniphila. It got flagged years ago as one of the MS bad guys: several studies reported more of it in people with MS, and in the foundational 2017 work it behaved as pro-inflammatory in human cells. A clean little villain story. But pile up more data and control for the confounders, and that story just comes apart. Elsewhere in the literature, the very same organism is cast as a beneficial, barrier-supporting microbe, the kind of bug wellness writing tells you to cultivate. Same species, reading as good or guilty depending entirely on the context around it.

This is the real lesson, and it is the one worth walking away with. There are no good bugs and bad bugs. A microbe that looks like a villain in one gut is a hero in another. It was never about the name on the bacterium. It is about what it is doing, and the context it is doing it in. This is also, quietly, an argument for humility about our tools. The 2026 study used a method called 16S sequencing, which reads a kind of barcode on each bacterium. It tells you roughly who is there. It does not tell you what they are making, or doing, or saying. Reading a guest list is not the same as knowing what happened at the party. The authors are candid about this: to measure function, you would need deeper sequencing they deliberately traded away in order to match their MS cohort against a nearly three-thousand-person population reference. Every mechanism discussed below is inference layered on top of that guest list, not a direct measurement.

---

What Survived: Fewer Bacteria, and a Gentler Gut in Milder Disease

If most of the famous markers evaporated, is anything left but noise? Actually, yes, and this is the part worth holding onto. When the researchers counted the bacteria properly, a couple of real things survived. This is a technical point that matters more than it sounds. Most microbiome studies report percentages: this bug is 12 percent of the community, that one is 3 percent. The problem with percentages is what you might call the pie-chart trap. If one bacterium blooms and takes over a bigger slice, every other slice looks smaller, even though nothing else actually changed. The fix is to count absolute cells, the actual number of bacteria per gram, rather than slices of a pie. It is the difference between saying a neighborhood is 40 percent one family and saying how many people actually live on the street.

When they counted that way, one finding held up that percentages had been hiding: people with MS tended to carry fewer bacteria overall. The ecosystem is thinner, not just rearranged. Depleted, in the sense we use a lot around here, the same way a monoculture field carries less life than a diverse one. A small, robust core of specific bacteria survived the correction too, most consistently a reduction in Bacteroides, but the headline is the thinning itself.

The second survivor is the one that starts to connect the ecology to how people actually feel. The people whose guts were dominated by a bacterium called Prevotella tended to have milder disease: lower disability, faster walking. And Prevotella is a maker of propionate, one of the short-chain fatty acids that, in laboratory work, nudges the immune system's calming cells. Propionate specifically runs low in people with MS, and lowest right after a first relapse (Duscha et al. 2020, Cell). So a picture starts to cohere: slow transit, a thinner ecosystem, fewer of the calming molecules, and a tilt toward inflammation. To be clear, this is an association, not a proven cause. But it is a clean bridge from the shape of an ecosystem to the experience of a person, and that is rare.

---

How Does a Gut Bug Talk to the Immune System at All?

To understand why any of this matters, you need the piece that sits underneath it, because the whole story runs through it. Your immune system has peacekeepers. They are a kind of cell called a regulatory T cell, or Treg, and their entire job is to walk around the body saying: that is us, do not attack that, stand down. They are the calm-down librarians. The cells that turn off the fire alarm when there is no actual fire. A great deal of health comes down to whether you have enough of them and whether they are doing their job, and a lot of autoimmune trouble is, at heart, a story about peacekeepers who never got properly trained.

Trained is the word that changes how you see the body. These peacekeepers are not simply born knowing what to tolerate. A large share of them are trained on the job, out in your tissues, and the single biggest classroom is your gut. The teachers are your microbes. This is not a fringe idea. A rationally selected mixture of human gut Clostridia can induce these peacekeeper cells and their calming IL-10 signal in mice (Atarashi et al. 2013, Nature), and a single purified molecule from one gut bacterium can do the same (Round and Mazmanian 2010, PNAS). The gut is where a good fraction of immune tolerance gets installed.

Which is exactly why a nerve disease with a gut clue stops sounding strange. If the classroom where immune peacekeepers get trained is the gut, and the teachers are the microbes, then a disrupted classroom is a plausible place for a failed education to begin. That brings us to the landmark study underneath all of this, the one that turned the gut clue in MS from a correlation into a mechanism: Egle Cekanaviciute and colleagues, publishing in PNAS in 2017.

---

The Heat-Killed Reveal: You Do Not Need Living Bacteria

Here is the part of the 2017 paper that is easy to skate past and should not be. To test whether these bacteria could move human immune cells, the researchers did not colonize anything with a living organism. They took the bacteria, killed them with heat, and sonicated them, physically rupturing the cells into debris. Then they dripped that dead-bacteria material onto human immune cells in a dish. And the immune cells changed their behavior anyway.

Sit with that for a second. The organisms were dead. Ruptured. And the lesson still got taught. That tells you the active ingredient was never the living bug. The lesson is in the molecules the bug leaves behind. This is the whole idea behind postbiotics, and here it is, quietly proving itself inside a study about a disease of the brain. The paper's own discussion says the active material may consist of any bacterial products, secreted or intracellular. The signal travels in the chemistry.

And now the twist worth holding onto, because it corrects a common oversimplification, and it rhymes with the no-good-bugs idea from the other direction. We tend to talk about dead bacteria, about postbiotic material, as though it were automatically gentle and calming. In this study, it was not. The dead material from the MS-associated ecosystem did not soothe the immune cells. It pushed them toward inflammation. So the molecules are not good or bad in the abstract either. They carry the state of the system they came from. A healthy ecosystem's leftovers can teach calm. A dysregulated ecosystem's leftovers can teach attack. Same principle, opposite direction, packed into the chemistry.

---

Which Molecule Is Doing the Teaching? A Single Bile-Acid Switch

Neither the 2017 anchor nor the 2026 study could name the exact molecule. The 2017 paper could only gesture at families: polysaccharides, short-chain fatty acids, and aryl-hydrocarbon-receptor ligands. But other work has gotten gorgeously precise about how a single microbial molecule can throw the switch between an inflammatory cell and a calm one.

One of the key papers is Hang and colleagues, 2019, in Nature, from Michael Fischbach's group and their collaborators. They found that a bacterial bile acid, a molecule called 3-oxoLCA, binds directly to one master switch inside a young, undecided immune cell and blocks it from becoming an inflammatory foot soldier. A related bile acid, isoalloLCA, pushes that same young cell the other way, toward becoming a calm-down peacekeeper. One molecular switch. Both directions. A microbe-made molecule throwing it.

The collaboration underneath this is lovely. Your liver makes the raw, parent version of these bile acids from cholesterol and pours them into your gut. Then your gut bacteria chemically rewrite them into the versions that do the teaching. The liver makes the raw material, the microbes edit it, and the edited molecule trains the immune cell. And this is real in humans, not only in mice: a 2022 follow-up in Nature identified the human gut bacteria and enzymes that make these bile acids and showed that the calming versions are depleted in people with inflammatory bowel disease (Paik et al. 2022). Newer work in 2026 is beginning to trace bile acids specifically in MS, still early, but part of the same widening story (J Neuroimmunol, 2026).

---

MS as One Continuum, Not Two Diseases

A 2026 review from the Mayo Clinic and the University of Iowa (Montini, Kantarci, and colleagues, in Frontiers in Immunology) pulled these threads into a frame worth borrowing. For a long time MS was described as two diseases: a relapsing form, with discrete immune attacks, and a progressive form, with a steady accrual of disability. This review argues that it is better understood as one continuum, a dial. At one end, inflammation predominates. At the other, slow neurodegeneration does. Both processes coexist in most people; the label just reflects which one is louder at a given moment.

And here is the part that matters for us: the gut appears to shift along that dial. Toward the relapsing, inflammatory end, the review describes a depletion of the calming short-chain-fatty-acid and bile-acid makers. Toward the progressive end, it describes a further loss of those protective molecules and, notably, a more oxidative gut environment, an ecosystem under more chemical stress, with measured drops in protective metabolites that track with atrophy and cognitive decline. That is the honest, on-topic home for the oxidative-stress thread in this story: not a borrowed idea from some other disease, but the continuum's own description of what the progressive gut looks like. The review's thesis is the one this whole episode keeps landing on. Stop cataloguing which microbes are present. Ask what they make.

---

Recent Mechanistic Work: What 2026 Added

Two more 2026 studies are worth flagging for readers who want the current edge of the mechanism, with the usual caveat that early mechanistic work is exactly that: early.

The first keeps drilling into the ask-what-they-make idea. Correale and colleagues (2026, Multiple Sclerosis Journal) looked at metabolites made along the vitamin B2 and B9 pathways by gut microbes, and found they can modulate a particular immune cell type, MAIT cells, taken from people with MS. It is a specific, molecule-level example of microbial chemistry tuning an immune population. It is also tissue-culture work, cells in a dish, not a demonstration of what happens in a living person, so read it as a mechanistic lead rather than an outcome.

The second connects the ecosystem back to how people feel. Meza and colleagues (2026, Brain and Behavior) ran a small pilot that paired a Wahls-style diet with gut-microbiome sequencing and used unsupervised machine learning to look for patterns predicting which MS-related symptoms shifted. It is a pilot, with the limits pilots carry, but it is a glimpse of where this field is trying to go: not one bug, not one molecule, but the whole ecosystem read as a system, with the diet and the symptoms in the same frame.

---

Does a Fecal Transplant Work for MS in People?

The 2017 mouse result was clean. Transfer the MS ecosystem, get worse disease. So the obvious next move is to run it in the good direction: put healthy bacteria into people with MS and see what happens. People have tried, through fecal microbiota transplantation. And this is where the confusion promised at the top comes due.

The one randomized human trial that has been published was small, nine patients, and it was stopped early. It found the procedure safe and tolerable, but it showed no significant benefit on disability scores or brain imaging (Al et al. 2022). Beyond that trial, the human record is a handful of uncontrolled single-patient case reports. There is no controlled evidence that changing the microbiome treats MS.

The tempting read is: so the gut does not matter after all. That reading is worth pushing back on, hard, and the reason ties back to the continuum. If MS is many roads to one destination, then some people's MS may be gut-influenced and some may not be at all, and we do not yet have the tools to tell those groups apart. The field's own consensus researchers now argue for exactly this framing, describing MS as a set of overlapping processes rather than one uniform mechanism (Kuhlmann et al. 2023, Lancet Neurology). Two people can carry the same diagnosis and have genuinely different underlying stories.

Inflammatory bowel disease is the useful parallel here. IBD is openly described as a continuum of disorders rather than a single condition (Cleynen et al. 2016, Lancet), and it is further along than MS at sorting its subtypes. The MS field does not yet have a validated way to identify, say, a gut-influenced subtype of the disease. So when you run a single fecal-transplant trial and pool everyone together, any signal from the people it might genuinely help gets washed out by the people it was never going to touch. This is not a hand-wave. There is a whole literature showing how much unmeasured patient variation can distort and even manufacture microbiome associations (Vujkovic-Cvijin et al. 2020, Nature), which is exactly why the largest MS microbiome study matched its controls by household, and why the 2026 anchor went to such lengths to correct for transit time. Both are the same defensive move against the same kind of noise.

---

Is There Anything Hopeful Inside the Confusion?

There is, and it is a correction to a story most of us absorbed without noticing: that a nervous system only ever moves in one direction, downhill. That is not the full picture. The central nervous system can, in fact, repair its own insulation. Myelin can be rebuilt by the brain's own repair cells, a well-documented biology (Franklin and Ffrench-Constant 2017, Nature Reviews Neuroscience). There is even a controlled trial showing that an existing drug can nudge that repair process along, measured as faster nerve signaling (Green et al. 2017, Lancet, the ReBUILD trial). Repair is real. It is also partial, age-limited, and does not undo established damage. Both things are true.

This is also where it is important to be careful, because hope is exactly where overclaiming creeps in. Terry Wahls is often invoked here, and her published research deserves to be described accurately. Her studies of an intensive diet-and-lifestyle program measured improvements in fatigue and quality of life in people with MS, in small studies, including one randomized trial comparing two dietary approaches (Wahls et al. 2021). Those studies measured fatigue and function. They did not measure a reversal of MS, and they cannot separate the diet from the exercise and stress-reduction bundled with it. Her own dramatic personal recovery is a single anecdote, not a trial result. The accurate sentence is: a multimodal diet-and-lifestyle program improved fatigue and quality of life in MS in small studies. Not: diet reverses MS.

It is worth knowing who she is, because the story is genuinely remarkable. Dr. Terry Wahls is a clinician-researcher at the University of Iowa who was diagnosed with multiple sclerosis, progressed to needing a reclining wheelchair, and then, after reorganizing her life around a nutrient-dense diet and a multimodal lifestyle program, regained a great deal of function. She lays out that approach in her book The Wahls Protocol: a modified Paleolithic, plant-and-nutrient-dense way of eating, paired with exercise, neuromuscular electrical stimulation, and stress reduction. It is important to hold two things separately here. The book is her clinical framework and her working hypothesis, not a clinical trial. What the peer-reviewed research has tested so far is that framework's effect on fatigue and quality of life, and those are the outcomes it moved, in small studies. Her own recovery, striking as it is, remains a single personal story, not evidence that generalizes.

The reason her work belongs in this episode is the mechanism she is implicitly betting on. A diet built on plant diversity and fiber is, among other things, a way of feeding a microbial ecosystem, which is exactly the lever the Meza pilot above was trying to read. Whether the gut microbiome is part of why some people feel better on a program like hers is an open, testable question, and a genuinely interesting one. That is the honest place to leave it: a compelling framework, a real fatigue-and-quality-of-life signal in small studies, a plausible microbial mechanism, and a set of trials still to be done. Hope grounded in curiosity, not a cure being sold.

The deeper idea underneath the hope, though, is well supported and worth keeping: the trajectory is more dynamic than a straight line down. You can move in both directions. The nervous system has more capacity for repair than the old model assumed, even if that capacity is limited and imperfect. That is biology and philosophy, not a treatment promise.

---

What It Means: The Signal Travels in the Molecules

Step back from MS specifically and look at the pattern that recurs at every level of this research, because the pattern is the real takeaway. At each step, the thing doing the immunological work is the diffusible, bacteria-made molecule, not the live microbe. The 2017 anchor moved human immune cells with dead, ruptured bacteria. The bile-acid papers move the same immune switch with single, purified molecules. The peacekeeper-training studies do it with defined bacterial products. And the 2026 anchor, for all its rigor, ends up pointing the same direction: stop obsessing over which species are on the guest list, because the list keeps contradicting itself, and start asking what the community is making.

This reframes what the microbiome is for. We have spent two decades cataloging who is present, sequencing the roster of species. But a roster does not tell you what is being said. The microbiome speaks eloquently in chemistry, and the chemistry is what your immune system, your nervous system, and the rest of your body actually listen to. The molecules are the message. The living bacteria are, in a sense, just the factory that makes them.

That is a hopeful reframing, because it tells you where the leverage is. If the signal is molecular, then the question of what a depleted or disrupted microbial ecosystem is failing to make becomes a question you can actually study, molecule by molecule. The literature is consistent with this picture. It is not a claim any single human trial has proven.

---

The Postbiotic Thesis: Delivering the Signals Directly

Here is where Thaena's work connects, and it connects at the level of the general principle, not at the level of any disease. Nothing in this article says or implies that a postbiotic treats, prevents, slows, or reverses multiple sclerosis, inflammatory bowel disease, or any autoimmune condition. It does not. The MS science above is the science of MS. What follows is a separate conversation about everyday microbial ecosystems and the logic of postbiotics.

The through-line of this research, the heat-killed reveal above all, is that living organisms are not required for a microbial community's molecular signals to have an effect. The signal is in the chemistry. That is the definition of a postbiotic: the molecular output of a microbial community, delivered without the live bacteria. And the 2026 finding that people with MS carried a thinner, depleted ecosystem is a reminder that when a community loses its members, it loses the chemistry those members made. When the microbes that produce a community's full spectrum of signals are depleted, delivering those signals directly is a logical thing to explore. Single strains, alone, cannot reproduce the emergent chemistry that only appears when a whole community is functioning.

ThaenaBiotic is one tool in a larger toolbox, framed as support within an ecosystem-restoration approach, never as a cure, a fix, or a complete protocol. The interesting science in this episode is not a product story. It is the deeper realization that we were never really one organism managing a bag of bacteria. We are an ecosystem, and the ecosystem speaks in molecules.

---

Ecosystems All the Way Down: Mitochondria and Microbes

There is one last image worth ending on, and it is the one that reframes the whole thing. The little engines inside every one of your cells, the mitochondria that make your energy, are by origin bacteria. Billions of years ago, a bacterium and another simple cell merged and never separated. The process is called endosymbiosis, and the evidence is textbook: mitochondria still carry their own small circular genome, their own double membrane, and divide in a bacterial style (Roger et al. 2017). Thank you, Lynn Margulis, for making that idea stick.

So you carry bacterial-descended machinery running your cells from the inside, and a whole ecosystem of bacteria running signals from your gut on the outside. Domesticated bacteria within, wild bacteria without, and both of them talking to the rest of you, your immune system, your metabolism, your nervous system, in the same chemical language. It is not a coincidence that mitochondrial trouble shows up in the axonal damage of MS (Dutta et al. 2006); the energy engines and the signaling engines are of one lineage. To be careful: the idea that microbial molecules tune your neurons' mitochondria specifically in MS is a bridge worth watching, not a proven pathway. The shown biology sits in gut and metabolic tissue, not MS neurons.

Still, the arc is honest and it is beautiful. You were never one thing. You are an ecosystem, all the way down. And a mysterious disease like MS might be part of what it looks like when the conversation inside that ecosystem starts to break down. Which is both humbling and, in its way, hopeful, because a conversation is something you can learn to support.

---

The Honest Limitations

The 2026 anchor is a rigorous study, and its rigor is the whole point, but it is important to be clear about what it can and cannot tell us. It is cross-sectional, a single snapshot in time, so it cannot show cause and effect. It used 16S sequencing, the guest list, not the deeper method that would reveal what the bacteria are actually doing. Its progressive-MS group was small. And every mechanistic link it hints at, transit to community to symptom, is an association, not a demonstrated pathway.

The same honesty applies across the story. The causal experiments, MS bacteria producing worse disease, are in mice. Medication and diet are real confounders: MS treatments reshape the gut community, so some of the difference between patients and controls is disease and some is the drugs. The one human fecal-transplant trial was small and stopped early. The bile-acid mechanism is precise but was mapped in mice and cell culture, and the human bile-acid work in inflammatory bowel disease is a different disease from MS. And a large fraction of the bacterial proteins doing the most interesting signaling work remain completely uncharacterized, genetic dark matter we cannot yet read.

The candor is not a weakness of the story. It is the story. Confounded associations, conflicting taxa, mouse-only causality, thin human data, and a mechanism whose signal we can see even when we cannot yet name every molecule carrying it.

---

Frequently Asked Questions

Why did earlier MS microbiome studies disagree so much?

A large 2026 study found that how fast the gut moves, its transit time, explained more of the difference between people's gut communities than an MS diagnosis did (Pauwels et al. 2026). Because MS often slows the gut, studies that did not account for transit kept mistaking the signature of a sluggish gut for a signature of the disease. Correcting for it made forty previously claimed MS-marker bacteria vanish.

Is Akkermansia a good bacterium or a bad one in MS?

Neither, cleanly. Akkermansia was flagged for years as an MS villain, but that story dissolves across more data and after correcting for confounders, and elsewhere the same species is described as beneficial. It is a clean example of why good-bug, bad-bug thinking keeps failing: what matters is what a microbe is doing in a given context, not its name.

Do you need living bacteria for the microbiome's benefits, or is it the molecules?

In the foundational 2017 study, heat-killed and ruptured bacteria, with no living organism left, still reprogrammed human immune cells in a dish (Cekanaviciute et al. 2017). The active signal is carried by the molecules a community makes. That molecular output, delivered without the live bacteria, is what the word postbiotic describes.

Is dead-bacteria or postbiotic material automatically good for you?

Not inherently. In the 2017 study, the dead material from an MS-associated ecosystem pushed immune cells toward inflammation, not calm. Microbial molecules carry the state of the system they came from, which is why the source ecosystem matters so much.

Why did the human fecal-transplant trial for MS come back inconclusive?

The one published randomized trial was small and stopped early, and it showed no benefit on disability or brain imaging (Al et al. 2022). MS appears to be many biological roads to one destination (Kuhlmann et al. 2023), and there is no validated way yet to identify which patients might have a gut-influenced form, so a pooled trial can wash out a subgroup signal. A larger trial is ongoing.

Did Terry Wahls prove diet reverses MS?

No. Her published studies measured improvements in fatigue and quality of life on an intensive diet-and-lifestyle program, in small studies (Wahls et al. 2021). They did not measure disease reversal, and her personal recovery is a single anecdote, not a trial result. The supported idea is that the nervous system is more dynamic than a straight decline, not that any diet cures MS.

What is a postbiotic, and how is it different from a probiotic?

A probiotic delivers live bacteria. A postbiotic delivers the molecular output of a microbial community, the signals, without live organisms. Because some of a community's most useful chemistry only emerges when many species interact, a full-spectrum postbiotic aims to capture what a single strain cannot. ThaenaBiotic® is a full-spectrum postbiotic designed to support everyday gut resilience.

---

The Bottom Line

Multiple sclerosis is still a mystery, and the microbiome did not solve it. What the 2026 work did was more interesting and more honest: it showed why a decade of studies kept disagreeing, and in doing so it dismantled the good-bug, bad-bug way of thinking that produced those disagreements. The villain bacteria dissolved. What survived was subtler and truer: a thinner ecosystem, a gentler gut in milder disease, and a field learning to ask what the community makes instead of who is on the guest list.

The thread that runs through all of it is that the signal travels in the molecules. Dead bacteria taught the lesson. A single bile acid throws the switch. The living organism was never the active ingredient, and the name of the bug was never the answer. That is the postbiotic idea, quietly proving itself inside a story about the brain, and it points at where the leverage in this biology might actually be.

Think of yourself not as one organism managing a bag of microbes, but as an ecosystem in conversation. The words being spoken, the molecules being made, are what the rest of you is listening to.

Stay curious. Take care of your ecosystem.

---

References

1. Pauwels A, Devolder L, Falony G, et al. Deconfounded, quantitative microbiome profiling identifies robust multiple sclerosis markers and clinical covariate associations. Gut Microbes. 2026;18(1):2681876. https://doi.org/10.1080/19490976.2026.2681876 · FREE FULL TEXT
2. Cekanaviciute E, Yoo BB, Runia TF, et al. Gut bacteria from multiple sclerosis patients modulate human T cells and exacerbate symptoms in mouse models. PNAS. 2017;114(40):10713–10718. https://doi.org/10.1073/pnas.1711235114 · FREE FULL TEXT
3. Montini F, Mangalam A, Zeydan B, Murray J, Kantarci OH. Gut microbiota and gut-derived metabolites in defining the multiple sclerosis phenotypic continuum. Front Immunol. 2026;17:1858047. https://doi.org/10.3389/fimmu.2026.1858047 · FREE FULL TEXT
4. Berer K, Gerdes LA, Cekanaviciute E, et al. Gut microbiota from multiple sclerosis patients enables spontaneous autoimmune encephalomyelitis in mice. PNAS. 2017;114(40):10719–10724. https://doi.org/10.1073/pnas.1711233114 · FREE FULL TEXT
5. Hang S, Paik D, Yao L, et al. Bile acid metabolites control T_H17 and T_reg cell differentiation. Nature. 2019;576(7785):143–148. https://doi.org/10.1038/s41586-019-1785-z · FREE FULL TEXT
6. Paik D, Yao L, Zhang Y, et al. Human gut bacteria produce T_H17-modulating bile acid metabolites. Nature. 2022;603(7903):907–912. https://doi.org/10.1038/s41586-022-04480-z · FREE FULL TEXT
7. Bile acid–microbiota interactions in multiple sclerosis: from experimental models to early clinical evidence. J Neuroimmunol. 2026;578898. https://doi.org/10.1016/j.jneuroim.2026.578898
8. Correale J, et al. Microbiome-derived metabolites from the vitamin B2 and B9 pathways modulate MAIT cells from multiple sclerosis patients. Mult Scler. 2026. https://doi.org/10.1177/13524585261447695
9. Meza LA, et al. Predicting dietary impact on multiple sclerosis-related symptoms with the gut microbiome: a pilot study using unsupervised machine learning. Brain Behav. 2026. https://doi.org/10.1002/brb3.71394
10. Atarashi K, Tanoue T, Oshima K, et al. T_reg induction by a rationally selected mixture of Clostridia strains from the human microbiota. Nature. 2013;500(7461):232–236. https://doi.org/10.1038/nature12331
11. Round JL, Mazmanian SK. Inducible Foxp3+ regulatory T-cell development by a commensal bacterium of the intestinal microbiota. PNAS. 2010;107(27):12204–12209. https://doi.org/10.1073/pnas.0909122107 · FREE FULL TEXT
12. Duscha A, Gisevius B, Hirschberg S, et al. Propionic acid shapes the multiple sclerosis disease course by an immunomodulatory mechanism. Cell. 2020;180(6):1067–1080.e16. https://doi.org/10.1016/j.cell.2020.02.035
13. Rothhammer V, Mascanfroni ID, Bunse L, et al. Type I interferons and microbial metabolites of tryptophan modulate astrocyte activity and central nervous system inflammation via the aryl hydrocarbon receptor. Nat Med. 2016;22(6):586–597. https://doi.org/10.1038/nm.4106
14. Bhargava P, Smith MD, Mische L, et al. Bile acid metabolism is altered in multiple sclerosis and supplementation ameliorates neuroinflammation. J Clin Invest. 2020;130(7):3467–3482. https://doi.org/10.1172/JCI129401 · FREE FULL TEXT
15. Al KF, Craven LJ, Gibbons S, et al. Fecal microbiota transplantation is safe and tolerable in patients with multiple sclerosis: a pilot randomized controlled trial. Mult Scler J Exp Transl Clin. 2022;8(2):20552173221086662. https://doi.org/10.1177/20552173221086662 · FREE FULL TEXT
16. Kuhlmann T, Moccia M, Coetzee T, et al. Multiple sclerosis progression: time for a new mechanism-driven framework. Lancet Neurol. 2023;22(1):78–88. https://doi.org/10.1016/S1474-4422(22)00289-7 · FREE FULL TEXT
17. Cleynen I, Boucher G, Jostins L, et al. Inherited determinants of Crohn's disease and ulcerative colitis phenotypes: a genetic association study. Lancet. 2016;387(10014):156–167. https://doi.org/10.1016/S0140-6736(15)00465-1 · FREE FULL TEXT
18. Vujkovic-Cvijin I, Sklar J, Jiang L, et al. Host variables confound gut microbiota studies of human disease. Nature. 2020;587(7834):448–454. https://doi.org/10.1038/s41586-020-2881-9 · FREE FULL TEXT
19. Franklin RJM, Ffrench-Constant C. Regenerating CNS myelin: from mechanisms to experimental medicines. Nat Rev Neurosci. 2017;18(12):753–769. https://doi.org/10.1038/nrn.2017.136
20. Green AJ, Gelfand JM, Cree BA, et al. Clemastine fumarate as a remyelinating therapy for multiple sclerosis (ReBUILD): a randomised, controlled, double-blind, crossover trial. Lancet. 2017;390(10111):2481–2489. https://doi.org/10.1016/S0140-6736(17)32346-2
21. Wahls TL, Chenard CA, Snetselaar LG, et al. Impact of the Swank and Wahls elimination dietary interventions on fatigue and quality of life in relapsing-remitting multiple sclerosis: the WAVES randomized parallel-arm clinical trial. Mult Scler J Exp Transl Clin. 2021;7(3):20552173211035399. https://doi.org/10.1177/20552173211035399 · FREE FULL TEXT
22. Roger AJ, Muñoz-Gómez SA, Kamikawa R. The origin and diversification of mitochondria. Curr Biol. 2017;27(21):R1177–R1192. https://doi.org/10.1016/j.cub.2017.09.015
23. Dutta R, McDonough J, Yin X, et al. Mitochondrial dysfunction as a cause of axonal degeneration in multiple sclerosis patients. Ann Neurol. 2006;59(3):478–489. https://doi.org/10.1002/ana.20736

---

This post accompanies the Lit Review Friday episode of Learn Something with Thaena.