The Gut Barrier and Your Microbiome

Between your bloodstream and tens of trillions of gut microbes sits a barrier that is one cell thick — sealed at the seams by tight junctions and blanketed in mucus. Watch fibre-eating bacteria ferment your food into butyrate, the fuel that feeds those cells, thickens the mucus and pulls the seams tight — then starve them of fibre and watch the same bacteria turn around and eat the mucus layer instead.

Try this: leave it on Healthy, high-fibre and watch gold butyrate molecules being absorbed and visibly tightening the junctions. Then hit Low-fibre / ultra-processed and watch the mucus get eaten away, the seams loosen, and red LPS begin leaking into the blood.

Diagram is illustrative — not to scale.
Small intestine upstream — food is still being digested here normally sparse: 10³–10⁴ bacteria/mL to colon → fermentation should NOT happen here — that job belongs downstream LUMEN 10¹¹–10¹² bacteria per mL of colon contents food, fibre, microbes MUCUS ≈ 150 µm EPITHELIUM ONE cell thick colonocytes renewed every 3–5 days LAMINA PROPRIA gut immune tissue (GALT) BLOOD → portal vein → liver Goblet cell secretes MUC2 mucin Tight junction occludin · claudins · ZO-1 integrity 97% Butyrate a short-chain fatty acid ~70% of a colonocyte’s fuel BARRIER: SEALED inner layer: no bacteria (as it should be) Treg — calmed by butyrate dendritic cell — samples the lumen plasma cell — secretes IgA macrophage — senses LPS

Live barrier readout

Microbial diversity · Shannon index
3.9 H′
richness 220 species detected
diverse, fibre-fed community
Butyrate production
18.0 mM in the lumen
20 mM 5
total SCFA 100 mM · acetate : propionate : butyrate ≈ 60:20:20
Mucus layer thickness
150 µm (inner + outer)
inner layer intact goblet cells secreting
Tight-junction integrity
97 %
dashed line = the point where the “leak pathway” opens
LPS / endotoxin crossing into blood
0 events/min (illustrative)
status: none detected
total crossed this session: 0
Systemic inflammation · hs-CRP
0.6 mg/L
<1 low · 1–3 average · >3 mg/L raised (dashed)
Upstream gas · breath test at 90 min
6 ppm H₂   3 ppm CH₄
positive if H₂ rises ≥20 ppm by 90 min, or CH₄ ≥10 ppm
Bloating 10%

What's happening

Fibre arrives in the colon. Bacteria ferment it into short-chain fatty acids — watch the gold butyrate molecules head for the epithelium…
fibre fermenters other commensals mucin-degraders gram-negative opportunists dietary fibre butyrate (SCFA) LPS / endotoxin secretory IgA H₂ / CH₄ gas cytokines (TNF-α, IL-6)

The Science in Plain Language

1. A barrier one cell thick

Unrolled, your gut lining covers roughly 32 square metres — about half a badminton court, not the tennis court you may have read about. Across that whole expanse, what separates your bloodstream from the contents of your gut is a single layer of cells, replaced entirely every 3–5 days.

What stops things slipping between those cells is a protein ring at their tops: the tight junction, built from occludin, a family of claudins, and the scaffolding proteins ZO-1, ZO-2 and ZO-3. It is not a weld but a gate, and there are three ways through: the pore pathway (tiny claudin channels — ions and water only), the leak pathway (larger, occludin- and ZO-1-dependent, widened by inflammatory signals), and the unrestricted pathway — the gaps where cells have been killed outright, which pass almost anything, including whole bacteria.

2. The mucus blanket — and the layer meant to be sterile

Bacteria never normally touch that cell layer, because of mucus. Goblet cells pump out a giant gel-forming protein, MUC2, which swells into a slippery blanket. In the colon it has two layers: a loose outer layer microbes are welcome to graze in, and a dense inner layer, firmly attached to the cells, which in health holds essentially no bacteria at all. Your body maintains a sterile no-man’s-land a few dozen micrometres wide and rebuilds it constantly — in mice, roughly every hour.

Thickness figures vary wildly by method, because mucus collapses the moment tissue is handled. The mouse inner colonic layer is about 50 µm; human estimates are larger and far more variable. Treat the readout as a scale, not a lab value.

3. Fibre → butyrate: the mechanism at the centre of this page

You cannot digest dietary fibre. Your bacteria can. Fermenting it releases short-chain fatty acids — roughly 500–600 mmol a day, in a rough molar ratio of 60 acetate : 20 propionate : 20 butyrate. Acetate and propionate mostly leave for the liver and muscle. Butyrate mostly never leaves, and that is the point:

One molecule, five jobs — and you get it by eating plants. Most people in Western countries manage about 15 g of fibre a day against a recommended 25–38 g. Variety matters as much as the total: different fibres feed different bacteria.

4. Starve the microbes and they will eat the house

Some gut bacteria — Akkermansia muciniphila, certain Bacteroides — can eat mucin, the protein mucus is made of. That is normal; they are ordinary residents, and at usual levels they are associated with good health.

But in gnotobiotic mouse experiments, removing fibre caused these mucin-degraders to expand and turn on the mucus layer itself as their food source. The barrier thinned, bacteria moved closer to the epithelium, and the mice became markedly more susceptible to an invading gut pathogen. Switch on Low-fibre / ultra-processed above and you are watching that experiment: the magenta mucin-degraders descend and graze on your barrier because you stopped feeding them anything else. The caveat matters — that is mouse work in a simplified microbiome, not proven to the same standard in people. But it cleanly explains what we do observe: low-fibre diets go with thinner mucus and less butyrate.

5. “Leaky gut” — what is real, and what is being sold to you

Increased intestinal permeability is real. It is measurable — classically by drinking two sugars (lactulose, large; mannitol, small) and measuring the ratio appearing in your urine, and more directly by confocal endomicroscopy during endoscopy. It is well documented in coeliac disease, Crohn’s disease and ulcerative colitis, cirrhosis, NSAID enteropathy, severe burns, trauma and sepsis. In coeliac disease the trigger is understood in molecular detail: gluten fragments prompt release of zonulin, which prises the junctions open. None of this is disputed.

“Leaky gut syndrome” is a different animal. The commercial version — blamed for fatigue, brain fog, acne, depression, autoimmunity and weight gain, treated with an expensive powder — is not an established clinical diagnosis. Three warnings:

The honest position: a leaky barrier is a real, measurable phenomenon with real disease associations — and it is a marketing engine. Both are true. Treating the underlying disease, eating fibre, and not poisoning your barrier with NSAIDs and alcohol are supported. Buying a powder because a website told you your gut is leaking is not.

6. SIBO — the right bugs in the wrong postcode

Your small intestine is deliberately kept nearly empty of bacteria — about 10³–10⁴ per mL, against 10¹¹–10¹² per mL in the colon. Four defences do it: stomach acid, bile and pancreatic enzymes, the ileocaecal valve, and above all the migrating motor complex — a housekeeping wave that sweeps the small bowel clean every 90–120 minutes between meals.

Break those — impaired motility, adhesions, diverticula, long-term acid suppression, scleroderma, diabetic autonomic neuropathy — and colonic bacteria creep upstream. Food is fermented too early, in a place with no capacity to absorb the resulting gas: hydrogen, and (from archaea, not bacteria) methane. Hence bloating, distension and pain. A breath test calls it — H₂ rising ≥20 ppm above baseline within 90 minutes, or CH₄ ≥10 ppm (intestinal methanogen overgrowth) — or a jejunal aspirate. Treatment is usually the poorly-absorbed antibiotic rifaximin plus fixing whatever broke the motility, because relapse is common otherwise. Note that in this scenario the barrier below stays largely intact: you can feel dreadful with a perfectly sealed gut. See the SIBO hub for the full picture.

7. Antibiotics, probiotics, and what actually helps

Broad-spectrum antibiotics collapse diversity within days. Most species rebound over 1–2 months, but follow-up studies have found some still undetectable at six months. Meanwhile the empty niche fills: losing butyrate producers means losing colonisation resistance, which is how Clostridioides difficile gets its opening. Antibiotics are often necessary and save lives — but they are not free, and recovery is slower and less complete than people assume.

Probiotics deserve an honest hearing. Most strains in most people are transient: they appear in your stool while you take them and vanish within days to weeks after you stop. They are visitors, not settlers, and whether they colonise at all is highly person-specific. Effects are strain-specific and dose-specific — “probiotics” is no more one intervention than “drugs” is. The evidence is good in a few defined settings (preventing antibiotic-associated diarrhoea, shortening infectious diarrhoea in children, pouchitis, preventing necrotising enterocolitis in preterm infants) and thin-to-absent for the general “boost your gut health” claim. Fermented foods may be the better bet: a randomised trial found a high-fermented-food diet raised diversity and lowered inflammatory markers over ten weeks — promising, and still needing replication.

NSAIDs and alcohol injure the barrier directly, and this is not controversial. A single NSAID dose can measurably raise small-intestinal permeability within hours, and capsule studies find small-bowel mucosal breaks in the majority of long-term users. Alcohol — via its metabolite acetaldehyde — disrupts occludin and ZO-1 and raises circulating endotoxin after a binge.

So the best-evidenced, largest-effect, cheapest intervention is the least exciting one: eat more plant fibre, from more different plants. Your bacteria make the butyrate; the butyrate feeds the cells, thickens the mucus and tightens the seams. That loop is the whole animation.

↑ Back to the animation

Connections