TL;DR: A 2026 Cell Reports study found loss of the ALS/FTD-linked gene C9orf72 in myeloid immune cells altered microbial glycogen-driven immune training, producing blood-brain-barrier leakage and T-cell entry into the brain.
Key Findings
- Digesting gut glycogen rescued the mice: Enzymatically breaking down bacterial glycogen in the gut improved survival and reduced microglial reactivity in C9orf72-deficient mice — the strongest causal link in the experiment.
- 10 unrelated bacterial strains triggered the response: Cytokine release in C9orf72-deficient myeloid cells did not depend on one bacterial family. The trigger was a shared microbial product, not a single species.
- Bacterial glycogen was the trigger: Metatranscriptomics pointed at glycogen biosynthesis as the pathway separating inflammatory bacteria from less inflammatory ones.
- Parabacteroides merdae drove neuroinflammation in the model: Germ-free C9orf72-deficient mice colonized with this glycogen-producing strain showed monocytosis, BBB breakdown, and T cell entry into the CNS.
- 15 of 22 ALS samples carried the inflammatory glycogen form: Vs. 4 of 12 healthy controls, and 1 of 1 C9ORF72-FTD sample — small numbers but consistent with the mechanism.
- The bridge is gene-by-microbe, not microbiome alone: C9orf72 sets the host vulnerability; bacterial glycogen supplies the trigger; together they shift immune behavior in the gut and brain.
Source: Cell Reports (2026) | McCourt et al.
C9orf72 is usually introduced as an ALS and frontotemporal dementia gene, but this study focused on myeloid immune cells. Researchers linked C9orf72 loss to an exaggerated response against microbial glycogen, a bacterial carbohydrate that can train inflammatory immunity.
The finding is more specific than a broad microbiome-and-disease association. It identifies a microbial product, a host vulnerability, and mouse interventions that reduced the downstream brain inflammation pattern.
C9orf72 Limited Microbial Glycogen Inflammation in Myeloid Cells
C9ORF72 repeat expansions are a major inherited cause of ALS and frontotemporal dementia. ALS damages motor systems; FTD affects behavior, language, and executive control.
The diseases share important biology, and C9orf72 is often discussed through neurons because neurodegeneration is where the clinical failure is visible.
C9orf72 also functions in myeloid cells — the broad immune family that includes macrophages, monocytes, and microglia. Those cells decide how aggressively the body responds to microbes, debris, and tissue damage.
Set their threshold too low, and ordinary microbial material starts looking like a threat that needs a stronger inflammatory response.
Researchers tested a gut-first hypothesis: C9orf72 may be doing inflammatory restraint work in the gut long before symptoms surface in the brain.
Ten Unrelated Bacterial Strains, One Shared Trigger
The team screened gut bacterial strains to see which provoked cytokine release in C9orf72-deficient immune cells. Cytokines are signaling proteins that immune cells use to amplify inflammation, recruit reinforcements, and coordinate defense.
Ten phylogenetically diverse strains drove cytokine release in a C9orf72-dependent way. The cross-family pattern was the first clue.
The trigger was not a quirk of one bacterial lineage. It was something many gut bacteria share.
Metatranscriptomics — which measures which microbial genes are actively expressed — pointed at the glycogen biosynthesis pathway. Glycogen is a stored carbohydrate, best known as an energy reserve in human liver and muscle.
Bacteria make their own forms of it, and the inflammatory version of bacterial glycogen behaved like a microbial input that C9orf72-deficient myeloid cells handled poorly. Without the C9orf72 brake, that carbohydrate read as a stronger inflammatory cue than it should.
C9orf72 Myeloid Cells Limited Microbial Inflammation in Mice
Germ-free mice start without resident microbes, which lets researchers ask what a single bacterium does in a controlled host. The team colonized germ-free C9orf72-deficient mice with Parabacteroides merdae — a species that produces inflammatory glycogen — and watched what happened.
The vulnerable mice tipped into a broader inflammatory state across three connected systems:
- Monocytosis: blood showed more monocytes, the immune cells that amplify inflammation and infiltrate tissues.
- Blood-brain barrier breakdown: the protective boundary between blood and brain became more permeable.
- T cell entry into the CNS: adaptive immune cells appeared in brain and spinal-cord territory where they can drive inflammatory damage.
That sequence connects the gut-derived immune trigger to the nervous system without pretending the gut alone explains neurodegeneration. The more careful reading is that microbial glycogen raises inflammatory pressure in a genetically vulnerable immune system — not that it produces ALS.

Antibiotics and Glycogen Removal Reduced the Brain Inflammation Pattern
The most actionable result was the rescue. When researchers enzymatically digested bacterial glycogen in the gut, C9orf72-deficient mice showed better survival and less microglial reactivity in the brain.
Microglia are the brain’s resident myeloid cells. They help clear damage when they activate appropriately, but chronic or excessive reactivity drives neuroinflammatory injury.
The intervention moved the result from description toward mechanism. Changing the gut-derived trigger changed the brain outcome. The microbial carbohydrate was a lever, not a passenger.
Human ALS and FTD Samples Supported the C9orf72 Immune Link
The human fecal data give the mouse work a clinical foothold. Inflammatory glycogen forms appeared in:
- 15 of 22 ALS samples
- 1 of 1 C9ORF72-FTD sample
- 4 of 12 healthy control samples
These numbers are too small for diagnostic claims. The FTD group has a single subject, and fecal microbiome findings can be shaped by diet, medication, disease stage, geography, and sampling methods.
What the comparison does support is that the pathway exists in real human gut contents, not only in engineered mouse experiments.
C9orf72 Myeloid Cells May Reframe ALS and FTD Inflammation Research
The strongest interpretation is gene-by-microbial-product. C9orf72 shapes how myeloid cells respond. Bacterial glycogen supplies the inflammatory trigger.
The mouse model connects that combination to barrier breakdown and neuroimmune activation. None of those pieces alone would justify the conclusion. Together, they describe a credible gut-brain pathway.
That is a real mechanism, not a license to sell microbiome fixes for ALS. The human data are small. The intervention evidence is animal-model work.
ALS and FTD are far more complex than any one bacterial carbohydrate. But the result gives the field a much more disciplined way to think about heterogeneity.
2 people with related genetic risk can carry different microbial exposures, different barrier states, and different immune thresholds — which helps explain why disease timing and inflammatory features vary so much across patients.
The next experiments are obvious. Measure inflammatory glycogen in larger, better-characterized cohorts.
Pair the measurement with diet, medication, gut symptoms, genotype, and systemic inflammation markers. Test whether the pathway changes inflammation before it changes disease.
That is a much sharper target than the question the field has been stuck with for years — whether “the microbiome” is generally good or bad for ALS.
Citation: DOI: 10.1016/j.celrep.2025.116906; McCourt et al; C9orf72 in myeloid cells prevents an inflammatory response to microbial glycogen; Cell Reports; 2026.
Study Design: Mechanistic gut-immunity study using bacterial screening, C9orf72-deficient immune models, germ-free mouse colonization, enzymatic glycogen digestion, and human fecal sampling.
Sample Size: Human fecal survey: 22 ALS, 1 C9ORF72-FTD, 12 healthy controls. Plus cell and mouse experiments.
Key Statistic: Inflammatory glycogen forms detected in 15 of 22 ALS fecal samples vs. 4 of 12 healthy control samples. Glycogen digestion improved survival and reduced microglial reactivity in C9orf72-deficient mice.
Caveat: Intervention evidence is animal-model only; human survey too small for diagnostic or treatment claims.






