Hypothesis

Age‑associated shifts in the gut virome—specifically an increase in lytic bacteriophages targeting beneficial commensals—trigger inflammaging by releasing bacterial DNA that activates the cytosolic DNA sensor cGAS‑STING in enteroendocrine cells. This activation alters serotonin signaling to the brain through vagal afferents, amplifying neuroinflammation and accelerating cognitive decline.

Mechanistic rationale

1. Phage expansion with age – Metagenomic studies show a rise in Caudovirales richness in older adults, correlating with reduced Lactobacillus and Akkermansia abundances (1).
2. Lysis‑derived DNA – Lytic phage infection causes bacterial cell death, liberating genomic DNA into the lumen. A fraction traverses the epithelium via microfold cells or transient barrier leaks.
3. cGAS-STING in enteroendocrine cells – Cytosolic DNA binds cGAS, producing cyclic GMP‑AMP that stimulates STING, leading to type‑I interferon and NF‑κB signaling. Enteroendocrine cells express high levels of cGAS and release serotonin upon STING activation (2).
4. Serotonergic vagal signaling – Altered serotonin release modulates vagal afferent firing, shifting brainstem nuclei toward a pro‑inflammatory state and reducing anti‑inflammatory cholinergic tone (3).
5. Amplification loop – Vagal dysregulation further increases gut permeability, allowing more LPS and phage particles to translocate, sustaining cGAS‑STING activation.

This cascade provides a bidirectional leak: virome changes influence brain state, and brain‑derived stress hormones (e.g., cortisol) can modulate gut mucosal immunity, affecting phage‑bacterial equilibria.

Testable predictions

• Prediction 1: In a longitudinal cohort (n ≥ 200) aged 60‑80, baseline stool virome diversity (Shannon index of phage taxa) will inversely predict change in Mini‑Mental State Examination (MMSE) score over 24 months, independent of bacterial alpha diversity (5).
• Prediction 2: Plasma levels of circulating cyclic GMP‑AMP (cGAMP) will positively correlate with both virome dysbiosis scores and CSF markers of neuroinflammation (e.g., YKL‑40).
• Prediction 3: Oral administration of a defined lytic phage cocktail targeting Enterobacteriaceae in aged mice will increase serum cGAMP, reduce colonic serotonin, and impair performance in the Morris water maze; conversely, fecal virome transplantation (FVT) from young donors will normalize cGAMP, restore serotonin, and rescue memory deficits.

Experimental approach

1. Human observational arm – Collect quarterly stool, blood, and cognitive assays; perform shotgun metagenomics for phage quantification, measure plasma cGAMP via ELISA, and assess CSF cytokines via lumbar puncture subset.
2. Mouse interventional arm – Use 20‑month‑old C57BL/6 mice; administer phage cocktail or sterile filtrate FVT for 8 weeks; sample lumen DNA, enteroendocrine cGAS‑STING immunostaining, colonic serotonin (HPLC), vagal nerve electrophysiology, and behavior.
3. Intervention test – Treat a subgroup with a STING inhibitor (e.g., H-151) to determine whether blocking the axis abrogates phage‑induced cognitive decline.

Falsifiability

If virome diversity shows no relationship with cognitive trajectories, or if plasma cGAMP fails to correlate with virome shifts and neuroinflammatory markers, the core mechanistic link is unsupported. Likewise, if phage modulation does not alter cGAMP, serotonin, or cognition in mice, the hypothesis would be refuted.

By targeting the neglected viral component of the gut microbiome, this hypothesis expands the inflammaging framework beyond bacteria, offering a concrete, quantifiable target for both diagnostic biomarkers and therapeutic intervention.