Hypothesis

Mitochondrial DNA heteroplasmy in intestinal epithelial cells drives tissue‑specific selection of mtDNA variants that alter mitochondrial metabolite export (succinate, citrate). These exported metabolites reshape the gut microbiome composition, which in turn changes the production of neuroactive microbial metabolites (butyrate, TMAO). The shifted microbial metabolite profile modulates nuclear histone acetylation and DNA methylation in the brain, influencing expression of nuclear‑encoded mitochondrial genes and thereby establishing a gut‑mitochondria‑nucleus feedback loop that determines the rate of cognitive aging.

Rationale

Recent work shows somatic mtDNA mutations accumulate mainly through replication errors and positive selection, not as a primary aging driver 1. However, positive selection for specific mtDNA variants occurs in different tissues during aging, with functional decline only after a critical heteroplasmy threshold is crossed 2. This suggests mtDNA variants can be selected for metabolic properties that affect cellular export. It's been shown that microbiome metabolites such as butyrate cross the blood‑brain barrier and boost brain mitochondrial ATP production via HDAC inhibition 3, while TMAO damages hippocampal mitochondria and accelerates cognitive aging 4. Importantly, mitochondrial dysfunction can alter microbiome composition 5. These observations don't specify which mitochondrial properties are being selected by the microbiome or how microbial metabolites feed back to nuclear gene expression. We propose that the selected mtDNA variants change the flux of TCA cycle intermediates out of mitochondria, providing a direct chemical signal to the gut lumen that shapes microbial communities.

Predictions

We'll test that mice engineered to carry a heteroplasmic mtDNA mutation that increases succinate export from intestinal epithelial cells will show a reproducible shift in gut microbiome composition toward succinate‑utilizing taxa (e.g., Desulfovibrio spp.) within 4 weeks. This microbiome shift will correlate with decreased fecal butyrate levels and increased circulating TMAO. Reduced butyrate will lead to lower histone H3K27 acetylation in the hippocampus and decreased expression of nuclear‑encoded mitochondrial genes (e.g., Nrf1, Tfam), measurable by ChIP‑seq and RT‑qPCR. Cognitive performance in hippocampal‑dependent tasks (Morris water maze) will decline proportionally to the heteroplasmy load. Supplementation with sodium butyrate will restore hippocampal acetylation, rescue nuclear mitochondrial gene expression, and ameliorate cognitive deficits without altering the mtDNA heteroplasmy level.

Experimental Design

• Model: We'll generate a knock‑in mouse line expressing a pathogenic mtDNA mutation (e.g., m.5024C>T in mt‑Nd2) specifically in intestinal epithelial cells using Villin‑Cre‑driven mtDNA editor (DddA‑derived cytosine base editor) to achieve graded heteroplasmy (0%, 20%, 40%, 60%).
• Microbiome profiling: 16S rRNA sequencing of fecal samples at baseline and 4, 8, 12 weeks; we'll quantify succinate‑utilizing bacteria via qPCR.
• Metabolite analysis: Measure fecal succinate, butyrate, and plasma TMAO by LC‑MS.
• Brain assays: Hippocampal histone acetylation (ChIP‑seq for H3K27ac), RNA‑seq for nuclear mitochondrial genes, ATP production via Seahorse assay.
• Behavior: Morris water maze and novel object recognition test.
• Intervention: Cohorts receive oral sodium butyrate (2% w/v in drinking water) starting at week 8; we'll compare to vehicle controls.

Potential Outcomes and Falsifiability

If the predictions hold, the data will support a gut‑mitochondria‑nucleus axis where mtDNA‑driven metabolite export shapes the microbiome, which then feeds back to nuclear epigenetic regulation of mitochondrial function in the brain. Failure to observe any of the predicted links—e.g., no microbiome shift despite increased succinate export, or butyrate supplementation not rescuing cognitive decline—would falsify the hypothesis and suggest that mtDNA heteroplasmy influences aging through cell‑autonomous mechanisms rather than via microbiome‑mediated signaling.