Within-Family Mendelian Randomization Reveals Gut Microbiota Effects on Longevity Are Independent of Epigenetic Age Clock Changes

Hypothesis

Causal effects of specific gut microbiota taxa on human longevity are mediated primarily through pathways that do not involve changes in epigenetic biological age clocks, meaning that clock reversal is an epiphenomenal marker rather than a driver of microbiota‑induced lifespan extension.

Rationale

Recent Mendelian randomization (MR) studies have linked taxa such as Coriobacteriaceae and Oxalobacter to increased longevity odds, while Fusobacterium nucleatum shows negative associations [1]. However, summary‑level MR is vulnerable to sample structure confounders (population stratification, cryptic relatedness, sample overlap) that can generate spurious causal inferences [2]. Moreover, epigenetic clocks like GrimAge reflect a mixture of damage and repair signals, and setting back the clock may suppress vital repair mechanisms [3, 4]. Until we disentangle whether microbiota‑driven longevity operates through clock modification or independent mechanisms, using clocks as primary endpoints for longevity interventions remains questionable.

Testable Predictions

1. After adjusting for sample structure with MR‑APSS (or within‑family MR), the causal estimate of microbiota taxa on longevity will remain significant.
2. The same microbiota taxa will show either no causal effect on GrimAge acceleration or an effect that is opposite in direction to their effect on longevity.
3. Multivariable MR will demonstrate that the microbiota‑longevity association is not attenuated when GrimAge is included as a covariate, indicating mediation independence.
4. Instrumental variable‑based mediation analysis (e.g., two‑step MR with G‑estimation) will reveal a negligible indirect effect of microbiota on mortality through epigenetic age change (<10% of total effect).

Experimental Design

Data sources: Summary statistics from large GWAS of gut microbiome composition (e.g., MiBioGen consortium), longevity (parental lifespan or UK Biobank survival), and GrimAge acceleration (from epigenome‑wide association studies).

Step 1 – Control for sample structure: Apply MR‑APSS or conduct within‑family MR using sibling pairs from the UK Biobank to eliminate bias from population cryptic relatedness and overlap.

Step 2 – Primary causality: Perform inverse‑variance weighted MR (with MR‑Egger and weighted median sensitivity) to estimate the effect of each taxa on longevity.

Step 3 – Clock association: Repeat MR for the same taxa on GrimAge acceleration.

Step 4 – Mediation analysis: Use a two‑step MR approach where the exposure is microbiota taxa, the mediator is GrimAge, and the outcome is longevity. Implement G‑estimation of a structural nested mean model to estimate the natural indirect effect, conditioning on measured covariates and testing the null hypothesis of no mediation.

Step 5 – Validation: Cross‑validate findings with the deep learning BA‑CA model [6] to ensure results are not clock‑specific.

Potential Outcomes and Interpretation

• If predictions hold (significant microbiota‑longevity effect, null or opposing microbiota‑GrimAge effect, negligible indirect effect), we conclude that microbiota influence lifespan through mechanisms such as mucosal barrier integrity, systemic immunomodulation, or metabolite‑driven hormesis that operate independently of—or even in opposition to—epigenetic clock state. This would challenge the notion that clock reversal is a necessary mediator of longevity and caution against using clocks as surrogate endpoints in microbiota‑targeted interventions.
• If indirect effect is substantial (e.g., >30% of total effect), it would support a model where specific taxa alter DNA methylation machinery (e.g., via butyrate‑mediated inhibition of HDACs or folate‑mediated one‑carbon metabolism) to shift clock state, thereby contributing causally to lifespan extension. In this scenario, clocks could be valid biomarkers, but only after confirming that the direction of change aligns with beneficial repair rather than damage.

Either outcome resolves a critical gap in aging biomarker research: it clarifies whether epigenetic age clocks are causal mediators or epiphenomenal correlates of microbiota‑driven longevity, guiding the selection of appropriate endpoints for future geroprotective trials.