Hypothesis

Targeted augmentation of slow-wave sleep (SWS) via closed-loop auditory stimulation will increase glymphatic clearance of synaptic tagging proteins (e.g., phosphorylated PSD‑95 and synaptophysin fragments), thereby reducing downstream microglial activation and SASP‑driven epigenetic aging as measured by DunedinPACE.

Rationale

Recent work shows that sleep-dependent CSF influx is an active triage system that removes metabolic waste and decides which neural architectures persist ([1][2]). Disruption of SWS elevates CSF amyloid‑β ([2]) and impairs autophagy of synaptic remnants ([3]), while poor sleep quality accelerates epigenetic aging clocks ([4][5][6]). Noradrenaline suppression during SWS is required for aquaporin‑4 polarization and glymphatic inflow ([7]), yet the specific cargo cleared beyond soluble metabolites remains undefined.

We propose that the glymphatic system does not merely flush bulk solutes; it selectively extracts synaptic tagging proteins that accumulate at synapses during wakefulness as markers of synaptic usage and plasticity. These tags, when not cleared, are taken up by perivascular microglia, triggering a low‑grade proinflammatory state (SASP) that drives DNA methylation changes captured by GrimAgeAccel and DunedinPACE ([4][5]). Thus, the link between sleep loss and epigenetic aging may be mediated by failed synaptic tag clearance → microglial activation → SASP → epigenetic drift.

Novel Mechanistic Insight

1. Synaptic Tag Hypothesis – During wakefulness, activity‑dependent phosphorylation of postsynaptic density proteins (e.g., PSD‑95) creates soluble fragments that diffuse into the extracellular space. These fragments act as "tags" signaling synaptic history.
2. Glymphatic Selectivity – The SWS‑linked CSF influx, facilitated by suppressed noradrenaline and astrocytic aquaporin‑4 polarization, preferentially convects these tagged fragments due to their size and charge profile, analogous to size‑exclusion in the lymphatic system.
3. Microglial Sensing – Uncleared tags bind microglial TREM2 receptors, activating NF‑κB and SASP secretion (IL‑1β, IL‑6). SASP cytokines then alter DNA methyltransferase (DNMT) activity and histone acetylation patterns, accelerating epigenetic aging.
4. Feedback Loop – Elevated SASP further disrupts SWS by increasing cortical noradrenaline tone, creating a vicious cycle of poor sleep, tag accumulation, and aging.

Testable Predictions

• Prediction 1: In older adults (60‑80 y), two nights of closed-loop auditory stimulation timed to up‑states of slow oscillations will increase CSF flux of phosphorylated PSD‑95 and synaptophysin fragments by ≥30% compared to sham stimulation (measured via lumbar CSF sampling).
• Prediction 2: The same intervention will reduce CSF levels of microglial activation markers (sTREM2, YKL‑40) and SASP cytokines (IL‑1β, IL‑6) by ≥20% relative to baseline.
• Prediction 3: Participants exhibiting the greatest increase in synaptic tag clearance will show the largest reduction in DunedinPACE acceleration (Δ DunedinPACE ≤ -0.05) after a 2‑week nightly stimulation regimen, independent of changes in total sleep duration.
• Prediction 4: Pharmacological blockade of noradrenaline reuptake (e.g., with atomoxetine) during stimulation will abolish the clearance effect, confirming the necessity of low noradrenaline for glymphatic selectivity.

Experimental Design

• Crossover, double‑blind, sham‑controlled trial (n=40). Each participant receives 2 nights of active stimulation and 2 nights of sham, separated by ≥1 week.
• Measurements:
  • High‑density EEG to verify slow‑wave enhancement.
  • Pre‑ and post‑sleep lumbar CSF for tagged synaptic proteins (ELISA/MS), amyloid‑β, tau, sTREM2, cytokines.
  • Blood draws before and after each night for DunedinPACE calculation using CpG‑based algorithm.
  • Actigraphy and subjective sleep quality to control for total sleep time.
• Analysis: Mixed‑effects models testing interaction between condition (active vs sham) and biomarker changes, mediation analysis to test whether CSF tag reduction mediates the effect on DunedinPACE.

Falsifiability

If active stimulation fails to elevate CSF synaptic tag clearance, or if clearance does not correlate with reduced SASP markers and DunedinPACE improvement, the hypothesis would be refuted. Conversely, a dissociation—where clearance improves SASP but not epigenetic age—would suggest alternative pathways dominate sleep‑linked aging.

Implications

Confirming this mechanism would position glymphatic‑mediated synaptic pruning as a reversible, low‑risk lever for epigenetic rejuvenation, complementary to but safer than cellular reprogramming. It would also explain individual variability in sleep‑dependent resilience: those with efficient tag clearance sustain youthful epigenomic profiles despite equivalent sleep duration.

Key References

[1] https://doi.org/10.1101/2024.08.30.610454 [2] https://doi.org/10.1093/brain/awx148 [3] https://elifesciences.org/articles/64140 [4] https://pmc.ncbi.nlm.nih.gov/articles/PMC11253334/ [5] https://doi.org/10.1016/j.bbi.2015.08.024 [6] https://www.aging-us.com/article/203372/text [7] https://pubmed.ncbi.nlm.nih.gov/31605901/ [8] https://doi.org/10.1101/2025.06.03.657596