Hypothesis

Transient senescent fibroblasts release extracellular vesicles (EVs) enriched in PDGF‑AA and specific microRNAs that reprogram infiltrating macrophages toward a pro‑repair, anti‑fibrotic phenotype. Senolytic drugs, while clearing senescent cells, also abolish this EV‑mediated signaling, shifting macrophage polarization toward a pro‑fibrotic M2‑like state and thereby exacerbating scar formation despite reduced senescent cell burden.

Mechanistic reasoning

1. EV cargo specificity – Young senescent fibroblasts show heightened expression of PDGF‑AA and miR‑21/miR‑29b within secreted EVs (PMC4349629; doi.org/10.1101/2025.06.08.658533). These molecules are known to activate fibroblast proliferation and simultaneously inhibit TGF‑β‑driven collagen transcription in macrophages.
2. Macrophage re‑education – EV‑delivered PDGF‑AA engages PDGFR‑α on macrophages, triggering STAT6‑dependent up‑regulation of ARG1 and downregulation of iNOS, while miR‑29b suppresses COL1A1 transcription in the same cells. This creates a macrophage population that secretes MMPs and limits fibroblast‑to‑myofibroblast transition.
3. Senolytic interference – Dasatinib + quercetin (D+Q) and navitoclax induce apoptosis in senescent cells but also disrupt vesicle biogenesis by inhibiting the neutral sphingomyelinase pathway required for exosome release (PMC12190739). Consequently, the EV burst that normally follows injury is blunted.
4. Phenotypic shift – Loss of EV signaling skews macrophages toward a classic M1/M2 hybrid that produces elevated TGF‑β1 and PDGF‑BB, promoting fibroblast activation and collagen deposition, thereby explaining the paradox of increased fibrosis after senolytic treatment observed in some models.

Testable predictions

• Prediction 1: In wounded mouse skin, D+Q treatment will reduce senescent cell numbers (p16^INK4a^+ cells) by >70% but concomitantly decrease EV‑associated PDGF‑AA and miR‑29b levels in wound interstitial fluid by >50% at 24 h post‑injury (measured by ELISA and small‑RNA‑seq).
• Prediction 2: Exogenous supplementation of purified senescent fibroblast‑derived EVs (or recombinant PDGF‑AA + miR‑29b mimic) to D+Q‑treated wounds will rescue macrophage ARG1 expression, reduce collagen deposition at day 15, and restore wound closure rates to those of untreated controls.
• Prediction 3: Conditional knockout of nSMase2 (Smpd3) specifically in fibroblasts will phenocopy the senolytic effect: normal senescent cell accumulation but impaired EV release, leading to exacerbated fibrosis despite unchanged senescent cell burden.

Falsifiability

If EV PDGF‑AA/miR‑29b levels remain unchanged after senolytic clearance, or if EV supplementation fails to modulate macrophage phenotype and fibrosis, the hypothesis would be refuted. Conversely, observing the predicted changes would support the notion that senescent cells act as EV‑based chaperones whose loss—not their presence—drives maladaptive repair.

Broader implication

This reframes senolytics not merely as senescent‑cell assassins but as inadvertent blockers of a regenerative vesicular communication network, suggesting that therapeutic success may require preserving or mimicking the EV senescent secretome while eliminating persistent, damaging senescence.