Wound healing

The conserved multicellular program that restores tissue barrier function and, where possible, tissue architecture following physical injury. Skin is the canonical context — most mechanistic detail and all classic four-phase models derive from dermal/epidermal wound studies — but the same general program operates in cornea, gut mucosa, vascular intima, cardiac muscle, and skeletal muscle, with phase timing and cell-type composition varying by tissue.

Wound healing deteriorates with age across all studied organisms, contributing to the clinical burden of chronic non-healing wounds in elderly populations. Its intersection with cellular-senescence is mechanistically distinctive: transient senescence is required for optimal acute healing (via PDGF-AA paracrine signaling), whereas the chronic senescent-cell accumulation characteristic of aged tissue impairs healing — a canonical instance of antagonistic-pleiotropy.

The four phases

Phase	Duration	Dominant cell types	Key signals	ECM produced
Hemostasis	Minutes–hours	Platelets, endothelial cells	Thrombin activates fibrinogen → fibrin; platelet α-granules release PDGF, TGF-β, VEGF	Fibrin/fibronectin provisional matrix
Inflammation	1–3 days	Neutrophils (first wave, hrs–day 2); macrophages M1 then M2	IL-1β, IL-6, TNF-α (M1); TGF-β, IL-10 (M2 switch); reactive oxygen species (neutrophil oxidative burst)	Provisional matrix retained; proteolytic debridement
Proliferation	3–14 days	Fibroblasts, myofibroblasts, keratinocytes, endothelial cells	TGF-β1 drives myofibroblast differentiation (α-SMA+); PDGF recruits/activates fibroblasts; VEGF drives angiogenesis	Collagen III-enriched granulation tissue; col3a1 predominates early
Remodeling	Weeks–>1 year	Myofibroblasts (then apoptose), macrophages	lox cross-links collagen fibers; MMP/TIMP balance governs remodeling	Collagen III → col1a1 replacement; tensile strength asymptotes ~70–80% of uninjured skin at ~1 year ¹ no-fulltext-access (Singer 1999 closed-access; figure widely cited but not verified against full text)

Phase notes

Hemostasis is the fastest phase and the only one not primarily cellular — thrombin converts circulating fibrinogen to fibrin, which traps platelets and forms the provisional matrix. Platelet α-granules simultaneously release PDGF (both -AA and -BB isoforms), TGF-β1, TGF-β2, and VEGF, establishing the early chemotactic gradient that recruits inflammatory cells. The provisional fibrin matrix also serves as a scaffold for subsequent cell migration ².

Inflammation requires coordinated neutrophil-to-macrophage handoff. Neutrophils clear bacteria and debris via elastase + MPO-mediated oxidative burst but also secrete proteases (MMP-8, MMP-9) that begin matrix debridement. Macrophage arrival (from circulating monocytes and tissue-resident pools) shifts the environment from pro-inflammatory M1 (microbicidal, phagocytic) toward pro-healing M2 (anti-inflammatory, growth-factor-secreting). The M1→M2 transition is obligatory for productive healing; failure to complete it is a hallmark of non-healing chronic wounds ².

Proliferation involves three overlapping subprocesses:

Fibroblast activation: fibroblasts from the wound margin and dermis migrate into the fibrin matrix, proliferate, and remodel it into granulation tissue. Key signal: PDGF-BB from platelets and macrophages stimulates fibroblast recruitment via PDGFRβ.
Myofibroblast differentiation: a subset of wound fibroblasts differentiates into contractile α-SMA+ myofibroblasts under TGF-β1 (canonical signal), mechanical tension, and specific matrix cues. Myofibroblasts generate the contractile force that closes the wound gap.
Re-epithelialization: keratinocytes from wound margins and hair follicles migrate centripetally beneath the fibrin scab, driven by EGF, KGF (FGF-7), and TGF-α; they proliferate to restore the epidermal barrier.
Angiogenesis: capillary sprouting driven by VEGF-A and FGF-2 supplies the metabolically active granulation tissue. New capillaries later regress during remodeling.

Remodeling replaces the mechanically weak, collagen-III-rich granulation tissue with a more ordered type-I collagen matrix. lox (lysyl oxidase) catalyzes covalent cross-links between collagen fibers and between elastin chains. Myofibroblasts apoptose as the mechanical work is completed; residual scar tissue lacks follicular appendages. Maximum tensile strength (~80% of original) is never fully recovered because the highly ordered fibril architecture of uninjured dermis is not reconstructed ¹. no-fulltext-access (Singer 1999 is closed-access; exact figure not verifiable against full text from local archive) needs-human-replication for quantitative LOX-deficiency phenotypes in human aging.

Transient senescence in wound healing

The role of cellular-senescence in wound healing is context-dependent: transient, acute senescent cells are beneficial; chronic, accumulated senescent cells are pathological.

Demaria 2014: PDGF-AA-secreting senescent cells are required for optimal healing

Using the p16-3MR transgenic mouse (which allows GCV-mediated selective elimination of p16-expressing cells), Demaria et al. showed ³:

Within 3 days of a 6 mm punch biopsy, p16+ cells appear at wound edges and peak days 3–6, then resolve — a transient population distinct from chronically accumulated aged-tissue senescent cells.
Both fibroblasts and endothelial cells undergo wound-site senescence; endothelial cells are described as the primary senescent population (Discussion: “wounding induces senescence in resident fibroblasts and primarily endothelial cells”).
GCV-mediated elimination of p16+ cells (25 mg/kg i.p., days 1–6 post-wounding) delays wound closure, with peak delay at day 6.
p16/p21 DKO mice (unable to senesce) show the same delay, ruling out off-target GCV toxicity as the cause.
The key SASP factor driving healing is PDGF-AA — not PDGF-BB, IL-6, or VEGF. Senescent-cell conditioned medium promotes myofibroblast (α-SMA+) differentiation in neighboring non-senescent fibroblasts; anti-PDGF-AA antibody blocks this effect; anti-PDGF-BB does not.
Topical recombinant PDGF-AA fully rescues the GCV wound-closure delay, confirming PDGF-AA is the rate-limiting paracrine signal and that senescent cells do not drive healing via direct contraction.
Senescent cells themselves do NOT upregulate α-SMA — their role is purely paracrine (PDGF-AA → neighboring fibroblasts → myofibroblast differentiation).

Mechanism summary:

Wound stimulus
  → Transient senescence in wound-site fibroblasts + endothelial cells (primarily endothelial)
  → SASP: PDGF-AA secreted paracrine
  → PDGFRα signaling on neighboring fibroblasts → α-SMA+ myofibroblast differentiation
  → Myofibroblast contraction + granulation tissue → wound closure
  → (Youth) Immune clearance of senescent cells → tissue resolution
  → (Aging) Incomplete clearance → persistent SASP → tissue damage

This paper provides the canonical molecular-level evidence that cellular senescence is subject to antagonistic-pleiotropy: the same SASP program is acutely adaptive (PDGF-AA-mediated healing) and chronically pathological (pro-inflammatory, pro-degradative senescent accumulation in aged tissue).

Implications for senolytics timing

Timing of senolytic administration matters. Senolytics administered during the acute wound-healing window (days 1–6 post-injury) could theoretically deplete PDGF-AA-secreting wound-site senescent cells and impair healing. This is the pre-clinical precaution established by Demaria 2014. Conversely, for chronic wound contexts where pathological senescent-cell accumulation is the dominant problem (diabetic foot ulcers, venous stasis ulcers), senolytic + senomorphic approaches may be beneficial.

A 2026 review in J Biochem (Ueda et al.) focuses specifically on this dialectic in diabetic skin, noting that senescent dermal fibroblasts, macrophages, and adipose-tissue cells accumulate in diabetic wounds and sustain a pathological SASP that prevents the normal M1→M2 macrophage transition and impairs fibroblast contractility ⁴. This is consistent with Demaria 2014 but mechanistically distinct: in diabetic wounds, the senescent-cell burden is chronic and pre-existing, not acutely beneficial.

needs-human-replication — No human RCT has specifically tested wound healing as a primary endpoint for any senolytic compound. Context (acute healing vs. chronic wound) is likely critical for any future trial design.

Wound healing in aging

Multiple concurrent mechanisms reduce wound-healing capacity with age ²:

Mechanism	Hallmark link	Evidence strength
Reduced fibroblast proliferative and migratory capacity	cellular-senescence (accumulated p16+ fibroblasts)	In-vitro; some in-vivo mouse
Impaired keratinocyte migration and interfollicular stem-cell stiffness loss	stem-cell-exhaustion	Mouse + early human data ⁵
Reduced angiogenesis (↓ VEGF response, vascular rarefaction)	mitochondrial-dysfunction / endothelial aging	Predominantly mouse
Dysregulated macrophage M1→M2 transition (prolonged M1)	chronic-inflammation (inflammaging)	Mouse + ex-vivo human macrophage data
Chronic SASP from accumulated senescent cells → pro-degradative + anti-proliferative	cellular-senescence	Mouse (in-vivo); human correlative
Reduced growth factor signaling (PDGF, TGF-β, EGF axes)	Multiple	Mouse + human observational

Aged skin shows a paradox: chronic senescent-cell burden contributes to a pro-inflammatory environment that impairs healing, while the acute wound-site transient senescence program may itself be impaired (reduced capacity to induce protective senescence acutely). The two phenomena are not mutually exclusive — the ratio of functional acute-senescent-cell induction to chronic pathological senescent-cell accumulation may determine the wound-healing outcome. no-mechanism for the quantitative contribution of each arm in humans.

A 2026 study (Miny et al., Exp Dermatol) found that interfollicular stem cells (ISCs) of the human epidermis lose a distinctive biomechanical stiffness signature with age — interpreted as a functional deterioration of the epidermal stem-cell pool that drives barrier renewal and re-epithelialization ⁵. This is consistent with the stem-cell-exhaustion hallmark operating at the epidermal basal-layer level.

Therapeutic interventions

Topical growth factors

Becaplermin (Regranex) — recombinant human PDGF-BB. FDA-approved 1997 for lower-extremity diabetic neuropathic ulcers. Mechanism: PDGFRβ-mediated fibroblast and monocyte/macrophage chemotaxis and proliferation. PDGF-BB is distinct from the wound-site-senescent-cell isoform (PDGF-AA); both signal through fibroblasts but via different receptor subunits (PDGFRβ vs PDGFRα). needs-replication — direct comparison of PDGF-AA vs PDGF-BB topical efficacy in aged wound models has not been performed post-Demaria 2014.

Peptide therapeutics

bpc-157 — synthetic pentadecapeptide; preclinical evidence for accelerating wound healing in rat models via upregulation of growth hormone receptor expression and EGF/VEGF pathways. Human evidence limited. needs-human-replication
tb-500 — synthetic fragment of thymosin β4 (Tβ4); promotes actin polymerization, keratinocyte and endothelial cell migration; preclinical wound-healing evidence (rat models). Human evidence absent for wound-healing indication. needs-human-replication
ghk-cu — copper-binding tripeptide (Gly-His-Lys); endogenous plasma peptide; promotes fibroblast migration and proliferation, collagen and elastin synthesis, and wound healing in rodent and in-vitro models. Human evidence limited to cosmetic formulation studies. needs-human-replication
ll-37 — cathelicidin antimicrobial peptide; dual antimicrobial + wound-healing function; induces keratinocyte and endothelial migration; some evidence for role in re-epithelialization; topical LL-37 studies in diabetic wound models (preclinical). needs-human-replication

Senolytics (timing-dependent)

senolytics (fisetin, dasatinib + quercetin) are in principle candidates for improving wound healing in aged skin by reducing the chronic senescent-cell burden — but must be timed to avoid the acute healing window (days 1–14 post-injury), where PDGF-AA-secreting senescent cells are required per Demaria 2014.

A 2026 in-vivo paper in diabetic mice (Numani et al., Adv Wound Care; DOI: 10.1177/21621918261426580) reported that topical fisetin reduces cutaneous senescent-cell burden and improves wound closure — a context where the chronic pathological senescent-cell program predominates. needs-human-replication and the generalizability to non-diabetic aged wounds requires separate evaluation.

Pathological wound-healing states

Chronic non-healing wounds — diabetic foot ulcers, venous stasis ulcers, pressure ulcers. Common features: arrested in the inflammatory phase, failure of M1→M2 transition, biofilm colonization, excessive MMP activity degrading growth factors and matrix. Senescent-cell accumulation is a candidate contributor (Ueda 2026) ⁴.

Hypertrophic scarring and keloids — pathological fibroproliferative responses driven by TGF-β overactivation, myofibroblast persistence, and failure of apoptosis-mediated myofibroblast clearance during remodeling. Mechanically distinct from normal scarring: keloids extend beyond the original wound boundary.

Fetal scarless healing — fetal wounds (before the third trimester) heal without scar formation. Proposed mechanisms: (1) collagen-III-enriched fetal wound milieu enabling more ordered matrix deposition; (2) attenuated inflammatory response (reduced TGF-β1/TGF-β2 signaling; relatively elevated TGF-β3); (3) sterile, low-oxygen amniotic environment. A target for scar-reduction research.

Extrapolation from mouse models

The Demaria 2014 PDGF-AA findings rest entirely on the p16-3MR mouse model.

Dimension	Status	Notes
Pathway conserved in humans?	yes	Senescence, PDGF-AA/PDGFRα signaling, and myofibroblast biology are conserved in human skin; PDGF-AA promotes human fibroblast-to-myofibroblast differentiation
Phenotype conserved in humans?	partial	p16+ senescent cells appear at human wound sites (histological evidence); whether acute depletion impairs human healing is untested; aged human wounds do heal more slowly and show reduced PDGF signaling — consistent but not causal
Replicated in humans?	no	No human equivalent of the p16-3MR GCV experiment exists; no senolytic wound-healing RCT has been completed

See model-organisms/_extrapolation-guide.md for evaluation rubric.

Limitations and gaps

No human RCT directly testing the transient-senescence / PDGF-AA axis. The entire mechanistic framework is mouse-derived. needs-human-replication
Senolytic timing window in human acute wounds is undefined. Preclinical caution established; human wound-specific senolytic trial data absent. dose-response-unclear for senolytic timing
Macrophage M1→M2 transition mechanism in aged human wounds. Mouse data support impaired transition; human ex-vivo evidence limited; interventional evidence absent. needs-human-replication
Relative contributions of fibroblast vs. endothelial senescent populations to PDGF-AA secretion at wound sites were not separately quantified in Demaria 2014. needs-replication
Peptide therapeutics (BPC-157, TB-500, GHK-Cu) lack adequate human RCT data for wound-healing indications; most evidence is preclinical rodent. needs-human-replication
Quantitative LOX-deficiency contribution to age-related scar quality — human data absent; extrapolated from mouse and in-vitro. needs-human-replication
col3a1 page — referenced in granulation-tissue ECM context but no wiki page yet. stub

Cross-references

Related processes: sasp, cellular-senescence, replicative-senescence, autophagy

Key cell types: dermal-fibroblasts, keratinocytes

Tissues: skin, dermis, epidermis

Molecules: col1a1, col3a1 (stub), lox (stub), pdgf-aa (stub), tgf-beta (stub)

Hallmarks: cellular-senescence, chronic-inflammation, stem-cell-exhaustion, antagonistic-pleiotropy

Interventions/compounds: bpc-157, tb-500, ghk-cu, ll-37, fisetin, dasatinib, quercetin, senolytics

Phenotypes/pathology: skin-aging

Key study: demaria-2014-senescent-cells-wound-healing

Footnotes

doi:10.1056/NEJM199909023411006 · PMID 10471461 · Singer AJ, Clark RAF · N Engl J Med 1999;341(10):738–746 · review · model: human — classic review of cutaneous wound healing phases; tensile strength asymptote (~70–80% at 1 year) and four-phase model definitions; closed-access — not in local archive; tensile strength figure not verified against full text no-fulltext-access ↩ ↩²
doi:10.1038/nature07039 · Gurtner GC, Werner S, Barrandon Y, Longaker MT · Nature 2008;453:314–321 · review · model: human + mouse — comprehensive review of wound repair and regeneration biology including aging context; end-to-end verified 2026-05-19 (local PDF). Note: Gurtner uses three stages (inflammation / new tissue formation / remodelling); the four-phase model including hemostasis as a distinct phase follows Singer 1999 convention. ↩ ↩² ↩³
demaria-2014-senescent-cells-wound-healing · n≥4/group (wound closure); n=3–5 (molecular assays) · in-vivo + in-vitro · p<0.05 (wound delay; PDGF-AA rescue) · model: p16-3MR transgenic mus-musculus; 6 mm dorsal punch biopsy; GCV senescent-cell ablation; R39 verified 2026-05-19 ↩
doi:10.1093/jb/mvag006 · Ueda N et al. · J Biochem 2026 · review · model: diabetic skin (mouse + human correlative) — role of cellular senescence (fibroblast, macrophage, adipose-tissue) in diabetic wound-healing impairment; senolytic/senomorphic therapeutic approaches; not OA; not end-to-end verified ↩ ↩²
doi:10.1111/exd.70268 · Miny S et al. · Exp Dermatol 2026 · in-vitro + ex-vivo · model: human epidermal biopsies (atomic force microscopy) — interfollicular stem cells lose biomechanical stiffness signature with age; consistent with stem-cell-exhaustion contribution to impaired re-epithelialization; DOI confirmed via PubMed; not locally downloaded ↩ ↩²

🧬 Aging Wiki

Explorer

wound-healing