The Wound Healing Process: Four Overlapping Phases, Explained Cell by Cell

The wound healing process is how the body closes a breach in its skin: four overlapping phases — hemostasis, inflammation, proliferation and remodelling — each run by different cells on a different clock. In adult skin it ends in repair rather than regeneration, which is why a closed cut leaves a scar that never regains the full strength of the skin around it.

Below: each phase cell by cell, why chronic wounds stall, how tissue repair is studied, and what peptide research has and has not shown.

What the wound healing process is

The wound healing process is the coordinated cellular response that restores the skin barrier after injury. Guo and DiPietro describe four programmed phases of wound healing — hemostasis, inflammation, proliferation and remodelling — that must occur in the proper sequence and time frame for a wound to close well. Textbooks call them the stages of wound healing, but Landén, Li and Ståhle stress that the phases are sequential yet overlapping: cells of one phase are still at work when the next begins.

In adults the end product is usually a fibrotic scar rather than restored skin, whereas early-gestation fetal tissue can regenerate without scarring (Gurtner and colleagues, 2008).

Primary versus secondary intention

When a wound’s edges are closed directly, as in a sutured incision, it heals by primary intention. A wound left open heals by secondary intention, “from the base upwards, without direct closure”: granulation first, then re-epithelialization and contraction (Potluru and colleagues, 2025). Secondary intention relies far more on granulation tissue and contraction.

Hemostasis: the first minutes

Hemostasis (US: hemostasis) begins the moment a vessel is cut. Exposed collagen sets off the clotting cascades, and aggregating platelets trigger vasoconstriction (Landén). Platelets adhere to the injured vessel wall within seconds, the receptor GPIIb/IIIa cross-links them, and fibrin is deposited into the growing thrombus (Eisinger and colleagues, 2018). The clot also serves as a provisional matrix, a scaffold for cells that migrate in later.

Platelets also send the first repair signals. Their alpha granules release TGF-β1; PDGF-BB, a strong mitogen and chemoattractant for mesenchymal cells; and VEGF, which drives endothelial sprouting (Eisinger). PDGF and TGF-β from platelets, and later from macrophages, draw fibroblasts into the wound (Landén), and platelets release thymosin beta-4 at sites of injury (Treadwell and colleagues, 2012).

Inflammation and the macrophage switch

Inflammation starts within hours and dominates roughly the first three days. Neutrophils arrive first, recruited by damage- and pathogen-associated molecular patterns (Krzyszczyk and colleagues, 2018), and in an uninfected wound usually stay two to five days (Landén). Monocytes follow and become macrophages: one review places their arrival at about day 3 (Landén), while in mice they are recruited from about 24 hours after wounding (Krzyszczyk).

Early macrophages are mainly pro-inflammatory, releasing TNF-α, IL-1β, IL-6, reactive oxygen species and nitric oxide (see our NF-κB pathway explainer). Later ones secrete IL-10, TGF-β, PDGF, IGF-1 and VEGF. Clearing apoptotic neutrophils, called efferocytosis, helps drive the switch (Landén). The M1/M2 labels are, in one review’s words, “grossly oversimplified”: wound macrophages form a spectrum (Krzyszczyk). When the transition stalls, inflammation persists.

Proliferation and granulation tissue

From about day 4 to day 21, new connective tissue fills the defect, vessels grow into it and the epidermis closes over it (Landén).

What granulation tissue is made of

Granulation tissue is the temporary, vascular connective tissue that fills a wound from its base. Fibroblasts migrate in from the dermis, proliferate and deposit collagen, fibronectin, proteoglycans and hyaluronic acid; with new vessels and macrophages, that matrix is granulation tissue. The first collagen is largely type III, made quickly but weaker than the type I that later replaces it, and fibroblasts also secrete matrix metalloproteinases (MMPs) that degrade the provisional clot matrix (Landén). VEGF from platelets and macrophages drives the capillary sprouting covered in our angiogenesis and VEGF guide.

Re-epithelialization

Keratinocytes at the wound edge start migrating within hours to a day, and from about day 2 or 3 basal keratinocytes and nearby stem cells proliferate behind them (Landén). In human skin, re-epithelialization and granulation tissue are what drive closure (Galiano and colleagues, 2004).

Contraction and the myofibroblast

Some fibroblasts become myofibroblasts under mechanical tension and TGF-β; they express alpha-smooth muscle actin (α-SMA), pull the edges together and undergo apoptosis once repair is complete (Landén). Gabbiani, Ryan and Majno first described them at the University of Geneva in 1971 as “modified fibroblasts” in granulation tissue. A 2012 review led by Boris Hinz of the University of Toronto’s Faculty of Dentistry, with Gabbiani as senior author, covers how α-SMA expression is controlled and the cells’ role in fibrosis.

The remodelling phase and the scar

The remodelling phase begins as granulation tissue formation ends and can continue for a year or more. Type III collagen is gradually replaced by type I, which is deposited more slowly but is stronger, and vessel numbers fall (Landén). Myofibroblasts and vascular cells are removed by apoptosis, leaving a collagen-rich scar, and MMPs from fibroblasts and macrophages degrade matrix throughout (Krzyszczyk; Landén). Healed skin reaches a maximum of roughly 80% of its original tensile strength, and after serious injury hair follicles and sweat glands may not regenerate (Landén).

Scarring versus regeneration

Adult mammalian skin repairs; some tissues and organisms regenerate, restoring the original architecture (Gurtner).

• Fetal skin. Human fetal skin from 15 to 22 weeks’ gestation, grafted onto athymic mice, repaired incisions without a scar when placed beneath the host’s skin but scarred when exposed to air; scarless repair needed neither fetal serum nor amniotic fluid (Lorenz and colleagues, 1992).
• Spiny mice. African spiny mice (Acomys) regenerate hair follicles in back-skin wounds, and their ear holes regenerate completely, with follicles, sebaceous glands, dermis and cartilage (Seifert and colleagues, Nature, 2012).
• Axolotls. Adult axolotls close full-thickness skin wounds without a scar, with a limited clot, fewer neutrophils, far fewer α-SMA-positive myofibroblasts than mouse wounds, little fibronectin and abundant tenascin-C; the authors present these differences as hypotheses to test (Seifert and colleagues, PLoS One, 2012).

Chronic wounds: repair that stalls

A chronic wound fails to progress through orderly repair to durable closure within an appropriate time, usually taken as about three months (Zhao and colleagues, 2016). It is the normal process stuck, most often in a self-perpetuating inflammatory stage: neutrophils release reactive oxygen species and proteases such as elastase and MMP-8, elastase degrades PDGF and TGF-β, and many wounds carry polymicrobial biofilms (Zhao). Macrophages stay pro-inflammatory, and chronic wound fluid has shown TNF-α and IL-1β about 100-fold higher than acute wound fluid (Krzyszczyk).

• Diabetic foot ulcers. Neuropathy, ischemia and trauma, plus glycation end-products, low-grade inflammation and microcirculatory dysfunction (Zhao); high glucose also impairs macrophage clearance of dead cells (Krzyszczyk).
• Venous leg ulcers. Venous hypertension lets macromolecules and red cells leak into tissue and attract leukocytes; these ulcers make up more than half of lower-limb chronic wounds (Zhao). Iron-overloaded, pro-inflammatory macrophages have been found in them (Landén).
• Pressure injuries. Sustained pressure or shear over bony prominences such as the sacrum, hips and malleoli causes ischemia; tissue dies from hypoxia and ischemia-reperfusion injury, sometimes after as little as two hours of immobility (Zhao).

Factors that slow repair

Guo and DiPietro list oxygenation, infection, age, sex hormones, stress, diabetes, obesity, medications, alcohol, smoking and nutrition as factors that can disrupt one or more phases; a bacterial burden above about 10⁵ organisms per gram of tissue is linked to impaired repair (Zhao).

How tissue repair is studied

Every model captures part of the wound healing process and misses the rest.

Geoffrey Gurtner’s group described splinting in 2004, noting that no rodent model closely parallels human healing (Galiano); a 2013 methods paper co-written by Edward Tredget of the University of Alberta’s Wound Healing Research Group in Edmonton sets it out step by step (Wang and colleagues). In any peptide experiment, matched controls separate sequence-specific from non-specific effects (see scrambled peptide controls), and a conclusion’s weight depends on whether it comes from cells, animals or people (see how to read a peptide study).

Peptides in wound-repair research

Each peptide below is labelled by the strongest evidence we could confirm.

LL-37: small human trials, mixed results

LL-37, the C-terminal fragment of the human cathelicidin hCAP18, is better known as an antimicrobial peptide. Heilborn and colleagues (2003) found hCAP18 upregulated after skin injury, peaking around 48 hours, in the inflammatory infiltrate and migrating epithelium, but absent from the epithelium at chronic-ulcer edges; in organ-cultured human skin wounds, anti-LL-37 antibodies inhibited re-epithelialization in a concentration-dependent way.

A first double-blind trial in 34 patients with hard-to-heal venous leg ulcers and three concentration arms reported a healing-rate constant about six times placebo’s in the lowest arm (p = 0.003); the middle arm’s difference was not significant and the highest arm did not differ (Grönberg and colleagues, 2014). The phase IIb HEAL LL-37 trial in 148 patients found no significant difference from placebo overall; a post hoc large-wound subgroup differed, which the authors say needs an adequately powered trial (Mahlapuu and colleagues, 2021). LL-37 was well tolerated in both (see our LL-37 research guide).

Thymosin beta-4: rodent data, small human trials

In a 1999 rat full-thickness wound study, thymosin beta-4 increased re-epithelialization by 42% over saline at day 4 and up to 61% at day 7, with more contraction, collagen and angiogenesis (Malinda and colleagues). A 2012 review reports faster punch-wound closure in several rodent models and refers to two phase 2 trials, in venous stasis and pressure ulcers, without patient numbers or significance tests in its abstract (Treadwell). See thymosin alpha-1 vs thymosin beta-4 and BPC-157 vs TB-500.

BPC-157: rodent studies, almost no human data

Much of BPC-157’s wound literature comes from one University of Zagreb group, whose 2021 review covers skin wounds, burns, diabetic ulcers, fistulas and musculoskeletal injuries, mostly in rats and mice. It mentions earlier ulcerative colitis and multiple sclerosis trials without results and describes no human wound trials (Seiwerth and colleagues). Our BPC-157 safety review covers the human data.

GHK-Cu: cell and rodent work

In a 1993 rat study, implanted wound chambers that received the tripeptide–copper complex GHK-Cu accumulated more collagen, glycosaminoglycans, DNA and protein than saline chambers, in a concentration-dependent way, while a control tripeptide had no significant effect (Maquart and colleagues). That is matrix accumulation in a rodent chamber, not human wound closure; see our GHK-Cu research overview.

Becaplermin: the approved growth-factor example

Becaplermin, a form of PDGF, was in 1997 the first growth factor approved by the US FDA for a wound indication (diabetic foot ulcers); one review calls its benefit modest (Zhao). It is not a research peptide.

What’s established and what isn’t

Sourcing repair-research peptides in Canada

For wound-model work, the usual documentation questions apply, plus two that matter more in repair biology.

• Ask for a batch-matched certificate with HPLC purity and a mass-spectrometry identity check, and match its batch number to the vial (see our peptide COA guide).
• Ask about endotoxin, because wound models lean on inflammatory and macrophage readouts (see endotoxin in research peptides).
• Know which thymosin molecule you have. TB-500 and TB-500 Fragment are separate products.
• Store peptides cold, dry and dark, as in our peptide storage guide, and see our overview of research peptides in Canada for judging a supplier.

LL-37Cathelicidin fragment studied in re-epithelialization and ulcer trialsView product TB-500 (Thymosin Beta-4)Thymosin beta-4; dermal wound data mainly from rodentsView product BPC-157Pentadecapeptide with wound studies mostly in rodentsView product

Frequently asked questions