GHK-Cu

GHK-Cu is a fragment of the extracellular matrix before it is anything else. The glycyl-histidyl-lysine sequence sits inside the alpha-2 chain of type I collagen, and proteases at an injury site can cut it loose; a second matrix protein, SPARC, releases the same motif when it is degraded. Molecules like this are called matrikines — the matrix signalling its own repair through its own breakdown products. It was isolated from human serum as a growth-modulating factor in 1977, and three years later shown to carry copper, with a histidyl-lysyl motif that mirrors the copper transport site on albumin. Almost everything the compound is sold for descends from those two facts.

What GHK-Cu does in repair tissue

The best-replicated action is on matrix synthesis, and it happens at concentrations far below what the word "supplement" usually implies. In fibroblast culture, collagen synthesis begins responding between picomolar and low nanomolar concentrations and peaks around one nanomolar, with no change in cell number — the cells are not multiplying, they are building. Glycosaminoglycan synthesis rises on the same curve, and the curve is biphasic: push the concentration higher and the effect falls back toward baseline. More is not better in this literature.

Which matrix molecules go up depends on where you look. In fibroblast culture the peptide preferentially stimulates dermatan and heparan sulfate, the skin-matrix glycosaminoglycans. In living wound tissue the picture is wider: repeated injection into rat wound chambers enhanced the accumulation of chondroitin sulfate as well as dermatan sulfate, raised decorin messenger RNA and lowered biglycan. Decorin is the interesting one for a joint reader — a small proteoglycan that organises collagen fibrils, present in cartilage as well as skin.

The compound also modulates the enzymes that take matrix apart, and this is where the mechanism becomes genuinely two-directional. In fibroblast culture it raised MMP-2 alongside its inhibitors TIMP-1 and TIMP-2; in rat wound chambers it prolonged MMP-9 expression into the late remodelling phase and raised active MMP-2 at days 18 and 22, while leaving interstitial collagenase untouched. In an ischemic rat wound the direction reversed entirely, with treated wounds holding significantly lower TNF-alpha, MMP-2 and MMP-9 than controls. MMP-2 and MMP-9 are the gelatinases that degrade cartilage. A molecule that moves them in different directions in different tissues is not one you can extrapolate into a joint on mechanism alone.

The in-vivo wound literature is the part that is genuinely large. The foundational experiment injected GHK-Cu into implanted rat wound chambers and found concentration-dependent increases in dry weight, total protein, collagen and glycosaminoglycan content, with collagen stimulated twice as much as non-collagen protein and no significant change in DNA. It replicates across species and models: ischemic rat wounds closed 64.5 percent by day 13 against 45.6 percent for vehicle and 28.2 percent untreated; full-thickness rabbit wounds contracted faster than untreated at days 7, 14 and 21; injected dog pad wounds carried significantly more type I collagen at day 14; mouse scald wounds healed in 14 days with liposomal GHK-Cu, which beat the free peptide on angiogenesis. Diabetic-rat wounds treated with a biotinylated-GHK collagen matrix contracted faster with more granulation collagen and higher tissue antioxidants, and hydrogel and nanofiber deliveries have reproduced the closure effect in diabetic and healthy mice.

Cell-level work fills in why. GHK-Cu shortened the doubling time of human dermal fibroblasts, including fibroblasts taken from previously irradiated patients, which recovered to roughly the rate of untreated normal cells while secreting more basic FGF and VEGF. Applied after red-LED photoirradiation it raised procollagen output and collagen-I messenger RNA well above light alone. Immobilised in a scaffold with the RGD adhesion motif and copper, it drove endothelial proliferation and angiogenic signalling far beyond either peptide by itself. In keratinocytes it raises integrin alpha-6 and beta-1 and the basal stem-cell marker p63 — and copper-free GHK reproduces that, which matters, because several other signature effects belong to the metal rather than the peptide.

Beyond matrix, the anti-inflammatory and antioxidant work is broad and recent: lower TNF-alpha-driven IL-6 in fibroblasts, reduced reactive oxygen species and cytokines in murine acute lung injury through NF-kappa-B and p38 suppression, attenuated silicotic lung fibrosis via peroxiredoxin 6, improved colitis through a SIRT1/STAT3 route, reduced immune-cell migration in zebrafish, preserved muscle mass and grip strength in smoke-exposed mice, and extended lifespan in nematodes. The cell-free chemistry is narrower than the word antioxidant suggests — GHK-Cu blocks lipid peroxidation only when the iron source is ferritin, and has no meaningful superoxide-dismutase or ceruloplasmin activity — but the tissue-level anti-inflammatory results are consistent across five organ systems.

The cartilage and joint studies

There is a cartilage literature, and it comes almost entirely from one laboratory at the Institute of Rheumatology in Prague in the 1990s.

In chick embryo chondrocytes, GHK supplementation raised proliferation and the collagen synthetic rate. What the cells built depended on what they were built on: cultured on cartilage collagen they synthesised predominantly type II, the cartilage collagen; cultured on denatured skin collagen they made types I and III. The authors' own conclusion was that GHK might be useful for preparing cartilage implants. The same group then grew pig chondrocytes in three-dimensional gel with cartilage collagens, aggrecan and GHK, and attributed the retention of chondrocyte phenotype to aggrecan rather than to the peptide.

They took those constructs into animals. Minipigs received a 3.5 mm cylindrical defect in the medial femoral condyle, filled with an autologous chondrocyte implant. All three implant types stimulated new hyaline cartilage at 8 to 12 weeks, and the implant conditioned with GHK produced tissue containing more cells. That is an in-vivo cartilage-defect experiment with a positive directional result, and it is the only one. It is also a Czech-language report with an English abstract, without stated animal numbers, randomisation, blinding, a histological grading scale or mechanical testing, and the peptide conditioned the implant in culture rather than being delivered into the joint. Nobody has repeated it in thirty years.

The one modern joint experiment is a rat ACL reconstruction. Seventy-two rats were randomised to saline or intra-articular GHK-Cu at 0.3 or 3 mg/mL, weekly for four weeks beginning at week two. At six weeks the treated knees had smaller side-to-side laxity differences than saline and the lower-dose grafts were stiffer; by twelve weeks neither difference survived, and ultimate load, gait and histology never differed at any point. The lower dose outperformed the higher one, consistent with the biphasic curve in culture.

Bone sits alongside rather than inside this. A collagen-glycosaminoglycan gel containing GHK accelerated healing of guinea-pig femoral defects, and GHK-Cu conjugates and coatings raise osteogenic and angiogenic markers in vitro. But in osteoblast culture the peptide increased cell attachment while slightly inhibiting alkaline phosphatase and osteocalcin — attachment up, differentiation markers down.

The human evidence

No human trial of GHK-Cu with a joint, cartilage, tendon or ligament endpoint has been published. Three separate 2026 reviews reach that conclusion independently — a PRISMA scoping review that searched cartilage and meniscus explicitly, a sports-medicine primer for orthopaedic surgeons, and a review in the orthopaedic literature — and all three note that indication, dose, frequency and duration remain undefined.

The human literature that does exist is dermatological, and it is more mixed than either its promoters or its critics tend to say. The largest randomised trial gave 0.4 percent copper-tripeptide cream to patients with chronic venous stasis ulcers, against silver sulfadiazine and an inert placebo, with 86 evaluable patients and blinded evaluators. Silver sulfadiazine beat both; the copper tripeptide did not differ from placebo. A randomised trial in patients recovering from CO2 laser resurfacing found no difference in erythema resolution, wrinkles or skin quality on blinded and computerised assessment — the only endpoint favouring GHK-Cu was the patient questionnaire. In pattern hair loss, a six-month placebo-controlled trial of a combined 5-aminolevulinic-acid and GHK product raised hair count in both treated arms against their own baselines, which is what its significance markers report — the counts themselves never separated from placebo. The one between-group result that cleared the line is a ratio of change, on the lower of the two doses, in the arm that had started with the fewest hairs of the three. Hair length and thickness did not move at any visit. The remaining human reports are a ten-patient uncontrolled pilot in striae and a single-patient case report with biopsies. A 2025 review of topical GHK, written from inside the cosmetic delivery field, records a surprising absence of clinical studies given how widely the ingredient is sold.

One human measurement is worth separating from the trials. Plasma GHK was measured by chromatography in nine patients with COPD and eleven age-matched healthy controls: roughly 70 against 133 nanograms per millilitre, correlating within the nine patients with pectoralis muscle area, TNF-alpha and an antioxidant enzyme. That is a disease contrast in twenty people whose correlations rest on nine. The age-decline curve quoted in marketing copy does not appear in the indexed literature in that form.

The gene-expression screens

The claim that GHK modulates thousands of human genes has a real source, and knowing what that source did makes the number readable. Investigators profiled 64 lung-tissue samples from eight lungs with COPD, graded for regional emphysema severity by micro-CT, and found 127 genes whose expression tracked severity. They then queried the Connectivity Map — a database of transcriptional signatures produced by dosing cultured cell lines with compounds — for something whose signature reverses that pattern. GHK came back.

A Connectivity Map query is a signature-matching exercise in cell lines at a single high dose. It nominates candidate compounds. It measures no clinical outcome, no dose-response and no tissue, and the gene counts it produces are counts of transcripts crossing a fold-change threshold in a screen, not counts of effects. What that paper did next is the part worth citing: GHK applied to human lung fibroblasts reproduced TGF-beta-induced expression, organised the actin cytoskeleton, raised integrin beta-1, and restored collagen-I contraction by fibroblasts taken from COPD lungs. That is a specific, testable finding about fibroblast behaviour.

The peptide and the metal

A recurring result in this literature is that the copper, not the tripeptide, carries the effect. The induction of MMP-2 and its TIMP inhibitors in fibroblasts was reproduced by copper ions alone and did not occur with copper-free GHK. Suppression of TNF-alpha-driven IL-6 was reproduced by copper chloride. The reduction in TGF-beta secretion was reproduced by copper chloride. A biotinylated-GHK wound dressing raised copper concentration at the wound site ninefold, and its authors read the healing effect as part copper delivery and part peptide signalling.

Against that, copper-free GHK reproduces the keratinocyte integrin and p63 results, and the angiogenic activity of the SPARC-derived motif was sequence-specific and did not require bound copper. Both things are true: some of what GHK-Cu does is the peptide, and some of it is a well-tolerated way to deliver copper into tissue.

Two other results run against the consensus and belong here. Transforming growth factor beta, which the compound is often said to activate, went down rather than up in two independent fibroblast studies. And one in-vivo wound study found slower skin reorganisation and delayed fibroblast activation with GHK-Cu in guinea pigs, with reduced cell reproduction in the accompanying culture arm at a concentration above the optimum. A separate irradiated-rat-flap experiment, the in-vivo counterpart to the positive irradiated-fibroblast culture work, found no difference in ischaemia, vessel number, vessel area or VEGF.

Regulatory status

Injectable GHK-Cu is not approved by the FDA for any indication; what is sold online for injection is an unapproved gray-market product, and a 2026 review in Sports Medicine addresses that market directly, alongside the placebo effect as a mediator of perceived benefit. GHK-Cu is not named on the WADA Prohibited List, but as a non-approved substance, injected use by athletes plausibly falls under the S0 catch-all.

Safety

Topical cosmetic use appears well tolerated, and there is a mechanism for it: in a keratinocyte irritation model GHK-Cu was not cytotoxic and moved none of the standard irritation biomarkers, while copper chloride and copper acetate at the same concentrations raised several of them. Chelating the copper inside the peptide lowers its irritancy. The six-month hair-growth trial reported no adverse events in 45 men, and the silicosis mouse study reported no significant systemic toxicity at effective doses.

Systemic and injected safety in humans is a different question and is essentially unstudied. No human parenteral dose has been validated. The copper consideration is specific rather than vague: the molecule's proposed function is copper transport, transdermal application delivers copper in measurable quantities, and a peptide-loaded dressing raised local copper ninefold, so repeated dosing is also repeated copper dosing. Gray-market injectables add the usual identity, sterility and dose-accuracy risks. For reference, the rat ACL study used 0.3 to 3 mg/mL intra-articularly and the lower dose performed better; in-vitro effects peak near one nanomolar and decline above it. This entry is research appraisal, not dosing guidance.

What would change the tier

The cartilage work that exists is thirty years old, from one group, and positive. Repeating it with modern endpoints is a small ask: glycosaminoglycan and hydroxyproline content, aggrecan and COL2A1 expression, ICRS-II histology and indentation stiffness in a chondrocyte or explant model would say more in one paper than the whole dermal literature says by analogy. A chondrocyte study resolving which way GHK-Cu moves MMP-2 and MMP-9 would settle whether the compound is plausible in a joint at all, since the dermal literature answers that question both ways. And a replication of the rat ACL result at the lower dose with sustained endpoints would tell us whether the six-week benefit was early wound-phase angiogenesis or the start of something durable. Any of the three would move this entry.