GHK-Cu Research: What the Published Literature Reports

Research Use Only. This article is written for laboratory and research professionals. The compounds discussed are supplied strictly for in-vitro laboratory research. They are not drugs, dietary supplements, or foods. They are not for human or veterinary use, and not for diagnostic or therapeutic use. Nothing here is medical, legal, or regulatory advice.

The short answer

The GHK-Cu peptide literature is a body of work on a copper(II)-binding tripeptide studied in fibroblast and keratinocyte cultures, in cell-free chemical systems, in reconstructed skin-equivalent models, and in rodent implant models. Its principal measured endpoints are collagen and glycosaminoglycan synthesis, matrix metalloproteinase and TIMP expression, proteoglycan messenger RNA levels, endothelial cord formation in culture, copper chelation behaviour, and transcriptional signatures. Every result below is an observation in a defined model system. None of it describes an outcome in a person, and this article does not extrapolate to one.

How to read this page

A note on names: the same compound is indexed as GHK-Cu, as copper peptide GHK-Cu, and as GHK-Cu (copper peptide). The papers below use the full chemical name. Each section here names the compound form tested, the model, and the measurement. Most secondary writing reports conclusions without the model, which is how a fibroblast culture result turns into a claim about people somewhere between the paper and the product page.

The model is the finding. A result in cultured rat dermal fibroblasts generates a hypothesis about tissue; it does not test one.

The compound form matters. Free GHK and the preformed complex are different reagents, and culture medium already contains copper and competing ligands. Several papers below compare peptide, complex and copper salt, the only way to attribute an effect.

Much of the core work comes from a small number of groups. The foundational biochemistry runs through Loren Pickart's laboratory and the Reims group led by François-Xavier Maquart. Independent replication across many laboratories is thinner than the citation counts suggest.

Where the literature starts: isolation, 1973

Pickart and Thaler reported the tripeptide in Nature New Biology in 1973, in the paper that first described it. The material was an activity isolated from human plasma by successive fractionation and later resolved to Gly-His-Lys; the model was cultured liver tissue; the measurements were cell survival and protein synthesis behaviour. It matters historically rather than mechanistically: the sequence occurs in human plasma, and it was found by tracking a culture activity.

Domain 1: collagen synthesis in fibroblast culture

The most-cited experimental result comes from Maquart and colleagues, FEBS Letters, 1988: Stimulation of collagen synthesis in fibroblast cultures by the tripeptide-copper complex.

• Material: GHK-Cu, the preformed complex.
• Model: fibroblast cultures, in vitro.
• Measured: collagen synthesis, plus cell number as a control.
• Reported: a response beginning between 10⁻¹² and 10⁻¹¹ M and reaching a maximum near 10⁻⁹ M, independent of any change in cell number.

Two details are worth holding onto. The concentration range is extraordinarily low, picomolar to nanomolar, where ligand-specific effects live rather than bulk chemical ones. And the cell-number control separates more collagen per cell from more cells.

The same paper noted that a GHK triplet occurs within the alpha-2 chain of type I collagen, and proposed that proteolysis at a site of tissue damage could release it locally. That is a hypothesis about origin, stated as one.

Domain 2: extracellular matrix accumulation in a rodent implant model

Maquart's group followed the culture work into animals: In vivo stimulation of connective tissue accumulation in rat experimental wounds, Journal of Clinical Investigation, 1993.

• Material: GHK-Cu at several concentrations, against saline controls and against a control tripeptide, L-glutamyl-L-histidyl-L-proline.
• Model: stainless steel wire mesh cylinders surgically implanted in rats, a standard wound-chamber preparation in which newly formed connective tissue can be recovered and assayed. Material was introduced into the chamber at intervals.
• Measured: dry weight, total protein, collagen, DNA, elastin, glycosaminoglycans, and messenger RNA for collagens and TGF-beta in the recovered chamber contents.
• Reported: a concentration-dependent increase in dry weight, DNA, total protein, collagen and glycosaminoglycan content; collagen synthesis rising roughly twice as much as non-collagen protein; increased type I and type III collagen messenger RNA; no increase in TGF-beta messenger RNA; an increased proportion of dermatan sulfate. The control tripeptide had no significant effect.

The negative result carries weight. TGF-beta is the obvious candidate intermediate for a matrix effect, and it did not move at the transcript level, which constrains the mechanistic story. The control peptide argues the effect is sequence-dependent rather than a generic peptide or copper artefact.

This is a rat model, not a human finding, and the compound went directly into an implanted chamber, a configuration with no counterpart outside the experiment.

Domain 3: metalloproteinases and their tissue inhibitors

Siméon and colleagues, Life Sciences, 2000: The tripeptide-copper complex stimulates matrix metalloproteinase-2 expression by fibroblast cultures.

• Material: GHK-Cu, compared against copper ions alone and against the free tripeptide.
• Model: dermal fibroblast cultures.
• Measured: MMP-2 in conditioned medium, MMP-2 messenger RNA, and secretion of the tissue inhibitors TIMP-1 and TIMP-2.
• Reported: increased MMP-2 protein and transcript, and increased TIMP-1 and TIMP-2 secretion. The effect was reproduced by copper ions but not by the free tripeptide.

That attribution is the most useful sentence in the paper. In this assay the metal, not the ligand, is the proximate signal, with the peptide acting as its carrier. Any account that treats the peptide as an independent signalling molecule has to reckon with it.

The direction of the finding also repays attention. A matrix-degrading enzyme and its inhibitors rising together describes increased turnover, not one-way deposition.

Domain 4: glycosaminoglycans and small proteoglycans

Siméon and colleagues again, Journal of Investigative Dermatology, 2000: Expression of glycosaminoglycans and small proteoglycans in wounds.

• Material: GHK-Cu.
• Model: the same rat wound-chamber preparation, plus rat dermal fibroblast cultures.
• Measured: dry weight and total protein; hydroxyproline for type I collagen; uronic acid for glycosaminoglycans; electrophoretic separation of glycosaminoglycan classes; northern blot for decorin and biglycan messenger RNA.
• Reported: enhanced accumulation of chondroitin sulfate and dermatan sulfate; increased decorin messenger RNA and decreased biglycan messenger RNA in the chamber; the decorin increase reproduced in fibroblast culture, while biglycan was unchanged in culture.

Decorin and biglycan are structurally related proteoglycans that moved in opposite directions: a specific, falsifiable observation rather than a global statement about matrix production.

Domain 5: peptides released from SPARC and endothelial cord formation

Lane and colleagues, Journal of Cell Biology, 1994: SPARC is a source of copper-binding peptides that stimulate angiogenesis.

• Material: SPARC protein, synthetic peptides from its cationic 113 to 130 region, and KGHK.
• Model: endothelial cell culture assays, plus in-vivo angiogenesis assays.
• Measured: cord formation, angiogenic response, and Cu(II) binding by the peptides.
• Reported: proteolysis of SPARC yields KGHK, GHK and longer peptides containing those sequences; KGHK accounted for most of the activity in their mapping, and substituting the histidine reduced it. The peptides bound Cu(II) with high affinity, but prior incubation with Cu(II) was not required, and a second Cu(II)-binding peptide from SPARC had no effect.

An independent group, a different endpoint, a partly dissenting result: copper binding alone did not predict activity here.

Domain 6: keratinocytes and reconstructed skin-equivalent models

Choi and colleagues, Journal of Peptide Science, 2012, examined copper-free GHK in keratinocyte and skin-equivalent models.

• Material: copper-free GHK.
• Model: monolayer cultures of normal human keratinocytes and three-dimensional skin-equivalent constructs.
• Measured: keratinocyte proliferation; basal cell morphology; immunostaining for alpha-6 and beta-1 integrin along the basement membrane; counts of p63-positive and PCNA-positive cells; integrin expression by western blot.
• Reported: increased proliferation, more cuboidal basal cells, stronger and more linear integrin staining, and higher counts of the two marker-positive populations.

The construct is an engineered tissue model, not skin: it restores three-dimensional architecture that monolayer culture lacks, but has no circulation, no immune compartment and no appendages.

Domain 7: redox behaviour in cell-free systems

Copper is redox active. Unbound Cu(II)/Cu(I) cycling drives Fenton-type chemistry and generates reactive oxygen species, while chelated copper behaves according to ligand and geometry. This literature asks which description fits GHK-Cu.

Miller, DeSilva, Pickart and Aust examined the effects of chelated Cu(II) on ferritin-dependent lipid peroxidation in Advances in Experimental Medicine and Biology, 1990.

• Material: the chelated Cu(II) complex.
• Model: a cell-free biochemical system using ferritin as an iron source and a measurable lipid peroxidation readout.
• Measured: iron release from ferritin and lipid peroxidation.

Cell-free systems isolate chemistry from biology, which is their strength and their limit. They establish what a complex does to a defined reaction mixture, and say nothing about concentration, compartment or competing ligands inside a cell.

Domain 8: transcriptional profiling

The most-quoted modern claims come from analyses of gene-expression signature databases rather than new bench experiments. In a 2014 analysis, Pickart and colleagues describe using the Broad Institute's Connectivity Map, a library of transcriptional signatures produced by exposing cell lines to a panel of small molecules, and report that GHK's signature involves a large fraction of measured transcripts.

• Material: GHK, as represented in an existing signature dataset.
• Model: cultured cell lines profiled by the database, then computational comparison.
• Measured: transcript-level changes above a threshold, expressed as counts and percentages of profiled genes.

Read these numbers as what they are: a reanalysis of a public signature library, at one concentration, in cell lines chosen by the database designers, filtered by a fold-change cutoff. Large transcript counts are common in such datasets; they generate hypotheses rather than demonstrate a phenotype.

What the literature reports about copper handling, chelation and cytotoxicity in cell models

People search for GHK-Cu side effects. The published record contains no safety profile for a person, and nothing here should be read as one. It does contain cell-model data on how copper-bearing compounds behave at defined concentrations, which is a different question with real answers.

Copper salts versus the peptide complex in keratinocytes. Li and colleagues, Scientific Reports, 2016, ran a keratinocyte-based comparison of copper compounds at cellular, transcript and protein level.

• Material: GHK, GHK-Cu, copper(II) chloride, and copper(II) acetate.
• Model: HaCaT keratinocytes.
• Measured: viability by tetrazolium reduction at 24, 48 and 72 hours; lactate dehydrogenase release across seven concentrations from 0.0058 µM to 5800 µM; quantitative PCR and ELISA for IL-1α, IL-8, HSPA1A and FOSL1.
• Reported: the two simple copper salts significantly reduced viability at 580 µM and 5800 µM and raised the four biomarkers at 58 µM and 580 µM after 24 hours; GHK and GHK-Cu at 58 µM and 580 µM produced no significant change in those transcripts or proteins.

Ligand identity, not copper content alone, governs what happens at a given copper concentration. The scope is a single immortalised keratinocyte line under defined exposure conditions.

Chelation activity does not predict cytoprotection. Ha, Choi and Boo, Antioxidants, 2023, compared copper chelation activity and cytoprotective effect across 40 amino acids and amidated amino acids in HaCaT keratinocytes exposed to copper sulfate.

• Reported: CuSO₄ at 0.4 to 1.0 mM caused concentration-dependent cell death. Of the compounds tested at 1.0 mM, only histidine and histidinamide prevented that death. Cysteine and cysteinamide chelated copper potently and did not protect. EDTA and GHK-Cu, used as reference compounds, showed no cytoprotective effect either.

That last clause is why this citation belongs here. A strong chelator is not automatically protective, and GHK-Cu did not behave as one in that assay. Reporting the result that cuts against the story is the point of reading primary sources.

Concentration is the whole argument. The responses reported in fibroblast culture sit at 10⁻¹² to 10⁻⁹ M. The copper cytotoxicity thresholds above sit at 10⁻⁴ to 10⁻³ M. Six to nine orders of magnitude separate them, which is why "how much, in what, for how long" is the only version of the question the literature answers.

Limits of this evidence base

• Much of the core work is 25 to 50 years old, predating current standards for blinding, randomisation and preregistration.
• Several key findings originate from a small number of groups with commercial interest in the molecule, disclosed in the papers and worth weighing.
• Culture media contain copper and competing ligands, so the operative species in any given experiment is often uncertain unless the design explicitly controls for it.
• Reviews restate primary findings without carrying the model forward, which is how "in fibroblast cultures" quietly disappears.
• Transcriptional signature analyses are hypothesis-generating and are routinely cited as if they were outcome data.

None of this makes the chemistry uninteresting. It makes the chemistry the defensible part, which is where the structural and mechanistic explainer picks up.

Reproducing any of this

Working from these papers requires knowing what is in the vial: identity by mass spectrometry, purity by reversed-phase HPLC, copper content by an elemental method, counterion and net peptide content, all traceable to a lot number. Without that, a picomolar concentration-response curve is uninterpretable, because the actual molar concentration is unknown.

Research-grade GHK-Cu sits alongside other compounds in the peptide catalog, supplied strictly as a laboratory reagent for in-vitro research use only.

Working with this material

The certificate of analysis for GHK-Cu, from Horizon Analytical, is on the product page and on Lab results. Our own lot records appear at the lot lookup as lots are released. Supplied for laboratory research use only. Not for human or veterinary use.

Compliance notice. Nothing in this article should be read as guidance for administering any compound to a human or animal subject. "Research Use Only" describes the conditions under which a material is supplied; it is not a determination of legal status and confers no exemption from the Federal Food, Drug, and Cosmetic Act. Researchers are responsible for compliance with all applicable federal, state, and institutional requirements, and for obtaining any approvals their work requires.

Frequently asked questions

What has GHK-Cu actually been studied for in the published literature?

Collagen synthesis in fibroblast culture, extracellular matrix accumulation in a rat implant model, MMP-2 and TIMP-1/TIMP-2 expression in dermal fibroblasts, glycosaminoglycan and proteoglycan messenger RNA levels, endothelial cord formation by SPARC-derived peptides, copper chelation chemistry in cell-free systems, and transcriptional signature comparisons. All are model-system endpoints.

Are GHK-Cu research findings human results?

No. The published work is in cell culture, cell-free chemical systems, reconstructed skin-equivalent constructs and rodent implant models. Those models generate hypotheses about tissue biology. They do not establish outcomes in people, and this article does not extrapolate to any.

What does the literature report about GHK-Cu and copper handling in cell models?

In HaCaT keratinocytes, copper(II) chloride and copper(II) acetate reduced viability at 580 µM and above and raised IL-1α, IL-8, HSPA1A and FOSL1 at 58 µM and 580 µM, while GHK and GHK-Cu at the same concentrations did not change those markers. In a separate keratinocyte study, GHK-Cu and EDTA used as reference chelators showed no cytoprotective effect against copper sulfate. These are cell-model observations, not a safety profile for any person.

At what concentrations is GHK-Cu studied in vitro?

The collagen synthesis response reported in fibroblast culture begins between 10⁻¹² and 10⁻¹¹ M and peaks near 10⁻⁹ M. Copper cytotoxicity thresholds in keratinocyte models sit far higher, in the 10⁻⁴ to 10⁻³ M range. Those two regimes are separated by roughly six to nine orders of magnitude.

Is copper-free GHK studied differently from the copper complex?

Yes, and the comparison is informative. One fibroblast study found the MMP-2 response was reproduced by copper ions but not by the free tripeptide. A keratinocyte and skin-equivalent study reported effects of copper-free GHK on proliferation and integrin expression. Culture medium contains copper and competing ligands, so attributing an effect to the ligand, the metal or the complex requires explicit controls.

Which paper is the original collagen synthesis result?

Maquart and colleagues, FEBS Letters, 1988, "Stimulation of collagen synthesis in fibroblast cultures by the tripeptide-copper complex glycyl-L-histidyl-L-lysine-Cu2+". It reports a response beginning between 10⁻¹² and 10⁻¹¹ M and maximal near 10⁻⁹ M, with no change in cell number, and notes the GHK triplet within the alpha-2 chain of type I collagen.

What are the main limits of the GHK-Cu evidence base?

Much of the foundational work is decades old, several key findings come from a small number of groups with commercial interests disclosed in the papers, culture conditions make the operative species uncertain without explicit controls, and signature analyses are often cited as though they were outcome data. The chemistry is well characterised; the extrapolations built on it usually are not.

Primary sources

The remaining literature cited above, for direct verification:

• Expression of glycosaminoglycans and small proteoglycans in wounds: https://pubmed.ncbi.nlm.nih.gov/11121126/
• SPARC is a source of copper-binding peptides that stimulate angiogenesis: https://pmc.ncbi.nlm.nih.gov/articles/PMC2120067/
• copper-free GHK in keratinocyte and skin-equivalent models: https://pubmed.ncbi.nlm.nih.gov/23019153/
• effects of chelated Cu(II) on ferritin-dependent lipid peroxidation: https://pubmed.ncbi.nlm.nih.gov/2244543/
• a 2014 analysis: https://pmc.ncbi.nlm.nih.gov/articles/PMC4180391/
• keratinocyte-based comparison of copper compounds: https://pmc.ncbi.nlm.nih.gov/articles/PMC5124859/
• copper chelation activity and cytoprotective effect across 40 amino acids and amidated amino acids: https://pmc.ncbi.nlm.nih.gov/articles/PMC10135049/