In vivo human plasma measurements. Researchers measured GHK concentrations in the blood of healthy human donors across different age groups.

GHK levels average ~200 ng/mL at age 20 and drop to ~80 ng/mL by age 60 — a roughly 60% decline.

High for establishing biological relevance, but it doesn't prove topical supplementation works. The fact that GHK-Cu is endogenous (your body makes it) and declines with age is a legitimate scientific rationale for supplementation. It's a much stronger starting point than a completely synthetic compound with no natural role. But "levels decline with age" is also true of hundreds of molecules — testosterone, collagen, NAD+, melatonin — and declining levels don't automatically mean that putting more on your scalp fixes hair loss. This is a necessary but not sufficient piece of evidence.

• Pickart L (1973). Original isolation of GHK from human plasma
• Pickart L, Thaler MM (1973)
• Pickart L (2008). Review of GHK-Cu biology

In silico — Connectivity Map (cMap) analysis. Researchers treated human cancer cell lines (MCF7 breast cancer, PC3 prostate cancer, HL60 leukemia) with GHK-Cu and used high-throughput gene chips to measure changes in mRNA expression across ~20,500 genes. The Wnt/beta-catenin claim comes from observing that Wnt pathway genes were among those modulated.

GHK-Cu treatment altered the expression of genes in the Wnt signaling pathway in these cell lines.

Low. This is computational analysis of gene expression in cancer cell lines — not hair follicle cells, not skin cells, and not in a living person. Cancer cells have fundamentally altered gene expression compared to normal cells. The cMap is a powerful hypothesis-generation tool, but it doesn't tell you whether a specific Wnt signaling change actually happens in a human hair follicle on a human scalp when you apply a serum. Wnt/beta-catenin is genuinely important for the hair cycle (it's the pathway that tells follicles to enter anagen), which makes this an exciting lead — but it has not been validated in a hair-specific model. No one has published data showing GHK-Cu activating Wnt signaling in dermal papilla cells or hair follicle stem cells.

• Pickart L et al. (2012). Oxid Med Cell Longev PMID PMC3359723
• Broad Institute (2012). Connectivity Map (cMap) database

Two models, both strong for what they measure: (1) In vitro — Human adult dermal fibroblasts (HDFa). Researchers treated cultured human fibroblasts with GHK-Cu and measured mRNA expression via RT-PCR and collagen production via colorimetric assays. (2) In vivo — Rat Shilling-Hunt wound chamber model. Wire mesh cylinders were implanted under rat skin to create controlled wound environments, then GHK-Cu was injected into the chambers.

In the HDFa model: GHK-Cu significantly increased mRNA for collagen types I, III, IV, and VII, plus elastin and glycosaminoglycans. In the rat wound chambers: decorin (a proteoglycan that organizes collagen fibers) mRNA increased by 302%. GHK-Cu also balanced MMP-2 (breaks down damaged matrix) and TIMP-1/TIMP-2 (protects new matrix), producing organized remodeling rather than scar tissue.

Moderate — this is real and well-documented, but it's a wound healing story, not a hair loss story. The ECM remodeling data is among the strongest evidence for GHK-Cu in any context. Multiple research groups have replicated it across human cells and animal wounds. The relevance to hair is indirect but reasonable: perifollicular fibrosis (scarring around the follicle) is associated with advanced androgenetic alopecia, and a compound that promotes healthy matrix remodeling could theoretically help maintain a hospitable environment for follicles. But "could theoretically help" is doing heavy lifting here. No study has directly shown GHK-Cu reversing perifollicular fibrosis in human scalps or correlating ECM remodeling with hair regrowth.

• Simeon A et al. (2000). J Invest Dermatol
• Simeon A et al. (1999). Life Sci
• Pickart L et al. (2015). BioMed Res Int

Two models: (1) In silico — cMap gene expression analysis (same cancer cell line platform as Claim 2). Showed modulation of TGF-beta superfamily genes (TGF-beta1, TGF-beta3, TGFBI, TGFBR2, TGFBR3). (2) In vitro — Human primary lung fibroblasts from COPD/emphysema patients. Researchers cultured fibroblasts taken from the actual diseased lung tissue of COPD patients and treated them with 10 nM GHK-Cu.

The cMap analysis showed TGF-beta pathway genes were among those modulated. In the COPD fibroblast model, GHK-Cu at 10 nM switched the gene expression pattern from tissue destruction (emphysema phenotype) to tissue repair. It restored TGF-beta pathway function and reorganized the actin cytoskeleton for collagen gel contraction (a standard measure of tissue remodeling capacity).

Low to Moderate. The COPD fibroblast finding is genuinely impressive — it used primary human cells from actual patients (not cancer cell lines), which is much more relevant than the cMap data. But here's the critical gap: these were lung fibroblasts, not scalp fibroblasts, not dermal papilla cells, and not hair follicles. TGF-beta1 does play a real role in hair loss (it's one of the signals that pushes follicles from anagen into catagen), so the biological logic is sound. But no one has shown GHK-Cu suppressing TGF-beta1 specifically in the hair follicle environment. The claim that TGF-beta1 "contributes to follicle miniaturization" is true, and GHK-Cu does modulate TGF-beta — but the two facts haven't been connected experimentally in hair tissue.

• Pickart L et al. (2012). Oxid Med Cell Longev PMID PMC3359723
• Campbell JD et al. (2012). COPD fibroblast study
• Pickart L (2014). BioMed Res Int

In vitro — Human mesenchymal stem/stromal cells (MSCs) embedded in an alginate gel carrier, and separately, human umbilical vein endothelial cells (HUVECs).

GHK-Cu produced a dose-dependent increase in secretion of VEGF (vascular endothelial growth factor) and bFGF (basic fibroblast growth factor) from human MSCs. These are pro-angiogenic factors that stimulate new blood vessel formation.

Low to Moderate. VEGF is important for hair follicle health — follicles need blood supply, and minoxidil's mechanism includes improved follicular blood flow. The fact that GHK-Cu stimulates VEGF from human stem cells is meaningful. But this was measured in cells in an alginate gel, not in scalp tissue. There's a significant leap from "stimulates VEGF in a lab dish" to "improves blood flow around your hair follicles when applied as a serum." The stratum corneum barrier, the question of whether enough GHK-Cu reaches the follicle, and whether the VEGF increase is large enough to matter in vivo are all unanswered. Also worth noting: if your main goal is follicular vascularization, minoxidil does this through a proven, well-characterized mechanism and has decades of RCT data behind it.

• Pickart L et al. (2015). BioMed Res Int
• Pollard JD et al. (2005). Arch Facial Plast Surg

Human randomized, double-blind, placebo-controlled trial (the gold standard). 45 men with androgenetic alopecia were split into three groups: high-dose ALAVAX (100 mg/mL), low-dose ALAVAX (50 mg/mL), or placebo. Applied once daily for 6 months. Total hair count, length, thickness, and adverse events were measured at months 1, 3, and 6.

Hair count increased by 52.6 hairs (high-dose) and 71.5 hairs (low-dose) vs. only 9.6 hairs in placebo after 6 months (p<0.05 for both treatment groups). Zero adverse events across all 45 patients.

Low, despite the gold-standard design. On paper, this is the highest-quality evidence type (RCT, double-blind, placebo-controlled), which matters enormously. But the details undermine the headline numbers in several ways: It wasn't GHK-Cu. The treatment was ALAVAX — a synthetic complex of 5-aminolevulinic acid (5-ALA, a photosensitizer with its own biological activity) chemically bonded to GHK peptide (not GHK-Cu, the copper-bound form). You cannot attribute the results to GHK alone, much less to GHK-Cu specifically. The standard deviations swallow the results. The 6-month hair count changes were 52.6 +/- 45.7 (high-dose), 71.5 +/- 44.9 (low-dose), and 9.6 +/- 45.1 (placebo). When your SD is +/-45 and your effect is +52 to +72, the 95% prediction interval for an individual treatment patient stretches from losing hair to gaining 160+ hairs — and roughly 18% of placebo patients likely gained 50+ hairs by chance alone. The patients didn't think it worked. Only 26.7% of the high-dose group rated their satisfaction as "good" or "excellent." In the low-dose group — the one with the best hair count numbers — only 14.3% did, barely above placebo's 7.1%. If patients are gaining 70+ hairs and can't tell, either the measurement is capturing noise or the hairs aren't clinically meaningful. No photographs were published. For a hair loss trial, this is a glaring omission. Every credible AGA trial (Olsen 2006, Kaufman 1998) includes standardized global photography. Its absence here means the results can't be visually verified. No one has replicated it in 10 years. A genuinely effective hair treatment published in 2016 would have generated follow-up research by now. There are no confirmatory studies, no commercial development of ALAVAX, and no independent labs pursuing this formulation. Taken together — a combination product tested on 15 people per arm with SDs that overlap between treatment and placebo, low patient satisfaction, no photos, and zero follow-up — this study does not provide meaningful evidence that GHK-Cu grows hair.

• Lee WJ et al. (2016). Efficacy of a Complex of 5-Aminolevulinic Acid and Glycyl-Histidyl-Lysine Peptide on Hair Growth. Ann Dermatol PMID PMC4969472

The primary publication for this specific trial could not be located. The claim appears in several secondary sources and product marketing materials but the original study (author, journal, PMID) is not readily identifiable in PubMed or major databases.

As reported in secondary sources: a 0.02% copper tripeptide lotion applied over 16 weeks produced a 7% increase in hair count in a Japanese trial.

Low — not because it's wrong, but because it's unverifiable. A 7% increase over 16 weeks is modest (for comparison, 5% minoxidil typically produces 15-25% increases over similar timeframes). The 0.02% concentration is extremely low. But the bigger issue is that without access to the actual paper, you can't evaluate the study design: Was it placebo-controlled? How many subjects? What was the measurement method? Was the 7% statistically significant? If you're evaluating peptides with the rigor this topic deserves, unverifiable claims — even plausible ones — shouldn't carry much weight in your decision.

• Unknown (2025). Not independently verified. Cited in secondary sources as a Japanese clinical trial.

In silico — Connectivity Map (cMap). Human cancer cell lines (MCF7, PC3, HL60) were treated with GHK-Cu. Gene expression was measured using GeneChip HT Human Genome U133A arrays covering ~20,500 genes. The "31.2%" figure represents genes showing at least a 50% expression change.

Over 4,000 of ~20,500 genes (~31.2%) showed at least a 50% change in expression when these cancer cell lines were exposed to GHK-Cu.

Very Low for hair loss decisions; interesting for basic science. This finding is real and scientifically interesting — it positions GHK-Cu as a broad genomic modulator rather than a single-pathway drug. But as a consumer, you should know three things: (1) This was measured in cancer cell lines, which have radically altered gene regulation compared to normal cells. The same experiment in normal dermal papilla cells or hair follicle keratinocytes might produce completely different numbers. (2) "Modulates 31% of genes" sounds transformative, but many of those changes may be tiny, transient, or irrelevant to hair growth. A 50% change in a gene that codes for a liver enzyme doesn't help your scalp. (3) The cMap is a screening tool for generating hypotheses. It tells you "this compound touches a lot of gene pathways." It does not tell you "this compound regrows hair." The 31.2% figure is often the headline in GHK-Cu marketing, but it's actually the weakest category of evidence (computational prediction in cancer cells), not the strongest.

• Pickart L et al. (2012). Oxid Med Cell Longev PMID PMC3359723
• Broad Institute (2012). Connectivity Map (cMap)