GHK-Cu is the copper(II) complex of the tripeptide glycyl-L-histidyl-L-lysine, and it is sold for research use only. Papers that profile its effect on gene expression often quote large, tidy numbers, such as the share of genes that "changed." Those numbers depend on how the study was built, so we look at the design instead of the headline. Everything below describes published cell-culture and database analyses, not outcomes in people or animals.
Introduction: why design decides what a gene list means
A gene expression profile looks like a result, but it's closer to a question. It tells researchers where to look next. How much weight it deserves depends on the cells, the concentration, the normalization and where the line for "changed" was drawn.
GHK-Cu is a good example because its best-known gene data come from a public reference set rather than a purpose-built experiment. The peptide's structure and naming are covered on the GHK-Cu reference page: the formula for the complex is C14H22CuN6O4, and the peptide alone is C14H24N6O4. We take four design questions in turn: what a profile measures, how the reference set was built, how thresholds shape the numbers and how researchers move from a transcript to a function.
What a gene expression profile measures
In short: it measures messenger RNA levels, and nothing downstream of that.
The basic design
Cultured cells are exposed to a compound, and the amount of mRNA from thousands of genes is compared with untreated cells. Each gene gets a fold change, which is often converted into a percentage up or down. The comparison with untreated cells is the whole experiment, so the choice of cells and the way the readout is processed matter a great deal.
What it cannot tell you
An mRNA change does not guarantee a change in protein level or activity. Authors of the GHK analysis say as much: in their words, there is "not enough information to translate gene profiling data into biological effects" (PMC4180391). A review of the same work adds that gene changes cannot always be traced to a protein product or a functional effect (PMC3359723). Keep that in mind for every number that follows.
The Connectivity Map profiles of GHK
In short: the widely quoted GHK signatures were pulled from an existing library, not run as a new study.
The reference set
The Connectivity Map (cMap) is a public library of transcriptional responses to known perturbing compounds. It evaluated five cancer cell lines against 1,309 small molecules, and GHK was tested at 1 micromolar (PMC6073405). GHK was run on only two of those five lines. Two signatures came from PC3 cells and one from MCF7 cells (PMC4180391). Both are cancer lines, which matters when the results are later discussed in terms of skin, lung or nerve biology.
Platform and processing
The signatures were generated on the GeneChip HT Human Genome U133A array. In the analysis by the peptide's reviewing authors, the raw CEL files were processed with MAS5 and background correction in GenePattern, then compared to get a fold change per probe set. Because several probe sets can point to one gene, the fold changes were averaged per gene. The array has 22,277 probe sets representing 13,424 genes, a ratio of about 1.66, and the authors used that ratio to estimate gene counts at each cutoff (PMC4180391). So the gene counts are estimates built on an averaging step. They aren't direct tallies.