A study combining animal, cellular, and computational experiments can offer several kinds of evidence. Those layers should not be treated as if they all directly prove the same mechanism. Primary study
What was reported
GHK-Cu improved several disease and tissue measures in DSS-induced mouse colitis. Cell co-culture experiments examined epithelial repair and junction-associated proteins. Computational work proposed SIRT1 as a target, while measured SIRT1 and phosphorylated STAT3 changed. STAT3 knockdown altered the repair-related response, but suppression of inflammatory factors was not fully explained by that pathway. Primary report
Our appraisal: different endpoints can have different dependencies
Docking is a prediction of a possible molecular interaction. It does not establish binding affinity in the relevant biological environment or show that the interaction is necessary for the observed effect. Direct binding and functional perturbation address additional questions.
Knockdown can strengthen a causal hypothesis, but interpretation depends on knockdown efficiency, off-target effects, baseline changes, and appropriate controls. A rescue experiment can help distinguish pathway involvement from a nonspecific loss of cell function.
The endpoint-specific pattern is important. If repair-related effects and inflammatory-factor effects respond differently to a perturbation, a single-pathway explanation is incomplete. The study should not be simplified to “all benefits occur through SIRT1/STAT3.”
Changes in a marker associated with a cell population also do not automatically count that population. Direct cellular measurements would be needed before asserting a quantified change in immune-cell abundance.
This report offers preclinical hypotheses for further testing. Chemically induced mouse colitis and a co-culture system do not establish treatment efficacy in human ulcerative colitis. Full review should inspect allocation, blinded pathology, timing, protein normalization, perturbation controls, and independent replication. The most useful interpretation preserves both the favorable observations and the evidence that more than one mechanism may be involved.
Frequently asked questions
Was human clinical efficacy tested?
No.
Did the study include computational predictions?
Yes.
Does docking prove biological binding?
No.
Was STAT3 experimentally reduced?
Yes, using a knockdown approach.
Did every endpoint show the same pathway dependence?
No.
Does a population-associated protein directly count cells?
No.
What should the mechanism claim acknowledge?
The different evidence levels and the possibility of additional pathways.