Fermented cosmetic actives and nutricosmetics

Two shifts sit under this heading: producing established cosmetic ingredients by fermentation instead of extraction, and taking actives orally instead of topically. They are unrelated technically and are best judged separately.

Fermentation replaces the source, not the molecule

Squalane was historically obtained from shark liver oil, where squalene is abundant. The fermentation route runs sugar through engineered yeast to farnesene, which is dimerised and hydrogenated to squalane. The product is the same molecule, and its advantage is supply: a defined, reproducible, contaminant-controlled stream that does not depend on a wild population. This is a sourcing improvement, not a performance one, and the honest framing says so.

Hyaluronic acid made the same move from rooster comb extraction to Streptococcus and later engineered Bacillus fermentation. Here the gain is real and technical as well as ethical: fermentation gives control of molecular weight distribution, which matters because HA’s behaviour is size-dependent — high-molecular-weight chains form viscous, occlusive surface films, while short fragments behave differently in tissue and are associated with different signalling. Extraction gives you whatever the tissue contained.

Recombinant collagen and the hydroxylation problem

Collagen’s defining structure is a triple helix of three chains in a repeating Gly-X-Y pattern. Its thermal stability depends on 4-hydroxyproline in the Y position: the hydroxyl group supports a water-mediated hydrogen-bonding network that holds the helix together near body temperature. Without it the helix melts well below 37 °C.

That hydroxylation is performed post-translationally by prolyl 4-hydroxylase, an enzyme requiring iron, 2-oxoglutarate and ascorbate as cofactors. Yeast and bacteria do not have it. A host expressing a collagen sequence alone therefore produces unhydroxylated, thermally unstable chains, and full-length triple-helical recombinant collagen requires co-expressing the human hydroxylase subunits.

This is why much of what is sold as recombinant collagen is a fragment or a designed peptide rather than the full protein — chosen for solubility and for a specific binding motif rather than for structural collagen. That is a legitimate product; it is simply not the same thing, and the distinction is invisible in an INCI name.

The oral route, argued properly

Ingested collagen is protein. Gastric and pancreatic proteases hydrolyse it, and what crosses the intestinal epithelium is free amino acids and small peptides. Intact collagen does not reach the dermis, and no plausible mechanism would take it there.

The argument that survives is narrower. Collagen is unusually rich in proline and hydroxyproline, and certain dipeptides — prolyl-hydroxyproline and hydroxyprolyl-glycine — are resistant to further hydrolysis and have been detected in human plasma after ingestion of collagen hydrolysate. The proposed mechanism is that these fragments act as signals to dermal fibroblasts, not as building material. Cell studies report fibroblast responses to Pro-Hyp; clinical trials on skin elasticity and hydration exist and are frequently industry-funded, modestly sized and heterogeneous in dose and hydrolysate.

Stating the claim correctly is most of the analysis. “Eat collagen to make collagen” is wrong on digestive grounds. “A specific hydrolysis-resistant dipeptide reaches circulation and may act as a fibroblast signal” is the defensible version — it names a molecule, a route and a target, and it can be tested. The evidence for it is suggestive and not yet strong.