Glycine

Glycine is the reason collagen-rich food is interesting at all. Every third residue in a collagen triple helix is glycine — it has to be, because it is the only amino acid small enough to sit at the crowded interior of the helix. If anything in a bowl of stock is going to matter for cartilage, this is the leading candidate.

The shortfall argument

A metabolic-flux analysis put numbers on it. Endogenous synthesis from serine supplies about 3 g of glycine a day and is stoichiometrically capped — the enzyme that makes it can't produce more glycine than the one-carbon units it generates alongside. Diet supplies another 1.5–3 g. Total collagen turnover in a 70 kg adult needs roughly 10 g/day.

That leaves a gap of several grams, and the authors' conclusion is that glycine should be treated as semi-essential rather than non-essential.

The argument is internally coherent and demographically plausible — modern diets really are low in the collagenous cuts humans historically ate. It is also modelled arithmetic rather than a measured human requirement, and it has never been tested prospectively in anyone.

The cell result

The same group's follow-up is the strongest single piece of evidence here. Bovine articular chondrocytes were cultured across a range of glycine, proline and lysine concentrations, with type II collagen measured by ELISA every 48 hours for 15 days.

Proline and lysine both increased type II collagen synthesis, but their effects plateaued below 1.0 mM — proline from 0.6, lysine from 0.85. Glycine kept going, still climbing at 7 mM. At 1.5 mM it raised type II collagen synthesis by about 225 percent over control, with proline at half that effect and lysine at a third. Aspartate and isoleucine, run alongside as controls, did nothing at any concentration, which is what makes the result specific to these three.

Two caveats travel with that number permanently, and the authors state the first themselves. The proline and lysine effects occurred within the physiological range; glycine's required a much higher one. And the conflict-of-interest declaration is unusually consequential — the senior author is president of the institute that produced both foundational papers, that institute runs a clinic selling amino acids, and it holds a pending patent on using glycine to treat osteoarthritis. None of that makes the result wrong. It does mean independent replication matters more than usual, and after seven years there is none.

The animal evidence

The best dietary-glycine joint study fed rats a glycine-containing diet before reactivating a peptidoglycan-induced arthritis. Ankle swelling rose 2.1 mm in controls and only 1.0 mm on glycine, with less inflammatory infiltration, less edema, less synovial hyperplasia, and TNF-α mRNA undetectable in the glycine-fed joints.

That is a real effect on a real joint from dietary glycine. But read the mechanism: it was blocked by strychnine and by removing chloride, and it works through a glycine-gated chloride channel on macrophages, damping NF-κB translocation and cytokine release. This is immunomodulation, not collagen synthesis — and the model is inflammatory arthritis, not osteoarthritis.

So the two strongest pieces of glycine evidence support two different mechanisms, and neither has been shown in a human.

How much is in food

Wildly variable, which is the recurring theme of this section. Only two bone stocks in the literature have been properly analysed, and neither paper reports glycine as a concentration — one weighs 11,760 mg into a 300 mL serve, the other sits at the bottom of the same literature. Divide the first out and put them side by side and the range is 150 to 39,200 mg per litre, a 260-fold spread. That division is this site's, not either paper's, and what it describes is how far apart two analysed pots were rather than what is in yours. Gelatin is about 22 percent glycine by weight, so the modelled 7–10 g shortfall works out to roughly 30–45 g of gelatin a day, or a very large amount of collagenous meat.

Those are arithmetic, not dosing guidance. No trial has validated any glycine dose against any connective-tissue outcome.

A complication for the whole substrate theory

The most interesting counter-evidence doesn't come from the glycine literature at all. In the chondrocyte work on collagen peptides, a mixture of free proline, hydroxyproline and the tripeptide Pro-Hyp-Gly did nothing — while the intact dipeptide Pro-Hyp tripled glycosaminoglycan deposition and doubled aggrecan expression.

If free amino acids are inert where the dipeptide is active, then "eat more building blocks and the body builds more cartilage" may be the wrong model of how collagen-rich food works, even where it does work. The glycine hypothesis and the peptide-signalling hypothesis are competing explanations, not complementary ones, and the site treats both as unresolved.

The collagen amino acids, side by side

Glycine is not the only thing a collagen food delivers, and putting the others beside it is what makes the substrate argument testable. "In the food?" here means a bone stock. The evidence column is each molecule's own case against cartilage, independent of how much of it is in your pot.

The last two rows are the whole argument of this entry in miniature. The amino acid you can measure in a stock does nothing by itself; the dipeptide that does something has never been measured in one. That is why the glycine hypothesis and the peptide-signalling hypothesis compete rather than combine.

Safety

Not the constraint here. Glycine is well tolerated at multi-gram doses — 3 g before bed is separately studied for sleep, and sustained dosing at 0.8 g/kg/day has been reported as well tolerated in a psychiatric trial. It is a cheap, low-risk thing to eat more of. What is missing is any reason to believe a specific amount does a specific thing for cartilage.

What this topic needs is simple and absent from every registry checked: a trial of glycine, or a glycine-rich eating pattern, with any connective-tissue endpoint at all.