Type I, II and III identify different collagen proteins. Undenatured, gelatin and hydrolysed describe what has happened to their structure. Two human studies illustrate the difference between finding fragments in the blood and measuring changes in knee symptoms.
Two words on one label, answering two different questions
A collagen label might say "Type II" next to "undenatured" or "hydrolysed". The two words are easy to read as one idea, but they answer different questions. The type number says which protein it is: which chains it's built from and which body tissues it's normally linked to. The processing word says what has happened to that protein's structure on its way from animal tissue to powder or capsule. A review of collagen supplements for joint health makes the same point: the single word "collagen" covers products with different structures, compositions and origins, so they can have different properties and possible effects.
Keeping the two questions apart also makes research easier to read. A type number on its own doesn't show that a product does anything, and neither does evidence that some collagen fragments reach the blood. A study result belongs to the specific formulation it tested, the people who took part and the outcomes it measured.
Collagen is a family of three-strand ropes
Collagens are a family of proteins that strengthen and support many tissues, including cartilage, bone, tendon and skin. Each collagen molecule starts as three protein chains wound together into a rope-like structure called procollagen. Enzymes then process it into mature collagen. The mature molecules line up into long, thin fibrils, and these fibrils link to one another (cross-link) in the spaces between cells to form very strong fibres. Scientists call the wound three-chain structure "triple-helical" collagen. Picture three strands twisted into a rope, with many ropes bundled into cables.
The type number refers to which chains make up the rope, and each chain is built from instructions in a particular gene. Type I collagen has two chains of one kind, α1(I), and one of another, α2(I), so it draws on two genes, COL1A1 and COL1A2. Type II is made of three identical α1(II) chains. Type III is made of three copies of its own chain. "Type II" isn't a grade or a strength rating; it names a particular molecule.
Each type has its usual places in the body. Type I is the most abundant collagen in the human body and is plentiful in skin, bone and tendon. Type II is found mainly in cartilage, especially hyaline cartilage, which covers the ends of bones in joints and makes up the growth plates of long bones. Type II is also part of the inner ear, the clear gel inside the eyeball, and the soft centre of the discs between the spinal bones. Type III is found in skin, lungs, intestinal walls and blood-vessel walls.
Tissues don't each hold just one type. Skin appears in both the Type I and Type III lists, and Type III fibrils cross-link with other collagen types as well as with each other. The tissue lists describe where a type usually occurs. They don't promise a benefit for those tissues, and they don't tell you exactly what a particular animal raw material contains.
What processing does: unfolding versus cutting
Native or undenatured collagen keeps its natural folded structure. In the rope picture, the three strands are still wound together. One researched example is undenatured type II collagen (sold as UC-II), a supplement ingredient made from chicken sternum cartilage.
Gelatin is a water-soluble protein made from heat-denatured collagen. Denaturing means the ordered three-strand arrangement has been lost, so the rope has come loose. That describes a change in shape. It doesn't, by itself, say how much the chains have also been cut.
Hydrolysed collagen goes further: the chains themselves are cut into short pieces called peptides. One review describes hydrolysed collagen as a group of low-molecular-weight peptides made by enzymes working in acid or alkaline conditions at a set temperature. If gelatin is the loosened rope, a hydrolysate is those strands chopped into short bits. The two steps can follow one another. A gelatin hydrolysate, the material used in the human study described below, is gelatin that has been further cut into peptides.
Losing the helix shape (denaturing) and breaking the chains into fragments (hydrolysis) are separate events, and one of the body's own enzymes shows this. MMP-1 is a collagenase, an enzyme the body uses to break down collagen when tissue is remodelled. The opening in the enzyme where collagen has to fit is too narrow to take the intact triple helix. Researchers found that the enzyme first unwinds the helix locally, and only then cuts the bonds that hold each chain together. That work concerned tissue remodelling, not supplement manufacturing or digestion.
"Hydrolysed" doesn't mean identical, either. The same review names the source the collagen is extracted from as one of the main factors behind a hydrolysate's properties, including peptide size, solubility and functional activity. The word tells you the chains have been cut into short peptides, not the exact mix of peptides in a particular product.
Source, type and form: three separate pieces of label information
A collagen ingredient can be described with three separate pieces of information. Source is the animal and tissue it came from. Type is the protein family, such as I, II or III. Form is the structural state: native or undenatured, gelatin, or hydrolysed peptides. These are connected, but none of them fully tells you the others. Source tissue influences which types are present: cartilage is mainly Type II, skin contains both Types I and III, and collagen from sea cucumber body wall was identified as type I. Neither source nor type settles the form, though. Hydrolysed collagen can come from bovine or porcine sources and from the skin, scales and bones of marine animals. Undenatured type II collagen has been made from chicken sternum cartilage. In one human study, gelatin hydrolysates were made from porcine skin, chicken feet and cartilage.
Type II shows how this plays out. Cartilage is rich in Type II collagen. That starting material can be kept with its collagen still wound, as with the undenatured ingredient above, or it can be turned into a gelatin hydrolysate, as in the absorption study described below. A hydrolysed product contains fragments of whichever proteins the starting material held, not the intact molecule. So a "Type II" label alone doesn't tell you whether the collagen is undenatured or hydrolysed.
Scientific writing sometimes blurs the two questions too. The joint-health review quoted earlier calls native and hydrolysed collagen "the most studied collagen types for joint health". There, "types" means forms, not Types I, II or III. When you read about collagen, it's worth checking which question a word is actually answering.
Form also matters because the two main forms are studied for different proposed reasons. The same review suggests that native collagen works through the immune system, which would need to recognise particular features of the intact structure. Hydrolysed collagen, it suggests, "may contain" active peptides that can reach joint tissues. Both ideas are hypotheses, not settled facts. The review's published summary presents no human mechanism data, and the authors of the UC-II trial discussed below say its mechanism still needs to be worked out.
What one study saw after people swallowed gelatin hydrolysates
A human absorption study in the Journal of Agricultural and Food Chemistry asked a basic question: when people eat hydrolysed collagen, which collagen peptides turn up in their blood? Healthy volunteers fasted for 12 hours and then ate 9.4 to 23 g of gelatin hydrolysates, meaning gelatin already cut into peptides, made from porcine skin, chicken feet or cartilage. The researchers tracked hydroxyproline in peptide form, that is, hydroxyproline still joined to other amino acids. Hydroxyproline is one of the building blocks collagen is rich in. Before the meal, the volunteers' blood held almost none of this peptide-bound hydroxyproline.
After the meal, peptide-bound hydroxyproline rose significantly. It peaked one to two hours after eating and had fallen to about half that peak by four hours. The main peptide identified was Pro-Hyp: just two amino acids, proline and hydroxyproline, joined together. Smaller but significant amounts of other two- and three-amino-acid peptides were also found, such as Ala-Hyp and Pro-Hyp-Gly.
The volunteers ate material that had already been cut into peptides, and the hydroxyproline-containing peptides found in their blood afterwards included fragments only two or three amino acids long, with Pro-Hyp the main one reported. The study therefore establishes that particular short, collagen-derived peptides can be absorbed into the bloodstream within a few hours of eating a hydrolysate. It isn't a complete inventory of the blood, because the researchers followed specific hydroxyproline-containing peptides.
According to its published summary, it followed a single meal in healthy, fasting adults and measured peptides in the blood. It didn't track whether those peptides reached joints or skin, whether they changed any tissue or symptom, or what happens with daily intake over time. Because no intact or undenatured collagen was given, it can't tell us how those forms behave after they're eaten. Nor does it support the popular idea that swallowed collagen arrives in the body as ready-made "replacement parts". A peptide in the blood shows absorption, not benefit.
Reading one Type II trial for what it actually tested
A multicentre, randomised, double-blind, placebo-controlled trial published in Nutrition Journal studied 191 volunteers with knee osteoarthritis. Participants were split into three groups. One took 40 mg of UC-II daily, one took glucosamine hydrochloride (1500 mg) with chondroitin sulfate (1200 mg), and one took a placebo, all for 180 days. The main outcome was the change in the total WOMAC score, a knee osteoarthritis symptom measure with subscales for pain, stiffness and physical function. Secondary outcomes included another function index (the Lequesne index) and a pain rating scale.
At day 180, the UC-II group's total WOMAC score had dropped significantly more than the placebo group's (p = 0.002) and the glucosamine-chondroitin group's (p = 0.04). Pain and stiffness subscale results were significant compared with both other groups, and physical function was significant compared with placebo. Safety outcomes didn't differ between the groups. The authors concluded that UC-II improved knee symptoms and was well tolerated, and that more studies are needed to explain how it works. The published summary gives p-values but not the size of the score changes, so it shows the differences were statistically significant without showing how much better participants felt.
This result belongs to one named undenatured ingredient, taken at 40 mg a day by adults with knee osteoarthritis for 180 days and measured with symptom scores. It doesn't automatically extend to hydrolysed Type II collagen, which rests on a different proposed rationale. Nor does it extend to other Type II ingredients, other doses, people without osteoarthritis, or finished products with different ingredients. Two products can share the words "Type II" and still differ in exactly the ways this trial pinned down.
Symptoms are also a different question from cartilage structure. The results reported in the trial's published summary are symptom and function scores. Less pain or stiffness doesn't demonstrate that cartilage regrew or was repaired. It doesn't rule out an effect on structure either; these reported results simply don't address it.
Questions to ask of a collagen study
Which source, type and form were used, and was it a named ingredient or a finished product? Who took part, at what dose, and for how long? What was measured: peptides in the blood, symptom scores, or tissue structure?
Sources
- COL1A1 gene: MedlinePlus Genetics — page text
- COL2A1 gene: MedlinePlus Genetics — page text
- COL3A1 gene: MedlinePlus Genetics — page text
- Identification of food-derived collagen peptides in human blood after oral ingestion of gelatin hydrolysates. — abstract
- Efficacy and tolerability of an undenatured type II collagen supplement in modulating knee osteoarthritis symptoms: a multicenter randomized, double-blind, placebo-controlled study. — abstract
- Hydrolyzed Collagen-Sources and Applications. — abstract
- Collagen Supplementation for Joint Health: The Link between Composition and Scientific Knowledge. — abstract
- Extraction and characterization of collagen and gelatin from body wall of sea cucumbers Stichopus horrens and Holothuria arenicola. — abstract
- Collagenase unwinds triple-helical collagen prior to peptide bond hydrolysis. — abstract