Culture medium is not a neutral solvent, and for an unmodified peptide it is closer to a digestion. Serum carries peptidases. Most cell lines display more on their surface. A native sequence added to a well can be substantially gone within minutes, and the concentration at the readout bears little relation to the concentration at the pipette. Nothing in the assay reports this. The peptide simply disappears, and a weak or absent result gets attributed to the compound rather than to its absence.
What is doing the cutting
| Enzyme | Where | What it cuts | Peptides exposed |
|---|---|---|---|
| Dipeptidyl peptidase-4 | Serum; cell surfaces (CD26) | After Pro or Ala at position 2 | Native GHRH, GLP-1, GIP and analogues without Aib |
| Neprilysin | Cell surfaces, especially kidney and neural lines | Before hydrophobic residues | Kisspeptin, enkephalins, many short sequences |
| Aminopeptidases | Serum; cell surfaces | From the free N-terminus | Anything uncapped |
| Carboxypeptidases | Serum | From the free C-terminus | Anything unamidated |
| Serum endopeptidases | Serum | Internal sites | Long, unstructured sequences |
Two of these are membrane-bound, which means the cells themselves are part of the problem. A peptide added to a dish of cells expressing CD26 is being cleaved by the very cells whose response is being measured, and the rate depends on how many cells are present and which line they are.
Does heat-inactivated serum solve this?
No, and the assumption that it does is widespread. Heat inactivation at 56 °C for thirty minutes was designed to destroy complement, and it does. Many peptidases survive it: DPP-4 in particular retains most of its activity. Heat-inactivated serum is complement-free, not protease-free, and a peptide added to medium containing it is still being degraded — more slowly than in untreated serum for some enzymes, at nearly the same rate for others. The only serum that does not degrade peptides is no serum.
How fast does it actually happen?
For a native, unmodified sequence with a DPP-4 site, the half-life in serum-containing medium is on the order of minutes. Native GLP-1 is the textbook case, at roughly two minutes in vivo and not much longer in a dish. Shorter unmodified peptides with free termini — kisspeptin-10, PE-22-28, the dipeptide bioregulators — are in the same range. A readout taken an hour after addition is measuring the response to a concentration that fell to a small fraction of nominal within the first ten minutes.
Why the engineered peptides survive
Every stability modification in the catalogue is an answer to one of the enzymes above. Aib at position 2 removes the DPP-4 site. A D residue at a cleavage site is not recognised. N-terminal acetylation blocks aminopeptidases; C-terminal amidation blocks carboxypeptidases. A Pro-Gly-Pro tail resists both. Cyclisation removes the free ends entirely. The four extension strategies map almost exactly onto the table, which is no coincidence: they were designed against it.
The practical consequence is that the same experiment behaves differently for a native peptide and its stabilised analogue, and the difference is not potency. Sermorelin and tesamorelin compared in serum-containing medium is largely a comparison of degraded peptide against intact peptide; compared in serum-free medium with no cells, they look alike.
What are the options for an unmodified peptide?
Four, in rising order of cost. Run in serum-free medium for the exposure window, which removes the soluble enzymes but not the membrane-bound ones. Add a protease inhibitor cocktail, which is effective and introduces a mixture of compounds with their own effects on cells. Shorten the window: a readout that is fast — calcium, cAMP, receptor phosphorylation — can be taken before much degradation has occurred. Or replace the compound at intervals to maintain something like a steady concentration, accepting that the concentration is now a schedule rather than a number.
How would I know how much peptide is left?
By measuring it, which is rarely done and always informative. Sample the medium at intervals after addition and run it by LC-MS: the intact peptide's peak falls and the cleavage products' peaks rise, and the half-life in that exact system is read off the curve. This takes one afternoon and replaces an assumption with a number. For any experiment where the exposure time matters, it is the control that determines whether the exposure happened at all. Without it, a negative result cannot distinguish an inactive compound from an absent one.
Does adsorption make this worse?
Yes, and the two are easily confused because both remove peptide from solution over time. Adsorption takes a fixed amount onto the plastic quickly and then stops; degradation takes a proportion continuously and does not. A measured time course distinguishes them: an early drop that then plateaus is the wall, a steady exponential decline is enzymes. In a serum-containing well with a dilute unmodified peptide, both are operating, and the concentration at ten minutes may be a fifth of nominal for two unrelated reasons.
What should the methods section say?
The medium and its serum content; whether the serum was heat-inactivated, since that is often misreported as protease-free; the cell line and its density, because they set the membrane-bound enzyme load; the time from addition to readout; and, ideally, a measured half-life in that system. A concentration–response curve built without knowing the exposure is a curve against nominal concentrations that were never present, and its EC50 is a property of the degradation as much as of the receptor.
All products referenced here are supplied for laboratory and research use only. They are not drugs, foods, supplements or cosmetics, and are not for human or veterinary use.
