Peptide gels in the vial come from self-assembly, not chemical damage

Does a gel in the vial mean the peptide is chemically damaged?

A gel by itself is not evidence of chemical damage: gelation is physical instability, a rearrangement of intact chains, and the act of gelling breaks no covalent bond. Formulation science splits instability into two categories that behave differently and mean different things.

Chemical instability changes covalent bonds: deamidation, oxidation, hydrolysis of the backbone, disulfide scrambling. Atoms move, and the molecule that comes out is not the molecule that went in.

Physical instability changes how intact molecules arrange themselves relative to each other: unfolding, adsorption to surfaces, aggregation, precipitation, gelation. Every covalent bond survives. What changes is the arrangement.

That split is the organizing frame of the standard reviews of protein pharmaceutical stability ^[1]⁠^[2]. Gelation sits squarely on the physical side. A peptide that has gelled has not been chopped up or oxidized by the act of gelling; it has associated.

A gelled vial is not automatically fine, either. A gel says the chains rearranged; it says nothing, in either direction, about whether the chains are intact, and a great deal of internet advice conflates the two. How peptides degrade through deamidation and oxidation covers the chemical side.

How does a peptide turn into a gel?

Peptide gelation is understood to run in three steps: chains pair up by hiding their hydrophobic faces, hydrogen-bond into beta sheets, and the fibers entangle into a network spanning the container.

Picture the peptide as a string of beads. Some beads are comfortable in water. Others, the hydrophobic ones (valine, leucine, isoleucine, phenylalanine, tryptophan), would rather touch each other than touch water. The three steps apply when a string carries a run of those water-avoiding beads on one face.

1. Chains pair up by hiding their hydrophobic faces against each other. In the best-studied model system, a designed beta-hairpin called MAX1, the fibril core is a valine-rich hydrophobic sandwich. Reducing the ability to make lateral hydrophobic contacts measurably slows the kinetics of assembly and lowers the mechanical rigidity of the resulting gel ^[4].
2. The paired chains hydrogen-bond into beta sheets running along the growing fiber. Aromatic rings, where present, can stack on top of that. In one ultrashort designed tetrapeptide, backbone hydrogen bonding and aromatic stacking together stabilize the sheet ladders ^[5].
3. The fibers entangle and branch into a percolating network. Rheology and dynamic light scattering both show power-law behavior with comparable critical exponents at the percolation threshold, which is the signature of a network spanning the whole container ^[6].

At step three, the liquid stops being a liquid. Nothing left the vial, and nothing was added.

Real-time high-speed atomic force microscopy in liquid has imaged parts of this process directly for one model beta-sheet peptide, KFE8 ^[3]. That study is explicit that complete assembly pathways have not yet been resolved, and that different morphologies form at a solid-liquid interface than in bulk solution ^[3].

Why does a peptide gel hold its water instead of separating?

A peptide gel holds its water because the fibrils form a continuous scaffold through the liquid, the opposite geometry from precipitation, in which the solid separates and settles. A gel is not the peptide falling out of solution.

In a gel, the liquid stays where it is, held by capillary forces in the mesh, by hydration of the fibril surfaces, and by the osmotic cost of squeezing solvent out of a network that spans the container.

The proportions make the point. Self-supporting peptide hydrogels form at low single-digit percentages of peptide by weight. One well-characterized short sequence gels at a critical gelation concentration of 20 mg/mL and above, which is 2 percent w/v ^[5]; the threshold is sequence-specific, and that is one measured value, not a general figure. Everything else in that gel is water: the vial contains the same peptide and the same water it contained a minute earlier, reorganized.

The same behavior is why the peptide-hydrogel field exists at all. Injectable drug-delivery scaffolds, tissue-engineering matrices and wound materials are built by deliberately engineering the behavior that shows up uninvited in a reconstitution vial ^[7]. In that field a gel is not a defect. It is the product.

Do AOD-9604 and kisspeptin-10 gel in the vial?

No published measurement shows that AOD-9604 or kisspeptin-10 gels, although they are the two peptides most often named as gel-formers. PubMed searches for gelation, aggregation, fibril formation, self-assembly and formulation behavior returned no rheology study, critical gelation concentration, or structural characterization of a gelled state for either ^[27].

For kisspeptin, the aggregation literature that exists concerns kisspeptin's relationship to other peptides' amyloid, not kisspeptin assembling itself. For AOD-9604, the published work is anti-doping detection and metabolism, not physical chemistry. The exact queries and their hit counts are listed in full, so anyone can rerun them ^[27]. A failed search is not proof of absence: the measurement could sit in a formulation patent, a supplier technical file, or a journal outside PubMed's scope ^[27].

Every source found asserting that these two peptides gel is a vendor page, a vendor's help-center article, or a blog monetized by peptide traffic. None presents primary data. Several state confidently that gelling indicates high purity, a claim with no mechanism behind it and an obvious commercial interest in front of it.

The evidence grade is commonly reported by sellers, mechanistically plausible from the sequences, not demonstrated in published measurement.

Do self-assembly design rules apply to hormone fragments like kisspeptin-10?

Applying self-assembly design rules to kisspeptin-10 or AOD-9604 is an extrapolation, because every system in the cited self-assembly literature is a peptide designed to assemble.

Those systems are MAX1 and its variants ^[4]⁠^[6]⁠^[9]⁠^[11]⁠^[23], KFE8 ^[3], the VEK octapeptides ^[10], Phg4 ^[5], the naphthalene-dipeptide ^[22], and (FKFE)2 ^[25]. Not one is a naturally occurring hormone fragment. The design rules those systems yield are real, but they were extracted from molecules selected for that behavior.

A hormone fragment was never selected for anything of the kind. Carrying the rules over to one is the weakest joint in the plausibility argument for kisspeptin-10 and AOD-9604, and it is a larger limit than the missing measurement.

What does kisspeptin-10's sequence predict about gelling?

Kisspeptin-10's sequence makes gelling plausible, not proven: four of its ten residues are aromatic, and its net charge near neutral pH is roughly +2.

Kisspeptin-10 is residues 112 to 121 of the KISS1 precursor, sequence Tyr-Asn-Trp-Asn-Ser-Phe-Gly-Leu-Arg-Phe, carried as the C-terminally amidated form ^[8]. Its four aromatic residues are one tyrosine, one tryptophan and two phenylalanines.

Aromatic stacking is a real lever on beta-sheet assembly, and a strong one. In one designed tetrapeptide, changing only the geometry of the rings, swapping flexible phenylalanine for rigid phenylglycine at constant aromatic content, converted a non-assembling sequence into a hydrogelator ^[5]. That study varied ring orientation, not aromatic count: it establishes that aromatic stacking can be decisive, not that four aromatics out of ten is itself predictive.

The broader proposition, that aromatic side-chain interactions drive early self-assembly recognition, is described in the primary literature as controversial rather than settled ^[25].

The +2 charge counts the arginine and the free N-terminus, with the C-terminal amide removing the usual negative. A net charge modulus of 2 sits above the threshold at which a systematic charge study found self-supporting gels form, which was a charge modulus greater than 1 ^[10]. That study tested only moduli of 0, +2 and minus 2, so it establishes a floor and not a window: nothing in it says a higher charge would prevent gelation.

What does AOD-9604's sequence predict about gelling?

AOD-9604's sequence points both ways: low net charge favors association, but a disulfide loop resists beta-sheet formation and its hydrophobicity takes the weaker, aliphatic route.

AOD-9604 is a tyrosine attached to the front of human growth hormone residues 177 to 191 ^[12]. Mapping that onto the somatotropin sequence gives Tyr-Leu-Arg-Ile-Val-Gln-Cys-Arg-Ser-Val-Glu-Gly-Ser-Cys-Gly-Phe. The two cysteines are the ones that form a native intramolecular disulfide in the parent hormone, so the fragment carries a small constrained loop ^[13].

That loop cuts against gelation. A disulfide bridging Cys182 and Cys189 in mature growth-hormone numbering constrains eight of the fragment's sixteen residues into a cycle, and a cyclized backbone cannot readily adopt the extended conformation an intermolecular beta sheet requires. Which form a given batch carries, closed disulfide loop or free thiols, changes the prediction, and nothing in the published literature settles it.

AOD-9604's hydrophobicity is mostly aliphatic rather than aromatic: leucine, isoleucine and two valines concentrated in the first half, with only two aromatic residues in sixteen (the added tyrosine and the terminal phenylalanine). The aliphatic route to assembly is real but weaker.

A systematic study replacing the phenylalanines of a model amphipathic beta-sheet peptide found that aromatics are not strictly required, but the natural aliphatic substitutions (alanine, valine, leucine) retained assembly competency only with attenuated kinetics ^[25]. Enhanced hydrogelation appeared with cyclohexylalanine, a non-natural residue AOD-9604 does not contain ^[25]. The aliphatic case for AOD-9604 is the weaker half of that result, not the stronger half.

AOD-9604's net charge near neutral pH is low, on the order of +1, and low net charge removes the electrostatic repulsion that keeps chains apart. In the systematic charge study, decreasing charge modulus increased lateral aggregation of the fibers, producing denser, stiffer networks ^[10]. In a separate kinetics study, point substitutions that reduced total peptide charge made gelation faster ^[11]. Neither study measured a failure to dissolve.

In one study, growth hormone fibrils needed zinc

Growth hormone itself forms cross-beta amyloid fibrils in the presence of zinc ions, and the same work found growth hormone co-localized with zinc in pituitary tissue, suggesting it may be stored in an ordered assembled form and release monomer on demand ^[14]. That is context for why ordered peptide self-assembly is not automatically pathology.

The zinc dependence matters. The finding is not evidence that growth hormone self-assembles unprompted, a reconstitution vial supplies no zinc, and the mechanism does not transfer to a fragment sitting in water.

What makes BPC-157 the structural opposite of a gel-former?

BPC-157 packs four prolines, four charged side chains and no aromatic residue into fifteen positions, which is close to the design brief for a peptide that will not form a beta sheet.

Peptides that reliably dissolve clear tend to be the structural opposite of the gel-formers. BPC-157's sequence is Gly-Glu-Pro-Pro-Pro-Gly-Lys-Pro-Ala-Asp-Asp-Ala-Gly-Leu-Val ^[15]. Its four charged side chains are one glutamate, one lysine and two aspartates, on top of the charged termini.

Proline is the canonical beta-sheet breaker: it lacks the backbone amide hydrogen a sheet needs, and its ring sterically hinders the extended conformation. That is why proline substitution is used as a deliberate strategy to disrupt amyloid assembly ^[16].

The reasoning is sound and general. Applied to any specific vial, it is still an inference, not a measurement.

What decides whether a peptide gels or stays liquid?

Four variables decide whether the same peptide gels or stays liquid: concentration, pH, ionic strength and temperature. That conditional behavior is not a quirk of a few sequences; it is the central finding of the hydrogel literature.

The pH row is the one most often turned into a rule of thumb. For a peptide with a basic isoelectric point, a mildly acidic diluent raises net positive charge and increases electrostatic repulsion between chains. For an acidic peptide, the same move does the opposite. Which case a given sequence falls into is calculable from the sequence, and nothing in the published literature establishes the answer for AOD-9604 or kisspeptin-10.

The ionic-strength row has its own limit. The study that gels at pH 7.4 with 150 mM sodium chloride held salt constant ^[9]. It establishes that gelation occurs at physiological ionic strength, not that varying salt is itself the lever. The lever claim rests on the kinetics study that did vary it ^[11].

Which variable matters most for peptide gelation?

Concentration, set by the volume of diluent at reconstitution, is the variable most tightly coupled to gelation in the measured literature.

Self-supporting gelation has a concentration threshold; below it, no percolating network forms ^[5]⁠^[10]. Of the four variables, concentration is the one fixed at the moment of reconstitution, and it is also the least discussed, because it is unglamorous. That is a statement about which variable the physics is most sensitive to, not a recommendation about what to put in a vial.

Should a stubborn or gelled peptide vial be shaken?

Shaking a peptide vial is the wrong move: it manufactures air-liquid interface, where proteins and peptides adsorb, unfold and aggregate. The handling findings describe how liquid behaves in a container; they are physical chemistry, not instructions for use, and not guidance about administering anything to anyone.

Shaking creates and destroys bubble surfaces thousands of times, dragging molecules onto fresh interfaces over and over. The effect is quantified, not asserted. Across a panel of sixteen antibodies, the initial rate at which surface pressure rose on adsorption to the air-liquid interface strongly predicted how much each one aggregated under agitation stress, while hydrophobicity and equilibrium surface pressure did not ^[17].

Orbital shaking of vials is a standard forced-degradation test because it reliably produces aggregation, and the amount depends on shaking orbit, frequency, and whether the vial is upright or on its side ^[18]. That dependence does not reduce to a simple ranking of interface area: turbidity rose in vertical orientation only at a 3 mm orbit and in horizontal orientation only at a 30 mm orbit, and computational fluid dynamics of how air and liquid disperse predicted the pattern better than interfacial area alone ^[18]. Where surfactants protect a formulation, they work by competing for that interface ^[19].

All three of those studies are on monoclonal antibodies or larger proteins; none studied a short peptide. The stated mechanism in all three is interfacial unfolding, and a ten-mer or sixteen-mer with little tertiary structure to unfold is a materially different case. The direction of the finding is conservative either way: agitation has a documented cost in the systems where it has been measured, and no documented benefit anywhere. Why peptide vials should not be shaken covers the agitation evidence in more depth.

Some beta-sheet peptide gels are shear-thinning and self-healing: they flow under applied shear and immediately re-form a solid gel when the shear stops, a behavior that depends on how much the fibril network branches ^[23]. Mechanical force can temporarily liquefy some gels. That is not an argument for shaking, since the interfacial aggregation penalty is separate from and additive to whatever the shear does, and the gel re-forms anyway.

What do approved peptide labels say about mixing and inspection?

The Cetrotide and Humatrope labels both say not to shake, and both say not to use a reconstituted solution that is cloudy or contains particles ^[20]⁠^[21]. Those labels are consistent with the direction of the agitation data, and they are quoted as handling and storage-science precedent only.

• The Cetrotide (cetrorelix) label, for an approved decapeptide supplied as a lyophilized powder, instructs the user to "gently swirl the vial until the solution is clear and without residues. Avoid forming bubbles," warns "Do not shake or you will create bubbles in your medicine," and states "Do not use a Cetrotide solution if it contains particles or if it is not clear" ^[20].
• The Humatrope (somatropin) label states "Do not shake. The reconstituted solution should be clear," and "Inspect visually for particulate matter and discoloration. If the resulting solution is cloudy or contains particulate matter do not use" ^[21].

What the measurements support is narrow: minimal agitation and no foam generation reduce interfacial turnover, which is the variable the aggregation data track. Slow addition of solvent down the vial wall is widely advised and fits the same mechanism, but no study measuring it has been found, so it remains plausible practice rather than a demonstrated result.

Cetrotide and Humatrope are approved drugs with real formulation development behind them. Research peptides, including AOD-9604, kisspeptin-10 and BPC-157, are not FDA-approved drugs. The labels' approval does not extend to them, and the handling conventions in those labels do not transfer to unapproved material as recommendations, endorsements or a read-across.

Will a peptide gel loosen if the vial is left to stand?

Waiting does not reverse an assembling peptide gel, because gel networks measurably stiffen as they age ^[22]. For a lyophilized cake that has not finished wetting, time helps; for a peptide that has already started to self-assemble, time does the opposite.

In a well-characterized dipeptide gelator, thin low-persistence fibers form first, then laterally associate into thicker bundles, and the initially weak hydrogel becomes stronger with time ^[22]. "Give it time and it will relax" is standard reconstitution folklore, and the self-assembly literature does not support it for a gel that is already forming.

Does warming a peptide vial dissolve a gel?

Warming triggers hydrophobically driven peptide self-assembly instead of reversing it; one designed peptide was tuned specifically to undergo thermally triggered hydrogelation at 37 degrees Celsius ^[9].

"Warm it up, warmth dissolves things" holds for most ordinary solids. Heat is not a solvent for a hydrophobically driven assembly. This is one of the clearest cases where an intuition imported from dissolving sugar gives the wrong answer for peptides.

Is a cloudy peptide solution the same as a gel?

A cloudy peptide solution is not a gel: only uniform gelling has the physical self-assembly explanation, and approved labels treat cloudiness as a discard signal ^[20]⁠^[21].

Cloudiness is where the physical explanation and the approved-label instruction diverge, and the divergence should be stated rather than smoothed over. The Cetrotide and Humatrope labels tell the user to discard a cloudy solution ^[20]⁠^[21], and for an approved product with a known formulation and a validated release specification, that is the correct rule. Nothing in the gelation mechanism overrides it.

A uniform gel and a cloudy solution are not the same observation, and what cloudiness in an injectable vial can mean is a separate question from gelation.

When is a strange-looking peptide vial a real problem?

A strange-looking peptide vial is a real problem when it shows discrete particles, fibers, foreign material or a color shift, none of which is gelation. The decision rule is a boundary, not a reassurance.

A gel is a plausible physical state. Smooth, uniform, no discrete inclusions, no color that was not there before. For a sequence with the structural features of a self-assembler, that is consistent with self-assembly and is not, by itself, evidence of chemical degradation.

Visible particulates and discoloration are a different signal entirely. Discrete particles, fibers, flakes, glass, angular debris, or a color shift are the failure mode that injectable-product standards exist to catch. Injections are required to be essentially free from visible particulates ^[24], and the approved labels tell the user to discard on exactly those findings ^[20]⁠^[21].

The accurate rule is not "a cloudy vial is always fine." It is this: uniform gelling in a peptide whose sequence supports self-assembly has a physical explanation; cloudiness, visible foreign material and discoloration do not, and those are where the explanation runs out.

Does a peptide that gels have higher purity?

Gelling says nothing about purity: identity and purity are analytical questions answered by chromatography and mass spectrometry, while appearance after reconstitution is a physical state.

What the peptide is, and how much of the material is that peptide, are answered by chromatography and mass spectrometry on the material itself, against a reference standard, before anything is dissolved. Identity and purity are numbers on an analytical report. How the solution looks after diluent is added depends on concentration, pH, ionic strength, temperature and the sequence's own assembly behavior.

A high-purity peptide with a self-assembling sequence can gel, and will above its critical gelation concentration. A low-purity peptide with a non-assembling sequence will dissolve clear. Appearance after reconstitution does not rank purity in either direction, and the vendor claim that gelling demonstrates high purity has no mechanism behind it. It is a sales line wearing a lab coat.

The corollary matters more than the correction: appearance will not tell anyone what is in a vial. Only analysis will, and the guide to reading a peptide certificate of analysis covers where those numbers appear.

Does gel behavior say anything about whether AOD-9604 or kisspeptin-10 works?

Gel behavior answers a physical-chemistry question, not whether AOD-9604, kisspeptin-10 or BPC-157 does anything in a person, and it supports no position on that. The gel chemistry is general information only, not medical advice, dosing guidance or instructions for use.

A 2026 narrative review of peptides marketed as modulating the growth-hormone and IGF-1 axis stratifies that class across evidence tiers, running from regulatory-grade randomized trial data down to a complete absence of human studies, and places AOD-9604 toward the thin end ^[26]. Kisspeptin-10 and BPC-157 fall outside that review's scope and carry no evidence tier from it. AOD-9604's human trial record is covered separately.

What is kisspeptin-10's compounding status in the United States?

At its October 29, 2024 meeting, the United States Pharmacy Compounding Advisory Committee voted against adding kisspeptin-10 to the 503A bulk drug substances list ^[28].

The committee voted the same way on ipamorelin, ibutamoren mesylate and L-theanine ^[28]. That is a compounding-eligibility decision. It places the compound and has no bearing on any physical-chemistry claim about gelation.