HCG (Human Chorionic Gonadotropin): A Research Overview

Where the synthetic-peptide playbook stops working

The habits a lab builds around research peptides are all downstream of one manufacturing method. A synthetic peptide is assembled residue by residue on a solid support, so it has a written sequence, a mass that can be calculated from that sequence before anyone measures anything, and a well-understood family of manufacturing defects: deletion sequences from incomplete coupling, truncations, incomplete deprotection, residual scavengers and cleavage reagents. The standard document set exists because it maps onto that method. A chromatogram catches the synthesis-related impurities, a mass spectrum catches the sequence errors, a content assay catches the salt and water that the balance also weighed.

None of that machinery was built for a protein that a cell made. This molecule is not chemically assembled at all. Two separate gene products are expressed, folded, disulfide-bonded and glycosylated by cellular enzymes, then the two finished chains associate with each other through non-covalent contacts. The resulting dimer is the species of interest; the individual chains, on their own, do not do the thing the assembled molecule does. Nothing in that description can be verified by calculating a mass from a printed sequence.

The practical failure this creates is quiet. A lab that files the vial in the same inventory category as its peptides will apply the same acceptance template, ask the supplier for the same two figures, receive them, and conclude that the lot is documented. The figures will look normal. They will also be answering questions that do not apply. A purity percentage generated on a denaturing reversed-phase separation is describing whatever survived the column, and the column pulls a non-covalent dimer apart while it runs. An intact mass is a distribution rather than a number, so a single reported value has already involved a choice about what to report. Neither figure says anything about the one property that distinguishes usable material from material that has failed, which is whether the dimer is still assembled.

The correct mental adjustment is to stop treating the vial as a peptide with unusual labeling and start treating it as a biologic that happens to be sold next to peptides. Biologics are characterized by a panel of orthogonal methods rather than by two headline numbers, their specifications are written as comparability against a reference lot rather than as a single threshold, and their documentation names the manufacturing route because the route determines the impurity profile. That is a heavier document set, and asking for it is not pedantry. It is the minimum that would let a reader say anything defensible about what is in the container.

The same questions asked of a synthetic peptide and of a glycoprotein hormone

Read the right-hand column as a list of things a peptide certificate is not obliged to contain and therefore usually does not. The gap between the two columns is not a quality judgment about any particular supplier. It is a statement that the ordinary document set was designed around a different manufacturing problem, and that carrying it across unmodified leaves the most important questions unasked. Everything in the rest of this page is an expansion of one row or another in that table: the amount row becomes the question of what an activity unit means, the identity row becomes the question of why there is no single mass, the destruction row becomes the question of how material fails without changing chemically, and the certificate row becomes the question of what to ask a supplier for.

Potency stated in units of biological activity

An International Unit is not a quantity of substance. It is a statement of how much biological response a material produces relative to an agreed physical reference preparation, measured in an agreed assay. The reference preparation is a real stock of material held by an international standards body, assigned a unit content by convention rather than by measurement, and distributed so that laboratories anywhere can calibrate against the same physical object. Successive standards are established as earlier stocks run out, each one bridged against its predecessor by comparative assay. The unit therefore describes behavior, and it inherits everything about the assay and the standard that produced it.

The immediate consequence is that no arithmetic converts units into milligrams. This is not a case of the conversion factor being hard to find. There is no factor, because the two quantities measure different things and their ratio is a property of a particular preparation rather than of the molecule. Two lots can be matched on stated units and still differ in the mass of protein they contain, because potency per unit mass depends on the glycoform distribution, and the glycoform distribution depends on the source and the process. Two lots can also be matched on protein mass and differ in units for the same reason, running in the other direction.

The second consequence is that a unit figure is only as good as the two things behind it. The first is the assay: a classical bioassay against the reference preparation and a cell-based assay reading a second messenger in a receptor expressing line are both legitimate potency methods, but they are different experiments with different variability, and a figure quoted without naming the method cannot be compared with a figure from a different supplier. The second is the standard: a potency value is meaningful relative to a named reference preparation and its edition. A certificate that reports units without naming what they were calibrated against has reported a number with no anchor.

There is a bench-level version of this that comes up constantly. Laboratory work in vitro is usually set up in mass or molar concentration terms, and a vial labeled in activity units does not supply that number. The honest way through is to measure protein content directly by a method appropriate to a glycoprotein, record that measurement as its own figure with its own method attached, and keep it separate in the records from the stated unit content. Recording two independent numbers side by side is correct. Deriving one from the other with a factor picked up from another product is where the error enters, and it is usually invisible afterward because the resulting figure looks like a measurement.

What an activity unit and a mass figure each establish

Any units-per-milligram figure encountered in a datasheet, a forum post or a supplier email is a property of one preparation measured one way. It is not a physical constant and it does not travel. Carrying such a figure across from a different product, or from an older preparation of the same product, silently converts a measured quantity into an assumed one, and nothing downstream in the records will show that the substitution happened. The safe habit is to treat the two quantities as belonging to separate columns in the inventory record that are never allowed to populate each other, and to note beside the unit figure which assay and which standard produced it. A number without that annotation is not wrong so much as unreadable, and it becomes harder to reconstruct with every month that passes after the lot was received.

Glycoform distribution and the missing single mass

The polypeptide part of this molecule is fixed: two chains, two gene products, the same amino acids every time. The carbohydrate part is not. Both chains carry N-linked glycans, and one of them additionally carries a cluster of O-linked glycans on a C-terminal extension. Each of those attachment sites can be occupied or unoccupied, and each attached structure can differ in branching, in how far the chain has been extended, and above all in how many terminal sialic acid residues it carries. The product of all those independent variations is not a molecule. It is a population.

That is why there is no correct molecular weight to print. An intact-mass measurement by electrospray returns an envelope of species spread over a range, and reporting it as a single deconvoluted number requires either picking the most abundant species or enzymatically removing the glycans first and reporting the polypeptide mass instead. Both are defensible; neither is what a reader of a peptide certificate expects when they see a mass field, and a certificate that prints one number with no explanation has hidden the choice it made.

The same heterogeneity blurs every separation that depends on physical properties the glycans influence. A reversed-phase run produces a broad feature rather than a sharp peak, because slightly different glycoforms elute at slightly different times. A gel produces diffuse bands rather than tight ones, for the same reason. Charge-based methods, isoelectric focusing and ion-exchange chromatography, are the exception that turns the problem into information: since sialic acid residues are negatively charged, the sialylation distribution resolves into a pattern of species, and that pattern is about as close to a fingerprint of the glycoform population as routine methods get.

Two practical points follow. The first is that lot comparability for this material is a comparison of distributions, not of numbers. The right question to ask about a new lot is whether its charge profile and its size profile overlay those of a lot that behaved acceptably, not whether a percentage cleared a threshold. Writing an acceptance criterion that way takes more work up front, because it requires holding a reference lot and a reference profile, but it is the only formulation that can actually fail a bad lot.

The second is that glycosylation is not analytical decoration. The sugar structures affect how long the molecule persists in a biological system and how it reads in assays that depend on recognition, which means the glycoform distribution shows up in the potency figure discussed above. An analytical difference between two lots and a potency difference between the same two lots are frequently the same observation seen through different instruments.

What each method returns for a heterogeneously glycosylated protein

The last row is worth holding on to. Removing the glycans destroys exactly the property that makes the molecule what it is, and that is precisely why the resulting map is such a clean identity test: with the variable part stripped away, what remains is two defined polypeptide chains that can be checked against their known sequences. Identity and heterogeneity are answered by different experiments here, and neither substitutes for the other. That split has a scheduling consequence as well: the identity question is a one-time check on a lot, while the heterogeneity question is a comparison that only means anything if a reference profile was captured earlier and retained. Labs that never captured the reference profile discover the omission at the worst possible moment, which is when a new lot behaves differently and there is nothing to compare it against.

Activity can be lost with composition intact

For a synthetic peptide, degradation and covalent change are effectively the same event. A methionine oxidizes, an asparagine deamidates, a bond hydrolyzes, and in every case the mass changes and a sufficiently careful analysis sees it. That equivalence is why a peptide lab can treat composition-based tests as a proxy for whether the material is still good.

The equivalence breaks here, and it breaks in the direction that is hardest to notice. The two chains are held together by non-covalent contacts, so they can come apart under conditions that leave both chains chemically untouched: low pH, chaotropic conditions, elevated temperature, prolonged storage in solution, extreme dilution, and air-liquid or solid-liquid interfaces. Dissociated subunits are still the same molecules they were. Their masses are unchanged, their sequences are unchanged, and a great many of their antibody-binding surfaces are unchanged. What they no longer do is act as an assembled dimer.

Two further routes lead to the same place without covalent change. Partial unfolding can leave the dimer nominally intact while altering the surface that matters. Adsorption removes material from solution entirely, and at low concentration in an untreated container a glycoprotein can lose a substantial fraction of its content to the walls, which is why carrier protein in the buffer is a methodological requirement in this space rather than an optional refinement. Aggregation is the mirror image: material leaves the monomeric or dimeric population upward instead of onto the wall, first as soluble oligomers that no visual inspection will reveal and eventually as haze or particulates that one will.

The consequence for document review is specific. A composition-based test can pass on material that has failed. An intact-mass measurement on a dissociated preparation returns the same chain masses it always did. A denaturing gel returns the same two bands, since the gel dissociated the dimer anyway. An immunoassay reports how much epitope is present, and epitopes survive dissociation better than function does, so the analyte reads as present at close to the expected concentration while the property of interest has declined.

What would actually address the question is a small, specific set. Size exclusion under non-denaturing conditions asks directly what fraction of the material is assembled dimer, what fraction is free subunit and what fraction is aggregate, which is the single most informative measurement available for this material. A potency method, whether a classical bioassay against the reference preparation or a cell-based readout in a receptor-expressing line, asks whether the material still does the thing. An immunoassay built on an epitope that exists only on the assembled dimer sits usefully between the two, and is worth asking about specifically, because a supplier reporting immunoreactive content may or may not be using an antibody pair that distinguishes assembled from dissociated material.

Failure modes, and whether composition-based testing can see them

Four of those six rows are invisible to the two tests that dominate a peptide certificate. That is the whole argument of this page compressed into one observation. A document set that reports mass and area percent for this material is not reporting badly; it is reporting on a set of failure modes that are not the ones this material is prone to. The asymmetry also runs one way, which is what makes it dangerous: composition-based tests can pass on failed material, but they do not fail on good material, so nothing about a clean-looking certificate ever prompts the question. The prompt has to come from knowing what the molecule is, which is why the structural point at the top of this page is practical rather than academic.

Two manufacturing routes, two impurity profiles

Material of this kind reaches the bench by one of two routes, and the routes are not interchangeable descriptions of the same thing. The classical route recovers and purifies the hormone from pooled human urine. The recombinant route expresses the two subunit genes in a cultured mammalian cell line, commonly Chinese hamster ovary cells, and purifies the assembled product from the culture medium. The polypeptide chains are the same gene products in both cases. Almost everything else about the two materials, and about the documentation that should accompany them, differs.

The impurity profiles diverge first. A urinary-derived preparation begins from a complex biological fluid, so the process-related impurities are other proteins that survived purification, including related gonadotropin activities that are structurally similar enough to be difficult to remove completely. The relevant questions are about donor material control, about the steps that reduce adventitious agents, and about how much of the co-purified protein burden remains. A recombinant preparation begins from a defined cell line in a defined medium, so the impurity questions are entirely different: residual host cell protein, residual host cell DNA, medium components, and any affinity ligand that leached from a purification column.

The glycosylation diverges second, and this is the difference with the longest reach. Sugar structures are built by the enzymes of whatever cell made the protein. Human trophoblast tissue and a hamster cell line do not build identical glycans, so the two routes produce different glycoform distributions from identical polypeptide chains. Since the glycoform distribution influences clearance behavior, recognition in immunoassays and the readout of a potency assay, materials from the two routes can behave measurably differently even when matched on stated units. Comparative work exists, and it is a good deal smaller than the volume of casual assertions that the two are equivalent.

The regulatory point should be stated plainly rather than left as an inference. Approved pharmaceutical products containing this molecule exist and are manufactured, tested and released under a regulatory framework designed for biologics, with the documentation, facility controls and lot release testing that framework requires. A research-grade vial sold for laboratory use is not that article. It was not made under those controls, it is not released against those specifications, and its documentation should not be read as though it were. That is a factual statement about two different categories of product, and it is the reason a research buyer has to ask for characterization data explicitly instead of assuming a compliance framework is supplying it in the background.

How the two manufacturing routes differ in what they require of documentation

The last row is identical on both sides on purpose. The route determines which impurities to look for and which questions the supplier should be able to answer, but it does not change the core characterization burden. Whichever route produced the material, someone still has to show that the dimer is assembled, that the potency figure is anchored to a named standard, and that the glycoform distribution resembles the lot the method was developed on. It is also worth recording the route in the lab inventory rather than only in the purchasing file, because a route change between orders is the most likely single cause of a material behaving differently while every printed figure on the certificate stays comfortably within its usual range.

A document set built for a biologic, not a peptide

A bare purity percentage establishes remarkably little here, and it is worth being precise about what it fails at rather than dismissing it. It does not establish that the two chains present are the correct ones, because area percent never inspects composition. It does not establish that they are assembled, because the separation that produced it dissociates them. It does not establish that the material has activity, because area is not potency. It does not establish how much protein is in the container, because area percent is a ratio and says nothing about absolute quantity. And it cannot be compared against a percentage from another supplier, because for a heterogeneous glycoprotein the integration boundaries are a matter of judgment rather than a matter of peak picking.

What replaces it is not one better number but a small panel, and the panel is familiar to anyone who has worked with biologics. Identity is established at the polypeptide level, by mapping after the glycans have been removed, or at minimum by a reduced gel showing two chains of the expected apparent sizes. Assembly state and aggregate content come from size exclusion run under conditions that do not pull the dimer apart. Glycoform distribution comes from a charge-based method reported as a profile against a reference lot rather than as a single value. Potency comes from an assay named explicitly and calibrated against a named international standard of a stated edition. Process-related impurities come from assays chosen to match the manufacturing route.

Two administrative fields carry unusual weight for this material. The first is the manufacturing route itself, which many research-grade documents simply omit, and without which a reader cannot tell which impurity questions even apply. The second is the identity and edition of the reference standard behind the potency figure, without which the unit value on the label is unanchored.

None of this is a demand that a research supplier produce a biologics license dossier. It is a claim about which questions are load-bearing. A supplier who cannot say how the material was made, what standard the unit figure was calibrated against, or what fraction of the material is assembled dimer has not withheld a fine detail. Those three answers are the ones that determine whether anything else on the certificate can be interpreted, and all three are questions that never arise for the synthetic peptides sitting next to this vial on the same shelf.

Document elements and what each one establishes

A useful way to run that table is as a request list rather than a scorecard. Each absent row converts into one specific question to a supplier, and most of them can be answered by sending a file that already exists. The rows that come back unanswered are the informative ones, because they mark the boundary of what the supplier actually knows about the material rather than the boundary of what they were willing to print. Keep the exchange itself with the lot record. Months later, when a result looks unlike the last one, the question of what was known about the material at goods-in is far easier to settle from an archived email thread than from memory, and the answer frequently turns out to be the explanation.

Questions this material raises at goods-in

Where to read next

• Peptide purity versus peptide identity for research labs the two-number logic this material does not fit
• How to review a peptide certificate of analysis the standard field-by-field review, as a baseline to depart from
• IGF-1 LR3: a research primer the other recombinant protein in the catalog, single-chain and unglycosylated
• hCG 5000 IU product record specification table and the long-form structural background
• Third-party lab testing and COAs