Cell Line Choice Decides the Answer
Before any question about a compound can be asked, there is a prior question about the culture: is it what the paper says it is? A substantial body of published work has been performed on misidentified or cross-contaminated lines, cells carrying one name in the methods section and a different genome in the flask. Authentication is cheap and available, and still not universal practice.
Mycoplasma is the companion problem. It cannot be seen, it alters metabolism and signaling, and it is prevalent enough that data from an untested culture stands a genuine chance of being data about the contaminant. Neither issue is exotic. Both are routine enough that a methods section saying nothing about them has left something unresolved, the same reading discipline applied to the controls in vehicle controls.
Why a finding cannot be quoted without its system
Suppose identity is beyond doubt. The line is still not a neutral vessel in which a compound is watched. It is one particular population carrying one expression profile, one metabolic state and one history, and whatever emerges is jointly a property of the compound and of that system.
Hence the difference between two sentences. “Compound X increases proliferation” is not a result. “Compound X increases proliferation in this line under these conditions” is. Predictably, the first is the version that travels.
Species mismatch hiding in plain sight
Sequences frequently differ between species, and the receptors diverge further still. Test a human sequence against a rodent receptor and you have run a cross-species experiment, whether or not anyone labels it that way.
Occasionally the sequences coincide and the concern evaporates. Frequently they do not, and that divergence is exactly what makes the comparison impossible to interpret. Kisspeptin illustrates it, as set out in kisspeptin-10 and KISS1R, where the human and rodent peptides are not the same molecule at all.
Receptor presence, level and subtype
A peptide working through a receptor can do nothing where that receptor is absent. Whether any particular line carries it is an empirical matter whose answer often surprises, and expression levels wander with passage number and culture conditions.
Failure runs in both directions from there. Test in a line that lacks the receptor, obtain nothing, and the nothing describes the model rather than the compound. Test instead in a line engineered to carry far more receptor than any natural tissue and a solid response can appear at concentrations where physiological expression would yield no measurable signal, displacing apparent potency in the sense examined under Ki, IC50 and EC50.
Presence alone is not sufficient either, because subtype carries equal weight. Wherever a family of related receptors exists, discriminating among them is the entire question, as with the melanocortin receptors treated in MC1R and MC4R selectivity. A line expressing the wrong member of the family has answered some other question competently.
Three kinds of culture, three kinds of question
| Immortalized | Division continues indefinitely because growth control has been broken somewhere. That breakage is what makes the line usable and simultaneously what makes it abnormal, and the abnormality frequently sits inside the very pathways being investigated. |
| Primary | Nearer to the tissue of origin, less consistent in behavior, variable from donor to donor, and progressively altered by passaging. |
| Engineered | Delivers a clean answer to one narrow question, namely whether the compound acts at this receptor, while saying nothing whatever about whether the receptor occurs anywhere it would count. |
No entry in that table is the right answer in general. Each suits a different question, and the usual mistake is interpreting output from one category as though it had come from another.
Variables that shift the answer while the line stays the same
Serum concentration governs how much compound remains free rather than bound, moving effective concentration in the manner described under why a concentration is not a dose. Confluence alters signaling behavior. Passage number alters expression. Atmospheric oxygen, which is what a standard incubator delivers, sits well above what most tissues actually encounter.
One variable belongs specifically to this compound class. Peptidase activity in serum degrades peptide during the incubation, so real exposure differs between serum-containing and serum-free conditions even when the nominal concentration is identical in both.
The record that makes a result portable
Approaching somebody else’s paper, seven things are worth locating: the line, its source, the passage at which it was used, whether the receptor of interest is expressed and at what level, whether authentication was performed, whether contamination testing was performed, and what the serum conditions were. A study that supplies those can be set against another study that supplies them. A study supplying none can be set against nothing.
Generating the work rather than reading it, the identical list is what allows anyone else to repeat it. Place it beside the material-side record described in inventory labels and records and both halves are documented, compound and system alike. That is the floor for a result that still means the same thing twelve months from now, and the entry point to the broader difficulty examined in why two laboratories get different results.
