A Reporter Signal Is Several Steps From the Receptor

Three quite different compounds can each throw a clean, concentration-dependent curve in a reporter assay while having nothing whatsoever to do with the receptor under study. This is not an exotic scenario. It is the ordinary hazard of the format, and seeing why requires counting backwards from the number, because a reporter assay attaches a figure to receptor activation without ever measuring receptor activation.

Curves that come from somewhere else

Three routes lead to a persuasive artifact.

• The compound acts on the enzyme itself. The reporter is generally a luciferase, and luciferase is a protein that things can bind. Certain small molecules inhibit it outright. Others stabilize it, lifting output by lengthening the protein half-life while transcription stays flat. Peptides are less inclined than small molecules to shut down an enzyme directly, though nothing rules it out, and the control costs almost nothing to run. Use a line carrying the reporter behind a promoter that runs constantly and answers to no pathway; anything that shifts signal there is working on the reporter system rather than the receptor.
• The compound suppresses the denominator. A second reporter driven by a constitutive promoter is typically co-transfected, and dividing experimental signal by it corrects for uneven cell numbers and transfection efficiency across wells. That job it does. What it also does is open a fresh failure mode, since anything touching the normalizing reporter warps the ratio. Depress general transcription or translation and the control channel falls, the ratio climbs, and mild toxicity has been dressed up as activation. Inspect the raw numbers from both channels instead of trusting the ratio, and run a viability measure alongside, for reasons covered in what a viability assay measures.
• The element does not belong to your receptor. A response element takes its name from the transcription factor that docks there, never from whatever receptor sits upstream. Several receptors and several pathways funnel into identical elements, and cyclic-AMP response elements are particularly indiscriminate integrators, gathering input from a great many sources. A construct therefore reports pathway activity, full stop. Pinning that activity on one receptor calls for evidence the reporter cannot furnish: a selective antagonist that wipes out the response, or the same construct run in cells lacking the receptor. Absent either, the defensible sentence is that the compound activated the pathway in that cell line, which is a good deal weaker than what usually gets written.

Counting back from the photons

The instrument records light. Trace it backwards: photons arise from substrate turnover, substrate is turned over by enzyme molecules, enzyme comes from transcripts, transcripts come from a gene switched on when a transcription factor engages a response element, and that transcription factor was roused by a cascade the receptor set in motion.

Assembly runs the opposite direction. A response element, meaning a short stretch of DNA that one particular transcription factor recognizes, is positioned upstream of a gene encoding a protein that is easy to quantify. The construct is introduced into cells expressing the receptor of interest. At the end, substrate is added and light output is read.

Gain has a price

Every link in that chain multiplies. One activated receptor generates many second messengers; those activate many transcription factor molecules; each yields many transcripts; each transcript yields many enzyme molecules; each enzyme molecule turns over many substrate molecules.

Sensitivity is the payoff, and a response too faint for a direct measurement registers here without difficulty. The bill arrives as non-linearity. Signal relates to receptor occupancy in a manner that is neither proportional nor consistent from one system to the next, so apparent potency in a reporter assay commonly looks better than potency from a readout nearer the receptor. That is receptor-reserve behavior, and it is why potency figures depend on the system, as covered in the difference between Ki, IC50 and EC50.

The practical upshot: a midpoint from a reporter assay and a midpoint from a binding assay are different kinds of quantity. Setting them side by side implies a comparison that neither one supports.

Expression is never the same twice

Transient transfection hands every well a different copy number, so receptor and reporter levels drift between wells, between passages and between days. Much of the reason reporter assays travel badly between laboratories traces to exactly this, as laid out in why two laboratories get different results.

Stable lines flatten that variability and substitute a distortion of their own. Expression sits fixed at whatever level the clone happened to land on, frequently far above anything physiological, with the potency consequences described under cell line choice.

Hours, not seconds

Transcription takes time, translation takes time, and protein has to pile up. A readout typically lands six to twenty-four hours after treatment, which makes the assay an integrator.

Two things are lost. Kinetics vanish, since a brief strong activation and a longer weaker one can accumulate to indistinguishable signal. Exposure also stops being fixed, because the compound may break down across that window, so what the cells actually saw is not what went in at the start. The peptide-specific form of that problem appears in how peptides degrade.

Fit for purpose

The strengths are real, and they explain why the format turns up everywhere. The final weakness is the one most often skated over: receptor involvement has to be shown with a receptor-negative line or a selective antagonist, and a responding reporter implies neither.