Reporter gene assays guide — Part 10: High-throughput screening
Reporter Assays for High-Throughput Screening
A reporter assay that works beautifully in a 96-well format with carefully-pipetted reagents and a graduate student paying close attention can fall apart when you ask it to run as 50,000 wells in a screening facility. HTS imposes constraints that simply do not exist in low-throughput work: liquid handling is automated, incubation times are tightly controlled but not adjustable, plate readers are integrated with robotics, and the data must be robust enough that no human will look at most of the wells. The chapter covers the design choices that determine whether a reporter assay is fit for HTS.
What makes an assay HTS-ready
The technical bar for a screening assay is defined by the Z' factor, the same metric discussed in the normalisation chapter. Z' combines the dynamic range and the variability of the positive and negative controls into a single number that predicts how well the screen will separate hits from non-hits. The standard thresholds are:
Z' > 0.5: Acceptable for HTS
Z' > 0.7: Excellent
Z' < 0.5: Marginal, requires further optimisation
Z' < 0: Worse than random, not usable
Z' is measured on a pilot plate under the actual screening conditions: same plate format, same reagents, same liquid handler, same plate reader, same timing. The pilot plate is the single most important quality-control step in the entire screen. Optimisation that does not improve Z' is wasted effort.
Cell-based reporter assays in HTS are usually expected to achieve Z' > 0.5. Biochemical assays (luciferase enzyme alone, for example) routinely achieve Z' > 0.8. Cell-based assays are limited by cell-to-cell variability, transfection efficiency, and the inherent noisiness of biological systems. Getting a cell-based reporter assay above 0.5 is real work; getting it above 0.7 is exceptional.
Plate format considerations
HTS is almost always done in 384-well or 1536-well format. The choice between them depends on the screening centre's infrastructure, the cost of reagents, and the biology.
384-well plates are the most common format for cell-based screens. They have wells of approximately 100 to 150 μL working volume, can be handled with standard 96-channel pipetting heads (with an adapter), and are compatible with most plate readers. They are a good balance between throughput and assay robustness.
1536-well plates give 4-fold higher throughput per plate but require specialised equipment. Working volumes are 5 to 15 μL, evaporation is a major problem, and cell number per well is very small. Cell-based reporter assays in 1536-well format are possible but require careful optimisation of cell density, transfection conditions, and liquid handling. They are most often used for biochemical screens or for very well-validated cell-based assays.
6-well and 24-well plates have no place in HTS. They are too low-throughput, too expensive per well, and incompatible with automated handling.
The choice of plate also affects the choice of plate reader. Most modern readers handle 96, 384, and 1536-well plates, but the integration time and the optics are different for each format. Optimise the reader settings for the specific plate format you are using.
Cell handling in HTS
Cells for HTS need to be uniform, healthy, and at the right density. This is more demanding than for a single experiment because a screen can run 50,000 wells over several days, and the cells at well 50,000 need to behave like the cells at well 1.
Cell line authentication. Verify the cell line identity by STR profiling or by a relevant functional marker. Misidentified or cross-contaminated cell lines are a real problem in HTS because the screen generates a lot of data on the wrong cell.
Mycoplasma testing. Test cells for mycoplasma before the screen and during long screens. Mycoplasma contamination causes high variability, low signal, and erratic responses: exactly the failure modes that destroy Z'.
Cell banking. Prepare a large, well-characterised bank of cells at a defined passage number and use these cells for the entire screen. Thaw fresh vials at regular intervals rather than maintaining a continuously-passaged culture. The cost of preparing a cell bank is trivial compared to the cost of repeating a screen because the cells drifted.
Cell density optimisation. A cell density titration is a standard part of assay optimisation. Plate cells at 2,500, 5,000, 10,000, 20,000, and 40,000 cells per well in the screening format, and identify the density that gives the best Z'. This depends on cell type, plate format, and incubation time. The right density is not necessarily the highest density that gives a good signal.
Transfection method. For cell-based reporter assays in HTS, transfection is a major source of variability. Reverse transfection (mixing DNA and transfection reagent in the well before adding cells) is more reproducible than forward transfection. Automated transfection is available on several platforms. The transfection reagent, DNA amount, and reagent:DNA ratio all need to be optimised in the screening format.
Reagent and assay considerations
Substrate stability. HTS runs can take hours or days, and the substrate must remain stable throughout. Firefly luciferase substrate is reasonably stable at room temperature for a few hours. Renilla substrate (coelenterazine) is much less stable and must be protected from light. NanoLuc substrate is also light-sensitive. Prepare substrate working solutions immediately before use, or use automated dispensers that prepare fresh substrate continuously.
Compound handling. Compound libraries are usually stored in DMSO at -20°C or lower, in 384-well or 1536-well master plates. The screening facility's automated systems transfer compounds to assay plates, typically at a 1:100 or 1:200 dilution. The final DMSO concentration in the assay is therefore 0.5 to 1%, which is the upper limit for most cell-based assays. Use DMSO-tolerant cell lines and validate that 1% DMSO does not affect the reporter.
Liquid handling. Automated liquid handlers (Hamilton, Tecan, Beckman, PerkinElmer) are standard in screening facilities. The choice of tips, the aspiration and dispensing speed, and the mixing parameters all affect assay performance. Validate the liquid handler for your specific assay before committing to a screen.
Plate reading. Most HTS-grade plate readers can read a 384-well plate in under a minute. The read time is rarely a bottleneck, but the read mode (luminometer, fluorometer, absorbance) and the integration time need to be set correctly. For luminescent reporters, integration times of 0.1 to 1 second per well are typical. For fluorescent reporters, the excitation and emission wavelengths need to be set with the right bandpass filters or monochromators.
Reporter-specific considerations for HTS
Firefly luciferase is the workhorse of HTS because of its mature chemistry, low cost, and broad compatibility with screening equipment. The main weaknesses in HTS are the cost of the substrate at very high throughput, the sensitivity of the signal to pH and temperature, and the relatively slow kinetics (10 to 20 seconds for full signal development). For HTS, firefly works well in both 384 and 1536-well formats with proper optimisation.
NanoLuc is increasingly the reporter of choice for HTS because of its high sensitivity, glow-type kinetics, and small size. The signal develops within 1 to 2 minutes and is stable for tens of minutes, which simplifies timing in automated screens. The main cost is the substrate, which is more expensive than D-luciferin, and the tendency to oversaturate the detector if the signal is too high. Plan for a dilution of the substrate or lysate in the screening protocol.
Dual-luciferase in HTS is challenging because of the two-step substrate addition and the timing required. Many screening facilities have moved to a single-reporter format with cell number normalisation by an imaging-based readout or by a separate viability assay in parallel. The dual-luciferase format is still the gold standard for mechanism-of-action studies and validation, but it is less common in primary screens.
Fluorescent protein reporters in HTS are rare because of the lower dynamic range and the susceptibility to compound autofluorescence. They are mainly used in cell sorting-based screens where the single-cell resolution is the priority.
Quality control throughout the screen
Plate-level QC. Calculate Z' for each plate in the screen. Plates with Z' below 0.5 should be flagged for investigation. The Z' of the entire screen should be plotted as a function of plate number to detect drift over time.
Edge effect detection. Compare the signal in edge wells to interior wells. If the edge wells are systematically different, exclude them or correct for the effect.
Trend detection. Plot positive and negative control values as a function of plate number. A trend (gradual increase or decrease) indicates drift in the assay: usually a reagent, cell, or instrument issue. Stop the screen, diagnose, fix, and resume from a fresh plate if possible.
Hit confirmation. Hits from a primary screen are confirmed by re-testing in triplicate at the same concentration. The confirmation rate is typically 30 to 70% for a well-optimised screen. Hits that do not confirm in the re-test are usually artefacts.
Dose-response. Confirmed hits are tested in 8- to 10-point dose-response. This gives an EC50 value and confirms the dose-dependence of the response. The dose-response is the standard output of a screen and the basis for hit prioritisation.
Counter-screening. Confirmed hits are tested in a counter-screen that filters out the most common false positives: cytotoxicity, luciferase inhibition, autofluorescence, and off-target pathway effects. The counter-screen is run in parallel to the primary screen from the start, not after the screen is finished.
Storage of HTS data
HTS generates a lot of data: typically tens of thousands of data points per plate, hundreds of plates per screen. The data needs to be stored in a structured format (usually a database), annotated with metadata (plate layout, reagent lot, cell passage, date, instrument), and analysed with a defined pipeline. Most screening facilities have a standard informatics platform for this. If you are running a screen at a screening centre, ask about the data storage and analysis before you start.
For a small academic screen (a few thousand compounds), Excel or a similar spreadsheet is often adequate, but the data should still be archived with the metadata. Future-you (or the next student on the project) will want to know what was in well H12 of plate 47 of the screen.