High Throughput Screening: Focusing on Assays in Drug Discovery

By Jennifer L. Clarke, John P. Pezacki, PhD and Helene Letourneau

The current approach to drug discovery involves screening hundreds of compound libraries against a designated biological target to obtain potential drug candidates. The compounds are usually derived from natural products, synthetic analogues and combinatorial chemistry methods, and their large-scale screening represents a first step in the drug-discovery process. To verify their therapeutic efficacy, compounds that successfully interact with the target molecule undergo secondary screening, followed by preclinical testing with animals, and then clinical trials with humans. Altogether, the time from initial compound testing to drug approval takes an average of 12 years. The process of identifying and validating a potential compound is thus time-consuming and expensive, and does not always yield a marketable drug. In fact, 50-70% of lead compounds fail from Phase II to III in clinical trials, (1) and overall, three-quarters of the cost of drug development is attributed to failure.(2) Weeding out ineffective or toxic compounds at the primary or secondary screening stage would therefore be valuable in terms of saving time and money, as well as reducing the amount of animal and human testing required.

Used in the early stages of drug discovery, high throughput screening (HTS) is a tool that could meet this goal. It combines miniaturization and automation to rapidly screen compound assays, reducing both the cost of reagents involved as well as the time taken to determine compound-target interaction. But in order to be as effective as possible, HTS should be done with targets in a biological environment that is close to their native state. This would enable early determination of compound toxicity, as well as exclusive interaction with the target in the presence of other biomolecules. The future of HTS therefore lies in its application to the screening of cell-based assays, where these assays best represent in vivo conditions, but remain amenable to automation with reasonable throughput.

From Biochemical- to Cell-based Assays

Traditionally, compounds have been tested for target interaction using biochemical-based assays. These assays are simple, well-defined systems that use purified proteins and reagents. They have few variables and are quick to screen, making them easy to analyse using a high throughput or ultra high throughput format. Yet they have proven to be poor indicators of compound clinical efficacy because cross-reactivity, among other factors, is not accounted for. Consequently, compounds that appear to be effective based on primary and secondary screens progress into expensive late-stage drug testing, where they inevitably fail. The use of cellular rather than biochemical-based assays enables testing of compounds on more representative systems, where there are important features such as membranes and cellular metabolism. This means that in addition to testing compound-target interaction, cell-based assays function as physiological indicators, since the overall cell response can be monitored.

However, with physiological relevance comes the complexity of biological systems. Unlike simple biochemical assays, there are many variables present in cell-based assays. The compound must enter the cell and maintain structural integrity during the course of the experiment. It must also interact with the target in the presence of potentially interfering molecules, and often requires further testing to determine its mechanism of action. Cell-based assays tend to be more variable due to temperature, CO2, confluence and passage number sensitivity, and can involve long incubation steps that affect data reproducibility. All of these factors make cell-based assays difficult to automate and execute in a high throughput format.

Regardless, industry is increasingly turning to cell-based assays in the screening process because cellular assays are more informative, as they enable assessment of cell growth and viability, in addition to monitoring many compounds and targets at once. The challenge lies in making them reproducible and compatible with HTS.

Smaller Assays, Automated Systems

The goal of HTS is to replace manual, low-throughput assaying and screening techniques with miniaturized, automated robotic systems. Hence, to make an assay high throughput, it must be automation-friendly. While a common approach for cell-based assays is to combine personnel labour with robotics, more sophisticated systems are available.

Most high throughput methods typically use a plate reader for end-point scanning, in combination with a liquid-handling robot, for transferring reagents in a well-plate format. To increase throughput, the trend is to use a higher number well plates, where the difference between a 96-, 384-, 1,536- and 3,456-well plate means 300, 50, 10 and 2 ml per well, respectively. The result is reduced screening time, more data, and decreased compound usage due to the smaller volumes required. In terms of cell-based assays, this also means fewer cells per well.

Further miniaturization of cell-based assays has been explored with volumes as small as picolitres. Diaz-Quijada et al. have demonstrated formation of wells with two-picolitre volumes by growth of Hek293 cells on aluminum micropatterned poly(dimethylsiloxane) (PDMS) surfaces (Fig. 1).(3) Miniaturization to this extent would enable HTS of compounds in a physiologically relevant environment.

Microlitre-sized cell-based assays have been run effectively on a high throughput scale using systems such as Molecular Devices Corp.’s (Sunnyvale, CA) SpectraMax® M5 multi-detection microplate reader integrated with its SynchroMax™ ET plate-handling robot. Other examples include Caliper Life Sciences’ (Hopkinton, MA) Allegro workstations for HTS, and the Kalypsys Inc. (San Diego, CA) uHTS System.

Alternatively, systems employing microscopy can be used to attain higher throughput capability. This involves imaging cellular assays to capture the additional information they provide.

High-Content Screening: Developments in Microscopy

High-content screening is a high throughput approach applicable to cell-based systems that refers to the analysis of cellular assays using automated, image-based technology. This enables monitoring of multiple assay parameters, as well as capturing cellular information in one step — including cell shape and viability, target movement and interaction of the compound with other biomolecules.(4) The difference is a few data points per well using a typical end-point analysis versus thousands per well using a 2-D imager. Thus, a high-content approach can reduce the cost of cell-based screening because many cellular features can be tracked at once.

Imaging devices usually combine microscopy with fluorescence or luminescence detection to extract more content from cellular assays. Using many objectives in parallel, markers targeting unique components of the cell can be used to monitor multiple biomolecules simultaneously (Fig. 2). Two examples of such integrated screening systems are Molecular Devices’ FLIPR® (Fluorescent Image Plate Reader) and FLIPRTETRA™, which combine internal plate handling and liquid transfer with multi-wavelength detection.
More highly automated imagers include GE Healthcare’s (Piscataway, NJ) — formerly Amersham Biosciences Corp. — IN Cell Analyzer 3000, a high-resolution confocal imaging system capable of autofocus and online image analysis. Simultaneous three-colour imaging is made possible with the use of three laser lines and three charge-coupled device cameras. Although it does not match the FLIPR’s throughput or multi-channel fluorescence capability, the IN Cell Analyzer can image as few as 100 cells in a well, compared to the FLIPR resolution of only 10,000 cells. Higher resolution levels allow monitoring of cellular processes, including enzyme activation and apoptosis. This is useful for drug screening, as it enables detection of unwanted compound interactions within the cell, as well as monitoring the cell’s response to drug addition.

Other technologies permit single-cell imaging. This capability has many applications, such as detecting particular cells in a heterogeneous mixture, looking at transfection efficiences, and catching effective compounds with low potency or signal.(4)

One example is Cyntellect Inc.’s (San Diego, CA) LEAP™ (Laser Enabled Analysis and Processing), which combines optical scanning with real-time image analysis and multi-wavelength targeting, making it capable of analysing and manipulating individual cells. It provides high throughput analysis, where greater than 100,000 cells can be imaged per second using low magnification. Intracellular detail can then be detected using higher magnification.

Cells can also be monitored independently using flow cytometry. This technique is higher throughput than simple fluorescence microscopy systems, with the advantage that it can purify and recover cell populations. Using fluorescence scatter detection, cells can be sorted as quickly as one million per hour. Systems such as Novasite Pharmaceuticals Inc.’s (San Diego, CA) flow cytometer or Luminex Corp.’s (Austin, TX) Luminex®100™ IS System offer high throughput, single-cell detection.

Due to the optimal sensitivity of individual cell detection, these systems offer advantages in the drug-discovery process, since both fewer cells of interest and less compound are required.

Image-based systems for cellular assays will continue to become faster and more sensitive, and will integrate various technologies to extract more information. This development will be useful for testing compounds on more representative systems: 3-D cellular assays.

Tissue-based Screening: 3-D Cellular Assays

It is becoming apparent that cellular assays need improvement, since current in vitro systems are not always good predictors of compound clinical efficacy.(5) These systems use 2-D cultures, which have fewer cell-cell interactions than those found in tissue. In contrast,
3-D cell cultures can attain the structure and intercellular signalling mechanisms found in vivo,(5) providing more accurate information about compound-target interaction and overall cellular response. This is important for determining compound mechanism of action, and also early detection of compound toxicity and unwanted reactions.

It has been shown, for example, that 3-D liver cell cultures exhibit better maintenance of cell function than monolayers, because they better reflect the liver tissue organization.(6) For the purpose of drug discovery, 3-D liver cell cultures would therefore be useful in preliminary screening to determine compound metabolism and toxicity. Such early screening is important, since half of the drugs that pass toxicity tests in 2-D culture turn out to be toxic during animal testing, and even more fail in the clinical trials. Worse still is the fact that one-third of drugs are withdrawn from the market due to hepatotoxicity. Drug withdrawals cost companies millions of dollars — money that could have been saved by predicting toxicity from cell culture. A high throughput 3-D liver cell assay is therefore desirable for identifying compound efficacy early in the drug-development process.

Obstacles in Liver Cell Culture

The current practice is to use hepatocytes as cultured monolayers — or liver slices — to assess drug metabolism and toxicity. However, neither primary cells nor liver slices store well, and human samples are difficult to obtain.(7) In the case of primary cells, isolation from their source results in the down regulation of some enzymes and loss of functionality due to dedifferentiation in culture.(7,8) Altogether, maintenance of functionality and viability are the main problems with primary cells and limit their use in drug screening.
Viability can be increased by co-culture of two main cell types that interact in the liver: hepatocytes with nonparenchymal cells. It has also been found that functionality is better maintained in 3-D hepatocyte culture, where viability also increases.(8) Methods of 3-D liver cell culture with high throughput potential include scaffolding techniques and the formation of multi-cellular spheroids (MCS) (Fig. 3).

3-D Cell Culture Methods

The basis of scaffolding is the use of polymers or other materials as matrices for cell growth. These artificial supports can be 2- or 3-D structures — such as collagen gel sandwiches and sponge forms — and serve as platforms for cell growth by enhancing cell adherence and development into 3-D culture. Scaffold-based 3-D cultures could, in principle, be used in a high throughput format, but reproducibility from well to well may be difficult to achieve.

MCS have been investigated with hepatocytes and grown on surfaces such as poly(lactic acid), positively charged polystyrene-8 and in roller bottles or spinner flasks. The roller bottle and spinner flask prevent cell plating on the sides through constant bottle rotation and constant stirring, respectively. The surfaces provide platforms for initial cell aggregation, after which the aggregates release into solution and assemble into spheroids.(7)

The spheroids formed can be several hundred microns in diameter, and have a viability gradient from core to surface, with the more proliferative cells contained in the outermost layer. This results in a diffusion-limited model that could be useful for drug penetration and toxicity studies.(5) MCS take longer to culture than 2-D methods, but they show improved maintenance of hepatocyte function over monolayers. And while these methods can result in spheroids of variable size, the use of filter devices has enabled sorting of these cultures for a particular size and shape.

High Throughput Challenges

Grown under the same conditions and refined with size-exclusion filters, MCS can be made that are similar in structure and microenvironment,5 providing small models representative of in vivo conditions. MCS therefore have great potential as 3-D liver cell cultures that can be used in HTS for early determination of compound toxicity and possibly metabolism. To reach this goal, their method of culture must be standardized to achieve reproducibility, and limitations — such as longer culture times, consistency of spheroid composition and method of detection — must be addressed. Confocal microscopy is a technique that has been previously used to study MCS and will likely find future application in the screening of these assays.

While 3-D liver cell assays will increase the efficacy of the drug-development process, compound testing on animals and humans will still be essential. However, the more cellular assays approximate in vivo conditions, the more potential they have to increase early detection of inappropriate drug leads. This would then decrease the amount of late-stage drug testing required, saving both time and money.

Thanks to Sylvie Belanger and Dr. Angela Tonary for their help.

Jennifer Clarke, M.Sc. is a researcher at the National Research Council of Canada’s Steacie Institute for Molecular Sciences (NRC-SIMS) (Ottawa, ON). Also at NRC-SIMS, Helene Letourneau is a communications officer, and John Pezacki, PhD is an assistant research officer with the biomolecular imaging and sensing group.

References
(1) Hodgson, J. 2001. ADMET-turning chemicals into drugs. Nat. Biotech. 19: 722.

(2)     Haberman, A. et al. 2001. Cellular Assays Help Ease Target-Validation Bottleneck. Cambridge Healthtech Institute, Molecular Med Monthly. Issue 9, August.
(3)     Diaz-Quijada, G.A. et al. Picoliter Wells from Selective Growth of HEK293 Cells on Chemically Modified Surfaces. Submitted to Biotechnology Progress. December, 2004.
(4)     Shah, A. 2004. In Automated Microscopy, Image is Everything. Drug Discovery & Development. January.
(5)     Kunz-Schughart, L.A. et al. 2004. The Use of 3-D Cultures for High-Throughput Screening: The Multicellular Spheroid Model. Journal of Biomolecular Screening 9(4): 273.
(6)     Landry, J. et al. 1985. Spheroidal Aggregate Culture of Rat Liver Cells: Histotypic Reorganization, Biomatrix Deposition, and Maintenance of Functional Activities. The Journal of Cell Biology 101(September): 914.
(7)     Farr-Jones, S. and S. Fox. 2003. High-Throughput ADMET Screening: Improving the Efficiency of Drug Discovery. Drug Discovery and Design: Decision Resources, Inc. April 3: 4-1.
(8)     Riccalton-Banks, L. et al. 2003. Long-Term Culture of Functional Liver Tissue: Three-Dimensional Coculture of Primary Hepatocytes and Stellate Cells. Tissue Engineering 9(3): 401.