Cell-based assays are poised to assume an increasingly important role in drug screening and discovery. It is becoming apparent that simple binding assays and in vitro experiments are grossly insufficient for extrapolating the efficacy and side-effects of potential drugs.

The explosion of information arising from completion of the human genome has led to a remarkable increase in the number of potential drug targets and avenues for disease treatment, as well as a new degree of appreciation for the complexity of biological systems. While current methods of high-throughput screening are inadequate in this context, whole-organism studies are too uncontrolled, complex, tedious and expensive to be applicable to the large universe of drug leads and targets that we have today.

The natural evolution beyond binding assays is to screen for the actions of drugs on biological cells, the functional units in living systems. Information obtained through such cell-based studies provides a direct insight into the physiological response of cells and tissues to pharmacological agents. The toxic side-effects of potential drugs, a key factor that underlies the costly nature of drug development, can only be glimpsed through functional studies, animal models and later in clinical trials.

The attrition rate of drugs in the later stages of drug development and in safety testing will be greatly reduced with the high quality data obtained in the context of living cells and tissues. The utility of cellbased assays is reflected in the rapid increase in the number of such screens that are carried out in pharmaceutical and biotechnology companies, where cell-based assays now represents almost half of all medium and high-throughput screens.

Future cell-based assays will undoubtedly go beyond single cells, in which complex cellular networks or even micro-engineered tissues and organs may be used in place of animal models. While the benefit of using living cells and tissues to screen for drugs may be evident, the development of a technology platform by which these screens can be carried out in a high-throughput fashion is non-trivial.

Categorically, the technology may be divided into two classes: methods to readout the response of the cell or group of cells and techniques to control the complex extracellular and intracellular environment of the cell. The reward, if such comprehensive platforms can be developed, is great. In the pharmaceutical industry, the increasingly severe bottlenecks and limitations of current technologies will fuel strong demand for these new generations of drug discovery tools.

Readout technologies

Optical methods based on fluorescence detection

The dominant optical method in cell-based assays is undoubtedly fluorescence, especially with the development and optimisation of new and better fluorophores, which range from proteins (green fluorescent proteins) to small organic molecules to nanometre-sized inorganic semiconductors (Quantum Dots). Fluorescence-based detection has become especially powerful given the wide range of available indicators to important biological signalling proteins and second messengers (notably calcium) and its inherent high sensitivity, dynamic range, ease of implementation and its amenability to high-throughput screening.

Examples of hardware systems that have been developed to address this area include well-plate formats such as the FLIPR instrument from Molecular Devices, Flow Cytometry methods such as the Fluorescence Activated Cell Sorters (FACS) from BD Biosciences and Beckman Coulter, and fluorescence microscopy-based imaging systems such as high resolution laser scanning confocal and epi fluorescence microscopes that are specifically designed for use with well-plates in a highthroughput format.

More important than the hardware are perhaps the reagents that have been, and are being, developed that report general cellular activities or probe for specific signalling events. One of the longest and most broadly used reagents may be the calcium indicator, which reports the activity of G-coupled protein receptors that form an important class of drug targets. Roughly 28% of drug targets today are G-coupled proteins.

In addition to calcium indicators, green fluorescent proteins have gained prominence over the past years, owing to the ease by which they can be manipulated genetically and co-expressed with the desired proteins. Besides fluorescent probes that report generally on the state of the cell from signalling to cell viability and apoptosis, a host of more specific assays are also being developed. The availability of such reagents together with the ease and rapidity of making fluorescence measurements will ensure it will be a mainstay in highthroughput screening.

Electrical methods based on patch clamp

The benefit of functional cell-based assay in general and electrical measurements in particular may be exemplified by the use of patch-clamp recordings on single cells for studying the action of potential drugs that interact with ion channels, a group of membrane proteins that constitute an important and large class of current drug targets. Ion channels play major roles in chronic and acute disorders such as cholera, cystic fibrosis, arrhythmia, melancholy, schizophrenia, diabetes, epilepsy, hypertension and several neurodegenerative diseases. In fact, 15% of the top 100 best selling drugs on the market today target ion channels.

One particular ion channel that has received much attention lately is the so-called hERG-channel, a voltage-gated potassium channel involved in the recharging of the ventricles of the heart so they can contract again. The hERG-channel has been linked to the prolonged QT syndrome and has been identified as a hot topic for safety testing since drugs unintentionally have shown to interact with the hERG-channel and causing arrhythmias that can be lethal in some instances.

Several drugs have been withdrawn from the market or in late clinical trials because they affect the hERG-channel, known to be quite promiscuous. Soon all new drugs entering the market will have to go through safety testing including assays for QT-prolonging effects according to upcoming FDA regulations. It is clear the demand for functional cell-based ion-channel screening assays will rapidly increase in the coming years.

Despite the availability of membrane-potentialsensitive dyes and the possibility of making highthroughput fluorescence measurements, patchclamp recording has remained the golden standard of assessing ion-channel activity because these optical methods as well as radiotracer efflux technologies yield indirect and low-information content data. Unfortunately, patch clamp recordings suffer severely from throughput restrictions since it is working at the single-cell level and requires skilled operators. At best using traditional methods and equipment, only up to 10 cells can be analysed per day. Naturally, this throughput challenge has led to intense efforts by both established and startup companies to implement patch clamp experiments in a high-throughput format.

The natural extension of single-cell patch clamping for high-throughput applications is parallelisation and automation. Figure 1 shows one commonly pursued scheme, known as planar patch clamp, in which the opening of the patch pipette is replaced by a hole fabricated in a planar substrate.

Indeed, this approach is pursued by a number of companies, including Axon Instrument, Molecular Devices, Sophion Bioscience A/S, Cytocentrics CCS GmbH, Flyion GmbH, Nanion Technologies GmbH and Cellectricon AB. However, only the Ionworks station from Molecular Devices and the PatchXpress from Axon Instruments are commercially available on the market currently. This general strategy of parallelisation makes logical sense, but there are a number of unsolved challenges associated with the quality, yield and cost of the actual measurement. These challenges originate from the difficulty in forming and maintaining high-resistance electrical seals between cells and microfabricated substrates.

Another strategy to achieve throughput is to exploit the rapid kinetics of ion channels. Because many types of ion channels can be activated and deactivated rapidly (milliseconds), a single patchclamped cell may be used to scan for many potential ligands in fast succession, provided the solution around the patch-clamped cell can be controllably exchanged in a similarly rapid manner. This strategy is implemented in the Dynaflow chip from Cellectricon (see Figure 2).

The conceptual breakthrough in this technique resides in taking full advantage of the peculiar behaviour of aqueous solutions flowing in micrometer-scale channel structures to exert complete control over the solution environment around the patch-clamped cell. We will describe this method in more detail in later sections, since this technique does not represent a new readout technology but a new way of perturbing the cell so that traditional readout technologies (in this case patch clamp) can be carried out much more efficiently.

Cellular imaging

In addition to optical and electrical readout methods, the morphology, or behaviour, of a cell may also be used to interpret the cellular response, which may be cell migration and growth or apoptosis and necrosis. Most common approaches are based on high-resolution microscopy that combines several modes of imaging, such as Nomarski, multi-colour fluorescence, spectral and lifetime imaging.

The core of these systems is often sophisticated image analysis, tracking and automation software to increase throughput and to facilitate interpretation of the acquired images. The growth in this area is exemplified by the host of companies that have launched high-throughput microscopy instruments that are designed to work with large arrays of well plates, such as the imaging stations from Amersham and Axon Instruments.

Perturbation methods

Cell-surface receptors

To successfully carry out cell-based assays, accurate readout of the cellular response by itself is insufficient. The perturbation that is delivered to the cell must also be accurately, controllably and reproducibly administered. In the case of delivering stimuli that act on the proteins at the cell surface, a number of straightforward methods are available. In the context of high-throughput and automation, the principal technology relies on the robotic handling and pipetting into well plates, such as the liquid-handling platform available from Tecan.

Although such platforms automates manual pipetting, washing and liquid dispensing, they in fact have very poor dynamic control over the immediate solution environment and the concentration of potential drugs around the cell, owing to the inherent slowness of diffusion at these lengthscales.

Attempts to control dynamically the immediate solutions and the concentration of bioactive molecules around cells are traditionally based on perfusion using micropipettes, some examples of which include instruments from Biologic Science Instruments, ALA Scientific Instruments Inc, and Automate Scientific Inc. Based on completely different principles and physical mechanisms, Cellectricon has recently introduced a microfluidics platform for controlling precisely with great dynamic control (milliseconds) the concentration and types of molecules to which a cell is exposed.

Figure 2 illustrates this Dynaflow platform, in which collimated fluid streams exiting a series of microchannels into an open volume act as virtual containers with the desired molecules at extremely well-controlled concentrations. The characteristics of these virtual containers permit the complete control over the solution environment and the molecules to which the cell is exposed.

One application area that this platform has made a big impact is in ion-channel screening and characterisation, and this technology applies well to several areas in the drug discovery process such as target validation, later phases of lead identification, lead optimisation, preclinical studies and safety assessment. The utility of this platform demonstrates the importance and benefit of the ability to control the environment to which the cell-surface receptors are in contact.

Intracellular receptors

To control the intracellular environment or to introduce potential drug molecules into the cell, these molecules must pass through the membrane barrier. There is a wide range of approaches to achieve this, from physical methods that use electric field, light, or temperature to chemical strategies that use detergents, lipids, or other chemicals (eg polyethylene glycol) to biological techniques that use viruses. Many of these methods, however, do not possess the qualities that make them amendable to automation and parallelisation, which is needed in highthroughput screening.

Electrical methods

The approach that best lends itself to parallelisation in a high-throughput format is perhaps electroporation (see Figure 3), which is one of the most commonly used and well-documented physical method for introducing molecules into cells, in which a transient electric field is applied across the cell to permeabilise the lipid membrane.

Although other physical methods exist and may also be widely used, such as microinjection or the use of gene guns, they either have low throughput or introduce excessive damage to the cells. The chemical methods usually are either too uncontrolled (eg polyethylene glycol) or lack sufficient generality (eg lipofection). In the context of these constraints, parallel electroporation seems currently to be the most promising platform for performing high-throughput applications related to drug screening, which range from binding studies with intracellular receptors, to screening experiments on intracellular enzymes such as phosphatases and kinases, to parallel transfection.

Currently, there is no technology platform or available instrument for introducing cell impermeant molecules into the cell in a general and highthroughput manner designed for cell-based assays. But given the presence of large numbers of important and unexplored targets within the cell, this area is sure to take centre stage in the near future.

Cell culture

Since the underlying material in functional cellbased assays is cells, the quality of the assay and the obtained information is intimately tied to the quality of the cell culture, the available cell lines, and the expression of the desired proteins. Such requirements will continuously drive the need for cell culture automation and the development of new cell lines or expression systems, if they are not already available. Traditionally, such automated cell culture systems have relied on a series of robotic instruments that will automate the different subtasks in cell culture. Ideally, such automated systems will not only increase efficiency and the ability to maintain hundreds of cell lines and well plates, but also result in reproducibility and accuracy of the process.

Technological opportunities

One of the most promising approaches to addressing the challenging issues in functional cell-based assays is by interfacing living cells into well-controlled environments created using techniques of micro- and nano-fabrication. To excel in the development of functional cell-based assays requires both a deep understanding of cell biology and an intimate working knowledge with cutting-edge technologies in fabrication and material science, technologies that are essential for interfacing with cells and for controlling the micro-environment with which cells interact.

Cells receive complex signals from the solution environment that surrounds them, often in the form of soluble biochemicals, such as proteins, peptides, small molecules and ions. This complex chemical landscape has distinct spatial patterns that evolve over time and which has direct impact on cellular physiology. To represent and to control dynamically such chemical landscape, for example, would require sophisticated methods in micro-fluidics and fabrication. These micro-technologies have already made a big impact in electronics and computing and are poised to affect an equally big paradigm shift in cell biology.

The sizes of biological cells, which are in the micrometer regime, are ideal for integration with micron-sized structures and devices created by micro-fabrication (see Figure 4).

This rich and emerging area of silicon/biology interface and functional cell-based assays will provide ample opportunities for growth and exploration for decades to come, to address important issues in both basic research and in the pursuit of practical treatments for diseases. DDW

—

Daniel Chiu is Assistant Professor in Chemistry at the University of Washington, Seattle. He did his post-doctorate in the group of Professor George M. Whitesides at Harvard and his PhD in the group of Professor Richard N. Zare at Stanford. He has authored and co-authored more than 50 scientific publications. He is a founder of Cellectricon.

Owe Orwar is Professor in Physical Chemistry at Chalmers, Sweden. He did his post-doctorate in the group of Professor Richard N. Zare at Stanford. He is also the Director of Bioelectronics and Bio computing program at Chalmers (MC2) as well as a founder of Cellectricon. He has authored and co-authored more than 75 scientific publications. In 1997 he was awarded the Roche Affinity Young Investigators Award and is a Research fellow of Swedish Royal Academy of Sciences.