Examination of libraries of biological information is increasingly seen as a potent tool in drug discovery. These collections may be genes, proteins, antibodies, RNAs, or chemical entities such as drug candidates. A decade ago, when the human genome and proteome were elaborated, it was predicted that these vast collections of data would offer up a wealth of innovative pharmaceutical targets. However, initial attempts at large-scale screening of gene and protein libraries faltered, and attempts to search out new candidates were unsuccessful.
Now the picture is changing, companies and individual investigators are taking advantage of elegant new technology and a much more sophisticated understanding of the manner in which target identification must be addressed. At a recent meeting of the “Society of Biomolecular Sciences in Montreal,” numerous presentations dealt with new approaches to library screening and the enabling instrumentation that makes these protocols possible.
An important application of new library screening approaches was discussed by Julian Davies, Ph.D., University of British Columbia, who considered the present dearth of new antimicrobial antibiotics. According to Dr. Davies, in the past 20 years the approval rate of new antibiotics has fallen precipitously, from 16 to 3 per year. Pharma companies have virtually abandoned the search for new antibiotics as costs have risen and the challenges prove ever more daunting. Dr. Davies pointed out that the complexity of the bacterial interactome, a web of protein-protein relationships, makes the detection of viable targets by straightforward screening for over- or underexpression in gene or protein libraries problematic.
While there has been a significant increase in antibiotic-resistant pathogens during this period, there have been few new treatment options. High levels of resistance in serious epidemic diseases such as drug-resistant tuberculosis, cholera, and parasitic and viral infections are now rampant throughout the world. Given the immense cost of discovery and regulatory uncertainties, large pharmaceutical companies are hesitant to commit to antibiotic discovery programs. The result is tens of millions of unwarranted deaths per year.
But even as the large pharmas have opted out of antibiotic discovery, Dr. Davies argues that unique opportunities are opening up for small biotechs. New instrumentation provides the ability to screen libraries of metabolites produced by soil-dwelling organisms such as Streptomycetes, which have the genetic capacity to produce many complex small molecules. While many of these compounds are produced by all living species, they have been mostly examined in microbes (i.e., bacteria, fungi) and plants. Up to 10,000 different species of Streptomyces have been identified with the capacity to produce a quarter of a million small molecules.
“In fact, Actinobacteria may produce tens of millions of low molecular weight bioactive compounds, and the chemical space covered by these compounds is immense,” Dr. Davies stated. While their natural function is to act as the alphabet of microbial communication, appropriate screening of libraries of these compounds will doubtless yield many antibiotic candidates.
Dr. Davies feels strongly that screening should, whenever possible, involve whole-cell systems, rather than screening libraries of compounds against isolated targets. “Back in the early days of genomics, companies screened huge libraries of chemicals against protein targets and they came up with many positive hits, but the compounds didn’t function as therapeutics,” he stated. “This was because they didn’t get into cells, so they were ineffective.
“A second problem with screening directly against targets is that they are always networked,” he continued. “This means that the response in vivo may be completely different from the isolated protein target, so once again a positive hit using isolated target screening will be meaningless in the context of the intact organism.”
Cellular Systems Biology
Screening against intact organisms is a critical pathway to drug discovery that has been recognized by a number of companies developing technology for screening libraries of compounds against whole-cell systems. Cellumen (www.cellumen.com) markets reagents to measure and manipulate genes and proteins, as well as coding and noncoding RNA.
The company’s approach uses a panel of fluorescent reagents that are tied to a range of cellular responses. The first category includes labels that allow localization and quantification of biomolecules such as organic dyes with reactive groups. Venerable old standards include fluorescein and rhodamine as well as more recently discovered cyanine and alexa dyes. The second category is noncovalent dyes with high affinity for the target such as DAPI, Hoechst, and propidium iodide, and the final category is the class of indicator dyes sensitive to a localized change within the cell such as pH or calcium levels. These compounds, when properly applied, are nontoxic to the cells, and changes in their emission will provide a measure of cellular response to members of a library of potential therapeutic molecules.
Another service offered by Cellumen, Cellular Models of Disease, is designed to speed drugs more rapidly toward clinical trials. This strategy involves the use of cells, assays, and profiles that closely recreate the disease state under consideration. In this fashion, large libraries of compounds can be rapidly screened, and candidates that demonstrate toxic fingerprints can be eliminated long before initiation of costly clinical trials.
“We employ high-content screening platforms with our panels of commercially available reagents to create targeted cellular systems biology assays. This provides a strategy to define pathways, identify disease biomarkers, and perform specially designed screens for clients,” reported Cellumen’s vp of business development, Justin Lancaster, Ph.D.
The company also offers an analysis of client compound libraries, the CellCiphr™ Cytotoxicity Profiling Services, and a panel of reagents, the CellCiphr Cytotoxocity Reagents, for use on high-content screening platforms. The CellCiphr technology includes positional biosensor pairs for purposes such as measuring the interaction between p53 and hdm2 in intact cells.
Kinase Library Screening
The kinome is a major focus in drug discovery and, once again, the approach of using whole cells to evaluate large compound libraries is a primary strategy. The term kinome refers to the 500 or so protein kinases in the human proteome that are known to be responsible for a complex array of overlapping cellular regulatory functions. As such, they represent appealing targets for drugs that may hold therapeutic potential in the treatment of cancer, cardiovascular disease, and a host of other disorders. In fact, a number of drugs already on the market work through inhibition of protein kinases such as Gleevec, so there is a demand for rapid screening systems for kinase inhibitors.
Whole cells are a much more physiologically relevant model than traditional cell-free approaches, which cannot provide information on specificity of compounds, membrane permeability, or their cellular toxicity. TGR BioSciences’ (www.tgr-biosciences.com.au) has developed an approach that combines its SureFire™ kinase assay kits with PerkinElmer’s (www.perkinelmer.com) AlphaScreen® assay technology.
The new technology allows the high-throughput detection of phosphorylated cellular proteins in a homogeneous assay format. One such product, the SureFire ERK kit, measures phosphorylation of the extracellular signal-related kinases or EKR protein, enabling researchers to easily screen GPCR activation in whole cells.
The AlphaScreen SureFire ERK assay offers a method for primary and secondary screening of GPCR targets, including those not optimally coupled through the calcium or cAMP pathways. Unlike other proximity-based technologies, AlphaScreen beads can be up to 200 nm apart, allowing for the detection of simple to complex biological interactions and substrates of almost any size.
Like all proximity-based assays, AlphaScreen depends on bringing together two molecules (in this case two distinct types of beads) whose proximity causes the production of a signal. The SureFire assay employs phospho- and non-phospho antibodies, specific to the protein of interest, which coat the beads. Only the phosphorylated protein interacts with both antibodies and brings the two sets of beads together, thus producing a response.
Michael Crouch, Ph.D., director of screening technologies at TGR, discussed the use of AlphaScreen technology as it applied to several projects. Dr. Crouch and his associates have demonstrated that the SureFire system is an effective screening tool for libraries of potential inhibitors of inflammatory mediators such as tumor necrosis factor. “We developed the SureFire technology and then partnered with PerkinElmer,” Dr. Crouch explained. “At present, ours is the only homogeneous high-throughput system for detecting cell-based phosphorylation.”
As Dr. Couch explained, TGR produces 12 kits for detection of inhibitors of various kinase pathways, and another dozen are in the works for release in the coming year. This will enable the screening of hundreds of thousands of drug candidates, and the use of the whole-cell assay allows compounds with toxic side effects to be eliminated early on in the screening process.
“Previously, screening was truly laborious, involving immunoblotting, ELISA, flow cytometry, or cellular imaging, technologies that did not measure the entry of the candidate into cells,” Dr. Crouch continued. “Now drug companies can narrow down a million compounds to find the one satisfactory contender.”
Microscope-based systems have long been employed for monitoring cellular responses to inhibitors through the use of coupled fluors. However, when carried out manually, this approach is overwhelmed in a primary library screening protocol that may involve 50,000 compounds or more. Blueshift Biotechnologies (www.blueshiftbiotech.com) provides an automated alternative for cellular imaging, IsoCyte™, that can scan plates in less than eight minutes per plate at any well density. Blueshift’s cellular and array-based assays provide instrumentation that images the entire well, cell by cell, at true high-throughput screening rates.
Conventional plate readers measure the total signal for the whole well, or less accurately, sample only a small area where the signal may not be uniformly distributed. Microscopes view only the field-of-view allowed by their objective lens, so they must focus and take multiple exposures at multiple filter wheel positions. The IsoCyte images the entire well, regardless of density, at micron resolution in four colors simultaneously.
“The IsoCyte system allows a simple, one-step protocol for GCPR-ligand binding assays. Imaging in the fluorescence anisotropy domain allows discrimination of bound and free label as well as providing a preferred method for detecting FRET in live cell protein-protein interaction screens,” said Chris Shumate, Ph.D., vp of business development.
Fragment Screening of Drug Targets
“Synthesizing sufficient proteins for refined analysis is our greatest bottleneck,” says Michael Hennig, Ph.D., of Roche Pharmaceuticals (www.roche.com). Exploration of drug targets by compound screening of large and random libraries is a key step in the early phase of a drug discovery project. An alternative approach is the use of focused libraries containing substances that are selected based on target or compound-specific properties.
Dr. Hennig recounted his experiences with fragment-based focused screening, in which thousands of candidates having molecular weights less than 300 kD were examined. The selection of the 2,000 compounds making up the fragment library is highly rational, based on the structural properties and chemical and biological similarities with the natural compounds that they displace from the drug target. Because of their small size, interaction with targets will display low affinity, and for this reason, Dr. Hennig needed a sensitive assay system.
In addition to the molecular weight criterion, Dr. Hennig and his colleagues placed a number of other restrictions on the fragments, including easy access, high solubility, availability of the compound in powder form, ease of synthesis, and excellent compound purity. Moreover, crystal structure analysis to define molecular interactions is required to aid in rapidly generating compounds with optimized biophysical and binding properties. Due to the use of biophysical assays, including SPR and protein crystallography, the application of this method is restricted to proteins that can be investigated in solution and produced in milligram quantities, currently no GPCRs.
The screening of compounds uses the SPR system available through Biacore (www.biacore.com), whose A100 instrument has the capability to screen about 1,000 compounds per day, the company reports. The principle behind the A100 is based on detecting changes in the mass of molecules in the aqueous layer bathing a gold sensor chip surface by measuring changes in refractive index. Thus the binding of a ligand to a protein molecule can be followed without labeling, and accurate Kon and Koff values are readily obtained. “Many companies use screening platforms, but ours is unique in that we use the Biacore system as a filter,” Dr. Hennig added.
Dr. Hennig described Roche’s program to screen compounds binding to the beta amyloid cleaving enzyme (BACE), a target in Alzheimer’s disease. It is believed that the release of a portion of the beta amyloid protein from the cell surface produces the raw material that will congregate in the plaques that clog neurons and prevent the proper transmission of signals in the brain.
According to this hypothesis, drugs that block the asp-protease BACE might prevent the initiation and progression of Alzheimer’s disease. In addition to the other properties mentioned, a viable candidate drug would have to cross the blood-brain barrier in order to have access to the cells of the central nervous system.
Dr. Hennig and his colleagues have screened their fragment library against BACE and have identified a number of candidates that bind to the enzyme. All but one of the positive fragments bound to the S1 pocket of the enzyme, which appears to be a key site for ligand binding. Based on the screening results and structural information, the tyramine fragment binding to the S1 pocket was selected for chemistry efforts and ortho-substituted molecules were synthesized exhibiting greatly improved Kd.
Dr. Hennig concluded the discussion with a cautionary note against irrational exuberance. “Library fragment screening with Biacore is one important step in drug development,” he said. “But from this point, we still have a long way to go on the path to the multidimensional optimization of the candidates.”
While fragment library screening represents a focused approach to new drug development, it will not provide new drugs in the course of a single experimental run. Rather it allows the discovery of ligand candidates that can serve as a starting point on the long road to a viable therapeutic modality.
In the last few years, our understanding of the basic biology has grown so that screening is a much more targeted and sophisticated process. By selecting particularly fertile areas of the genome such as the kinome and using rapid and economical screening technologies, small biotech companies are able to address challenges that previously only big pharma, with its immense resources, could manage. This means that less lucrative targets are fair game for drug development, and with the FDA’s fast-tracking authority, the search for therapies for incurable but infrequent diseases is now a feasible goal.