samedan logo
home > ebr > spring 2012 > screening libraries
European Biopharmaceutical Review

Screening Libraries

With the innovation and miniaturisation of various assays, screening libraries have become a cornerstone of modern drug discovery. The process of elucidating a new drug in order to ameliorate a particular disease usually involves a screen of candidate structures. Whether this involves a classic high throughput screen (HTS) of a synthetic small molecule library, a natural product library, focused library or a fragment-based library, each of these have resulted in programmes underpinning the discovery of clinical candidates. Although screens and libraries are a vital element of modern drug discovery, some require selective approaches with some assays being more amenable depending on the library type. The following technical review examines various types of libraries’ advantages and pitfalls, as well as the strategies for executing successful screens.

Modern drug discovery begins with the premise that certain enzymes, receptors, transporters and so on are involved in a particular disease state, and as such represent attractive ‘targets’ for pharmacologic intervention. The obvious question then becomes, what is the best way to elucidate novel agents that will become effective new therapies? Historically, traditional biochemical and pharmacologic methods were performed in no less than a 1mL reaction volume, requiring 5mg to 10mg of test substance, and these types of assays were often performed in individual glass tubes. This approach was limiting, in that only about 50 compounds could be screened per day. Reagents, such as receptors and enzymes, were also a limiting factor for throughput. With the advent of molecular biology and advances in detection parameters, however, the ability to screen larger sets of compounds has become possible.

Common assays in various screening paradigms include two major groups, biochemical assays and cell-based assays. The biochemical assays include classic radiolabelled type assays where conversion of a labelled substrate or binding ligand can be easily measured in a 96- or 384-well format. Fluorescence polarisation (FP) is a measurement of a fl uorescent molecule’s rotation from its excitation until its emission, and changes in the rotation by binding to a larger molecule results in a change in the amount of light emitted in a measured plane. With homogeneous time resolved fluorescence (HTRF), a signal is generated through fluorescent resonance energy transfer (FRET) between a donor and an acceptor molecule when in close proximity to each other. Interference is dramatically reduced by the dualwavelength detection, and the final signal is proportional to the extent of product formation/interaction. The HTRF assay is sensitive and robust so that it can be utilised in 384- and 1536-well plate formats.

Cell-based assays for HTS can be grouped under various types of assay measurements. Second messenger assays monitor signal transduction from activated cell-surface receptors. Second messenger assays typically measure fast, transient fluorescent signals that occur in a matter of seconds or milliseconds. Reporter gene assays monitor cellular responses at transcription/translation level. It indicates the presence or absence of a gene product that in turn points toward changes in a particular signal transduction pathway of interest. Quantification in these types of assays is carried out by measuring the enzymatic activity of an introduced reporter. Such reporters can either be transient in nature or stably expressed. Cell proliferation assays assess overall-growth or no-growth responses of the cell to a compound of interest. These assays are quick and easy to be employed for automation, and it should be noted that these assays are often performed as a counter screen.

Synthetic Small-Molecule Libraries

With the advent of high throughput screening (HTS), the ability to evaluate large numbers of compounds in either biochemical or cell-based assays has become a reality. One review describes the advent of small molecule screening at Pfizer Inc in the late 1980s (1). The idea soon took hold that if screening thousands of compounds per week was good, screening hundreds of thousands would be even better. This led to ultrahigh throughput screening (uHTS), where additional miniaturisation and high-speed automation resulted in the ability to screen millions of compounds (2).

Most recently, advances in drop-based micro-fluidics have allowed thousands of compounds to be screened per second. One example, using aqueous drops dispersed in oil as picolitre-volume reaction vessels, are able to screen 100 million compounds using less than 150 microlitres of reagent (3). However, the importance of screening ever greater numbers of compounds must be questioned. The idea is that quality of the compounds is more important than quantity. In other words, are the compounds being screened ‘drug-like’ in nature? This is in addition to the necessary robotic and liquid handling hardware necessary to screen large numbers of compounds.

As a result, some research groups and companies have focused on creating libraries of a few hundred thousand compounds that are both drug-like and chemically diverse, providing the hits resulting from a screen a better chance of being tractable for further medicinal chemistry efforts. Recently, a review examined screening libraries from 29 sources reporting on various drug-like properties, such as Lipinski characteristics, as well as diversity and exclusivity for each library (4). Further refinement based on scaffold and cluster analysis have resulted in an even more drug-like library where approximately 90 per cent of the compounds in the library satisfy all but one of the Lipinski rules (5). These compounds were analysed based on scaffolds and cluster analysis, resulting in the rapid development of structure activity relationships (SAR) based on the hits derived from a screen. Caveats to manually screening compound libraries include solubility issues, ‘frequent hitters’ or aggregators, and positional effects. Care must be taken to ensure that inter- and intraassay variability is kept to a minimum.

Virtual and Focused Libraries

Virtual screening involves the use of various computational algorithms to search large libraries (millions of samples) to identify structures that may interact with a single target. These types of screening can be generally categorised into two main groups. Chemogenomic design involves creating a model of the target, based on as little information as peptide sequence homology or as much as an x-ray structure, and then docking of virtual compounds to the target and applying a scoring function. The scoring function is an estimate of the probability that a particular compound will bind to the target with high affinity.

The other approach is ligand-based virtual screening. This method is based on searching for molecules with similarities to compounds known to interact with the target (ligands, substrates and inhibitors). Compounds selected based on the above methods become small focused libraries. One review presents a useful example of a focused-based kinase library and another provides an up-to-date review on the subject (6,7). It should be noted that limitations to virtual screening can be directly tied to the quality of the homology models and in addition may not take into account ligand flexibility, solvent effects which lead to the generation of negative results.

Fragment-Based Libraries

As another means of increasing the chemical diversity for potential pharmacologic agents, fragment based drug design (FBDD) was developed during the late 1990s. It is based on identifying small chemical fragments, which initially bind only weakly to the chosen target, and then combining disparate fragments or optimising a fragment to develop a lead with higher affinity. FBDD can be compared with HTS, where libraries with up to millions of compounds with molecular weights of around 500 are screened, and nanomolar binding affinities are the goal.

In contrast, in the early phases of FBDD, libraries of a few thousand compounds with molecular weights often less than 200 are screened mostly by NMR or x-ray crystallography, with high micromolar to millimolar affi nities being considered useful. In addition, while HTS libraries are often based on Lipinski’s rule of fi ve, FBDD screening sets mostly follow the rule of three where MW of the compounds is less than 300, and cLogP, hydrogen bond donors and acceptors are all less than three (8). Interestingly, the types of assays amenable for FBDD are primarily biophysical in nature. Two reviews demonstrate SPR and x-ray crystallography/NMR method based FBDD screening respectively (9,10).

Natural Products

Some of the earliest screening that has taken place in the pharmaceutical industry involved natural products. Natural products refer to compounds that arise from plant and animal tissues, including toxins, marine organisms and microorganisms (see Figure 2). Frequently, samples from the natural material are extracted based on polarity. These extracts can then be used in various pharmacologic assays. One caveat is that the extracts contain a multitude of organic compounds of diverse sizes and structural classes, including peptides, and as such effort is made to identify single active components from an active (hit) extract. Techniques utilised in combination include UV/ NMR/MS to parse the identifi cation of novel compounds or readily compare an isolated active compound to a database of known structures for identifi cation.

Natural product hits/leads often have complex structures and atypical properties that daunt medicinal chemists. However, in many cases the compounds can be prepared through fermentation technologies which can also deliver biosynthetic intermediates as feedstock for synthesis, delivering semi-synthetic analogues to a lead optimisation programme.

In summary, a robust assay is one that is not affected by minor variations in various reagents and laboratory conditions. There are typically three parameters used to establish a wellvalidated assay that may be used in any type of screen. The fi rst is the assay or signal window. One review defi nes the signal window in the following equation (10):

SW =   ms-mb – 3(ss+sb) 

Here, ms and mb are the averages of the positive and negative control values respectively and ss and sb are the standard deviations for the same signals. Using this equation, a recommended SW is greater than 2. Another simple parameter utilised is the Z’ factor or Z’ shown below:

Z' =   1- 3(ss +sb) 
           (ms -mb)

This is a statistical tool to measure if a particular assay is amenable to screen compounds (11). A Z’ equal to or greater than 0.5 is considered acceptable for a screening campaign. It should also be noted that the Z’ can be improved by increasing the signal window and by decreasing the standard deviation of the controls. With a validated assay, on hand investigators can then select the screening strategy most amenable, based on the criteria listed in Table 1.

It should be noted that, based on the target or target classes, some libraries are more suited than others for various targets or target classes. As drug discovery moves forward into the next decade, however, it is expected that fruitful therapies will be discerned using each of the library types mentioned and indeed, more than one type of library will be screened for a particular target.

Read full article from PDF >>

Rate this article You must be a member of the site to make a vote.  
Average rating:

There are no comments in regards to this article.

Jeffrey C Webster is a Senior Research Scientist in the IVB Department at AMRI in Bothell, Washington. In this capacity, Jeffrey is a project leader for a number of different client-based projects involved in therapeutic drug discovery. Previously, he was employed by TransTech Pharma, High Point, NC where he worked at increasing levels of responsibility to become Director, Department of Biology and Chair of the Joint Research Committee for TransTech Pharma’s collaboration with Pfizer in the area of Alzheimer’s disease. Prior to TransTech, Jeffrey worked for DuPont Pharmaceuticals/Bristol-Myers Squibb as a Senior Research Scientist in Inflammatory Diseases Research. Jeffrey earned his PhD in Biochemistry from North Carolina State University and received post-doctoral training in the Department of Physiology at the University of North Carolina. Jeffrey was also a research fellow with the National Institutes of Health. He is the author of numerous papers and several patents.
Jeffrey Webster at AMRI
Print this page
Send to a friend
Privacy statement
News and Press Releases

First cellularized collagen membrane implant to regenerate heart tissue in patients with ischemic heart disease

• 3P Biopharmaceuticals and Viscofan go ahead with this innovative project of regenerative medicine for cardiac use • Beginning of Phase I clinical trial in a group of 10 patients
More info >>

White Papers

Financial implications of off-site visits in clinical research, perception versus reality

Illingworth Research

The concept of patient centricity is still relatively new within clinical trials. For too long, the most important person in the trial has been overlooked… the patient. Here, we explore the benefits of conducting a more patient-centric trial for the patient, site and sponsor through mobile, or ‘off-site’ visits. These advantages are much greater than purely financial, although the financial implications of running a more patient-focused clinical trial may previously have deterred some from this approach. Home nursing in clinical trials continues to be thought of as a high value “premium” service, which many sponsors decline in an attempt to remain within budget for the trial. However, is the actual cost of off-site nursing more expensive than the traditional site model?
More info >>

Industry Events

CPhI Festival of Pharma

5-16 October 2020, Virtual

A virtual gathering for the pharmaceutical industry, the CPhI Festival of Pharma has been created to fill the gap left by in-person events at a time when travel and gatherings are off the cards.
More info >>



©2000-2011 Samedan Ltd.
Add to favourites

Print this page

Send to a friend
Privacy statement