ACS Publications Division - Journals/Magazines
About MDD - Subscription Info
September 2001, Vol. 4
No. 9, pp 34–36, 41.
To MDD Home Page
Focus: Molecular Modeling
Feature Article
Finding new targets


Modeling and docking are all well and good when you know the targets—but how will the new ones be found?

opening artFinding novel molecular targets for potential new therapeutics is perhaps one of the most important facets of drug discovery. Many consider the search for these “druggable targets” as key to providing new fodder for the type of molecular modeling and docking testing discussed in “2001: A dock odyssey” in this issue of Modern Drug Discovery). In part that is true, but new targets are critical for much more. They also provide fodder for standard pharmaceutical high-throughput screening systems, a testing ground for the vast numbers of combinatorial libraries already devised. Equally important, new targets can also provide new windows on disease treatments and metabolic process control.

The NIH is extremely interested in funding the search for new cancer targets, pushing researchers to reorganize what they call the “front end” or “gateway to drug discovery”. In its 2000 program, Molecular Target Drug Discovery for Cancer: Exploratory Grants, available to fund both private for-profit and nonprofit organizations, the NIH is encouraging investigators “to identify a novel molecular target, or to validate the target as a basis for cancer drug discovery, or to develop an assay for the target” (1).

The following are among the various approaches to finding new and potentially druggable targets:

  • mining the literature on basic metabolic biochemistry, especially as seen in known disease systems;
  • mining the human genome (and others) for plausible candidates in known classes of proteins, especially those that may be involved in disease states;
  • mining nature for plant, animal, and microbial products that can serve as a means of finding new keys in the screening for new locks or, on occasion, mining for new locks directly;
  • comparative genetics—the use of differential gene activation that is examined across species, tissues, or a variety of metabolic states to identify genes uniquely turned on or silenced in disease or stress conditions;
  • functional genomics, which applies to the use of genomic information to determine the “real-world” function of particular genes (known or unknown) and how interacting genes and gene products operate in normal physiology and disease; and
  • physiomics, perhaps the ultimate solution—knowledge of the complete physiology of an organism, including all interacting metabolic pathways, structural and biochemical scaffolding, the proteins and accessories that make them up, and the gene interactions and cues that control them.

Mining books
According to the NIH, even though researchers “may already have developed significant preliminary data on a signaling pathway, they nevertheless may have not focused on the point of greatest vulnerability in the pathway and, therefore, perhaps the optimal point of drug attack.” Survey and identification of known metabolic pathways with extant chemical and genetic information from the literature have the potential to provide guidance to potential target sites worthy of greater investigation, especially in light of the tremendous amount of genomic and protein research data constantly being generated. This is a clear case in which overspecialization may prevent molecular biologists from being fully aware of the wealth of physiological information produced over the decades by their peers who have studied metabolism or classical genetics.

Mining genes
Much of the genome mining is in the area of finding new targets analogous to already-known targets. One of the most important examples of this is the search for the genes for what are called orphan G-protein coupled receptors (GPCRs). GPCRs make up the largest single class of cell surface receptors and, according to Shelagh Wilson and Derk Bergsma of SmithKline Beecham pharmaceuticals, “the superfamily of GPCRs is [thus] one of the most important families of drug targets for the pharmaceutical industry. Of the top 200 best-selling prescription drugs, more than 20% interact with GPCRs, providing worldwide sales of over $20 billion.” These include drugs such as Cimetidine for ulcers, Losartan for hypertension, and Ropinerole for Parkinson’s disease (2).

More than 1000 members of the GPCR family have been cloned from various species, including over 160 distinct human subtypes with a known ligand and over 100 so-called orphan sequences (with homology to known receptors of 40%) for which the corresponding ligand is unknown (hence the designation “orphan GPCRs”). Attempts to characterize such orphan receptors and determine whether they are indeed novel potential drug targets rely on a process Wilson and Bergsma describe as “reverse pharmacology”, which starts with an unknown target protein to determine its function and ultimately attempts to find a drug that will modulate it. This approach is very similar to reverse genetics—moving from knowledge of a particular protein to elucidate the gene that produces it.

The reverse pharmacology strategy uses the orphan receptor as a “hook” to capture its own ligand from cells, and “the ligand is then used to explore the biological and pathophysiological role of the receptor. High-throughput screening is initiated on the receptor in parallel with the biological characterization in order to develop antagonists.” (2) The availability and success of these will help determine whether the receptor can be considered a druggable target.

When looking for totally unknown candidates, searching the genome for new targets is equivalent to searching for new locked treasure chests in a vast storehouse of mixed wealth, trash, and attic storage. In addition, the multiplicity of new genomes being sequenced is providing an overabundance of gene sequence information for mining as well as a remarkable reservoir of information for compare-and-contrast genomics. Bioinformatics is a key player in teasing out potential new targets from this mass of data (see box, “Informatics in action: The case of nicastrin”). The databases used in such analyses have been previously discussed (see “Databasics” in Modern Drug Discovery, Jan 2001, pp 31–35).

Mining nature
This can be as simple as screening test cells for general toxicity to new natural products or as complex as screening for unique metabolic changes. In either event, a search is then directed for the binding site or mode of action of effective molecules to (hopefully) discover new targets.

The NIH recognized the value of such an approach when it suggested that researchers in its new cancer target discovery grant program might “identify the function of a cellular target after first finding the target as a result of exploring binding patterns of natural products or other ligands to the novel target.” Similarly, plants, animals, or microbes with unique resistances or sensitivities to certain diseases, toxins, or stresses might provide potential targets if these traits are molecularly encoded and can be elucidated.

Comparative genetics
The genetic differences between expressed genes in different tissues (whether healthy or diseased), between healthy and diseased individuals, and even between species with varying responses to external stimuli can yield a wealth of information on potential drug targets.

On a technical level, one of the best ways to evaluate such differences is to compare DNA microarray pattern displays between tissues, cells, and species under different physiological or genetic conditions. This concept involves a comparative display of genes turned on and off that can differentiate between one cell type and another, one species and another, or between metabolic “sickness” and metabolic health. One example of the fruits of such a comparative study is the Cancer Genome Anatomy Project (, that has identified mutational sites in cancer cells (through careful comparisons with “normal” genotypes), that may prove to be unique to specific types of cancer and hence plausible areas to search for novel targets.

Much of the ultimate usefulness of this information involves the necessity of target validation—checking whether they are truly druggable sites. According to the NIH: “Clearly, if a prime target is fully validated, such that the target can be used to totally exploit a unique difference between a healthy and a cancerous cell, control of cancer would only await the discovery of a drug molecule through a systematic screening and/or drug design program” (1).

Pharm follows function?
Functional genomics has been variously defined, but it is perhaps best thought of as the culmination and synthesis of all the previous approaches, including “reverse pharmacology”.

For example, when SmithKline Beecham researchers approach the analysis of the function of novel GPCR-like gene sequences, they use the following schema (highly abbreviated here):

  1. Segments of GPCR-like sequences identified using exhaustive searches of DNA sequence databases are used to clone full-length cDNAs using extant libraries and PCR-based techniques.
  2. Differential libraries are used “to determine whether the receptor is expressed in tissues of therapeutic interest.”
  3. Appropriately cloned full-length cDNAs are expressed in mammalian cell libraries (typically in Chinese hamster ovary or human embryonic kidney 293 cells).
  4. “The expressed receptor is then screened in a variety of functional assays to look for an activating ligand.” Assays for GPCR-like activity typically involve monitoring changes in intracellular calcium or cAMP levels, which can be detected using automated fluorescent techniques.
  5. Once ligands are discovered that interact with the GPCR-like target, a variety of agonist and antagonist tools are used “to evaluate the physiological role of the receptor and its potential as a therapeutic target for drug discovery.”

Transgenic or knockout animals can provide the ultimate validation of reverse pharmacology (2). A noteworthy example of the success of such an approach is the discovery by Masashi Yanagisawa and his team at the Howard Hughes Medical Institute of two new neuropeptide hormones (orexins A and B) in mice that appear involved in sleep regulation. Both were discovered from the exploration of a particular set of orphan GPCRs. Orexin knockout mice showed cataleptic and sleep attacks similar to human narcolepsy (3). In humans, the number of brain cells containing orexin (also known as hypocretin) was found to be reduced by 85–95% in people with narcolepsy (4). The orexins also appear to have a significant role in influencing appetite. Mice with orexin delivered to their brains via catheters ate 8–10 times more than normal within a few hours.

Physiomics futures?
Ultimately, though, it is the quest for the totally unknown targets that will require the greatest effort and creativity among pharmaceutical scientists. As the known “orphan” potentials in the genome are all discovered and all the standard cell surface receptor-like sites are used up, all that will be left is a well of ignorance. The newest of the “-omics” may be the new well from which hope for novel targets springs.

Physiomics essentially refers to a systems approach to organismal physiology. Intracellular targets that modulate complex pathways, whether metabolic or gene control, are its great frontier. These targets cannot be accessed from simple knowledge of gene sequences, but rather through analysis of complex biochemical and physiological interactions, often involving modeling the behavior and interaction of a variety of cell types at the tissue, organ, and even whole-organism level. Currently, physiomics is a predicted science, a conglomeration of research approaches, rather than an extant field, as genomics was a decade ago. But the first wave of researchers and companies are staking their futures on this trend. When and whether authentically new druggable targets will arise using this approach remains a mystery.


  1. Molecular Target Drug Discovery for Cancer: Exploratory Grants,
  2. Wilson, S.; Bergsma, D. Orphan G-protein coupled receptors: Novel drug targets for the pharmaceutical industry. Pharm. News 2000, 7 (3). Available at
  3. Howard Hughes Medical Research Laboratory Institute,
  4. Siegel, J. Mystery of Human Narcolepsy Solved,

Mark S. Lesney is a senior editor of Modern Drug Discovery. Send your comments or questions regarding this article to or the Editorial Office by fax at 202-776-8166 or by post at 1155 16th Street, NW; Washington, DC 20036.

Return to Top || Table of Contents

CASChemPortChemCenterPubs Page