Genetic Information, Genomic Technologies, and the Future of Drug Discovery FREE

Despite many advances in medicine, disease burdens remain significant in both developed and emerging countries.¹ Effective drugs for treatment and prevention are needed for many disease areas, including cardiovascular disease, cancer, neurological disorders, infectious diseases, endocrinology, and inflammatory and chronic degenerative diseases. Therefore, there is excitement about the potential biological revolution that will emerge with understanding the human genome.²

The various genome initiatives have provided drafts of the chromosomal sequences of humans and other species. The enabling technologies for this accomplishment, such as transcript profiling, now provide new tools to examine complex biological systems at the level of essentially all expressed messenger RNA (mRNA) and corresponding proteins.

This explosion of biologic information about the proteins and pathways relevant to cellular physiology and disease has stimulated biotechnology and pharmaceutical researchers to assign top priority to identification and validation of key targets (known or novel) to develop therapies for the many remaining diseases. Hypothesis-based biological research is now supplemented with multidisciplinary approaches to systems and circuit-based biology that integrate bioinformatics, genomic databases, and cellular and molecular biology with the traditional drug discovery disciplines of physiological biochemistry, pharmacology, and medicinal chemistry. As a consequence, interventional strategies now include recombinant proteins, monoclonal antibodies, peptides, and small organic molecules as drug candidates. The goal is to expedite the testing of novel therapeutic hypotheses in humans and to develop strategies to identify optimal therapy for individual patients.

Successful drug treatments of today and in the past involve fewer than 500 targets or growth factors as of 1996,³ whereas the human genome contains 35 000 to 120 000 genes.^{4 - 5} At least 5000 of these genes should be important targets or produce therapeutic proteins, suggesting that only 10% of potential therapeutic strategies have been identified and exploited to date.⁶ This avalanche of genetic information, largely coming from genomic sequences and expressed sequence tag–based sequencing of complementary DNA (cDNA) libraries, initially represents a complete catalogue of component parts of the cell and tissues but does not yet provide insight into how all of the protein products of these genes interact or function within the cell.

The challenge for pharmaceutical research is to unravel the pathophysiology of human diseases and thus, make it possible to identify targets accessible to drug intervention. The new systems or circuit view of biology that has evolved from gene research must be considered in successful drug discovery.^{7 - 8} This perspective requires integration of various new technologies into the traditional toolbox of pharmaceutical research. This article focuses on this new genomically influenced multidisciplinary approach to contemporary drug discovery by examining genomic information for hypothesis and target generation and the technologies being developed to validate targets for new therapies.

Biochemistry, pharmacology, and medicinal chemistry will continue to play an essential role in identification of pharmacological targets in the posthuman genome era. Expertise in preclinical models and human biology is necessary to incorporate genomic information into a molecular systems approach to physiology. The traditional drug discovery process will now be supplemented with additional sources of genomic information at the level of chromosomal DNA, disease gene associations, mRNA transcript profiling of tissues, human genetic variance data, and animal and developmental models relevant to disease (Figure 1).

For example, if the disease target involves the human endothelium, genomic information on the target organ system can be mined with bioinformatics for the discovery process.^{9 - 10} DNA sequencing of a cDNA library of human endothelial cells or the use of transcriptional arrays can provide a view of gene transcription in normal human endothelium.^{11 - 12} Likewise, proteomics (or the systematic study of expressed proteins) can provide insight into the functional proteins in specific cells.^{13 - 14}

Additional uses of these technologies include the ability to compare normal and disease paradigms of human endothelium by forward and backward cDNA library subtraction technologies and differential display techniques to concentrate on transcripts associated with the disease hypothesis.^{15 - 16} These approaches will provide a high resolution view of the endothelial cell system with thousands of components catalogued as mRNAs of the corresponding proteins.

In the future, many of these component parts will still be proteins of unknown functions. Identification of the functions of these proteins will be part of biological research for years to come. The discovery of new pathways and new molecular interactions could form the basis of new pharmacological strategies. Protein-protein interaction maps are being constructed using strategies such as the yeast 2-hybrid system, which eventually will provide detailed insight into the pathways and networks operational in biological systems.¹⁷ Until all these interactions and pathways are mapped out, a pragmatic approach is to view the data in the context of drug target families.

A biological target is pharmacologically accessible (or "druggable") when an organic molecule (peptide, protein, or monoclonal antibody) can modulate the target's function. The predominant target of currently available drugs is the G protein coupled–receptor family of molecules.¹⁸ Other cell surface receptors and molecules, proteases, protein kinases, and phosphatases are also "druggable."^{19 - 21} In addition, secreted hormones, growth factors, chemokines, soluble receptors, and decoys may serve either as a drug substance or as targets for other biological strategies.^{22 - 24}

The sequence databases described for the example of human endothelial cells could be scanned with sequence searching algorithms programmed to recognize molecular signatures of these druggable families. This "filtered" genomic information from bioinformatics may identify pharmacologically accessible targets in the absence of clues to their biological relevance to the disease. Target identification could be routine with these technologies, whereas, at present, selection of the correct target is the key strategic challenge.

Ultimate biological validation of a pharmaceutical target comes with a successful phase 3 clinical trial and broad application of the therapy to large populations of patients. Intermediate stages of target validation are summarized in Figure 2. The initial stages involve relating the target hypothesis to the disease under study. The hypothesis may originate from a relevant animal model, such as the discovery that the molecular defect in ob/ob mice involved the hormone leptin.²⁵ This observation was linked to human obesity by identification of the human homolog and study of its expression in human obesity.

Mining druggable genomic targets as transcripts from various tissue libraries, prompts additional questions of whether the protein is expressed in vivo and whether the target protein is expressed in various physiological states. Human and animal tissue banks are an invaluable resource for genomic target validation. In situ and immunohistology studies make it possible to correlate mRNA and target protein levels with genetics, transcript profiling, and proteomic or biochemical pharmacology studies. Such analyses of target expression in all human tissues is critical to understanding the potential toxicology of the drug strategy under development.

Molecular and cellular biology strategies are frequently used to ablate, overexpress, or modulate the expression of an in vitro or in vivo target. Such studies can help provide evidence for the function of the target as well as models for drug screening and further pharmacological research when drug candidates have been identified.^{26 - 27} Species such as Drosophila melanogaster and Caenorhabditis elegans are assuming increasing roles for target validation.²⁸

At this point or earlier in the process, drug discovery often involves developing screening strategies to discover drug candidates that modulate the targets to test the therapeutic hypotheses. These agents essentially become validation tools in the process. Various analytical technologies have been used to develop highly leads for potential targets.^{29 - 30} The types of drug candidates screened are originally derived from synthetic organic chemistry, combinatorial chemistry, which reflects parallel synthesis on chemical templates; natural product chemistry; and use of chemical libraries.^{31 - 33} Additional technologies include development of monoclonal antibodies for the target (restricted to cell surface or soluble extracellular molecules), which may act as antagonists, agonists, or other modulating functions of the proposed target.³⁴ Candidate drugs can then be evaluated in preclinical pharmacological investigations. For example, in vivo models can be tested with the drug candidate to evaluate efficacy. If the results are promising, further optimization and evaluations of potency, efficacy, bioavailability, and toxicology can be conducted prior to testing in humans.

In the interim, the strategies used to find the target initially, such as transcript profiling, can be used to examine the effects of the drug in the biological system under study. An example of this approach involves recombinant human activated protein C(rhaPC), an anticoagulant protein that functions as a serine protease to regulate the activity of factors V and VII in the coagulation cascade (Figure 3).³⁵ Preclinical studies in models of sepsis suggested that rhaPC also has anti-inflammatory properties. A proinflammatory response was induced in human endothelium by exposing human umbilical cord endothelial cells to tumor necrosis factor, and rhaPC downregulated expression of several proinflammatory genes and up-regulated antiapoptotic transcripts. The effects are potentially relevant to the utility of the drug in acute inflammatory disease (D. Joyce, MD, unpublished data, 2001). While these studies were performed after preclinical pharmacology, novel pharmacological actions of rhaPC were uncovered. These strategies appear broadly applicable to investigation of disease and drug mechanisms, structure activity relationships, and toxicogenomics, and they should enhance the potential success of drug candidates in the future. In addition, the use of singular transcript profiling strategies in preclinical models may identify compensatory systems or circuits in biology that can be targets of a future pharmacological intervention.

These strategies will continue to be enhanced with the development of new biotechnologies and knowledge about the biological circuits of life. Although new validated targets will emerge in unprecedented numbers, several key questions also arise. For instance, which targets and interventions should receive high priority for clinical investigation? What drug modality can be used to rapidly evaluate the hypothesis in humans? When will the first new therapies from these new strategies emerge?

It is likely that the initial impact of genetic and genomic technologies for new therapies will involve therapeutic proteins and monoclonal antibodies.^{36 - 38} Numerous examples exist of novel therapeutic protein candidates derived from genomic databases and some of these agents are currently in clinical development.^{36 - 38} Innovative developments in monoclonal antibody technology such as engineered mice with human immunoglobulin repertoires have also provided researchers with the tools to rapidly identify and evaluate human monoclonal antibodies as drug candidates. Taken together, therapeutic proteins and monoclonal antibodies represent one of the largest classes of drugs in development, with estimates of more than 350 molecules under investigation in early 2000.³⁹ Most of these agents are directed toward cancer or inflammation targets for which biological drugs continue to provide significant innovation.^{40 - 41} These approaches accelerate human clinical trials, but some diseases are not amenable to protein therapies because of lack of accessibility of the target (eg, intracellular enzyme) or because of physiological constraints (eg, blood brain barrier).

However, a parallel explosion of many drug candidates of small molecular weight will occur in the next decade as research intensifies into the regulatory circuits of disease coupled with parallel drug discovery on every new validated target. It is likely that polygenic diseases will provide multiple targets for therapy, each of which will need clinical investigation. An example of a disease platform for which genetic and genomic technologies have had an impact is in obesity research, for which several intervention options have been identified recently.⁴² An alternate strategy may involve the in-depth understanding of a family of potential targets that are implicated in key biological activities. For instance, this strategy is apparent in developments in the pharmacology of excitatory amino acid receptors, which have potential therapeutic applications in several diseases.⁴³

The chemical "hit/lead" database will be combined with bioinformatics and structural genomics to expedite the development of new leads in the future.^{44 - 45} It seems inevitable that these various therapeutic intervention strategies would identify targets that today are not presently accessible to pharmacology. Gene therapy may also provide alternative pharmaceutical strategies as may cellular therapies, particularly with human embryonic and adult stem cells.^{46 - 47}

The genetic and biological revolution undoubtedly will change clinical trials and clinical practice in the future. The effect of human genetic variance on responses to therapy will influence drug-development clinical trials and the use of products in clinical practice.^{48 - 49}

The "resequencing" of the human genome will establish the frequency of single nucleotide polymorphisms in the genome and other types of genetic variance. DNA, which is now routinely obtained in many clinical trials, will be examined retrospectively for the association of genetic variance or mutations with outcomes. Smaller clinical studies prospectively biased toward the target, the drug, and the appropriate patient will be designed. These approaches and the accompanying data could make it possible to devise optimal strategies for future therapies. Biomarker/surrogate marker panels for human diseases will provide additional strategies to diagnose, monitor, and predict outcomes of various therapies.

The potential of genetics, genomic information, and genomic technologies for new treatments will only be realized if these new disciplines are integrated into the drug discovery process at every stage—from hypothesis generation to clinical evaluation. Translation of biological information to disease knowledge, validated target mechanisms, and new therapies will indeed make the coming century an era of biomedical revolution.