Bioinformatics in Drug Discovery and Development

Traditionally, pharmaceutical companies have employed a cautious mostly chemistry and pharmacology-based approach to drug discovery, and are finding it “increasingly difficult to find new compounds that will lead to new drugs” (Lim, 1997; Duggan et al., 2000). In the highly competitive  “winner takes all” pharmaceutical industry, the first company to patent a new chemical entity (NCE) for a particular therapy takes all the spoils, leaving other competitors to mostly wait for patent expirations to partake in the largesse (Duggan et al., 2000). Pharmaceutical companies therefore invest heavily in any processes that can accelerate any step of the drug development cycle (Lim, 1997; Van Arnum, 1998; Papanikolaw, 1999; Overby, 2001; Whittaker, 2003). The increasing pressure to generate more drugs in less time has resulted in remarkable interest in bioinformatics (Papanikolaw, 1999).

Although bioinformatics attained prominence because of its leading role in the storage, management and analysis of genomic data, its focus seems to be shifting due to the need of the life sciences to exploit these data (Whittaker, 2003). According to Overby (2001), technologies grouped under the umbrella of bioinformatics involve the use of computers to store, organize, generate, retrieve, analyze and share genomic, biological and chemical data for drug discovery. Several other writers have made the connection between bioinformatics and drug discovery. Whittaker (2003) posited that bioinformatics is used in “drug target identification and validation and in the development of biomarkers and toxicogenomic and pharmacogenomic tools to maximize the therapeutic benefit of drugs”. Ratti & Trist, (2001) suggest that “today’s [drug discovery and development] process … has been enriched by advances in technological developments in screening, synthetic chemistry, and by the increased number of possible targets due to the application of genomics and bioinformatics.” The traditional chemistry and pharmacology-based approach to drug discovery has recently given way to a more modernized information-based approach – bioinformatics (Lim, 1997). The drug discovery and development landscape has changed – for good, with the practice of bioinformatics becoming prevalent in the drug industry such that the drug industry is one of the major players guiding the development of the bioinformatics field (Van Arnum, 1998; Duggan et al., 2000; Attwood & Miller, 2003). Duggan et al, (2000) observed that “many (if not all) of the large pharmaceutical companies have established internal bioinformatics groups whose purpose is to beat the competition to solutions of a problem that may give their company that crucial edge in producing the next major drug.” bioinformatics has certainly come to stay and is now ubiquitous with drug discovery. According to Pollock and Safer (2001) “few if any (drug discovery) projects are computer free”. The impacts that bioinformatics has had and continues to have in the early stages of drug development are encouraging and would only lead to further bioinformatics investments (Ratti & Trist, 2001). Wentland (2004) provides a graphical illustration of the role of bioinformatics in drug discovery:

Read Also

One of the main thrusts of current bioinformatics methods is the finding of biologically active candidates (Whittaker, 2003). Drugs are usually only developed when the specific drug target for those drugs’ actions have been identified and studied (Lim, 1997). Until recently, drug development was restricted to a small fraction of possible targets since the majority of human genes were unknown. The number of potential targets for drug development is increasing dramatically, due mainly to the genome project (Lim, 1997). Mining the human genome sequence using bioinformatics has helped define and classify the genomic compositions of genes coding the target proteins, in addition to revealing new targets that offer potential for new drugs (Van Arnum, 1998; Southan, 2001; Foord, 2002).  This is an area where the human genome information is expected to yield big payoffs (Southan, 2001; Foord, 2002). Drug developers are presented with an unaccustomed luxury of choice as more genes are identified and the drug discovery cycle becomes more data-intensive (Lim, 1997).  Of the estimated 35,000 genes in the human genome, Zambrowicz & Sands (2003) contend that the 100 top-selling drugs have targeted only 43 of their encoded proteins. By enabling the identification and analysis of more and more drug targets, bioinformatics is expected to greatly increase the breath of potential drugs in the pipelines of pharmaceutical companies (Lim, 1997; Overby, 2001; Whittaker, 2003; Zambrowicz & Sands, 2003).

After drug targets – or better still, “potential” drug targets – have been discovered, there is an invaluable need to establish a firm association between a putative target gene or target protein with the disease of interest (Whittaker, 2003). The establishment of such a key relationship provides justification for the drug development. This process, known as target validation, is an area where bioinformatics is playing a significant role.  Target validation not only helps build the case that the drug modulation of such a target will result in beneficial effects on the disease, it also helps mitigate the potential for failure in the testing and approval phases (Ratti & Trist, 2001; Gilbert et al., 2003; Whittaker, 2003).

The current high cost of drug discovery and development is a major cause for concern among pharmaceutical companies (Papanikolaw, 1999). Along with increasing productivity, pharmaceutical companies always aim to reduce the high failure rate in the drug discovery process thereby increasing the number of drugs coming to market (Papanikolaw, 1999). The cost of clinical trials limits the number of drugs a pharmaceutical company can develop, and hence selecting the compounds with the best chances for approval is critical (Klein & Tabarrok, 2003; PhRMA, 2003; Wierenga & Eaton, 2004).  The costs of drug discovery and development – generally include total costs from discovery to approval (Klein & Tabarrok, 2003; PhRMA, 2003; Wierenga & Eaton, 2004) though some studies have included the costs of failed drugs and the costs for commercialization (Gilbert et al., 2003). There is also a cost associated with the elongated process, beginning from discovery all the way to final approval (Lim, 1997; Klein & Tabarrok, 2003; PhRMA, 2003). Advances in bioinformatics have allowed for marked efficiencies, beginning with target identification and validation, to assay development, and high-throughput screening (HTS) – all with the goal of identifying new chemical entities (Belkowski, 2003). With more efficient target discovery and validation, bioinformatics can help ensure that more drug candidates are successful during the approval process as well as shortening the discovery and development cycle, making it more cost-effective (Lim, 1997).

There are some other non-discovery/development costs – collateral costs – that plague the pharmaceutical industry. These costs include commercialization costs (Gilbert et al., 2003; Mullin, 2003), litigation and drug-recall costs (Klein & Tabarrok, 2003; Tait & Mittra, 2004), and general costs to society (Lazarou et al., 1998; Kohn et al., 1999).

Commercialization costs, estimated by Gilbert et al., (2003) to be about $250 million per approved drugs, are high mainly because most “new” drugs approved are essentially functional replicas of drugs that already exist (Koppal, 2004; Tait & Mittra, 2004). Because these mostly copycat drugs are being commercialized to abate ailments for which there are already drugs, there is a need for what Meyers & Baker (2001) describe as a high and “intense promotional noise” in their (pharmaceutical companies) efforts to attract the attention of both patients and physicians who already have access to similar medication. Bioinformatics, by enabling the more efficient discovery and identification of drug components and targets, will bridge the innovation gap, thereby allowing pharmaceutical companies the opportunity to efficiently discover and develop novel drugs and chemical entities (Whittaker, 2003). Commercialization costs are then expected to fall significantly, as drugs are commercialized not in competition with already existing equivalents, but in announcement and advertisement of new drugs that offer new therapeutic benefits for hitherto unmet medical needs (Meyers & Baker, 2001)