|
New technologies and partnerships are helping to meet the challenges of antibacterial R&D
Andrew B. Benowitz, Jennifer L. Hoover, and David J. Payne
Andrew B. Benowitz is a Chief Scientist in the Department of Medicinal Chemistry, Antibacterial Discovery Performance Unit, Jennifer L. Hoover is an Investigator in the Department of Microbiology, Antibacterial Discovery Performance Unit, and David J. Payne is Vice President of the Antibacterial Discovery Performance Unit, Glaxo-SmithKline, Collegeville, Pa.
Summary
· Many challenges contribute to the scarcity of novel antibacterial agents, including relative shortages of investment in this field as well as shortcomings in scientific approaches.
· Although bacterial genomics did not live up to early expectations, it is providing valuable insights as other technologies come into play in the search for antibacterial agents.
· Alternate approaches such as boron chemistry, natural products screening, and novel hit discovery platforms are being explored to increase the probability of finding new antibacterial leads.
· Both government and nongovernment organizations are supporting antibacterial R&D, while top-level officials who recognize the public health importance of antibiotic resistance are seeking new incentives to meet this challenge.
The lack of new antibacterials for the treatment of serious infections has been broadly recognized as a major unmet medical need for at least a decade. A recent survey of 96 intensive care doctors across Europe showed that 50% have treated at least one patient with a gram-negative pathogen that was totally or almost totally resistant to all antibiotics during the previous six months. This is a dire predicament, as it is expected that no entirely novel mechanism agent for the treatment of gram-negative infections will be launched before 2015. Furthermore, recent experience with H1N1 pandemic influenza emphasizes the urgency of the situation, as bacterial coinfection with MRSA was implicated in a surprising proportion of fatalities. This phenomenon is especially worrying, since few oral treatment options exist for MRSA pneumonia.
It is generally recognized that an overall lack of investment in this therapeutic area has played a major role in creating this situation, but scientific challenges have also proved to be a significant impediment to the discovery of new antibacterials. In this article, we would like to illustrate some of these challenges, as well as highlight some of the approaches that we have put in place to increase our chances of success.
The Genomics Promise Unfulfilled?
Bacterial genomic analyses figuratively exploded during the mid-1990s, boosting the field of antibacterial discovery by enabling investigators to scan genomes for new drug targets. Although this approach proved highly successful at its primary goal, exploiting those new targets to discover novel classes of drugs remains problematic. For instance, between 1995 and 2001, we and our colleagues at GlaxoSmithKline (GSK) evaluated more than 350 bacterial genes, nearly half of which were proven necessary for survival.
 Based on those findings, we developed and ran high-throughput screens (HTS), evaluating our collection of chemical entities for their activity against 67 new bacterial targets. However, we found hits for only 16 of those screens, and a mere 5 of them became leads for further development. The success rate for identifying novel antibacterial agents proved to be four- to fivefold lower than for other therapeutic areas at GSK. During this seven-year process, we learned two important facts: (i) novel targets can pose unexpected difficulties, and (ii) new methods are required to discover not only more, but also higher quality antibacterial leads.
New Targets, New Surprises
With access to the genome sequences of many pathogenic microorganisms, it is possible to identify hundreds of potentially novel bacterial targets based on their species conservation and differences with human homologs. As our next step, we validated these targets by "knocking out" the gene to perform essentiality testing in a single bacterial species.
However, despite similarities in sequences across different pathogens, some species could tolerate the lack of a particular gene much better than others. This was observed at GSK with at least six gene targets-trmE, ypuL, yaaJ, ykrA, yrrK, and ydiB-making it imperative that we evaluate the essentiality of a particular gene in multiple pathogens to enhance the possibility of discovering antibacterial agents with sufficient spectrum to cover all the likely causative pathogens for a specific infection or "indication."
This effort provided us with a list of target genes, whose protein products were shown to be essential for bacterial survival in vitro. However, simple gene knockouts do not indicate how fully the gene product must be inhibited to cause cell death. Consequently, we examined which genes are more sensitive to down regulation, as the protein products of highly sensitive genes would be expected to be good antibacterial targets. For example, when we titrated gene expression versus cell viability for two early targets, bacteria proved 10-fold more sensitive to down regulation of the methionyl tRNA synthetase (MRS) gene than to down regulation of the polypeptide deformylase (PDF) gene. By this criterion, MRS appeared to be a more appealing target. Although this method is attractive in its logical simplicity, many other factors, such as the mechanism by which potential agents inhibit a target (i.e., reversible or irreversible), must also be considered, further illustrating the complexities in selecting targets.
As we progressed, we encountered additional surprises. For example, the aro genes encode the chorismate biosynthetic pathway, which is critical to bacterial survival in vivo. We assumed that removing any aro gene would be lethal for bacterial cells. However, although aroB, aroD, and aroK proved to be essential for S. aureus survival, we were surprised to find that aroE is nonessential, making it a poor target for agents to inhibit this important pathogen. Surprises like this suggested the dangers of making seemingly logical assumptions, while demonstrating that each target requires rigorous validation.
 It took many months-and significant chemical and biological resources-to recognize another example of how antibacterial drug discovery continues to challenge us. Following our bioinformatic analysis and gene down-regulation studies suggesting that MRS would be a good target, we identified several MRS inhibitors that had weak antibacterial activity against S. aureus. After modifying those compounds into derivatives with very good antibacterial potency and subjecting them to more extensive minimal inhibitory concentration (MIC) testing, we began to notice a resistant S. pneumoniae subpopulation.
We soon learned that those resistant isolates contain two MRS genes, from which they can produce two distinct MRS enzymes. Unfortunately, our lead compounds inhibited only MRS1, and we were unable to optimize them to encompass both MRS1 and MRS2. A subsequent survey indicated that 46% of international S. pneumoniae isolates contain both of these MRS genes, leading us to realize that our advanced MRS inhibitors were not going to be effective against nearly half of all S. pneumoniae isolates. Since S. pneumoniae was an important pathogen for this program, we were forced to halt our discovery efforts in this area. This experience illustrates the downside of relying on target judgments based so heavily on genomelevel analysis. In this case, we missed an important issue until we tested seemingly robust lead compounds against a substantial number of pathogenic strains.
The work described above was conducted in the late 1990s, and in recognition of our limited success in discovering leads for targets derived from genomic analysis, we subsequently shifted our efforts to optimizing compounds that inhibited already well-validated targets, such as the ribosome and DNA gyrase. Although this shift in strategy led to the discovery and launch of the topical antibiotic retapamulin as well as to the entry of several novel mechanism agents into Phase I clinical studies, we still sought other, more innovative approaches to fuel our longterm success.
Can a Good Start Produce a Good Finish?
Our experiences at GSK suggest that traditional hit discovery efforts are not well suited for discovering novel-mechanism antibacterial agents. One reason that HTS methods have mostly failed to identify compounds that inhibit bacterial targets is that corporate compound collections, by design, are heavily biased toward mammalian targets. Mammalian targets encompass a wide range of proteins, including Gcoupled protein receptors, ion channels, and nuclear hormone receptors, whereas bacterial targets are mostly intracellular enzymes. Inhibitors of intracellular enzymes are expected to occupy a different physicochemical space than inhibitors of other protein targets, and creating a high-quality, discrete compound collection of sufficient size and diversity to cover both mammalian and bacterial targets is not easy. However, what if we could screen millions or billions of compounds against essential bacterial targets?
Recently, we began using a new method for synthesizing and selecting small-molecule combinatorial libraries through the use of doublestranded DNA molecular tags. This method, called encoded library technology (ELT), enables us to screen billions of small compounds against very small quantities of target proteins. Since each compound carries a unique DNA tag, sequencing the associated tag enables us to identify those that bind to the target. Compounds of interest are then reprepared without the tag for further evaluation in functional assays.
Although DNA encoding strategies for combinatorial libraries have been known since the 1990s, synthesizing and screening these libraries proved problematic when it came to meeting pharmaceutical discovery requirements. In contrast, the new ELT libraries enable the rapid screening of multiple protein targets against mixtures of billions of small molecules. This method is particularly advantageous because it allows us to interrogate a much larger portion of drug-like chemical space than had been accessible. Additionally, the output data is in the form of selection-enrichment plots, which can be used to identify likely structure-activity relationships (SAR) before running biochemical assays. This powerful technique is enabling us to revisit a large number of bacterial targets that we identified from our prior genomics research in the context of a new screening paradigm.
Fragment-Based Lead Discovery Reflects a "Small Is Beautiful" Strategy
The number of potential drug-like molecules has been estimated to be 1060. Thus, even the largest DNA encoded libraries, consisting of billions of compounds, sample only a small fraction of this chemical space. In contrast, the number of small chemical fragments is estimated as 107 different compounds. Therefore, it should be possible to sample a much larger proportion of this fragment- like chemical space than of drug-like chemical space. Furthermore, since high molecular weight and high lipophilicity are correlated with poor clinical performance, small, polar fragments should provide better hits for lead discovery. Such reasoning led to the strategy of fragment-based lead discovery (FBLD), an approach that both large and small pharmaceutical companies are quickly adopting.
Although FBLD is rapidly gaining momentum, the hits it generates are only starting points that require considerable follow-up development. Fragments, by definition, have low complexity, and fragment libraries typically consist of fewer than 20,000 compounds, making it likely that there will be low overlap between fragment hits and few observable SAR trends. Therefore, it is critical to have detailed structural information, such as protein-fragment X-ray or nuclear magnetic resonance data, combined with a robust molecular modeling effort to enable further development. Nonetheless, we believe that this low-molecularweight, low-lipophilicity chemical space is ideally suited to antibacterial hit and lead generation, and we are actively applying this strategy to our own antibacterial discovery efforts.
Driving Discovery with Innovative Alliances
 More than two-thirds of the antibiotics in clinical use are natural products or their semisynthetic derivatives. Although pharmaceutical companies have largely abandoned natural products in drug discovery, recent collection efforts in underexplored ecological areas, combined with the ability to look for unexpressed secondary metabolites through bioinformatics-based genome mining are again providing novel molecules to investigate.
Recognizing those advances, but mindful of the costs and infrastructure requirements involved in screening natural products, GSK established an alliance with Galapagos, a leader in natural products screening. In addition to its extensive natural product collection, Galapagos has developed a sophisticated screening and analysis method that substantially reduces the time required to identify promising molecular structures. Since this process is traditionally a bottleneck, we believe that this approach is uniquely well-suited to accelerate the discovery of natural product antibacterials. Since this GSK-Galapagos alliance began in 2007, several compound classes with antibacterial activity have been identified for further investigation.
Although we consider natural products to be an important reservoir of structural novelty for leads to new antibacterial agents, we continue to investigate a variety of approaches to increase structural diversity. For example, in 2007 we formed an alliance with Anacor Pharmaceuticals, where researchers are incorporating boron into drug-like molecules. Although many members of the pharmaceutical community consider boron-containing compounds too toxic for all but specialized use, boron is naturally found in many foods and is also an essential plant nutrient. Furthermore, extensive safety data indicate that boron is not inherently toxic. Indeed, boric acid has an LD50 similar to regular table salt. Therefore, incorporating boron into drug-like molecules remains a largely untapped area to explore in drug discovery.
To address this opportunity, Anacor is developing heterocyclic oxaborole rings as components for drug leads. These oxaborole rings are highly polar, and, depending on substituents on the ring, the boron atom can reversibly add water to form a boronate species. The boron atom can also interact with alcohol moieties in various enzymes in an analogous fashion, opening the possibility of creating reversible enzymeinhibitor complexes with wide applicability for drug discovery.
With Anacor, we are exploring the use of boron as a design element against a variety of antibacterial targets. The initial results of this strategy are very promising, and a novel mechanism, boron-containing antibiotic targeting gram-negative pathogens is currently progressing into development.
Although accessing structural diversity is a key driver for the discovery of agents with novel mechanisms, many currently available antibacterials have well-defined risk benefit profiles, but are compromised by resistance. This problem is especially acute among gram-negative pathogens. Perhaps some of these approved agents can be revitalized. For instance, efflux pumps are the major mechanism through which gram-negative bacteria protect themselves from antibiotics. Thus, efflux pump inhibitors (EPIs) could dramatically improve the clinical effectiveness of both existing and novel classes of antibacterial agents against these pathogens.
To investigate this strategy,GSKentered into an alliance with Mpex Pharmaceuticals, an acknowledged world leader in developing EPIs. The initial results of this collaboration show that EPI molecules offer the powerful ability to potentiate the action of existing classes of antibiotics against gram-negative pathogens. EPIs also significantly improve the gram-negative activity of some of our novel mechanism gram-positive candidate molecules, leading us to view EPIs as potentially transformational for tackling multidrug-resistant, gram-negative pathogens.
Measures To Stimulate Innovation
Given the scientific challenges associated with early stage antibacterial drug discovery, we concluded that we needed to place larger teams of scientists on some of our programs to accelerate the delivery of candidate compounds with novel mechanisms of action. To help resource these larger teams, we applied for and received external funding from the Wellcome Trust Seeding Drug Discovery Initiative (SDDI) and the Defense Threat Reduction Agency (DTRA), which is part of the U.S. Department of Defense.
SDDI supports areas of high unmet medical need where there is a low level of investment from the pharmaceutical industry. The driver for DTRA is to create antibacterial agents for use in countering biothreat pathogens. The DTRA funding is being used at GSK to develop agents with spectra covering both biothreats and multidrug-resistant, gram-negative pathogens that someday would be available as "dualuse agents" to treat infections caused by either biothreat or conventional pathogens. Those two organizations are also supporting other companies with similar goals.
Although such partnerships represent a very encouraging step toward increasing the level of antibacterial R&D, accelerating the delivery of a sustainable pipeline of new antibiotics requires a more comprehensive set of incentives if the overall investment in this area is to increase. The Incentives for Innovative Antibiotics conference held by the Swedish Presidency of the European Union (EU) in September 2009 and hosted by the Swedish Health Minister represented a landmark event to explore such incentives. Encouragingly, the conference ended with an EU Council decision to create a "comprehensive action-plan with concrete proposals to develop new effective antibiotics." Additionally, at the EUU. S. Summit held in November 2009, Presidentin- Office of the Council of the EU Fredrik Reinfeldt and U.S. President Barack Obama established a joint task force to combat antimicrobial resistance.
Further initiatives include the Infectious Diseases Society of America 10 x '20 campaign, in which the organization calls for 10 new antibacterials by the year 2020, and the British Society for Antimicrobial Chemotherapy's "The Urgent Need" initiative that is also looking at solutions for the discovery of new antibacterials. All of these stimulus plans are focused on minimizing the development of resistance, creating meaningful incentives, and raising the profile of the need for new antibiotics. Many different types of incentives under discussion are described as either pull incentives (create a more favorable commercial environment for innovative antibacterials), or push incentives (creating a more appealing environment for companies to conduct antibacterial R&D). Currently, this looks like the most promising climate to date for creating meaningful incentives with the potential to reinvigorate investment in novel antibacterials.
There are many factors that have contributed to the decrease in investment in antibacterial R&D and concomitant lack of novel mechanism antibacterials in clinical development. Our hope is that this review illustrates one aspect of the scientific challenges that face antibacterial drug discovery. A central theme to our approach has been to create alliances and innovative partnerships that enable new platforms and strategies to be applied to the problem. However, increased academic funding for antibacterial sciences and meaningful R&D incentives are also needed to create the increase in critical mass in this area that is vitally needed to ensure the delivery of a sustainable pipeline of novel antibiotics.
SUGGESTED READING
Baker, S. J., C. Z. Ding, T. Akama, Y.-K. Zhang, V. Hernandez, and Y. Xia. Therapeutic Potential of Boron-Containing Compounds. 2009. Future Med. Chem. 1:1275-1288.
Clark, M. A., R. A. Acharya, C. C. Arico-Muendel, S. L. Belyanskaya, D. R. Benjamin, N. R. Carlson, P. A. Centrella, C. H. Chiu, S. P. Creaser, J. W. Cuozzo, C. P. Davie, Y. Ding, G. J. Franklin, K. D. Franzen, M. L. Gefter, S. P. Hale, N. J. V. Hansen, D. I. Israel, J. Jiang, M. J. Kavarana, M. S. Kelley, C. S. Kollmann, F. Li, K. Lind, S. Mataruse, P. F. Medeiros, J. A. Messer, P. Myers, H. O'Keefe, M. C. Oliff, C. E. Rise, A. L. Satz, S. R. Skinner, J. L. Svendsen, L. Tang, K. van Vloten, R. W. Wagner, G. Yao, B. Zhao, and B. A. Morgan. 2009. Design, synthesis and selection of DNA-encoded small-molecule libraries. Nature Chem. Biol. 5:647-654.
Congreve, M., G. Chessari, D. Tisi, and A. J. Woodhead. 2008. Recent developments in fragment-based drug discovery. J. Med. Chem. 51:3661-3680.
Fischbach, M. A., and C. T. Walsh. 2009. Antibiotics for Emerging Pathogens. Science 325:1089-1093.
Gentry, D. R., K. A. Ingraham, M. J. Stanhope, S. Rittenhouse, R. L. Jarvest, P. J. O'Hanlon, J. R. Brown, and D. J. Holmes. 2003. Variable sensitivity to bacterial methionyl-tRNA synthetase inhibitors reveals subpopulations of Streptococcus pneumoniae with two distinct methionyl-tRNA synthetase genes. Antimicrob. Agents Chemother. 47:1784-1789.
Lepape, A., and D. L. Monnet on behalf of participating members of the European Society of Intensive Care Medicine (ESICM). 2009. Experience of European intensive care physicians with infections due to antibiotic-resistant bacteria, 2009. Eurosurveill. 14(45):4-6.
Miller, J. R., and G. L. Waldrop. 2010. Discovery of novel antibacterials. Expert Opin. Drug Disc. 5:145-154.
O'Shea, R., and H. E. Moser. 2008. Physicochemical properties of antibacterial compounds: implications for drug discovery. J. Med. Chem. 51:2871-2877.
Payne, D. J., M. N. Gwynn, D. J. Holmes, and D. L. Pompliano. 2007. Drugs for bad bugs: confronting the challenges of antibacterial discovery. Nature Rev. Drug Disc. 6:29-40.
 |
Facebook
del.icio.us
Digg
Furl 
Yahoo!My Web
StumbleUpon
Google
Technorati 
reddit
Twitter
Subscribe to this article's comment feed
Show/Hide comment form