Medicinal Chemistry

Future

Review

For reprint orders, please contact [email protected]

Medicinal Chemistry

Utilizing diversity-oriented synthesis in antimicrobial drug discovery

The development of resistance to existing antimicrobials has created a threat to human health that is not being addressed through our current drug pipeline. Limitations with the use of commercial vendor libraries and natural products have created a need for new types of small molecules to be screened in antimicrobial assays. Diversityoriented synthesis represents a strategy for the efficient generation of compound collections with a high degree of structural diversity. Diversity-oriented synthesis molecules occupy the middle ground of both complexity and efficiency of synthesis between natural products and commercial libraries. In this review we focus upon the use of diversity-oriented synthesis compound collections for the discovery of new antimicrobial agents.

Eamon Comer*,1, Jeremy R Duvall1 & Maurice duPont Lee IV1 1 Therapeutics Platform, Broad Institute of MIT & Harvard, 7 Cambridge Center, Cambridge, MA 02142, USA *Author for correspondence: [email protected]

Keywords: chemical complexity • diversity-oriented synthesis • drug discovery • hepatitis C virus • infectious disease • malaria • orthopoxviruses • Trypanosoma cruzi

We continue to view with apprehension the slender pipeline of therapeutics with new mechanisms of action to treat drug-resistant infections. The development of antibacterial agents is one of medicine’s great success stories; however, the emergence of resistance is steadily eroding progress made in this area. This situation is mirrored in other infectious diseases. For example, in malaria, resistance has developed to nearly every clinically approved therapeutic with the exception of artemisinin; however, reports of prolonged parasite clearance times with patients in southeast Asia have raised concerns that it might only be a matter of time before this too succumbs to resistance [1] . The response from the drug discovery community to this serious threat to human lives has had limited success. In the case of antibiotics, just 20 drugs have come onto the market since 2000, most of which are analogs of existing compounds and have known liabilities, such as toxicity or resistance [2,3] . Compounding this problem is the fact that many pharmaceutical companies are now leaving the antimicrobial field. While numerous factors

10.4155/FMC.14.111 © 2014 Future Science Ltd

have contributed to this exodus, it is worth noting that better returns on investment can be made in other disease areas [4] . Thus, we are left with a legacy of relatively few efficacious anti-infectious drugs, a record that, if left unchecked, will have catastrophic consequences for society. In 2007, GSK (PA, USA) reported an effort to utilize information about new targets that arose from sequencing the bacterial genome. The results of sequencing gave hope for the possibility of new classes of compounds that would arise from high-throughput screening (HTS) of the targets [4] . Disappointingly, out of 67 costly campaigns against specific targets only 15 resulted in hits and just five of those resulted in leads. This more than likely had the effect of discouraging further research in this area and various theories were used to explain these poor screening results. One explanation was that the biological targets screened were not ideal for modulation by a small molecule and were therefore ‘undruggable’. An alternative explanation posited by the research team was that they lacked the appropriate chemical matter to modulate

Future Med. Chem. (2014) 6(17), 1927–1942

part of

ISSN 1756-8919

1927

Review Comer, Duvall & duPont Lee

Key terms Malaria: Malaria is a major public health threat responsible for high mortality and morbidity burdens in endemic countries. It is caused by parasitic protozoans of the Plasmodium type. Diversity-oriented synthesis: Diversity oriented synthesis (DOS) is a strategy for the efficient generation of a set of molecules diverse in skeletal and stereochemical properties. Trypanosoma cruzi: Chagas disease is a tropical disease caused by the parasite Trypanosoma cruzi (T. cruzi), which is endemic to Central and South America. Approximately 13 million people are infected with the parasite, and 25–30% of infected patients suffer from irreversible damage to the heart and digestive tract resulting in disability and death within 20 years of infection. Orthopoxviruses: Orthopoxviruses are a class of DNA virus in the Poxviridae family. Well known examples include vaccinia, monkeypox virus and variola virus (the causative agent of smallpox). Hepatitis C virus: Hepatitis C is an infectious disease affecting primarily the liver, caused by the hepatitis C virus (HCV). While primarily asymptomatic until many years later, chronic infection will often lead to cirrhosis and eventually liver cancer or liver failure.

the targets, leading to the belief that a better outcome could have been achieved with the right compounds. Interestingly, the researchers did point to the lack of chemical diversity and compounds containing desirable chemical properties within their screening collection. They also noted “synthetic screening collections of different companies probably have substantial overlap, due to the use of similar chemical synthetic methods and acquisition of compounds from the same vendors. New sources of compounds are needed.” Although it should be pointed out that a recent analysis of two major pharmaceutical companies showed low overlap of their screening collections [5] . It is also clear that simply modifying existing drugs will not solve the long-term problem of infectious disease resistance. Resistance to an individual anti-infective frequently predisposes resistance to the whole class of compounds, as the underlying resistance mechanism will still be present [6] . This point can be appreciated when we take into account the fact that bacteria, for instance, are evolving on a timescale of hours [3] . This also has consequences for the use of privileged scaffolds as sources for antimicrobials. The latter strategy generates libraries based on biologically active motifs or core structures of natural products. For example, OZ439 is a synthetic antimalarial in clinical trials, which has the same endoperoxide bridge that gives artemisinin its potent activity [7] . While this compound addresses many of the liabilities of its parent, it is not unreasonable to assume that OZ439 might be less effective against artemisinin-resistant parasites should they arise.

1928

Future Med. Chem. (2014) 6(17)

Natural products have long been the main source of antimicrobials and most of the major classes of antibiotics in therapeutic use are natural products or analogs thereof. This is intuitive given that most natural products arise from microbial warfare and have undergone rounds of optimization via natural selection. Despite this, problems associated with using natural products as drug leads has gradually led to many pharmaceutical companies moving away from this source. The isolation, identification, optimization and manufacture of natural product-based drugs can be extremely challenging and truly practical syntheses are still uncommon. The latter point is especially relevant in neglected diseases, such as malaria and tuberculosis, which are epidemic in resource-deficient areas and where keeping manufacturing costs low is paramount. On the other hand, there are almost certainly many more as yet undiscovered natural products that would serve as useful antibiotic leads, and so the field runs the risk of missing out on a vast number of biologically valuable chemotypes. What is needed is a systematic and efficient strategy to discover compounds that exert their biological effect through unique mechanisms. HTS of synthetic compound collections in phenotypic assays is an attractive approach towards this goal. However, not all synthetic compounds are alike and sources of small molecules for HTS tend to consist of commercial vendor libraries, which have become popular as they are readily obtainable at low cost [8] . Most corporate compound collections are heavily populated with flat, sp2-carbonrich heterocycles that follow Lipinski’s ‘rule-of-five’ criteria and have minimal chemical complexity. These may not be the best starting points for new antimicrobials and it is worth considering that the majority of clinically used antibiotics are topographically complex compounds that fall outside of Lipinski’s rules [9] . Diversity-oriented synthesis (DOS) is a strategy for the efficient generation of compound collections with a high degree of structural diversity. DOS compounds have structural features such as high ratios of sp3-hybridized atoms, greater number of stereogenic elements and greater diversity of core skeletons relative to commercial collections. For this reason diversitybased screening collections may well have a greater role to play in future infectious disease programs as resistance grows to traditional chemotypes. It has been suggested that DOS molecules occupy the middle ground between the structural complexity of natural products and the efficiency of synthesis of commercial libraries [8] . The build/couple/pair (B/C/P) synthetic algorithm offers a systematic process for obtaining stereochemically and skeletally diverse (DOS) products in relatively few steps [10,11] . In addition to providing new

future science group


chemotypes, there is evidence that compounds from diversity collections have increased hit rates across several targets and increased selectivity for a given target [12] . It has also been suggested that the likelihood of developing an oral drug candidate increases with fewer aromatic rings [13] . Additionally, increasing sp3 count correlates favorably with decreasing clinical toxicity, the foremost cause of drug attrition in the clinic [14] . Despite the concept of DOS being relatively new, it has seen remarkable progress in the field of antimicrobials, which gives reason to believe it will have a major impact in this area. In the remainder of this review we focus on a few noteworthy examples of biologically active compounds derived from DOS pathways. We will concentrate initially on parasitic diseases with examples from Novartis (Singapore and CA, USA) and Broad Institute’s (MA, USA) advances in targeting malaria and Trypanosoma cruzi. The second section looks at DOS-based antibacterials and specifically anti-methicillin-resistant Staphylococcus aureus (MRSA) agents. Finally, we look at the successes in inhibiting orthopoxviruses and hepatitis C virus (hcv) replication using DOS compounds. DOS-based antiparasitic compounds The parasite T. cruzi is the underlying cause of Chagas disease, an ailment that has been continuously neglected despite affecting millions of people [15,16] . Chagas disease is mainly isolated to South America, where the transporters of the parasite, triatomine (also known as kissing) bugs reside. In the acute phase of Chagas disease, T. cruzi enters the system through broken skin and mucous membranes and non-lethal symptoms, such as fever, occur and subside on their own. Untreated, however, the chronic infection can persist and lead to more complicated symptoms, such as heart arrhythmia and heart failure. The parasite exists in multiple stages – the trypomastigote stage occurs during the infection, however, the more clinically relevant amastigote stage allows for multiplication of the virus. Current therapeutics for treatment of Chagas disease include benznidazole and nifurtimox, both of which have severe side effects and are less effective against the more harmful chronic phase. Based on this unmet clinical need and emerging resistance to the current therapeutics, novel treatments are needed. To this end, a high-throughput screen was run at The Broad Institute in an attempt to find novel chemical matter to inhibit the growth of T. cruzi [17] . The broad institute’s DOS collection of 100,000 small molecules was used in the HTS campaign, a phenotypic screen using a recombinant strain of T. cruzi cocultured in mouse fibroblast NIH/3T3. Due to the multimode assay format, growth inhibition of T. cruzi


Review

can occur in at least three different ways: inhibiting growth at the trypomastigote stage; inhibiting the invasion of T. cruzi into the cell; or, inhibiting growth at the amastigote stage (most clinically relevant). Compounds that were scored hits in the primary assay were repeated in dose in the primary assay format as well as against the host cell (fibroblast NIH/3T3) to remove any compounds with general toxicity. Triaging the 100,000 compounds through this screening cascade led to the discovery of a novel inhibitor of T. cruzi, having nanomolar activity in the ‘multimode’ assay described above [17] . The compound (Table 1) is a fused eight-membered lactam with three stereocenters, of which all eight possible stereoisomers were screened and showed some level of activity in the growth inhibition assay. Stereoisomers with a trans relationship at C5 and C6 (RR or SS, 1–4) were highly potent (1–5 nM), while stereoisomers with a cis relationship at C5 and C6 (5–8) showed moderate activity (16–450 nM). Since there was no strong preference for a single stereoisomer based on the activity profile, initial in vitro absorbtion, distribution, metabolism and excretion profiling was done. Interestingly, the C5C6 -trans stereoisomers showed a nearly 100-fold reduction in solubility (50

3

SM_A5B3_2P141

I

O(CH2) 4OH

H

25.7

>50

4

SM_A6B5_2P100

Br

Ph

CH2OH

6.6

>50

5

SM_A15B5

I

Ph

CH2OH

3.5

27.1

50

(∈…uM)

Reprinted with permission from [52]© American Chemical Society (2014).

1938

Future Med. Chem. (2014) 6(17)



While CMLDBU-6128 was unable to prevent host gene expression shutdown (a normal occurrence post-infection as viral gene expression rises), viral protein synthesis was generally reduced (Supplementary Figure 4A) . Analysis of reporter gene expression post-infection when administered at different time points determined that CMLDBU-6128 could arrest established late gene expression when administered 10 h post-infection. Also, interestingly, when CMLDBU-6128 was administered at T = 0 and then washed away at 10 h, there was no subsequent increase in reporter gene expression suggesting that viral recovery from an arrested infection does not occur. Finally, resistant mutants were generated through serial passage of the LV virus under CMLDBU-6128 exposure and sequencing of drug-resistant clones indicated a single coding change in J6R, which encodes the large subunit of RNA polymerase. It was hypothesized that since J6R is highly conserved across orthopox viruses, it was likely that CMLDBU-6128 would have cross-species activity. Indeed, in single-cycle growth assays of A549 cells infected with monkeypox virus Zaire 1979, cowpox virus Brighton red and vaccinia virus strain IHDJ, CMLDBU-6128 inhibited viral replication of all, reducing viral yield from 2.4 to 4.0 log units. The data suggest that CMLDBU-6128 has broad-spectrum activity against Poxviridae. It is estimated that between 150 and 200 million people worldwide are infected with HCV. While primarily asymptomatic until many years later, chronic infection will often lead to cirrhosis and eventually liver cancer or liver failure [46] . The primary form of treatment is pegylated interferon with ribavirin, often in concert with the protease inhibitors boceprevir or telaprevir. These therapies typically cure 50–80% of patients, but the treatment has a highly undesirable side effect profile and carries risk of infection due to immunosuppression [47] . The recently approved Sofosbuvir, an orally bioavailable nucleoside analog, represents a major step in the treatment of HCV infections [48] . It has a high cure rate and is a component of the first interferon-free treatment regimen, which improves the side effect profile. Some concerns remain, however, especially given that the cost of treatment is reported to be in excess of US$80,000 [49] . Since no vaccine for HCV is available, the high cost of treatment is still a barrier to distribution given the dearth of alternative therapies. This suggests that continued development and competition in this field should remain a priority. As part of their continued effort to identify compounds with high potency and improved side effect profile for treatment of HCV, a team of researchers from across the globe recently reported the screen-


Review

ing of a DOS compound collection for inhibition of HCV replication [50] . This HTS followed up some earlier work in which the researchers successfully identified anti-proliferative compounds from a collection of known bioactives, by screening against a cultured Huh7/RepFeo replicon cell line for viral RNA replication [51] , and similar methods were used for this screen. The compound library used by this group is a metacollection represented by more than 8000 compounds each from more than ten separate DOS collections obtained from various groups around the USA and Canada. The HTS identified more than 40 anti-proliferative hits, which were counterscreened for toxicity using Promega’s standard CTG cell viability assay. Interestingly, the researchers noted that half of all the hits from the primary screen contained an epoxide moiety, and notably these epoxide-containing hits came from two distinct libraries. The first library to provide hits was the BUCMLD epoxyquinol library [52] , whose members represented the most potent antivirals identified in the screen (Figure 7) . Secondary screening revealed that BUCMLD-B10A11 was the most potent hit compound (EC50