Antiviral Research and Development Against Dengue Virus Bruno Canard, PhD. [email protected] 1 Table of Contents Part 1. Antivirals 3 A short historical view on antiviral research and therapies 3 Lessons learned from recent viral diseases and pandemies 3 The methods used to discover antivirals 4 Infected cell assays 5 Knowledge-based methods 5 The source of anti-infectious molecules 6 Why has natural product screening been neglected in antiviral research ? 8 Challenges associated with natural products in antiviral research 8 What is a validated antiviral target ? 9 Animal models 9 Patient cohorts and clinical trials 10 Frequent arguments about antiviral therapy feasibility 10 The introduction of dengue as a druggable disease 11 Diagnostics, and what does it tells us for antiviral therapy ? 11 Current treatment 11 Part 2. Dengue 13 Preamble 13 The Dengue Virus 13 The DENV targets for antiviral research 13 Overview of genome organisation 14 Overview of the DV particle and DV proteins as targets for drugs 14 The structural proteins 14 The Non-Structural proteins 15 RNA structures 17 The dengue validated targets 17 The cellular targets for antiviral research against dengue 18 siRNAs as tools and/or therapeutic agents 19 Response modifiers 20 Monoclonal antibodies 21 Mechanical devices 21 Part 3. Academic and academy-associated research centers 22 Part 4. The current industrial network of AV discovery 31 Part 5. Mapping the dengue drug design effort and needs 38 Annex 1. References 42 Annex 2. Patents 43 2 Part 1. Antivirals A short historical view on antiviral research and therapies The first significant successes of anti-infectious disease treatments originated from the discovery and use of antibiotics. The discovery of many viruses preceded largely the discovery of the first antiviral molecule, which occurred at least 40 years after that of penicillin in 1928. The first documented description of an antiviral molecule, that of 5-iodo-2'-desoxyuridine, occurred in 1959. It was discovered active against Herpes ophthalmologic infections and followed by a series of related active molecules. The fight against herpes was the perhaps the earliest and most significant driving force of antiviral research. Herpes was the only significant viral disease for which all technical elements and systems required to develop an antiviral molecule first became available (i.e., in vitro infected cell systems, animal models, chronically infected patients,…). The antiviral drug field came of age in the next decades with the first antiviral molecule finding its way to the clinic: Gertrude B. Elion discovered acyclovir(2) a scientific breakthrough for which she was later awarded the Nobel prize in 1988. The subsequent emergence of AIDS in 1981, and the following pandemics drastically changed the field of antiviral research, allowing the widening of concepts, technical developments, rules, and business. Lessons learned from recent viral diseases and pandemies HIV and HCV: chronic invaders The most important lesson comes from the following great achievement: it is possible to control a chronic infection of a very sophisticated virus, such as HIV, that hides inside the chromosomes of the infected cell. Although the victory is not total yet, it has profoundly changed the fate of the pandemic victims, at least in western countries. After being inspired by other research fields, anti-HIV research has “infected” other field of antiviral research and will continue to do so. Remarkably, after the identification of HIV, the control of HIV through antiretrovirals originated from a collective effort on a wide variety of scientific and medical fields, including efficient transfer from academia to the corporate world. More recently, hepatitis C virus (HCV) research is now boosting the antiviral chemotherapy field. Viral polymerases and proteases are targets par excellence, validated by the use of inhibitors against HIV reverse transcriptase and protease, hepatitis B polymerase, and herpes virus polymerase. Anti-HCV protease and polymerase inhibitors are in various stages of clinical trials. Novel targets and cognate inhIbitors are adding to the list, such as the HIV integrase, and the HCV NS5A. HCV (genus Hepacivirus) and DENV (genus Flavivirus) belong to the same viral family Flaviviridae sharing similar genome organization and replication strategies. Initially, research conducted on dengue virus (DENV) was the actual starting and inspiration point for HCV research, when it became known 3 that HCV had a flavivirus-like genome. Presently and conversely, knowledge and strategies gained from the successful drug discovery and design process against HCV can now be translated back to the DENV research field. SARS and Influenza (H5N1 and H1N1): “hit and run” viruses The SARS pandemic was due to a novel coronavirus which emerged in 2003 from China. The virus took the world by surprise as coronaviruses were not known to cause life threatening pathologies. Coronaviruses were clearly neglected viruses from the scientific and the medical/veterinary point-of- view. The pandemic revealed blatantly our unpreparedness to such a problem: point-of-care in hospitals crowded with contagious patients, high toll for clinicians, tracing secondary contacts of taxi drivers and plane passengers, etc…The pessimistic say that nowadays viruses travel around the world in 3 days. The optimistic say that social networks and cell phones make information travel much faster. Perhaps the true challenge is elsewhere: making people believe and adhere to an “official” information, as exemplified with the recent H1N1 crisis and the unsuccessful vaccination campaign. In any case, this crisis has been the best advocate for antivirals as a complementary strategy to prEvention and vaccination. In the case of influenza, the size of the market has been the main booster of anti-influenza drug development. This includes the availability of patients for clinical trials, and the fact that a potential devastating pandemic would undoubtedly provoke stockpiling of antivirals in the time-window into which an appropriate vaccine would available. Advice to stockpile anti-influenza drugs has been recurrently advertised, mostly after 1995 when the 1918 spanish influenza strain genome was published(8). well before the H5N1 and the H1N1 fear hit the world. These two viruses do not produce chronic infections. These types of virus produce an infection (unnoticed, mild, or acute) which resolves with virus clearance. This transient nature of the infection has long been a problem to design an efficient therapeutic answer. Indeed, there are too many unpredictable parameters to build a drug-design program based on traditional planning and funding approaches. The two biggest problems are that it is impossible to evaluate precisely the market (and invest accordingly), and that there is an unpredictable number of patients available for clinical trials. The instructive aspect of these pandemics, however, is that they greatly contributed to re-shape antiviral research at large (how can we anticipate? how money is going to be invested? ). These recent crises have shaped considerably the grand public opinion towards the necessity to have broad- spectrum anti-influenza drugs ready. The methods used to discover antivirals The original method of discovery of antivirals was partially a knowledge-based method, centered around nucleobases and nucleosides (eg., uridine derivatives mentioned above against Herpes), known to be used by viruses for their replication. The advent of AIDS and the discovery of non-nucleoside 4 reverse transcriptase inhibitors opened the era of large-scale screening, which is entirely a trial-and- error procedure, not based on previous knowledge. Millions of compounds are tested as fast as possible (using high throughput screening (HTS) techniques), and only those showing activity are selected. Infected cell assays In both cases (Herpes and AIDS), infected cell cultures provided the antiviral read out, before purified targets were available and could be used. In these assays, compounds are tested individually to see if they either cure an infected cell, or protect it from infection, pathogenic effects. The process is simple, and relies on a cell culture system able to support virus growth. Not surprisingly, the discovery of antivirals parallels the establishment of a robust infected cell based assay. When this was difficult or even not possible (eg., HCV), the use of sub-genomic replicons or surrogate viruses has nevertheless allowed drug discovery and design. There are now a wide variety of assay systems specific for each virus. Robust dengue infected cell assays are available, highly efficient in terms of characterizing the potency of a drug candidate. One significant disadvantage is the cost associated with cell culture reagents and facilities, especially in low income countries. However, this method has an impressive record of success compared to other methods. Knowledge-based methods The general trend is to reduce this trial-and-error approach and inject knowledge as much as possible in the selection process so as to reduce costs and increase efficiency. Computer-aided structure activity relationship (SAR) studies facilitate a responsive and efficient management of research results and programs. Drug-resistance must be considered as part of the drug- design process, as drug resistance mechanisms are being increasingly characterized and drug combinations optimized, in order to avoid or delay resistance. The first large-scale effort to discover anti-DENV drugs is to be credited to the Novartis Institute of Tropical Diseases (Singapore), who conducted a complete screen of their proprietary chemical library against the DENV protease domain from non-structural protein NS3 (see below). Knowledge-based methods differ from classical cell based screening techniques in that they use screening or discovery systems characterized at the molecular and sometime atomic levels. The discovery system represents or approximates a given step of the virus life-cycle. The knowledge associated with the system reduces the number of putative targets, and is supposed to provide directly a mechanism of action of the compound or candidate drug. Examples of such systems are purified enzymes used directly in the drug discovery test. It is expected that inhibition of the enzyme by a compound in a test tube will mimic inhibition of the enzyme in the context of a viral infection. This is 5 of course not granted. For a good enzyme inhibitor, the most frequent reasons of failure to inhibit a virus in a cellular context are: • The compound does not penetrate inside the cell. • The compound is rapidly degraded/metabolized/transformed into an inactive compound • The compound is toxic and of poor selectivity, ie., when used in a cellular context, the compound will kill the infected cell and any direct effect on the virus is not apparent. Many compounds can be selected as good inhibitors of a viral enzyme. However, the majority will fail to convert into a candidate drug for the above reasons. However, the main advantages of the method are: • it discovers both a compound and its target at the same time. • Currently, increasing general medicinal chemistry knowledge allows a better pre-screening of compounds that have potential, ie., chemical libraries used as the source of molecules are each day better in terms of containing “drug-like” molecules. The future is the integration of both cell based assays and knowledge-based methods, to reduce the time involved in i) finding the target at the molecular level, ii) having a trustable molecular/atomic model to go quickly into hit-to-lead development by medicinal chemistry. The source of anti-infectious molecules Before their antiviral properties are discovered, antiviral chemicals or molecules either exist physically somewhere in the world (and are selected or discovered), or they do not exist, and are invented and subsequently synthesized. For molecules having a physical existence, they are either owned by someone, and generally organized in a chemical library (or repository), or they are in the wild, in plants, marine organisms, insects, etc…The issue of final ownership (ie., of a discovered molecule having interesting properties) is then much more complicated. This is an important distinction that has wide implications in drug discovery, from the ease of discovery to the final ownership and availability to patients. The source of antiviral molecules is indeed a key issue, particularly for dengue, for two main reasons: • The cost of a drug is going to be a main issue because dengue occurs majoritarily in low income countries. • Although the low income dengue-afflicted countries generally do not have screening and drug design facilities, most of the potential natural sources of drugs (mostly plants) are located in these countries. 6 Chemical libraries The availability of large collections of pure compounds that can be handled, tested, analyzed and whose compounds can be re-ordered has considerably evolved over the last decades. These large collections were initially exclusively found in large pharmaceutical companies. Over the years, these companies had accumulated compounds, assays, and know-how. The situation has drastically changed over the last decade, mainly because robotics and bio-chemo-informatics have penetrated academic modest-in-size labs and research structures. It has long been argued that screening was “an industrial job” best accomplished in a corporate setting, an observation that was true to a certain extent, because sophisticated robotics, engineering know how, and manpower was more easily mobilized there. The decreasing cost of screening-associated technology, the diversity of the screening needs (targets, pathways, organisms, pathologies,…), as well as the advent of proteomics and siRNAs (see below) has done that many labs have their own screening facility, often small scale, for the defined process or biological system they are studying. Likewise, many service centers and small companies are able to propose screening as a service. Many large chemical libraries can be bought, several million pure compounds are physically available to any purchaser. For a lab or company screening compounds against a given virus or biological system, the problem is more to have original libraries. Indeed, unique libraries that are not freely available minimize risks for a lab of being competed out, and simplify the intellectual property of discovered molecules. It looks that the tendency of screening very large collections of pure molecules is declining. This may be due to the fact that methods to pre-screen virtually these collections have evolved to a point where more focused libraries can be built, and small focused screens can be conducted on these “enriched” libraires. Likewise, the increasing availability of atomic models (mostly crystal structures) of targets make this preparation of enriched libraries much easier. It is certainly too early to draw conclusions about the justification of great hype and faith on HTS during the last two decades. There is a consensus to state that the number of drugs reaching the market has decreased sharply in this period of time, relative to previous periods where drug discovery relied much more on the screening and discovery of natural compounds (see below). The reasons are certainly more complex than a mere wrong direction of the whole drug discovery and design world. However, for a number of reasons, discussed below, the source of antiviral molecules will undoubtedly evolve towards more screening from natural resources, blending with a great deal of experience in high throughput techniques and medicinal chemistry expertise acquired in these past two decades. Natural sources Plants have been the traditional (and almost exclusive) source of active substances for most therapies. At the present time plants are the indirect or direct source of ~ 50 % of approved drugs. Anticancer drug research has been a leading force in natural product research and screening processes. From the 7 1940s to 2007, 73% of the 155 small molecules approved as anticancer drugs were from natural origin, directly derived from a natural product, or inspired by a natural product(4). To put antiviral research in perspective, in the 1940s, there was not a single molecule known having an antiviral effect, and the discovery and isolation of the first human viral pathogen was only 13 years away (Yellow fever virus, in 1927). It is not surprising that the concepts of natural product screening, established for cancer and inadequate for antiviral screening, had not entered the antiviral research field. Why has natural product screening been neglected in antiviral research ? First, many plant extracts are cytotoxic, a desired property for an anti-cancer drug. However, an extract that kills the cell does not allow the monitoring of virus growth or inhibition. One has to use extracts that are non-cytotoxic. By jeopardizing selectivity (ie, the ratio of inhibitory concentration for the virus over toxic concentration for the cell), cytotoxicity has stopped many compounds or extracts on their way towards antiviral pre-clinical trials. Second, when non cytotoxic crude extracts are used, almost all of them exhibit antiviral activity. This antiviral activity is due mainly to compounds that have no interest as antiviral drugs. These compounds are a wide variety of polymers, polyphenols, and tannins. Third, if one avoids the above traps, screening natural extracts yields a lot of true inhibitory molecules that are already known and characterized (and will fail in a composition-of-matter patent), and have low potential for chemical modification into a useful and unique pharmacophore. As consequence, during the past twenty years, the advent of combinatorial and parallel chemistry coupled with high-throughput screening techniques has led to a decreased emphasis on plants (or microbial, marine extracts) as a compound source. Nature has continued to inspire chemists and drug- designers during the development of natural product-based compounds (such as antiviral nucleoside analogues), but no natural product has actually been approved as an antiviral drug out of the 35 drugs approved up to 2002. The trend seems to change as at least 8 natural products are since in clinical trials in the field of virology (HIV and HCV), such as Calanolides A and B, DCK(PA-334B), 3,5-Di- O-caffeoylquinic acid, MX-3253, 4-Methylumbelliferone, Bevirimat, Sho-shaiko-to H09, and Sutherlandia frutescens(6). Challenges associated with natural products in antiviral research Whilst natural products as a source of drugs were falling out of favor of pharmaceutical companies, the interest of this source was growing dramatically in countries were these resources are located, ie., mostly low income countries of the developing world. The main incentive was the adoption of the Convention on Biological diversity, enforced in 1993. The challenges associated with this resource are technical and policy issues(3). On the technical side, the difficulty to deal with natural extracts has been developed above. There are now an increasing number of methods reporting how to prepare an extract suitable to specific needs, including antiviral research. Here also, technology has helped in the preparation of extract libraries 8 pre-cleaned from unwanted substances described above (tannins, polymers, …). A second difficulty is the variability of the source. Re-collection of the same plant may not give the same chemical composition of an extract (different season, different development stage, misidentification, etc…). The third difficulty is the resupply problem, particularly in large quantities. Over-harvesting may occur, although it is sometimes possible to find alternate sources of the compound (plant cell culture, other species, etc…). Last, isolation, re-synthesis of hemi-synthesis can be challenging, although science, technology, and know-how are advancing faster than policy issues. However, natural products collection and assays are located in the developing world, which is increasingly involved in finding primary activities of an extract. When the next step in engaged, large pharmaceutical companies having down-sized their natural product departments are not often ready to carry on. There is an increase need to build intermediate/small dedicated structures in the corporate of academic world. The policy issues associated with biodiversity exploitation address mainly the location of study of the natural product collection. Authorizations, contracts and agreements can vary from extremely slow to quite easy and diligent(3). What is a validated antiviral target ? A “validated target” is a cellular or viral component (protein, membrane, macroassembly,…) which, when bound to a drug, leads to virus control (growth inhibition, elimination, virostatic,…) in the infected cell, and hopefully, in patients eventually. Thus, any protein viral protein is not a validated target, for example if, when inhibited or destroyed, the virus can still grow using and alternating pathway compensating for the loss of its inactivated protein. Regarding the host cell proteins that can be used as targets (ie., cellular proteins required for virus growth), blocking those proteins must be both safe enough for the cell (ie., shown no toxicity at least for the uninfected cell) and the virus must be unable to use an alternate protein compensating for the unavailability of the blocked protein. Typical example are viral receptors. In many instances (eg., HIV), there are several cellular receptors for the same virus. Blocking one receptor forces the virus to rely on other receptors, sometimes with little effect on viral growth. Animal models In the case of dengue, the best animal model is the AG129 mouse, which has however a number of drawbacks such as low and short viremia. Efforts are ongoing, mainly in academic labs to improve this model. Surrogate systems exist that do not have these drawbacks (eg., the flavivirus Modoc virus), and chimeric dengue/modoc viruses may yield interesting systems. Also, other animal systems are being evaluated (golden hamster, macaques, …) with the increasing variety of dengue strains available. 9 Patient cohorts and clinical trials The first large scale clinical trial has been conducted in Viet Nam (chloroquine), and another is ongoing (see below, the Roche nucleoside analogue). This indicate that there is a sufficient and sustained number of patients to go into clinical trials for any promising molecule. Other sites for clinical trials will undoubtedly see the light, either in Asia or South America. Frequent arguments about antiviral therapy feasibility The most common argument opposed to antiviral therapy is that it would occur too late, ie., when the viremia is already declining, low, or the virus cleared. Diagnostic tests to rapidly detect DENV infection at an early stage (ie., early viremia) are currently available (see below), and it has been demonstrated that there is a direct correlation between high viral load and the development of the more severe, life-threatening form of dengue disease. The higher, the worst. Thus, a drug reducing viral load at an early stage would potentially prevent DF and DHF/DSS. Dengue viremia is short, being detectable only shortly before or concomitant to the onset of fever and lasts four to five days after. The ability to rapidly diagnose dengue disease is thus key to the successful implementation of antiviral chemotherapy. Although virus may not be detectable in plasma, viral replication may be occurring in other cell reservoirs, tissues, and body compartments where an antiviral drug could reach and target them. For example, it is suspected that in addition to plasma leakage, the life-threatening DSS may involve damage of organs such as the liver (DSS associated hepatitis) or heart (DSS-associated myocardiopathologies). In endemic outbreaks, prophylactic mass treatment around index cases would be essential. Rapid diagnostics would detect infected yet asymptomatic people. Another as yet unevaluated consequence of prophylactic treatment and decrease of viremia should appear in the vector-infection pattern. Decreasing viremia in humans should result in a decrease in infected vector population and thus impact on the transmission chain. Therefore, an efficient and safe drug, delivered early in the course of dengue disease, should not only save lives but also curb potential epidemics. An on-the-shelf drug allows a rapid response in the case of a sudden outbreak, and should not require cold storage, an advantage for use in developing countries. Lastly, in other deadly viral systems (Monkeypox virus as a surrogate system for Smallpox virus), 24h post exposure prophylaxis with the drug cidofovir has been shown to significantly reduce mortality of monkeys challenged with a lethal monkeypox infection, compared to 24h post exposure vaccination which had no effect(7). The second most common argument is that dengue occurs in low income countries where there is no sizeable nor predictable market. Up to now, it is true that the common models of drug development rested on analysis of an existing market, and financial planning and investment accordingly. It is 10