Development of Antiviral Drugs for Poliovirus
CHAPTER 2 OF THIS REPORT DELINEATED the importance of an effective antiviral drug in ensuring the eventual success of the polio eradication program. Chapter 3 identified a number of poliovirus targets at which drugs might be aimed. The goal of this chapter is to discuss the steps that a potential antiviral drug will have to pass through to be successful.
The first step is to identify what the drug must do and how it will be used, to write a “package insert,” as it were, for the drug before it even exists. In Chapter 2, the committee recommended that an antiviral drug be developed to be used principally as a prophylactic in the event of a poliomyelitis outbreak, with the capability of both preventing infection and preventing spread from those already infected. It identified a number of additional requirements:
Once, or at most twice, daily dosing
From the point of view of development, some of those requirements merit early attention. First, it is important to develop drugs that have low effective concentration (EC)—that is, are efficacious at very low plasma concentrations so that they can be administered in small doses. This will
both help to keep the cost low and will be likely to reduce safety risks. EC50 is defined as the drug concentration that inhibits 50% of virus replication; EC90 is the dose that provokes a response that is 90% of the maximum. The EC50 and EC90 should be in the low nanogram range. In addition to showing substantial antiviral activity, a molecule suitable for development must be amenable to large-scale, cost-effective synthesis. Difficulties in synthesis will be directly reflected in higher costs.
Chapter 2 further suggested that the most likely application of a polio antiviral drug would be in combination with inactivated poliovirus vaccine. If ring prophylaxis with a medication and vaccine is expected, the number of people requiring treatment would be exceedingly large and require large quantities of medication.
Successful development of an antiviral drug to prevent poliovirus transmission will require simultaneous attention to two challenges: active agents must be identified and optimized, and how the agent will be used must be carefully defined, because this will determine how it will be tested, regulated, and administered.
POTENTIAL HURDLES TO ADDRESS AT THE OUTSET
At the outset of a program designed to develop drugs to treat poliovirus infections, potential hurdles should be identified. Development of medications for therapy of poliovirus infections historically has not attracted support from the pharmaceutical industry. The committee does not envision industry support of development costs, so a fundamental question that needs to be resolved is: Who will fund the development of these medications? As clinical trials are envisioned, a parallel question will be: Who will hold the investigational new drug application? Those two questions should be addressed at the outset of any initiative. A credible clinical development plan that ensures the project could proceed beyond preclinical proof of concept will be critical to attract support for such a program. Because it is envisioned that these drugs will be used globally, it would be helpful to establish an international collaboration of responsible individuals, including regulatory authorities, at the outset of the program.
IDENTIFYING AND OPTIMIZING POTENTIAL POLIO ANTIVIRAL DRUGS
Exploitation of Existing Leads
Development of a polio antiviral drug can build on experience with other antiviral drugs. Furthermore, given that on average it takes $802 million and 10 to 15 years from start to finish to bring a drug to market (Frank 2003), the exploitation of existing, advanced leads may be crucial to this potential poliovirus antiviral effort (especially considering possible time constraints related to the eradication effort). The average cost and time stated above reflects many attempts and failures (and much basic research) for many diverse targets and indications. For poliovirus, the utilization of existing, advanced leads could be expected to reduce significantly the time and cost of development. A number of leads exist at various stages of advancement for poliovirus targets including the capsid, 3Cpro proteinase, and to some extent 2Apro proteinase and the RNA polymerase 3Dpol. Members of the picornavirus family—consisting of the rhinoviruses, hepatitis A, and a large number of potentially dangerous enteroviruses, including polioviruses 1, 2, and 3—are common but cause widely divergent clinical illness. Historically, efforts have focused on the development of antiviral therapeutics to treat rhinovirus infections, such as the common cold. Those efforts led to the elegant studies of structure-based activity of capsid-binding inhibitors, of which pleconaril is the best known example. Pleconaril was evaluated through Phase 3 clinical trials for treatment of the common cold. It had a modest degree of efficacy in the treatment of rhinoviral disease, but licensure by the Food and Drug Administration (FDA), as recommended by an advisory committee, was denied because of the induction of hepatic cytochrome enzymes that altered the metabolism of birth control pills and resulted in intramenstrual bleeding. As a consequence, the drug was abandoned for systemic administration.
Pleconaril has no significant activity against polioviruses. However, ViroDefense, Inc. has identified a member of the capsid-binding inhibitor class, V-037, that has significant in vitro activity against human poliovirus type 1 (Collett 2005). We present some of the details of the behavior of this specific molecule simply as an example of an existing lead in this target class for which a significant amount of development has already occurred. For human poliovirus type 1, in vitro cell culture activity at an EC90 was identified as less than 0.02 µM. V-037 is less active against human polioviruses
type 2 and 3. The strains tested in vitro were constituents of the oral polio vaccine. V-037 is also less active against wild-type polioviruses. Preclinical assessments demonstrate that it is Ames test-negative and chromosome-aberration negative. Given orally, it is bioavailable in a murine model and achieves concentrations in the central nervous system that are four to six times the plasma concentration. These levels were not corrected for protein-binding and thus do not necessarily reflect the level of free, or unbound drug. However, while protein binding is a fundamental concern in drug development, its application to the development of antiviral drugs is not well established. In vitro studies are performed using human blood, but their translation to probability of success in the clinic has not been established. For example, the highly protein bound bromovinyl arabinosyl uracil proved efficacious in the treatment of herpes zoster albeit it was not licensed for other reasons (Gnann et al. 1998). After cerebral inoculation in a murine model of poliovirus type 2, 100% of mice survived. The no-effect doses (the dose at which no toxicity is seen) were 100 mg/kg in the mouse and 750 mg/kg in the dog. At the highest doses, an engorged liver was observed. The latter finding is important because the lead molecule of this class, pleconaril, induced hepatic cytochrome activity and the finding may be a signal of potential toxicity. Other chemically distinct poliovirus capsid-binding ligands have been identified by ViroDefense and others (Andries et al. 1992, 1990), however, and would probably have different properties.
An enzyme essential to rhinovirus replication, 3Cpro proteinase, provided another logical target for the development of antipicornaviral therapeutics. That enzyme and its precursor polypeptide 3CDpro are responsible for proteolytic cleavage steps that are conserved across all entero- and rhinoviruses. A drug candidate, rupintrivir, was studied through Phase 1B/2 clinical trials with human rhinovirus infections but failed to demonstrate significant clinical activity. As a consequence, this lead molecule was abandoned. A second generation 3Cpro proteinase inhibitor has been developed by Pfizer and appears to have a pharmacokinetic profile that is satisfactory for the treatment of picornavirus infections; however, its activity against polioviruses has not been evaluated (Patick et al. 2005).
Existing capsid-binding and 3Cpro proteinase inhibitors are potential therapeutic leads in the development of therapies for poliovirus infections that already exist. The prior investigation of these two biological targets provides a well-established platform upon which therapeutics for polioviruses could be developed. In fact, the parallel development of two classes
of molecules with different mechanisms of action for the treatment of poliovirus infections is both indicated and desirable, as described below.
Identifying Leads by Screening Libraries of Molecules
Large libraries of molecules exist that could be screened against replication of polioviruses. Such libraries are available through the National Institutes of Health (the National Institute of Allergy and Infectious Disease [NIAID] and the National Cancer Institute [NCI]), the Southern Research Institute (Birmingham, AL), the pharmaceutical industry, and small biotechnology firms. A coordinated effort should be instituted to collect representative molecules of different classes for screening. In addition, target-focused libraries of known picornavirus capsid-binding and 3Cpro proteinase inhibitors and other similar compounds (identified through substructure and similarity searches) could be assembled and screened. The initial screening assays will probably involve live viruses and focus on plaque reduction. As screening efforts are developed, serious consideration should be given to which live virus strains should be used to demonstrate efficacy. Both vaccine-derived and wild-type polioviruses should be included. Initial assays can use whole virus replication with a plaque-reduction format. Ultimately, however, higher throughput assays should be developed to expedite the screening. Such high-throughput assays have been developed through sponsorship by the NIAID for other emerging infections, including West Nile virus and SARS coronavirus. Before the development of a suitable high-throughput assay, or if such an assay remains elusive, large libraries (up to millions of compounds) can be screened by computer analysis for binders to targets with known three-dimensional molecular structures (such as the poliovirus capsid or the 3Cpro proteinase), and relatively small numbers of molecules can be selected for experimental testing. Even if the three-dimensional structure of a target does not exist or the exact target is unknown, for any validated ligand (or series of ligands) it should be possible to screen large virtual libraries to identify other small molecules for testing that have overall shape and pharmacophore features similar to those of the query molecule.
As with other existing inhibitors, once molecules have been demonstrated to inhibit poliovirus replication in in vitro assays, mechanism-of-action studies should be initiated promptly. As a component of those studies, there should be attempts to develop resistant viruses. Such studies will further clarify the mechanism of action of these compounds and lead
to the synthesis of second and third generation molecules. For advanced leads, selective indexes (such as efficacy and toxicity) should be determined.
As an example of a possible screening paradigm, the important role that FDA’s Division of Antiviral Therapy played in identifying potential molecules to screen for anti-SARS activity should be considered. At the time of the SARS epidemic, FDA formally put out a request to industry and academic sources to supply compounds for testing that had some expected likelihood of being active against the therapeutic target. A number of companies participated in the effort. With the availability of known inhibitors and three-dimensional structures of at least two potential targets for poliovirus—the capsid and 3Cpro proteinase—a collaboration with FDA could lead to the identification of companies that have potential lead molecules.
The appropriateness of laboratory facilities to perform in vitro susceptibility testing and animal model studies also needs to be considered. With the imminent global cessation of polio immunization, the potential biological threat of an accidental or deliberate exposure of susceptible individuals to poliovirus should be recognized. Therefore, biocontainment laboratories of an appropriate level will have to be used.
Finally, a source of support for the screening of libraries should be identified at the outset of the endeavor. Historically, NIAID, through its Division of Microbiology and Infectious Diseases (Division of Virology), has initiated contracts for screening of antiviral drugs. Negotiations with NIAID should be entertained. Alternative sources of support from public and private foundations should also be considered.
Optimization of Lead Molecules
Medicinal chemistry is a fundamental component of any drug development program. Once lead molecules for potential poliovirus antiviral drugs have been identified, a medicinal chemistry program will need to be initiated to refine the molecules to optimize inhibitory properties. Contract medicinal chemistry companies (of which there are many with substantial expertise in the United States and abroad) could be employed to optimize lead molecules further. Fundamental to that effort will be structure-activity-based chemistry and the determination of structure-activity relationships among the active molecules. Given that three-dimensional x-ray crystal structures are available for both capsid-binding and 3Cpro proteinase inhibitors bound to their targets (Lentz et al. 1997; Hiremath et al. 1995,
1997; Grant et al. 1994), structure-based drug design approaches could be applied to expedite further development of lead molecules for these targets.
Criteria in the ongoing evaluation of molecules for development as drugs would include the ease and scalability of the synthetic route for a candidate drug. The “cost of goods” and complexity of manufacturing of any molecule as a medication will be directly reflected in its cost per dose. Manufacturing cost will be important in the development of any drug to treat poliovirus infections. In addition, the shelf life of the molecule will be important, given that it would probably be a component of the medical repository in place in the event of an outbreak of poliovirus infection. Finally, although the ability of any such molecule to cross the blood-brain barrier would be advantageous because it could potentially also be used to treat the symptoms of a neuropathic poliovirus infection, it is not a requirement, inasmuch as the drug will be developed primarily as a means of preventing the spread of infection.
The resistance profile of these molecules should be evaluated in depth early in development. As noted above, such knowledge can result in second and third generations of molecules that have a greater propensity to slow the emergence of resistance in addition to other improved characteristics.
As lead molecules are modified, the cost of synthesis, their stability and their propensity to elicit resistance will need to be kept in mind.
At the outset of polio antiviral drug development, the clinical development pathway should be defined. A key component in the development of potential therapeutic molecules will be a detailed consideration of whether vaccine-challenge studies or transmission studies can be performed. If efficacy studies cannot be readily performed in humans, the use of the FDA “Animal Rule” will need to be considered and discussed with the agency. The Animal Rule requires that an animal model system reflect human infection. At the present it is unclear if the FDA would be willing to license a therapeutic directed against polio utilizing the Animal Rule; the agency’s position on this matter has not yet been sought. Should the Animal Rule be adopted, three models could be considered. The first employs CD155 expressing transgenic mice. The second utilizes the humanized mouse model. Both of these models require validation. In addition, classically, the monkey has been used to test virulence of polio vaccines. This model might well serve as one to test therapeutics of poliovirus interventions.
Even if the Animal Rule is not utilized, animal model studies can be of immense value in assessing how an antiviral drug could influence disease pathogenesis and should be instituted in parallel with drug development. Knowledge of metabolism, distribution, and clearance of a drug in an infected animal model are all of importance.
If the Animal Rule is not adopted, human efficacy trials will need to be considered. These could include chronic shedders (i.e., immunocompromised hosts), but these individuals are uncommon and will be difficult to define. A more realistic choice would be OPV challenge models in either naïve or distantly immunized subjects. Since OPV is used in many countries, the duration of viral shedding would be an acceptable endpoint (placebo controlled).
Clinical trials thus far envisioned might involve populations in which mass vaccination has taken place (such as India) or in challenge studies of individuals not at risk (such as IPV-immunized individuals without children). In both cases, assessment of efficacy would depend on determining the effect of therapy on shedding of virus in the stool. A key component in the performance of a clinical trial will be the effect of medication on the quantity of virus in the stool. Investigative efforts designed to determine the quantity of virus excreted in the stool will need to be reassessed, and it will be necessary to determine the average amount of virus necessary to infect a contact.
Regardless of the regulatory path taken, safety would need to be established in a significant number of human subjects. The preclinical toxicology profile should anticipate a maximal period of administration, at least in the presence of an outbreak, as identified in Chapter 2, of 6-8 weeks. It is envisioned that a therapeutic would be administered to prevent transmission of poliovirus in a community and probably be co-administered with inactivated polio vaccine. Thus, the molecule must be bio-available if given orally once or twice a day, must be safe for children as well as adults, and ideally will have a toxicity profile acceptable for populations who have an inherent medical disability (for example, malnutrition or co-infection with other infectious agents).
The clinical trial development of molecules to inhibit human poliovirus replication should be considered in a traditional format. Phase I clinical trials will include normal human volunteers to assess pharmacokinetics and safety. Initial studies will be performed in adults; however, once safety has been established in adults, pediatric pharmacokinetic studies should be initiated as promptly as possible. That point cannot be emphasized enough
because it is envisioned that medication will be deployed primarily to pediatric populations. In the pharmacokinetic studies, special populations need to be considered, including pregnant women, immunocompromised hosts (such as HIV-infected people), and populations peculiar to developing countries (such as malnourished individuals).
As noted above, for the controlled clinical trial to be realistic for licensure, the “package insert” for the medication should be written before the clinical trial investigation. To that end, the primary use of a drug in treating poliovirus infections will be to prevent transmission from infected people to the susceptible individuals surrounding them. A potential secondary use would be to treat the very limited number of immunocompromised people who have been identified as chronic shedders of poliovirus. These individuals could be an ideal population for testing “the proof of principle” of the efficacy of the medication even if the ability to clear infection from them is not the ultimate goal, or measure of success, for these drugs.
Clinical trials could explore a wide array of populations, so discussions should focus on feasibility. To expedite the process, a product development committee should be established, including clinical investigators, representatives of the regulatory authorities, and representatives of industry. An important component of the product development committee will be representation of the countries involved; this is essential to ensure the ethical conduct of clinical trials.
THE IMPORTANCE OF DEVELOPING MORE THAN ONE ANTIVIRAL DRUG
The identification, optimization, and testing of drugs for the treatment of poliovirus infections is a complex process that is inherently unpredictable. Any potential lead could be proved unsuitable at any stage of the process. Furthermore, the infidelity of poliovirus replication potentially could lead to the emergence of viruses that are resistant to any individual therapeutic approach. Thus, simultaneous development of more than one drug, ideally substances that have alternative mechanisms of action, is the most prudent approach to a drug development campaign. Advanced leads that act against two targets—capsid-binding agents and 3Cpro inhibitors—are already identified and may therefore be most likely to yield success as a first approach.
Longer-term opportunities for the development of therapies directed against poliovirus infections require fundamental research on known critical
components in their replication cycle, many of which were identified in Chapter 2. The development of any of these targets would probably result in selective and specific inhibitors of viral replication but would require a long-term investment of 10-15 years. Thus, as an immediate strategy for the prevention of poliovirus transmission, those targets are less feasible than the near-term projects. However, continued basic research on poliovirus and continued monitoring of non-small-molecule approaches to antiviral treatment would constitute a wise investment and would mean that additional candidate molecules and antiviral approaches would be in the development pipeline in case the early approaches prove disappointing.
TIMELINES AND COSTS
One issue that deserves careful consideration is whether a polio antiviral drug can be developed in time to contribute to the global eradication program. If current plans are followed, OPV will continue to be used for up to 6 years after the last wild poliovirus infection is detected. This 6-year period is envisaged as being composed of two 3-year intervals. When 3 years have passed since detection of the last known case of wild polio infection, OPV use will be simultaneously ceased worldwide. During the following 3 years, plans call for intensive surveillance and rapid response to any cVDPV outbreak with monovalent OPV. As discussed in Chapter 2, the committee does not expect a polio antiviral drug to have great utility if the means of outbreak control is a live vaccine. Therefore, even under the most optimistic scenario, if the last case of wild poliovirus were to occur this year, it will be at least 6 years before the need for an antiviral might become apparent. If, as currently expected, it will be at least another year before wild poliovirus transmission can be interrupted in Nigeria, the time available for antiviral development stretches to 7-8 years.
The development of timelines for medications to treat poliovirus infections can be summarized briefly. Progress has already been achieved with ViroDefense’s V-037, but it remains to be determined whether it is fully optimized and free of toxicity for administration to large numbers of people. Optimization to achieve an acceptable risk:benefit ratio will be essential. Almost certainly, Pfizer’s Compound 1 will require further optimization; however, that too remains to be determined and obviously will depend on its untested activity in vitro against poliovirus replication. All other targets require further development, which will require a substantial investment of time and effort, probably at least a decade.
ViroDefense estimates that preclinical toxicology and good manufacturing practices (GMP) production of V-037 will take about 18 months. Clinical trials from Phase I through Phase III will require at least 3-4 years. Depending on the rapidity with which studies can be performed in the adult population, overlapping safety studies in children will need to be performed; however, a timeline for these studies cannot yet be assessed. Thus, if development were to proceed immediately, the ViroDefense compound could be ready in less than 6 years, with Compound 1 perhaps taking slightly longer. The committee recognizes that these are estimates and should be more carefully assessed before proceeding, but these estimates suggest that the timelines for drug development and the currently planned eradication program are not necessarily incompatible.
The development of “near-term” molecules would not only probably be faster, but will be less expensive than that of “long-term” molecules that require the identification of new targets. The committee found it difficult to assign costs because of distinct differences in the near versus long term potential development of molecules. Should a near term candidate be deemed viable for development, the cost to an Investigational New Drug (IND) application is estimated to be $5 to $8 million (Collett 2005). These costs would include GMP production, preclinical toxicology by a contract house (single dose, maximally tolerated dose (MTD), dose escalation studies, animal absorption, distribution, elimination and metabolism). Standard protocols would involve preclinical toxicology under the umbrella of 7, 14, and 28 day and 3 month drug administration. The end product would be a document that would serve as an IND application.
With the successful award of an IND from the FDA, a Phase I study could be initiated in normal human volunteers and would need to include both pediatric and adult pharmacokinetic studies, including drug interactions, and evaluation in high-risk populations. These studies typically recruit 30 to 50 patients, are managed by contract research organizations and are labor intensive. A typical study would be estimated to cost $5 million. In general for any compound taken forward, detailed efficacy-defining studies leading to registration would cost $20 to $40 million each. Experts present at the workshop estimated the cost of developing a near-term compound at $75 million over 5 years. In agreement with this number, one published report estimates the average out-of-pocket clinical period costs for investigational compounds to be $60.6 million in 2000 dollars (DiMasi et al. 2003).
For long-term targets, NIAID has developed partnership grants with
small biotechnology companies. The grants have averaged about $20 million, including co-support from industry, to allow a molecule to get to the IND stage.
Importantly, all of the above calculations are estimates predicated upon the rapidity at which an existing lead molecule can be advanced toward an IND. Should a previously unexplored target be chosen for drug screening, the costs assuredly will be greater. In such circumstances, the molecular biology required for drug development would significantly increase costs.
The projected cost of medication development includes only the cost for licensure and not for the subsequent manufacture and storage of the large number of treatment doses that might be required to respond to an outbreak. The development of a drug to treat poliovirus infections is not expected to be of great interest to the pharmaceutical industry. Therefore, support for this program will have to come from alternative funding sources. They could include the Bill and Melinda Gates Foundation or a new foundation dedicated specifically to the development of a polio antiviral drug as a component of the global polio eradication program. The Rotary Club does not have such a foundation in place, but it has invested over $650 million in the eradication of polio and might regard the development of an antiviral as a good backup to ensure the eventual success of its investment. Finally, consideration of support from the National Institutes of Health, the Centers for Disease Control and Prevention, and the World Health Organization is essential.