Assessment of Future Scientific Needs for Live Variola Virus (1999)

The most compelling reason for continued retention of live variola stocks is the identification and development of antiviral agents for use in the event of a large outbreak of smallpox. As noted earlier, virtually the entire human population is now susceptible to smallpox. Depending on the size, density, and mobility of the exposed population and the means by which the virus was introduced, millions of individuals could quickly become or be at risk of becoming infected.

There is no known drug therapy available that is effective in the treatment of smallpox. There is also no diagnostic test capable of detecting infected individuals during the incubation period preceding clinical symptoms. Vaccination can reduce the seriousness of the disease if administered within 4 days of infection [1]. However, the current U.S. vaccine supply is limited and may be deteriorating. Moreover, if an outbreak were to occur in a highly mobile population, widespread immunization within this narrow time frame would be logistically challenging. Antiviral medications with prophylactic and/or therapeutic properties—especially those that were safe, could be mass produced, exhibited good shelf life at ambient temperature, and could be taken orally—would therefore be critical in dealing with a large-scale outbreak of smallpox. Access to live variola virus would make it possible to test the activity of candidate antivirals in cell culture and could ultimately lead to the development of novel animal models systems for testing antivariola activity in vivo. It must be recognized however that neither of these preclinical approaches is able to substitute completely for trials conducted in infected patients. The combination of extensive preclinical studies with clinical pharmacokinetic data nevertheless provides a credible strategy for estimating the best dosing regimen to combat an outbreak of smallpox in unvaccinated humans.

The discovery of antiviral agents is currently an active field of research in academia and in the biotechnology and pharmaceutical industries [27]. More than 20 new chemical and biological agents have received U.S. Food and Drug Administration approval for treatment of human viral diseases in the last 10 years. Numerous compounds, recombinant proteins, and monoclonal antibodies are currently under active investigation as antiviral or immune-boosting agents. Some but not all antiviral agents have been tested for their ability to prevent variola virus infection of cultured cells in the Biological Safety Level 4 (BSL-4) biological containment facilities at the Centers for Disease Control and Prevention (CDC) in Atlanta or the State Center of Virology and Biotechnology (VECTOR) in Kotsovo.^*

Cidofovir, for example, is an antiviral, initially developed as a DNA polymerase inhibitor for the treatment of cytomegalovirus (CMV) retinitis, then found to be active in preventing variola infection of cultured cells (see also Chapter 5). While cidofovir's low oral bioavailability and potential for severe renal toxicity limit its clinical utility for the treatment of variola infection, numerous advances in drug discovery technology—such as combinatorial chemistry, molecular modeling, and high-throughput screening—are providing many new chemical entities that could be tested for their safety and efficacy against variola [27, 28]. Moreover, successful strategies for blocking the infectivity of other types of viruses may suggest new approaches for combating variola. If a large outbreak of smallpox is a credible threat, the infrastructure for testing the antivariola activity of existing and future antiviral agents must be retained by CDC on an ongoing basis. Should a backlog of promising agents develop, expansion of the infrastructure might well be considered.

Discovery of a new antiviral agent is a complex and costly process, typically requiring evaluation of many tens of thousands of candidates in several assays. The primary screening assay typically tests the ability of an agent to bind and inactivate a recombinant, cell-free target protein of viral or human origin, or inhibit replication of the virus itself. Intrinsic potency is determined at this stage in titration experiments, while measurement of potency against closely related proteins provides an indication of specificity. Since single amino acid changes can have dramatic effects on the potency and specificity of antiviral agents, it is imperative that authentic target proteins be tested. Authentication of the target protein includes sequence analysis of the corresponding DNA from several clinical isolates.

The potency and efficacy of an antiviral agent are influenced by many factors in addition to the agent's ability to bind to the target protein. These factors include

(but are by no means limited to) the intracellular and/or extracellular location of the target protein, the ability of agents to permeate host cells in the case of intracellular targets, the local concentration and turnover rate of the target protein within the infected cell, and conformational changes that may be induced within the target protein by neighboring proteins if the target is part of a protein complex. These latter considerations require that candidate antivirals be tested in cell culture assays employing live virus so that potency and efficacy in this more complex setting can be estimated. The concentration of agent required to block infectivity by 90 percent in cell culture (IC90) is often a valuable guide to determining the concentration of agent that must be continuously maintained in vivo in order to obtain the desired therapeutic effect. To the extent that animal model studies are problematic (see the following section), cell culture assays provide the only preclinical opportunity to evaluate the ability of an agent to block host cell infection. Moreover, even if animal model systems are available, cell culture assays provide the only preclinical opportunity to test the activity on human cells.

Whenever possible, orthopoxvirus family members other than variola should be used for routine testing involving live virus in order to reduce the hazards associated with handling of variola by laboratory staff. This is particularly true if the target protein is identical among orthopoxvirus family members. Another strategy would be to inactivate a gene that variola requires for replication. This gene would necessarily be unrelated to the gene whose protein product is the drug target. Because replication-defective virus is able to replicate only in host cells genetically engineered to express the viral gene, the risk to laboratory personnel and the likelihood of an outbreak originating from a laboratory would be significantly reduced.

Although useful for general drug screening, however, use of such recombinants has limitations. In particular, such genetically engineered host cell lines are not necessarily representative of cells infected by the virus in vivo. It is therefore essential that candidate agents be tested for activity and potency in a tissue culture assay employing clinical isolates of variola virus and recently isolated human cells to ascertain whether equivalent potency is obtained in the surrogate system.

Evidence supporting this contention comes from studies showing that a given antiviral agent can exhibit substantially different potencies against related viruses in tissue culture. For example, non-nucleoside reverse transcriptase inhibitors (e.g., nevaripine, delavirdine, and efavirenz) are active agents against human immunodeficiency virus 1 (HIV-1), but not HIV-2; likewise, amantadine and rimantadine are active against influenza A, but not influenza B, while sorivudine is very effective against herpes simplex virus I (HSV-1), but ineffective against HSV-2 [29–31]. Even more relevant is the observation that phosphonates related to cidofovir are about fivefold more active against variola in cul-

ture than against either monkeypox or cowpox in culture.^* Assay of these substances against either cowpox or monkeypox would therefore result in a substantial underestimation of their potency against variola. Underestimation of potency could lead to overdosage; greater risk of toxicity; and the need for larger, more costly inventories of antiviral drugs. On the other hand, overestimation of potency could encourage selection of escape mutants. In addition, it is highly desirable and accepted practice that in vitro testing be conducted with several clinical isolates of variola to ensure that results obtained with a given isolate are representative of the species, if not the genus of Orthopoxviridae.

The activity of antiviral agents is influenced by their pharmacokinetic and metabolic profiles, which are examined initially through studies conducted on uninfected animals. Regimens of increasing single and multiple doses are used to determine drug absorption and duration, tolerance, and maximum tolerated dose. The IC₉₀ determined in vitro using live variola virus and cultured human host cells as described above would be used, along with observed blood and tissue levels of antiviral agent in test animals, to design a dose-ranging study aimed at determining safety and tolerability in humans under an approved investigational new drug (IND) application. This is essentially the approach taken to develop antiviral agents for the treatment of HIV.

While experience with the development of anti-HIV drugs is instructive with respect to the development of antivariola agents, there are no patients infected with variola with whom to conduct pharmacodynamic and efficacy studies. Animal model studies therefore provide the only opportunity to study the pharmacodynamic effects of a candidate antiviral agent in a whole-animal setting. Specifically, it is important to determine prior to use during a large-scale human outbreak what effect the agent has on the natural history of the disease, the development of immunity, and the clearance of virus when given before and during various stages of infection. Once again, the lack of an animal model specific to variola virus means that use of other orthopoxvirus family members should be considered, especially if the agent has demonstrated similar activity against other family members in cell culture. Priority would reasonably be given to antivariola agents whose activity against other orthopoxviruses would permit their evaluation in surrogate animal model systems. To the extent that otherwise promising agents exhibited variable potency against different orthopoxviruses in

cell culture, it would be highly desirable to develop models using variola itself such that the pharmacodynamic properties of variola-specific compounds could also be assessed. Regulatory agencies such as the U.S. Food and Drug Administration should be active in the development of such drugs for use in emergencies.