In this three-part blog series, we present the typical workflow of an antiviral discovery program, from the initial identification of antiviral molecules in vitro (Part 1) to preclinical testing (Part 2) and assessing the onset of drug resistance (Part 3). These early stages of drug discovery are critical to bringing any compound to the clinic and, given the relatively lower costs of the R&D stages compared to clinical studies, it is important that every step of the process is optimised to select only the most promising compounds. In this blog (Part 2 of 3), we discuss lead optimisation and in vivo testing, keeping in mind that the final goals of preclinical studies are the progression to Phase I.
From hit compound to best compound: medicinal chemistry
After the initial identification of interesting chemical matter, it is important to explore the chemical space around the hit compound to identify the chemical groups that are critical for antiviral effect, as well as the modifications that can improve potency and efficacy. This can be achieved by purchasing or synthesising random analogues of the hit compound, or (when the pharmacological target and its structure are known) through rationally designed analogues. This entire process is best carried out with the help of a medicinal chemist, as potency and efficacy aren’t the only important parameters. Molecular weight, logD, polar surface area, solubility, stability, and many other factors need to fall into specific ranges, while also carefully considering the route of administration. Oral or topic regimens are generally preferred for regular administration of drugs by the patients themselves, while a parenteral route is often preferred in a hospital setting or when the patient is incapacitated. Working with these targets in mind from the start can significantly reduce costs and reduce friction during the approval process later down the line.
Toxicity, toxicity, toxicity.
Toxicity is one of the main reasons why many compounds never make it to the clinic and failing to determine the extent and the consequences of toxicity can have the most severe consequences. While in vitro studies can only partially address this issue, there is increasing pressure to undertake systematic in vitro toxicity studies early on, in order to determine important parameters, such as the metabolic degradation of the compound, its impact on the P450 system, and the non-specific interaction with unwanted targets (e.g., the hERG channel, famously known for causing cardiac arrhythmias). To this end, testing compounds against a panel of human receptors is recommended and, if the antiviral has been designed to inhibit viral components that have homologues in humans (e.g., the viral DNA polymerase), in vitro studies should ensure that the activity of the human homologue is not significantly affected. All of these studies are requested by regulators before clinical trials, and they are better performed early on by a GLP-compliant contract research organisation specialised in these types of studies.
Where, when, how much?
When all in vitro studies have been ticked-off, it is time to test the lead compound in vivo. The scope of in vivo studies is to determine the effectiveness in a living system, and also to determine its pharmacokinetics and pharmacodynamics. This second aspect is very important but sometimes underestimated. Aside from providing critical data on in vivo toxicity of a compound and of its metabolites, it provides an important framework to assess effectiveness by testing whether there is sufficient compound at the site of viral replication at the right time and for the right duration. A compound targeting a virus that causes encephalitis is likely to be required to cross the blood-brain barrier, while for viruses that replicate in the respiratory tract, a sufficient concentration of compound will have to be retained at that site to inhibit infection. Also, compounds that are cleared too fast or that are not easily absorbed are likely to be problematic or require frequent administrations. These initial studies also allow the first estimation of dosage, and to determine the best administration regime, organ toxicity, and reversibility, and the best way to monitor infection and disease progression; all precious information for clinical studies. It makes sense to assess its antiviral effectiveness only after a sufficient concentration of compound has been determined at the site of interest for the sufficient amount of time.
In vivo antiviral effectiveness
Measuring antiviral activity in a small animal model is an exciting confirmation of drug effectiveness, but it is not always possible, and regulators are aware of this. Often, there is no small animal model for the virus studied, and even when there is, it might not be representative of infection in human, making it difficult (if not impossible) to find correlatives of protection. In certain cases, the animal model can be too stringent and underestimate the effectiveness of a compound: it is important to be aware of this to avoid discarding a potentially useful drug too early. This is the case, for instance, when a virus that only causes non-lethal infection in human is instead lethal in mouse. An example is animal models for dengue or Zika, where high lethality even upon lowering of viral titres can mask the effect of the compound. It is, therefore, important to be aware of the limitations of these models and assess as many parameters as possible, including histological examination of multiple tissues, quantification of viral titers over time, isolation and characterization of resistant isolates in animals that experience viral rebound (more on this in the third part of this blog series), quantification of viral antigens and antibodies, and analysis of symptoms (e.g., behavioural, neurological, weight loss, and ruffling fur), as well as morbidity and mortality. Before clinical trials in human volunteers can begin, GLP-standard evaluation of toxicity is required in at least two animal species, generally a rodent and a non-rodent. The costs of testing in a non-rodent species are of course much higher, and they should only be undertaken when promising results from previous in vitro and in vivo studies have clearly highlighted the antiviral benefit.
What if the virus fights back?
It likely will. One of the main challenges of antiviral discovery is the development of resistant mutants. This will be the topic of our next blog!
