Key 3 – Implement 4R standards (right patient, right drug, right dosage, right timing)
1. Right patient
The idea of matching drugs to patients is a widely accepted paradigm in pharmacology. This personalized medicine approach emphasizes the numerous patient-level factors that can affect therapeutic targets, a medication’s PK, as well as the overall burden of disease.
Patient selection is complicated by heterogeneity in the clinical manifestations of COVID-19. About 80% of patients will only have mild/moderate flu-like or pneumonia symptoms.8 The majority of COVID-19 cases are self-limited and not life-threatening. A considerable portion of patients, up to 17.9% of RT-PCR-confirmed cases in some models,9 may be completely asymptomatic for the entire duration of infection, though some may have objective subclinical manifestations of disease10. However, around 20% of patients are deemed severe cases with significant dyspnea, hypoxia, or lung imaging findings that require supplemental oxygen or intensive care and current estimates of COVID-19 case fatality rates vary from 2-7%,8,11,12 though this figure varies by location.
A number of factors contribute to the disparate manifestations of COVID-19; the most well-studied factor is age. It has been reported in China, Italy, South Korea and United States that elderly patients comprised approximately 60-83% of all COVID-19 fatalities.13,14,15 A similar trend is seen with designation of “severe” cases as well as rates of hospitalization and ICU admission, all of which are more prevalent in older individuals.15
Another factor is cardiovascular disease (CVD). There is a high prevalence of CVD amongst severe COVID cases and COVID cases overall, with death rates of individuals with CVD reportedly over four times the overall mortality rate.11 The leading hypothesis for this is suggested to involve SARS-COV-2’s interaction with angiotensin-converting enzyme 2 (ACE2) and the renin-angiotensin-aldosterone axis. Chronic respiratory disease is also linked with severe and fatal COVID-19 cases,15potentially due to the decreased lung pulmonary reserve, dysregulated immune system, and disrupted lung microbiome in these patients.16,17 Diabetes is also a significant risk factor for severe COVID-19 infection, potentially due to similar mechanisms of immune dysregulation, alterations in ACE2 expression, and increased processing of the SARS-COV-2 spike protein.18 In addition, it has been suggested that common therapies for these conditions, such as angiotensin converting enzyme inhibitors, angiotensin receptor blockers, and immunomodulators like inhaled steroids have an effect on COVID-19 pathophysiology.19,20,21
In designing therapeutic drug trials for COVID-19, we may have to control for not only age, CVD, pulmonary disease, diabetes, and other patient-level factors like immunodeficiency, but also control for the therapies patients have been given.
2. Right Drug
Two approaches for developing scientific treatments for COVID-19 exist:
1) a “bottom-up” approach by repurposing already approved drugs or molecules under clinical development for other indications
2) a “top-down” approach targeting at new molecules and vaccines specifically designed for SARS-COV-2 which can be time-consuming, but more effective and safer.
COVID-19 is caused by SARS-CoV-2 that is a single-stranded positive-sense RNA virus.22 Since SARS-CoV-2 is a newly discovered pathogen, no specific drugs are currently available. Several existing drugs and new drugs that have potential therapeutic effects are summarized in Table 1. These therapeutics are divided into 4 categories: convalescent plasma or immunoglobulins; direct-acting antiviral agents (DAA); host cell internalization protein blockers; and anti-inflammatory drugs.
Convalescent Plasma or immunoglobulin fractions have been used for treatment of Ebola virus, Middle East respiratory syndrome coronavirus (MERS-CoV), and SARS-CoV infections.23,24,25 A clinical study of 5 critically ill patients with COVID-19 and acute respiratory distress syndrome (ARDS) showed that administration of convalescent plasma containing neutralizing antibodies improved some patient’s clinical status.26 Issues associated with this approach include donner identification, elimination of residual SARS-CoV-2 risk, off-target immunoglobulin binding, and dose estimation. Of special concern is the ability to scale this approach to the level required.
The potential targets of DAA against non-structural proteins include RNA-dependent RNA polymerase (RdRp), coronavirus main protease (3CLpro), and papain-like protease (PLpro). A potential target for a DAA against a viral structural protein was the glycosylated spike (S) protein.27 S protein mediates SARS-CoV-2 entry via binding to angiotensin-converting enzyme 2 (ACE2) located on the surface membrane of host cells following which host cell produced transmembrane protease serine 2 (TMPRSS2) was involved in S protein priming that facilitates the process of internalization.28 Thus, both host cell membrane proteins ACE2 and TMPRSS2 have become potential therapeutic targets. In addition, some recent studies also reported that S protein can also bind to the host cell receptor CD14729 and the glucose regulated protein 78 (GRP78)30 that may also mediate internalization.
The protease inhibitor combination lopinavir and ritonavir (Aluvia®), and two viral polymerase inhibitors, favipiravir and remdesivir, are non-structural SARS-CoV-2 protein targets currently in clinical trials. Galidesivir (BCX4430) is an adenine analogue RdRp inhibitor originally developed for the treatment of hepatitis C virus. It is currently undergoing safety testing in early clinical studies and is being evaluated for its efficacy in treating yellow fever. In preclinical studies, it exhibits activity against a variety of RNA viruses, including SARS and MERS.
Both SARS-CoV, responsible for severe acute respiratory syndrome (SARS), and SARS-CoV-2, responsible for COVID-19, are beta-coronaviruses (CoV) that share a structurally similar spike glycoprotein (S) complex surrounding the spherical viral particle that is comprised of a receptor binding domain (RBD) S1 subunit and a membrane fusion S2 subunit.31 The S proteins of SARS-CoV-2 have about 76% homology to those of SARS-CoV and both recognize and bind to ACE2.31
While Abidol is mainly used for the prevention and treatment of influenza virus infections, it may disrupt the binding of S proteins to ACE2 to prevent viral internalization and is currently in clinical trials (see Table 1). Soluble recombinant human angiotensin converting enzyme-2 (srhACE2) was initially proposed as a treatment for general ARDS but its affinity for the SARS-CoV-2 spike protein could enable a neutralization with the virus preventing the viral internalization. It is possible for shrACE2 to treat COVID-19 through a combination of preventing lung injury by reducing local angiotensin-II levels and preventing lung epithelial internalization of the virus.
Nafamostat and Camostat are two TMPRSS2 inhibitors with similar molecular structure that could block the virus internalization process. Camostat is currently in clinical trials for COVID-19 Infection (see Table 1).
Hydroxychloroquine (HCQ) / chloroquine (CQ) are anti-malarial drugs used for treating Lupus and forms of arthritis. HCQ is a derivative of CQ with therapeutic effects similar to those of CQ, but with reduced toxic side effects. HCQ/ CQ can inhibit the in vitro replication of several coronaviruses. Recent publications support the hypothesis that CQ can improve the clinical outcome of patients infected by SARS-CoV-2. CQ may interfere with ACE2 receptor glycosylation to inhibit SARS-CoV-2 binding to target cells. It may also inhibit cleavage of S proteins by acidifying lysosomes and may inhibit cathepsin activity. HCQ activates CD8 + T-cells that reduces the production of pro-inflammatory cytokines32,33,34 to limit lung inflammation.
SARS-CoV-2 may cause the rapid release of inflammatory cytokines resulting in ARDS and multiple organ failure. Anti-inflammatory drugs can relieve this response. IL-6, CCR5 and JAK kinase are potential targets for relieving SARS-CoV-2 caused inflammation.35 Tocilizumab (ACTEMRA®) is a recombinant human monoclonal antibody that specifically binds to soluble and membrane-bound IL-6 receptors (sIL-6R and mIL-6R) and inhibits inflammatory IL-6-mediated signal transduction. CytoDyn’s humanized CCR5 antagonist leronlimab (PRO140) has applied for a phase II clinical trial for adult patients with mild to moderate respiratory disease after infection with SARS-CoV-2. The JAK kinase inhibitors baricitinib and ruxolitinib can inhibit JAK-mediated inflammatory processes and are currently in clinical trials for COVID-19 (see Table 1).
3. Right dosage
New drug development paradigm : Dose is an important issue for antiviral drug development requiring years of work. The process starts at the interface between preclinical and Phase I when the first in human dose is calculated. Under today’s model-based drug development paradigm, this exercise is supported by translational PK-PD modelling using available data of nonclinical PK, drug metabolism and toxicology. Phase I dose escalation trials will then be conducted to examine safety, tolerability and PK in reference of in vitro susceptibility and in vivo animal data. Phase II studies will be further conducted in target patients to achieve goals of PoC and dose ranging using biomarkers for both efficacy such as viral load and safety. Dosage for Phase III trial will be optimized based on a target PK/PD metric (e.g., Cmin,ss/EC90) and population PK of the drug. In addition, experimental medicine studies will be conducted to study tissue penetration, and a series of clinical pharmacology studies will be conducted to examine intrinsic and extrinsic factors including age, body weight, the stage of the disease, the presence of co-morbidities, patients’ use of other medications for the purposes of adjusting dose in special populations.
Dose estimation for repurposed drugs : Drug development for COVID-19, thus far, is dependent on the indication extension of existing, approved drugs. These anti-viral drugs must have been through the full pre-clinical discovery and clinical development process with extensive pharmacological, PK, toxicology and manufacturing work done. It is essential that drug candidates for repurposing must have shown efficacy for their original intended indication such that regulatory approval is achieved or fully expected. Absent these requirements additional exploratory development work will be required prior to conduct of a pivotal efficacy trial for the new indication (e.g. COVID-19).
The only information that should be needed to extend an existing drug for use to the new indication is the characterization of its target interaction as nothing else should change. If the new viral target attacks a different tissue, then additional tissue distribution work may be required.
The need to characterize an existing drug’s interaction at a different target in the same tissue such as the RdRp for SARS-CoV-2 versus that of another RNA virus in the lung can be done at the in vitro level. If the EC50 value in this case is different, as it is more likely that the drug will be less potent, then a proportional dose adjustment is a useful approximation. In this case, a much greater effort must be made to estimate the effect of increasing the dose for the new indication. Do the human PK findings support a simple proportional dose adjustment? Do the safety findings support a higher exposure level? Small differences in potency make it likely that the drug can be used for the new indication but larger ones may well preclude this use. This is especially true if significant new clinical development efforts must be made that will substantially increase the time required to get the drug into patients. Assuming that the outcome supports repurposing the existing drug for the new indication it still must be tested in an adequately powered clinical trial for an appropriate patient population.
In COVID-19 drug trials, we found only two trials intended to explore different dosage regimens, Gilead remdesivir trials examine treatment durations of 10 days vs 5 days [NCT04292730, NCT04292899], and the PrEP trial by Washington University School of Medicine on HCQ employ low, mid and high dose regimens [NCT04333732]. None tried a dosage higher than its approved level.
Clinical investigators should collaborate closely with the innovator company to understand the PK, PD, biomarker and safety of drugs repurposed for the new indication and propose a dosage with highest possibility of suppressing the virus within the exposure range known to be safe. In most cases, safety data for levels much higher the than approved dose may be present in the development data file (i.e., single or multiple ascending dose studies). However, there can be no justification for pushing the clinical dose beyond the drug’s known therapeutic window.
The in vitro-in vivo translation of antiviral efficacy should be considered throughout the development process. There are many considerations when using in vitro derived EC50 or EC90 to predict an efficacious dose in vivo . For typical drugs, the free-drug hypothesis can be used to translate in vitro EC values to in vivo efficacious dose based on the theory that unbound drug in the circulation equilibrates with that at the target site under steady-state conditions. However, differences in metabolic activity, drug permeability or transporter expression between the model cell line and target tissue or organ can result in prediction inaccuracy. In addition, special attention should be paid for an anti-viral prodrug, as its active moiety is a metabolite which only stays intracellularly (e.g., remdesivir).36,37 An investigator using a cultured cell model needs to convince themselves that any such differences were characterized during the model development and can be accounted for.
Remdesivir and CQ were found to be relatively more potent (EC50 reported to be 0.77 and 1.13 μM, respectively) compared to ribavirin, penciclovir, nitazoxanide, nafamostat and favipiravir against a clinical isolate of SARS-CoV-2 in vitro.36 This study also demonstrated that remdesivir functioned at a stage after virus entry while CQ functioned at both entry, and at post entry stages of the SARS-CoV-2 infection in Vero E6 cells.36 In another study, the EC50 of HCQ is reported to be 0.72 µM.38
Remdisivir is an example of a proposed “repurposed drug” that appears not to meet the key standard on clinical efficacy described previously. An efficacy trial for Ebola where the drug, when combined with the triple monoclonal antibody ZMapp, failed to show efficacy relative to ZMapp alone.39 Remdisivir is active against SARS-CoV and MERS-CoV replication in cell models and shows efficacy against SARS-CoV infection in carboxylesterase 1c knockout mice37 and MERS-CoV infection in rhesus monkeys40.
CQ and HCQ act at both viral entry and post entry stages so that their in vitro and in vivo translation can be complicated by major differences in their lung to plasma concentration ratios. Both CQ and HCQ are highly distributed to the lung in male CD albino rats41 and HCQ42.
For srhACE2 to occupy viral S protein it must reach the interior of the lung where the virus is located. To block infection srhACE2 must be delivered to the site of infection in both the upper and lower respiratory track areas by inhalation therapy in an adequate amount to occupy S protein sites on the virus. Intravenous administration is unlikely to block infection since there is little virus in the blood but might serve to reduce lung injury though this was not demonstrated in ARDS patients.
4. Right timing
Timing is important to the administration of therapy during the COVID-19 disease course. While our understanding of the natural history of severe COVID-19 infection is incomplete, it is thought to progress from invasion of the respiratory tract and gastrointestinal mucosa, to dysregulation of the RAAS and immune systems, systemic spread to other organs, and then finally cytokine storm, sepsis, and acute respiratory distress syndrome.43
COVID-19 is best treated in its early stages when viral load is low and host physiology is relatively unperturbed. However, this is complicated by COVID-19’s relatively mild onset, with many patients being asymptomatic. A study showed that virus shedding was very high during the first week of symptoms (peak at 7.11 × 108 RNA copies per throat swab, day 4).44 Furthermore, once the disease has progressed to its end-stage and organ failure has occurred, the number of specific therapies available dwindles as drivers of pathology shift from direct viral effects to secondary systemic dysregulation of host physiology. Given that most of the trial drugs currently being studied act on the Sars-Cov-2 viral lifecycle, rather than the downstream manifestations of COVID-19, these medications may not have an effect later in illness. This can be seen in the treatment of influenza, where the anti-viral oseltamivir is only effective if given in the first 48 hours.45 After this period, there are no specific therapies that have shown benefit, and supportive care is the standard. Thus, it is important to define patients’ disease progression in clinical trials, as certain drugs may have different effects when given at different times.