Are revisions to remdesivir dosing needed for treating SARS-CoV-2?

From the Schiffer lab, Clinical Research and Vaccine and Infectious Disease Divisions

As SARS-CoV-2 variants continue to circulate and the severely ill seek medical care, physicians require a quantitative means to measure virus growth in order to assess treatment efficacy. Because SARS-CoV-2 infects both the lungs and the nasal cavity, testing virus presence in the nose is far easier and more comfortable for the patient than sampling the lungs. Yet, understanding the extent of lower lung disease is critical for successful treatment of SARS-CoV-2 infected patients. Therefore, when a clinical trial revealed that remdesivir treatment improved SARS-CoV-2 related symptoms without reducing virus abundance in the nose, some concern arose about how we are measuring recovery from SARS-CoV-2 infection and how this impacts validation of antiviral drugs. The Schiffer group at Fred Hutchinson Cancer Center approached these counterintuitive results by 1) re-analyzing published viral load experiments in rhesus macaques infected with SARS-CoV-2 treated with placebo or remdesivir and 2) using mathematical models incorporating the viral load data to aid in our interpretation of these findings in humans. Their models of SARS-CoV-2 growth kinetics suggested that the immune response to infection differs between these sites. In a fraction of cells within the lungs, the immune response induces resistance to infection, but this does not occur to the same extent in the nasal cavity. Additionally, modeling simulations predicted that high dose remdesivir treatment could overcome infection site specific differences and similarly limit virus replication in both the nose and lungs. Their findings were recently published in iScience.

SARS-CoV-2 is typically measured from nasal cavity or lung samples to determine viral abundance. Remdesivir and other drugs reduce illness and mortality in SARS-CoV-2 infected individuals, specifically when treating promptly following virus infection. In a clinical trial evaluating remdesivir’s antiviral potency in SARS-CoV-2 infected patients, a surprising discovery was made. The abundance of virus from nasal swab tests were the same for the placebo group as compared to the remdesivir treated patients. Yet, remdesivir treated patients were 85% less likely to be hospitalized or die as compared to the placebo group. These findings were quite confounding and revealed a gap in our understanding of SARS-CoV-2 growth kinetics.

To fill this knowledge gap, the Schiffer group first revisited published data on SARS-CoV-2 viral load kinetics in infected monkeys. The infected animals were either untreated or promptly treated with remdesivir. The amount of virus was measured daily from nose and lung samples using nasal swabbing and bronchoalveolar lavage, respectively, over a 7-day period. Like the clinical trial in humans, early remdesivir treatment reduced disease severity as compared to placebo or “vehicle”. Furthermore, the researchers discovered that virus abundance was reduced in the lungs of these animals as compared to the placebo group. These data are not feasibly acquired in human clinical trials because bronchoalveolar lavage sampling in humans is quite taxing on the patient and frequent sampling is needed to accurately follow viral kinetics. Despite reduced viral load in the lungs for remdesivir treated animals, SARS-CoV-2 abundance in the nose several days after initial infection, was surprisingly increased as compared to the placebo treated animals. These data in non-human primates illustrated again that remdesivir reduces disease related to SARS-CoV-2 infection, while failing to limit virus growth in the nose. With viral abundance data from both the nose and lungs of these animals, the Schiffer group, for the first time, developed mathematical models that included virus growth kinetics to explain these data.

Mathematical modeling with prior fitting to rhesus macaque #1 (RM1) viral load data predicts increasing viral load in the nose (NASAL, Treatment) and decreasing viral load in the lungs (BAL, Treatment) following treatment with remdesivir at 12h following infection as compared to the placebo (Vehicle) animal.
Mathematical modeling with prior fitting to rhesus macaque #1 (RM1) viral load data predicts increasing viral load in the nose (NASAL, Treatment) and decreasing viral load in the lungs (BAL, Treatment) following treatment with remdesivir at 12h following infection as compared to the placebo (Vehicle) animal. Image modified from figure in publication

Initial mathematical models were built to fit actual SARS-CoV-2 abundances over time for each animal as shown in the figure above for remdesivir (Treatment) and placebo (Vehicle) treated animals. Additionally, modeling included pharmacokinetic, pharmacodynamics and immune dynamics. The pharmacokinetic and pharmacodynamic models reflected the steps of remdesivir conversion into the active form of the drug that occurs within the animal and drug diffusion kinetics to the lungs as compared to the nose. For immune dynamics, innate immunity can convert cells susceptible to SARS-CoV-2 to refractory cells, or cells resistant to infection. The best model for explaining the differences in virus growth between the nose and lungs following remdesivir treatment raised three potential explanations of these data: 1) limited virus growth in the lungs occurs because refractory cells are generated in the lungs and stunt local virus spread, 2) low dose remdesivir has reduced potency against SARS-CoV-2 growth but even a incompletely potent antiviral can reduce SARS-CoV-2 related disease in the lungs, and 3) increased virus abundance in the nose could occur if remdesivir limits virus growth following initial infection, causing incomplete activation of the innate immune response, and allowance of persistent virus growth in the nose. In summary, “the model provides testable and plausible hypotheses for the counterintuitive finding that remdesivir lowers viral load in lungs but increases viral load in nasal passages of non-human primates that received remdesivir,” stated Dr. Schiffer.

By including actual viral load data in the mathematical modeling of SARS-CoV-2 growth in the lungs and nose during remdesivir treatment, the Schiffer group provided insight into this gap in knowledge. Mainly, “the model suggests one possible mechanism for viral rebound from SARS-CoV-2 directed antiviral treatment, which is that early and effective therapy blunts innate immune responses in the nasal passages thereby allowing prolonged shedding.” Additional modeling suggested that increasing the potency of the antiviral drug either by treating with a higher dose of remdesivir or using another, more potent antiviral would resolve the issue of persistent virus growth in the nose. In the future Dr. Schiffer hopes that this will be “tested [empirically] in follow up, non-human primate experiments in which viral load is carefully measured in both lung and nasal passages.”

The spotlighted research was funded by the National Institutes of Health.

Goyal A, Duke ER, Cardozo-Ojeda EF, Schiffer JT. 2022. Modeling explains prolonged SARS-CoV-2 nasal shedding relative to lung shedding in remdesivir-treated rhesus macaques. iScience. 25(6):104448.