key: cord-1032661-9sjdb7f3
authors: Pruijssers, Andrea J.; George, Amelia S.; Schäfer, Alexandra; Leist, Sarah R.; Gralinksi, Lisa E.; Dinnon, Kenneth H.; Yount, Boyd L.; Agostini, Maria L.; Stevens, Laura J.; Chappell, James D.; Lu, Xiaotao; Hughes, Tia M.; Gully, Kendra; Martinez, David R.; Brown, Ariane J.; Graham, Rachel L.; Perry, Jason K.; Du Pont, Venice; Pitts, Jared; Ma, Bin; Babusis, Darius; Murakami, Eisuke; Feng, Joy Y.; Bilello, John P.; Porter, Danielle P.; Cihlar, Tomas; Baric, Ralph S.; Denison, Mark R.; Sheahan, Timothy P.
title: Remdesivir inhibits SARS-CoV-2 in human lung cells and chimeric SARS-CoV expressing the SARS-CoV-2 RNA polymerase in mice.
date: 2020-07-07
journal: Cell Rep
DOI: 10.1016/j.celrep.2020.107940
sha: c7c594390203ea1b999c9da2012c33092d764d46
doc_id: 1032661
cord_uid: 9sjdb7f3

Summary Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent of the novel viral disease COVID-19. With no approved therapies, this pandemic illustrates the urgent need for broad-spectrum antiviral countermeasures against SARS-CoV-2 and future emerging CoVs. We report that remdesivir (RDV) potently inhibits SARS-CoV-2 replication in human lung cells and primary human airway epithelial cultures (EC50 = 0.01 μM). Weaker activity is observed in Vero E6 cells (EC50 = 1.65 μM) due to their low capacity to metabolize RDV. To rapidly evaluate in vivo efficacy, we engineered a chimeric SARS-CoV encoding the viral target of RDV, the RNA-dependent RNA polymerase, of SARS-CoV-2. In mice infected with chimeric virus, therapeutic RDV administration diminishes lung viral load and improves pulmonary function compared to vehicle treated animals. These data demonstrate that RDV is potently active against SARS-CoV-2 in vitro and in vivo, supporting its further clinical testing for treatment of COVID-19.

cause severe respiratory disease with respective mortality rates of 11% (Chan-Yeung and Xu, 2003) , 35% 23 (Arabi et al., 2017) , and an estimated 3% (Chen, 2020) . The development of effective broad-spectrum 24 antivirals has been hampered by viral diversity, the capacity of CoVs to adaptively overcome negative 25 selective pressures, and the ability to actively counteract drugs through the action of a proofreading 26 exoribonuclease. We previously reported that remdesivir (RDV), a monophosphoramidate prodrug of the 27 C-adenosine analog GS-441524, potently inhibits replication of a broad spectrum of pre-pandemic bat 28

CoVs and human epidemic CoVs in primary human lung cell cultures (Agostini et were quantified by plaque assay and RT-qPCR, respectively. RDV and GS-441524 potently inhibited 117 SARS-CoV-2 replication in a dose-dependent manner in both cell types ( Fig. 2; Table 1 ). In Calu3 2B4 118 cells, both compounds displayed dose-dependent inhibition of viral replication as determined by plaque 119 assay (Fig. 2B) and RT-qPCR (Fig. 2C) . RDV inhibited SARS-CoV-2 with an EC 50 = 0.28 µM and EC 90 120 = 2.48 µM. The parent compound GS-441524 displayed similar potency: EC 50 = 0.62 µM, EC 90 = 1.34 121 µM ( Fig. 2D; Table 1 ). EC 50 values determined by quantification of viral genome copies showed a 122 similar trend ( Fig. 2E; Table 1 ). Both compounds also displayed dose-dependent inhibition of viral 123 replication in Vero E6 cells as determined by infectious viral titer and genome copy number (Fig 2F) . 124 RDV inhibited SARS-CoV-2 with EC 50 = 1.65 µM and EC 90 = 2.40 µM. However, in this cell type 441524 was more potent (EC 50 = 0.47 µM, EC 90 = 0.71 µM) than RDV as determined by both infectious 126 viral titer ( Fig. 2G; Table 1 ) and genome copy number ( Fig. 2H; Table 1 ). Thus, RDV inhibits SARS-127

CoV-2 more potently in Calu3 2B4 than in Vero E6 cells, and the relative potencies of prodrug and parent 128 nucleoside are cell type dependent. 129 RDV is a highly potent antiviral inhibitor of SARS-CoV-2 in primary human airway epithelial 130 (HAE) cultures. SARS-CoV-2 is known to replicate in the upper and lower airways in humans (Adachi 131 et al., 2020; Wölfel et al., 2020) . In addition, we have recently shown that SARS-CoV-2 replicates in 132 human primary cells from the nasal to alveolar epithelium (Hou et al., 2020) . To demonstrate the antiviral 133 activity of RDV against SARS-CoV-2 in a human primary epithelial culture system, we performed 134 antiviral assays in HAE cultures, which are grown on air-liquid interface and recapitulate the cellular 135 complexity and physiology of the human conducting airway (Sims et al., 2005) . HAE cultures are the 136 model system in which we have amassed data for many other enzootic, emerging, and endemic CoV, 137 allowing for the comparison of SARS-CoV-2 data with previous reports (Sheahan et al., 2017 (Sheahan et al., , 2020a . In 138 RDV-treated, SARS-CoV-2 infected HAE, we observed a dose-dependent reduction in infectious virus 139 production, with >100-fold inhibition at the highest tested concentration (Fig. 3A) . Importantly, RDV 140 demonstrates potent antiviral activity with EC 50 values of 0.0010 and 0.009 µM in two independent 141 experiments (Fig. 3B) . Similar to previously reported studies, RDV did not cause cytotoxicity in HAE 142 across the dose range where we see potent antiviral effects (Fig. S4 ) (Sheahan et al., 2017) . Together, 143 these data demonstrate that RDV is potently antiviral against SARS-CoV-2 in primary human lung 144 cultures with a selectivity index of >1000. 145 Antiviral activities of RDV and GS-441524 correlate with RDV-TP metabolite levels. Cell type 146 specific expression of genes for enzymes that metabolize ribonucleoside analogs and their prodrugs can 147 have a profound impact on activity (Eriksson, 2013; Koczor et al., 2012) . Table 1 against SARS-CoV-2 is highly variable in different cell culture models. Both RDV and GS-441524 150 undergo intracellular conversion to the active metabolite RDV-TP involving several metabolic steps (Fig.  151   S5 ) and the efficiency of each step might differ between cell types. Therefore, to reconcile the differences 152 in antiviral activity of RDV and GS-441524 observed in our and other studies, we compared intracellular 153 RDV-TP concentrations in Vero E6, Calu3 2B4, and HAEs following compound incubation. RDV-TP 154 levels per million cells produced after 8-to 48-hour treatment with RDV were substantially higher in 155 primary HAE cultures than either Calu3 2B4 or Vero E6. (Fig 4; Table 1 ; Tables S1, S2). Given the 156 primary nature of HAE cultures, we used cells from two independent donors with similar demographic 157 profiles. RDV-TP was efficiently formed in both donor cultures following incubation with RDV with a 158 difference of < 3-fold from each other. The lowest levels of RDV-TP were observed following RDV 159 treatment of Vero E6 cells and were approximately 4-and 20-fold lower than those observed in Calu3 160 2B4 and HAE cultures, respectively. Incubation of Vero E6 cells with GS-441524 yielded 3.5-fold higher 161 RDV-TP levels compared to incubation with RDV corresponding to higher antiviral potency of GS-162 441524 relative to RDV. (Table S1 ). In conclusion, the RDV-TP levels in the different cell types directly 163 correlated with the antiviral potencies of RDV against SARS-CoV-2. HAE cultures produced 164 substantially higher levels of RDV-TP, which translated into markedly more potent antiviral activity of 165 RDV (Table 1) . Importantly, the metabolism of RDV in Vero E6 cells appeared less efficient particularly 166 in comparison with the HAE cultures, indicating that Vero E6 cells might not be an appropriate cell type 167 to characterize the antiviral activity of RDV and potentially also other nucleotide prodrug-based 168 antivirals. 169 RDV is active against the SARS-CoV-2 RdRp in vivo. SARS-CoV-2 does not bind the murine ortholog 170 of the human entry receptor (i.e. mouse angiotensin converting enzyme-2, mACE2) to enter cells (Zhou et 171 al., 2020b) . To rapidly assess the therapeutic efficacy of RDV against SARS-CoV-2 in vivo, we 172 constructed a chimeric mouse-adapted SARS-CoV variant encoding the target of RDV antiviral activity, 173 the RdRp, of SARS-CoV-2 (SARS1/SARS2-RdRp) (Fig. 5A) . Although other chimeric replicase ORF 174 recombinant CoVs have shown to be viable (Stobart et al., 2013) , this is the first demonstration that the 175 RdRp from a related but different CoV can support efficient replication of another. After recovery and 176 sequence-confirmation (Fig. 5B ) of recombinant chimeric viruses with and without nanoluciferase 177 reporter, we compared SARS-CoV and SARS1/SARS2-RdRp replication and sensitivity to RDV in Huh7 178 cells. Replication of both viruses was inhibited similarly in a dose-dependent manner by RDV (SARS-179

CoV mean EC 50 = 0.007 µM; SARS1/SARS2-RdRp mean EC 50 = 0.003 µM) ( Fig. 5C and D) . We then 180 determined the therapeutic efficacy of RDV against the SARS1/SARS2-RdRp in mouse models 181 employed for previous studies of RDV (Sheahan et al., 2017) . Mice produce a serum esterase absent in 182 humans, carboxyl esterase 1c (Ces1c), which dramatically reduces the half-life of RDV. Thus, to mirror 183 pharmacokinetics observed in humans, mouse studies with RDV must be performed in transgenic 184

C57Bl/6 Ces1c -/mice (Sheahan et al., 2017) . We infected female C57Bl/6 Ces1c -/mice with 10 3 PFU 185 SARS1/SARS2-RdRp and initiated subcutaneous treatment with 25 mg/kg RDV BID at one day post-186 infection (dpi). This regimen was continued until study termination. While weight loss and lung 187 hemorrhage did not differ significantly between vehicle-and RDV-treated animals ( Fig. 5E and F), we 188 found differences in pulmonary function as measured by whole body plethysmography (WBP) between 189 RDV and vehicle treated animals. The WBP metric, PenH, is a surrogate marker of pulmonary 190 obstruction (Menachery et al., 2015a) . Therapeutic RDV significantly ameliorated loss of pulmonary 191 function observed in the vehicle-treated group (Fig. 5G) . Importantly, RDV treatment dramatically 192 reduced lung viral load (Fig. 5H) . Taken together, these data demonstrate that therapeutically versus SARS-CoV-2 isolates used in the previously mentioned studies assessing RDV potency did not 222 reveal consensus changes in nsp12 sequence, suggesting that any isolate-specific variation in RDV 223 sensitivity is not likely due to differences in the RDV-TP interaction with the RdRp. Therefore, the 224 differences in EC 50 may be partially explained by intrinsic differences of SARS-CoV-2 virus isolates, 225 quantification methods, Vero cell lineages, and assay conditions such as incubation period and virus 226 input. 227

Although Vero E6 cells support robust replication of SARS-CoV-2 as illustrated here and elsewhere, our 228 study emphasized the caution that should be exercised when interpreting nucleoside prodrug potency 229 experiments performed using Vero cell lineages. Nucleoside analog potency is dependent on metabolism 230 into the active form. In contrast to the nucleoside GS-441524, RDV is a monophosphoramidate prodrug 231 with moieties that mask the negative charges of its monophosphate group which enhances its cellular 232 uptake. Intracellularly, RDV is then rapidly metabolized to its monophosphate, which is efficiently 233 See also Table S1 and Figure S5 . 

Further information and requests for resources and reagents should be directed to and will be fulfilled by 407 the Lead Contact Andrea Pruijssers (ardina.pruijssers@vumc.org) 408

Plasmids generated in this study are available upon request. 410

The published article includes all datasets generated and analyzed during this study. 412 drug addition to all cultures, cells were washed 3 times with ice-cold tris-buffered saline, scraped into 0.5 544 mL ice-cold 70% methanol and stored at -80°C. Extracts were centrifuged at 15,000 x g for 15 minutes 545 and supernatants were transferred to clean tubes for evaporation in a MiVac Duo concentrator (Genevac). 546

Dried samples were reconstituted in mobile phase A containing 3 mM ammonium formate (pH 5) with 10 547 mM dimethylhexylamine (DMH) in water for analysis by LC-MS/MS, using a multi-stage linear gradient 548 from 10% to 50% acetonitrile in mobile phase A at a flow rate of 300 µL/min. Analytes were separated 549 using a 50 x 2 mm, 2.5 µm Luna C18(2) HST column (Phenomenex) connected to an LC-20ADXR 550 (Shimadzu) ternary pump system and HTS PAL autosampler (LEAP Technologies). Detection was 551 performed on a Qtrap 6500+ (AB Sciex) mass spectrometer operating in positive ion and multiple 552 reaction monitoring modes. Analytes were quantified using a 7-point standard curve ranging in 553 concentration from 0.156 to 40 pmol prepared in extracts from untreated cells. For normalization by cell 554 number, multiple untreated Calu3 or Vero E6 culture wells were counted at each timepoint. HAE cells 555

were counted at the 24-h timepoint and the counts for other timepoints were determined by normalized to 556 endogenous ATP levels for accuracy. 557

Formulations for in vivo studies. RDV was solubilized at 2.5 mg/mL in vehicle containing 12% 558 sulfobutylether-β-cyclodextrin sodium salt in water (with HCl/NaOH) at pH 5.0. 559

In vivo efficacy studies. All animal experiments were performed in accordance with the University of 560

North Carolina at Chapel Hill Institutional Animal Care and Use Committee policies and guidelines. To 561 achieve a pharmacokinetic profile similar to that observed in humans, we performed therapeutic efficacy 562 studies in Ces1c -/mice (stock 014096, The Jackson Laboratory), which lack a serum esterase not present 563 in humans that dramatically reduces RDV half-life (Sheahan et al., 2017) . 17 week-old female Ces1c -/-564 mice were anaesthetized with a mixture of ketamine/xylazine and intranasally infected with 10 3 PFU 565 SARS1/SARS2-RdRp in 50 µL. One dpi, vehicle (n = 7) and RDV (n = 7) dosing was initiated (25 mg/kg 566 subcutaneously) and continued every 12 h until the end of the study at five dpi. To monitor morbidity, 567 mice were weighed daily. Pulmonary function testing was performed daily by whole body 568 plethysmography (WBP) (Data Sciences International) (Sheahan et al., 2017) . At five dpi, animals were 569 sacrificed by isoflurane overdose, lungs were scored for lung hemorrhage, and the inferior right lobe was 570 frozen at −80°C for viral titration via plaque assay on Vero E6 cells. Lung hemorrhage is a gross 571 pathological phenotype readily observed by the naked eye and driven by the degree of virus replication, 572

where lung coloration changes from pink to dark red (Sheahan et al., 2017 (Sheahan et al., , 2020a . For the plaque assay, 573 5 x 10 5 Vero E6 cells/well were seeded in 6-well plates. The following day, medium was removed, and 574 monolayers were adsorbed at 37˚C for one h with serial dilutions of sample ranging from 10 -1 to 10 -6 . 575

Cells were overlayed with 1X DMEM, 5% Fetal Clone 2 serum, 1× antibiotic-antimycotic, 0.8% agarose. 576 Viral plaques were enumerated three days later. 

Clinicopathologic and Immunohistochemical Findings from Autopsy 588 of Patient with COVID-19

Coronavirus Susceptibility to the Antiviral Remdesivir (GS-5734) Is 591 Mediated by the Viral Polymerase and the Proofreading Exoribonuclease

Hydroxycytidine Inhibits a Proofreading-Intact Coronavirus with a High Genetic Barrier to Resistance

Viral replication. Structural basis for RNA replication by the 598 hepatitis C virus polymerase

Middle East Respiratory Syndrome

Remdesivir for the Treatment of Covid-19 -Preliminary 604 Report

SARS-CoV-2 and SARS-CoV differ in their cell 607 tropism and drug sensitivity profiles

Broad spectrum antiviral remdesivir inhibits human endemic and 610 zoonotic deltacoronaviruses with a highly divergent RNA dependent RNA polymerase

SARS: epidemiology. Respirol. Carlton Vic 8 Suppl

Pathogenicity and transmissibility of 2019-nCoV-A quick overview and comparison 614 with other emerging viruses

Synthesis and antiviral activity of a series of 1′-substituted 4-aza-7,9-dideazaadenosine C-617 nucleosides

Remdesivir, lopinavir, emetine, and 620 homoharringtonine inhibit SARS-CoV-2 replication in vitro

Data, disease and diplomacy: GISAID's innovative 622 contribution to global health

Is the expression of deoxynucleoside kinases and 5'-nucleotidases in animal tissues 624 related to the biological effects of nucleoside analogs?

Well-differentiated 626 human airway epithelial cell cultures

Structure of the RNA-dependent RNA polymerase from COVID-19 virus

Structural basis for active site closure by the poliovirus RNA-630 dependent RNA polymerase

Remdesivir is a direct-acting antiviral that inhibits RNA-dependent RNA polymerase from severe acute 633 respiratory syndrome coronavirus 2 with high potency

The antiviral compound 635 remdesivir potently inhibits RNA-dependent RNA polymerase from Middle East respiratory syndrome 636 coronavirus

Compassionate Use of Remdesivir for Patients with Severe Covid-19

Early Release -Severe Acute Respiratory Syndrome Coronavirus 2 from Patient with 642 2019 Novel Coronavirus Disease

First Case of 2019 Novel Coronavirus in the United States

SARS-CoV-2 Reverse Genetics Reveals a Variable Infection 649 Gradient in the Respiratory Tract

Viral loads in clinical specimens and SARS 652 manifestations

Identification of 654 antiviral drug candidates against SARS-CoV-2 from FDA-approved drugs

Structure of the SARS-CoV nsp12 polymerase bound to nsp7 656 and nsp8 co-factors

The role of transporters in the toxicity of nucleoside 658 and nucleotide analogs

GS-5734 and its parent nucleoside analog inhibit 661 Filo

The ProTide Prodrug Technology: From the Concept 663 to the Clinic

New Metrics for Evaluating 665 Viral Respiratory Pathogenesis

A SARS-like cluster of circulating bat 668 coronaviruses shows potential for human emergence

SARS-like WIV1-CoV poised for human 671 emergence

Lusakibanza 673

Cell-line dependent antiviral activity of sofosbuvir against Zika 677 virus

Viral load of SARS-CoV-2 in clinical 679 samples

Severe acute respiratory syndrome

Lianhuaqingwen exerts anti-viral and anti-inflammatory activity against 683 novel coronavirus (SARS-CoV-2)

Pharmacologic Treatments for 685 Coronavirus Disease 2019 (COVID-19): A Review

Reverse genetics with a full-length infectious 688 cDNA of the Middle East respiratory syndrome coronavirus

Remdesivir and SARS-CoV-2: Structural requirements at both nsp12 692 RdRp and nsp14 Exonuclease active-sites

Broad-spectrum antiviral GS-5734 inhibits both epidemic and 695 zoonotic coronaviruses

Comparative therapeutic efficacy of remdesivir and 698 combination lopinavir, ritonavir, and interferon beta against MERS-CoV

An orally bioavailable broad-spectrum antiviral inhibits 701 SARS-CoV-2 in human airway epithelial cell cultures and multiple coronaviruses in mice

GISAID: Global initiative on sharing all influenza data -from vision 704 to reality

Severe acute 706 respiratory syndrome coronavirus infection of human ciliated airway epithelia: role of ciliated cells in 707 viral spread in the conducting airways of the lungs

Chimeric Exchange of Coronavirus nsp5 Proteases (3CLpro) Identifies Common and 710 Divergent Regulatory Determinants of Protease Activity

Receptor Recognition by the Novel 712 Coronavirus from Wuhan: an Analysis Based on Decade-Long Structural Studies of SARS Coronavirus

Remdesivir and chloroquine effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in 716 vitro

Therapeutic efficacy of the small molecule GS-5734 against Ebola 719 virus in rhesus monkeys

Clinical benefit of remdesivir in rhesus 722 macaques

Prophylactic and therapeutic remdesivir (GS-5734) treatment in the rhesus macaque 725 model of MERS-CoV infection

SARS and MERS: recent 727 insights into emerging coronaviruses

Virological assessment of hospitalized patients with COVID-730 2019

Structural basis for inhibition of the RNA-dependent RNA polymerase from SARS-CoV-2 by 733 remdesivir

Dynamic Innate Immune Responses of Human Bronchial Epithelial Cells to 736

Severe Acute Respiratory Syndrome-Associated Coronavirus Infection

Reverse genetics with a full-length infectious cDNA of severe acute 739 respiratory syndrome coronavirus

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a 742 retrospective cohort study

A pneumonia outbreak associated with a new coronavirus of probable bat origin