Identification of Potent and Safe Antiviral Therapeutic Candidates Against SARS-CoV-2

COVID-19 pandemic has infected millions of people with mortality exceeding >1 million. There is an urgent need to find therapeutic agents that can help clear the virus to prevent severe disease and death. Identifying effective and safer drugs can provide more options to treat COVID-19 infections either alone or in combination. Here, we performed a high throughput screening of approximately 1,700 US FDA-approved compounds to identify novel therapeutic agents that can effectively inhibit replication of coronaviruses including SARS-CoV-2. Our two-step screen first used a human coronavirus strain OC43 to identify compounds with anti-coronaviral activities. The effective compounds were then screened for their effectiveness in inhibiting SARS-CoV-2. These screens have identified 20 anti-SARS-CoV-2 drugs including previously reported compounds such as hydroxychloroquine, amlodipine besylate, arbidol hydrochloride, tilorone 2HCl, dronedarone hydrochloride, mefloquine, and thioridazine hydrochloride. Five of the newly identified drugs had a safety index (cytotoxic/effective concentration) of >600, indicating a wide therapeutic window compared to hydroxychloroquine which had a safety index of 22 in similar experiments. Mechanistically, five of the effective compounds (fendiline HCl, monensin sodium salt, vortioxetine, sertraline HCl, and salifungin) were found to block SARS-CoV-2 S protein-mediated cell fusion. These FDA-approved compounds can provide much needed therapeutic options that we urgently need during the midst of the pandemic.


INTRODUCTION
Novel coronavirus (CoV)-mediated disease (COVID-19) emerged as a major pandemic and has spread across the world in such a short period since December 2019. As of October 5, 2020, more than 35.1 million confirmed infections have been reported with approximately 1 million deaths (WHO, https:// covid19.who.int/). These numbers may be a vast underestimation as many of the infected patients may remain asymptomatic and can only be detected by antibody testing (1). Similarly, many of the deaths may not be accounted for due to a lack of testing. The disease is caused by a novel CoV termed SARS-CoV-2 which belongs to the Coronaviridae family and is the third major CoV pandemic in the last 20 years after Severe Acute Respiratory Syndrome (SARS) and Middle East Respiratory Syndrome (MERS) (2)(3)(4)(5)(6)(7). The lack of available therapeutic options is a major limiting factor in treating these infections, leading to excessive mortality.
Currently, there is an urgent and unmet need for effective antiviral therapy that can not only decrease the disease burden in the patient but can also decrease the ability of the person to infect others. It is not practical to develop a novel drug for urgent needs such as during the current pandemic, which may take years to confirm safety and efficacy. Alternatively, it sounds like a lucrative option to repurpose the US Food and Drug Administration (FDA) approved drugs for their efficacy against SARS-CoV-2. Earlier screens have found antiviral efficacy of approved therapies such as hydroxychloroquine (8)(9)(10)(11)(12)(13)(14); however, these therapies failed to provide any beneficial effects in COVID-19 due to their toxic side effects (13)(14)(15)(16)(17)(18)(19)(20). Finding efficacy of an approved drug against SARS-CoV-2 with minimal toxicity can provide much needed therapeutic option to treat COVID-19.
Here we screened approximately 1,700 US FDA-approved compounds to test their ability to inhibit SARS-CoV-2 replication. Here, we report 20 compounds that are highly effective in inhibiting SARS-CoV-2 replication at concentrations that were significantly lower than those having cytotoxic effects. We also investigated the possible mechanism of these compounds.
Human coronavirus (HCoV) strain OC43, which is a gift from Peking Union Medical College Hospital, was propagated in LLC-MK2 cells in 0.5% FBS MEM and virus titers were determined via TCID50 with LLC-MK2 cells. SARS-CoV-2 virus was isolated from the respiratory samples of patients in Wuhan of Hubei Province (3). SARS-CoV-2 virus was propagated in Vero cells and used in this study. All experiments with SARS-CoV-2 virus were conducted in a BSL-3 laboratory.

Antibodies
Mouse polyclonal against the OC43 N antibody was prepared in the laboratory. Rabbit polyclonal against SARS-CoV-2 N protein antibody was purchased from Sino Biological (Beijing, China). Alexa Fluor 488-conjugated goat anti-mouse IgG, Alexa Fluor 488-conjugated goat anti-rabbit IgG were purchased from Thermo.

Screening of FDA-Approved Drugs
US FDA-approved drug library which contains 1,700 compounds were purchased from TargetMol (Massachusetts, USA). LLC-MK2 cells were seeded at 2 × 10 4 cells per well in 96-well plates and incubated at 37°C and 5% CO 2 . The next day, LLC-MK2 cells were treated with the compounds at a concentration of 10 mM. After 1 h of treatment, cells were infected with OC43 at MOI of 1. At 48 h post-infection (hpi), cells were fixed with 4% paraformaldehyde for 20 min at room temperature. Immunofluorescence staining was performed using mouse anti-OC43 NP antibody, followed by anti-mouse Alexa Fluor 488 and DAPI (Sigma, St. Louis, MO). Images were captured by Operetta (PerkinElmer, Massachusetts, USA) at the magnification of 20× objective. The infection ratios were calculated using automated image analysis software (Harmony 3.5.2, PerkinElmer). Remdesivir and DMSO were used as positive and negative controls, respectively.
The positively identified drugs from this screen were used to perform dose-response curves against OC43 on LLC-MK2 and against SARS-CoV-2 using Vero cells as described below.  , cells were fixed with 4% paraformaldehyde for 20 min at room temperature. Immunofluorescence was conducted with mouse anti-OC43 N protein antibody or rabbit anti-SARS-CoV-2-NP antibody and followed by anti-mouse, or anti-rabbit Alexa Flour 488 and DAPI. Images were performed by Operetta with 20× objective. The IC 50 was calculated using automated image analysis software (Harmony 3.5.2, PerkinElmer).
For CC 50 , cells were pre-treated with each drug at concentrations 0.013, 0.041, 0.123, 0.370, 1.111, 3.333, 10, and 30 mM, respectively. After 48 h (LLC-MK2 cells) or 24 h (Vero cells) post-treatment, cell viability was evaluated by using a CCK8 kit (Yeasen, Beijing, China) according to the manufacturer's instructions. The selectivity index was calculated using the following formula: SI = CC 50 /IC 50. Graphpad Prism 7.0 was used for analyzing IC 50 and CC 50 .

Immunofluorescence
Cells were fixed with 4% paraformaldehyde for 20 min at room temperature, and permeabilized with 0.5% Triton X-100 for 10 min. Cells were then blocked with 5% BSA and stained with primary antibodies, followed by staining with an Alexa Fluor 488 secondary antibodies. Nuclei were counterstained with DAPI.

Quantitative RT-PCR
Vero cells were pre-treated with indicated concentrations of drugs for 1 h and incubated with SARS-CoV-2 at 0.1 MOI for 1 h. Then, cells were washed with opti-MEM for one time and incubated with indicated concentrations of drugs. At 24 hpi, supernatants were collected and viral RNA in the cell supernatants were extracted by using Direct-zol RNA MiniPrep kit (Zymo Research, CA, USA) according to the manufacturer's instructions. Viral copy numbers were measured by RT-PCR using primers and probe targeting the SARS-CoV-2 N gene. The reference standard was tenfold diluted from 1 × 10 9 copies to 1 × 10 4 copies. PCR amplification cycle was 50°C, 15 min, 95°C, 3 min; 95°C, 15 s, 60°C, 45 s+ Plate Read, 50 cycles. The amplification process, fluorescence signal detection, data storage, and analysis were all completed by quantitative fluorescence PCR and its own software (Bio-Rad CFX Manager). The copies of the virus were calculated according to the standard curve. The inhibition ratio was obtained by dividing the number of copies of the virus in the vehicle control group. The data were nonlinearly fitted by GraphPad 7.0 software to calculate the IC 50 of each drug.

Cell-Cell Fusion Assay
Cell-cell fusion assays were performed as described previously (20). Briefly, HEK-293T cells were co-transfected with SARS-CoV-2-S glycoprotein and eGFP. At 24 h post-transfection, cells were digested with trypsin (0.25%) and overlaid on a 50% confluent monolayer of 293T-ACE2 cells at a ratio of 1:1 which were treated with candidate drugs for 1 h. After 24 h incubation, syncytial images were captured with Operetta (PerkinElmer, Massachusetts, USA).

Inhibitory Potential of FDA-Approved Drugs Against hCoV OC43
Initial screening was performed using HCoV-OC43 due to its low biosafety concerns. OC43 is a hCoV that usually causes mild disease in humans and cattle (21). The experimental protocol is demonstrated in Figure 1A using LLC-MK2 cells infected with OC43 at an MOI of 1 for 48 h in presence of the US FDAapproved compounds. The inhibitory potentials of these compounds were measured with the treatment of compounds at 10 mM for 48 h. The viral presence was detected by immunostaining for the virus and DAPI staining for the cell nuclei. The inhibitory capacity was measured using the ratio of viral fluorescence to the DAPI and is depicted in Figure 1B. The  initial screen obtained 231 compounds that had the ability to inhibit OC43 replication >95%. The remdesivir was used as a positive control ( Figures 1C, D).
Calculation of IC 50 , CC 50, and SI of FDA-Approved Drugs Against OC43 Next, we sought to determine the effective concentrations of positively screened drugs in our initial approach and test the toxicity profile of these 231 drugs in relation to their viral inhibitory concentrations. Our data show that 56 of the positively screened drugs were effectively inhibited the viral replication at submicromolar concentrations including many of them can almost completely inhibit the viral replication at micromolar range ( Figure  2). Surprisingly, the effective drugs against the CoV belonged to a wide range of therapeutic groups including those used for neurological diseases, hormones, enzymes, and antimicrobial agents among others ( Table 1). The cytotoxic concentration to kill 50% of cells (CC 50 ) was noted for these drugs by measuring cell viability over similar concentrations. The selective index in our study was found to be >600 for 5 of the screened compounds. The SI Testing the Antiviral Effectiveness of Positive Drugs Against SARS-CoV-2 Finally, we sought to determine whether these 56 compounds can show efficacy against SARS-CoV-2, the causative agent of COVID-19. These 56 compounds from the initial screen were tested for their antiviral efficacy against SARS-CoV-2 in Vero cells. SARS-CoV-2 replicates within the Vero cells and causes cytopathic effects in these cells in the absence of any antiviral treatment. We generated the dose-response inhibition curves along with the cytotoxicity curves for these compounds in the presence of SARS-CoV-2 ( Figure 3). Remdesivir was used as a positive control. Our data show that 20 of these compounds show significant efficacy in inhibiting SARS-CoV-2 replication with sub micromolar IC 50 for many of these drugs such as  nilotinib, clofazimine, and raloxifene. The effects also confirmed by immunofluorescence assay (data not shown). These compounds also belong to a wide variety of classes including cardiac glycosides, anti-malarial drug hydroxychloroquine, cyclooxygenase-2 inhibitors, and ion channel blockers. The IC 50 , CC 50 , and SI of these compounds are shown in Table 2.

Five Candidate Drugs Inhibit Cell Fusion
Finally, to determine the mechanism by which the compounds inhibit SARS-CoV-2, we investigated these candidates' effects on virus entry. First, we constructed the cell-cell fusion assays. As indicated in Figure 4, SARS-CoV-2 S protein expression can initiate cell fusion with ACE2-overexpressed cells, but the control vector did not. Then, we detected the effects of these indicated drugs on S protein-mediated cell fusion. Indicated drugs were added to cells at 10 mM before the co-incubation of the cells. We found that fendiline hydrochloride, monensin sodium salt, vortioxetine, sertraline hydrochloride, and salifungin inhibited the SARS-CoV-2 S protein-mediated cell fusion ( Figure 4).

DISCUSSION
The current pandemic of COVID-19 is the third major outbreak in this century and the largest outbreak of the CoV in known human history. The three novel CoV outbreaks in such a short time are strong indicators of the potential threat posed by CoVs. While most respiratory viral infection research has been focused on influenza viruses that cause a huge burden of seasonal flu and occasionally pandemic outbreaks, CoV is likely to emerge as a similar or more severe pathogen than flu in long term. Given the scale and devastation of the current COVID-19 outbreak and the persistent threat of CoVs in causing human disease, there is an urgent need to find effective and safe therapies to treat these patients. Currently, there are no approved therapies for CoVs including SARS-CoV-2. The experimental therapies being used with known antiviral agents either show limited efficacy (remdesivir) or have high systemic toxicity (hydroxychloroquine), limiting their usefulness (22)(23)(24). Finding new therapies that are effective and safe are urgently needed. In this study, we have identified many FDA-approved therapies that are highly effective against CoVs, including 20 of the effective agents against SARS-CoV-2. This screening confirms previous reports demonstrating anti-SARS-CoV-2 activity of hydroxychloroquine, amlodipine besylate, arbidol hydrochloride, tilorone 2HCl, dronedarone hydrochloride, mefloquine, celecoxib, and thioridazine hydrochloride (8)(9)(10)(25)(26)(27), while identifying additional 12 drugs. However, there are seven drugs have been reported when our manuscript was underreview (28)(29)(30)(31)(32)(33)(34)(35)(36)(37). The underlying mechanisms of viral replication inhibition by these drugs are not clear. It is highly unlikely that these compounds will have similar antiviral mechanisms given the vast structural and pharmacological diversities of the effective antiviral compounds in our study. However, it is clear from other viral studies such as influenza or HIV, where antiviral drugs can affect various viral life cycle steps including attachment, entry, replication, assembly, and budding of viral progeny. Five drugs may inhibit S-mediated cell fusion as indicated by our data (Figure 4). Further studies are required to understand the precise mechanisms of each of the effective compounds found in this study.
Toxicity is one of the limiting factors in the therapeutic application of many drugs despite their known antiviral activities. Many of these drugs had SI of >600, showing promise of their usefulness at safe doses. For comparison, the SI of hydroxychloroquine was found to be 22 in our study while SI of amlodipine besylate was found to be~3, demonstrating a much lower safety profile of this drug. Similarly, other drugs known to have low selective index such as digoxin for their approved use, also show lower SI in our screen. Five of the drugs with SI of >600 include tyrosine kinase inhibitor nilotinib, antibiotics such as clofazimine and actidione, and selective estrogen receptor modulators raloxifene, and non-steroidal anti-inflammatory drug celecoxib.
Betacoronaviruses have raised great public health threats to human beings, as most known HCoVs including all the three virulent HCoVs (SARS-CoV, MERS-CoV, and SARS-CoV-2) and two seasonal HCoVs (OC43 and HKU1) belong to this species (3-7, 21, 38). It is of great value to identify antivirals against a broad spectrum of HCoVs, particularly the Betacoronaviruses, to tackle such threats by pharmaceutical interventions (39,40). To this end, we first screened the compounds which showed apparent activity of anti-OC43, the most prevalent HCoV circulates worldwide (41). We then narrowed down the candidates by the screening on SARS-CoV-2, resulting in the identification of 20 compounds which can inhibit both OC43 and SARS-CoV-2. Our study provides a foundation for subsequent anti-HCoVs drug screening of a broad spectrum. However, further tests are warranted to verify their efficacies.
In summary, our screen identified 14 previously unknown FDA-approved compounds that are effective in inhibiting SARS-  CoV-2 besides confirming the antiviral properties of 7 previously reported compounds, validating our approach. This screen identified five new compounds highly efficacious in inhibiting the viral replication of SARS-CoV-2 with SI >600. Further studies are needed to confirm these drugs' in vivo efficacy in COVID-19 relevant mouse models such as those with human ACE2 transgene (42) and human clinical studies.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.