Sertraline Is an Effective SARS-CoV-2 Entry Inhibitor Targeting the Spike Protein

ABSTRACT

The global spread of the novel coronavirus severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and the continuously emerging new variants underscore an urgent need for effective therapeutics for the treatment of coronavirus disease 2019 (COVID-19). Here, we screened several FDA-approved amphiphilic drugs and determined that sertraline (SRT) exhibits potent antiviral activity against infection of SARS-CoV-2 pseudovirus (PsV) and authentic virus in vitro. It effectively inhibits SARS-CoV-2 spike (S)-mediated cell-cell fusion. SRT targets the early stage of viral entry. It can bind to the S1 subunit of the S protein, especially the receptor binding domain (RBD), thus blocking S-hACE2 interaction and interfering with the proteolysis process of S protein. SRT is also effective against infection with SARS-CoV-2 PsV variants, including the newly emerging Omicron. The combination of SRT and other antivirals exhibits a strong synergistic effect against infection of SARS-CoV-2 PsV. The antiviral activity of SRT is independent of serotonin transporter expression. Moreover, SRT effectively inhibits infection of SARS-CoV-2 PsV and alleviates the inflammation process and lung pathological alterations in transduced mice in vivo. Therefore, SRT shows promise as a treatment option for COVID-19.

IMPORTANCE The study shows SRT is an effective entry inhibitor against infection of SARS-CoV-2, which is currently prevalent globally. SRT targets the S protein of SARS-CoV-2 and is effective against a panel of SARS-CoV-2 variants. It also could be used in combination to prevent SARS-CoV-2 infection. More importantly, with long history of clinical use and proven safety, SRT might be particularly suitable to treat infection of SARS-CoV-2 in the central nervous system and optimized for treatment in older people, pregnant women, and COVID-19 patients with heart complications, which are associated with severity and mortality of COVID-19.

INTRODUCTION

Coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is an ongoing public health emergency and has posed a severe challenge to health care systems globally (1). As of 27 October 2022, the World Health Organization (WHO) has declared more than 626 million confirmed cases and over 6.5 million deaths, in addition to huge secondary losses to the economy worldwide. SARS-CoV-2 belongs to the genus Betacoronavirus and has a genomic sequence that is closely related to its predecessor, severe acute respiratory syndrome coronavirus (SARS-CoV) of 2003 (2, 3). SARS-CoV-2 is transmitted mainly through the respiratory pathway and may cause severe pneumonia with acute respiratory distress syndrome (ARDS) and elevated plasma proinflammatory cytokines, which is known as the cytokine storm (4). ARDS is thought to be the cause of death in up to 70% of fatal COVID-19 cases (5).

Although vaccines are considered the most potent weapon to end the COVID-19 pandemic, there are still many concerns beyond vaccine efficacy and safety (6, 7). Vaccines might not be effective or safe for specific subgroups, such as pregnant women, the frail elderly high-risk population, individuals suffering from chronic health complications, and immunocompromised individuals (8). As an RNA virus, SARS-CoV-2 is constantly mutating, which also poses severe threats to the efficacy of vaccines and current therapy. Massive global efforts toward the identification of novel small molecules against SARS-CoV-2, which can enhance clinical efficacy, extend global drug supplies, and address potential emergence of viral resistance, are under way.

The repurposing of several approval antiviral therapies has been the focus of clinical treatment of COVID-19 (9). Currently, several drugs, such as hydroxychloroquine (HCQ) (10), ribavirin (RBV) (11), favipiravir (FVP) (12), lopinavir/ritonavir (LPV/r) (13), remdesivir (RDV) (14), and oseltamivir (15), have been suggested as effective treatments for COVID-19. However, there are controversial clinical results regarding their beneficial effects against SARS-CoV-2 (16–18). Recently, molnupiravir, an oral antiviral agent, has emerged as a promising new drug against COVID-19, due to its advantage over injectable drugs, such as remdesivir (19). However, molnupiravir induces RNA mutagenesis by the RNA-dependent RNA polymerase (RdRp) of SARS-CoV-2. It might possess host mutational activity as well (20). Long-term genotoxic side effects of molnupiravir should be carefully evaluated in future studies. Therefore, development of safe and novel anti-SARS-CoV-2 agents is urgent.

The SARS-CoV-2 spike (S) protein, presenting on the viral surface, plays critical roles in viral attachment, fusion, and entry, and it has been used as a target for the development of antibodies, vaccines, and entry inhibitors (21–23). The S protein comprises the S1 and S2 subunits and exists in a metastable prefusion conformation. The S1 subunit, which contains a receptor-binding domain (RBD) capable of functional folding independently, is responsible for viral binding to the receptor, human angiotensin-converting enzyme-2 (hACE2). Receptor binding is deemed to trigger significant conformational changes in the S complex which destabilize the prefusion trimer, leading to shedding of the S1 subunit and activate the fusogenic activity of the S2 subunit, which is responsible for membrane fusion (24). Membrane fusion occurs through plasma membrane or endosome, depending on the availability of host proteases. In the presence of transmembrane protease serine 2 (TMPRSS2), by which the S protein can be activated at the cell surface, fusion occurs between the viral membrane and the plasma membrane. In the absence of TMPRSS2, SARS-CoV-2 enters the cell after receptor binding via the endocytic pathway, where endosomal proteinases, such as cathepsin B/L, activate the S protein and prepare the virus for subsequent fusion (25, 26).

Early in our research on Ebola virus (EBOV) (27), we noticed that several cationic amphiphilic drugs (CADs) could act as entry inhibitors of several viruses, such as EBOV (28), rhinovirus (29), hepatitis C virus (HCV) (30), Zika virus, Dengue virus (31), etc. CADs represent part of the compounds approved by the FDA, which are widely used clinically in the treatment of mental illness, cardiovascular disease, allergies, and inflammation (32). CADs usually contain a hydrophobic aromatic ring or ring system and a hydrophilic side chain containing an ionizable amine functional group (32). They have the propensity to interact with different cell membranes, accumulate in acidic intracellular compartments, such as late endosomes/lysosomes, and affect the function of integral membrane proteins, such as acid sphingomyelinase (aSMase) (33), cholesterol, and the calcium channel (34). Thus, CADs display varied antiviral mechanism related to viral entry, replication, and budding, despite their similar structural properties (28). Interestingly, SARS-CoV-2 and EBOV rely on similar endocytic and proteolytic processes to gain entry into target cells (35). Therefore, we speculate that certain CADs might also inhibit the entry of SARS-CoV-2.

For the rapid development of therapeutic treatment of COVID-19, we screened several approved CADs based on pseudovirus (PsV) neutralization. We identified sertraline (SRT) as having a strong inhibitory effect on SARS-CoV-2 entry. SRT is a selective 5-hydroxytryptamine (5-HT) reuptake inhibitor used as a first-line treatment of major depressive disorder. We further characterize the mechanism by which SRT affects SARS-CoV-2 entry. We demonstrate that SRT exhibits protective effects against SARS-CoV-2 PsV infection in transduced mice in vivo, highlighting its clinical potential.

RESULTS

SRT inhibits SARS-CoV-2 PsV infection and S-mediated cell-cell fusion.

We first conducted a screen of FDA-approved CADs to identify potential entry inhibitor of SARS-CoV-2 using PsV (see Fig. S1 in the supplemental material). Compounds were considered active if they inhibited the pseudotyped signal by 50% and showed no or minimal effects on cell viability through a parallel antiproliferation screen in uninfected host cells.

We identified that SRT possessed the best inhibitory activities against SARS-CoV-2 PsV infection, with a 50% inhibitory concentration (IC₅₀) of 0.765 ± 0.446 μM, which was comparable to that of chloroquine (CQ), the positive control (Fig. S1). Then, we validated the antiviral activities of SRT in several cell lines, including ACE2/293T cells (Fig. 1A), Vero E6 cells (Fig. 1B), and Caco-2 cells (Fig. 1C); the IC₅₀s were 0.649 ± 0.128, 0.295 ± 0.062, and 1.344 ± 0.721 μM, respectively.

FIG 1.

Inhibition by SRT of SARS-CoV-2 PsV infection and S protein-mediated cell-cell fusion in vitro. (A to C) Antiviral activities of SRT against SARS-CoV-2 PsV infection were evaluated in 293T/ACE2 cells (A), Vero E6 cells (B), and Caco-2 cells (C). Results are from three independent experiments. (D and E) SRT inhibited SARS-CoV-2 S protein-mediated cell-cell fusion. Images were captured at 6 h after treatment with SRT to assess SARS-CoV-2 S protein-mediated cell-cell fusion (D). The syncytia were counted under an inverted fluorescence microscope, and the percent inhibition was calculated (E). Representative data from three fields were selected randomly from each sample.

We further investigated the potential role of SRT in SARS-CoV-2 S-mediated cell-cell fusion, as previously described (36). As the control, effector cells (293T/SARS-CoV-2-S/EGFP), which express SARS-CoV-2 S protein and enhanced green fluorescent protein (EGFP), and the target cells (Vero E6 cells [Fig. 1D and E] and Vero cells [Fig. S2A and B]), which express hACE2 receptor, were cocultured, allowing S-mediated fusion. In the presence of inhibitor at the indicated concentrations, SRT significantly inhibited the fusion between effector cells and target cells, resulting in a reduction in syncytium formation in a gradient concentration (Fig. 1E and Fig. S2B). Meanwhile, no fusion was observed between 293T/EGFP cells without S expression and Vero E6 cells or Vero cells (Fig. 1D and Fig. S2A). These results suggest that SRT could effectively inhibit SARS-CoV-2 entry and S-mediated cell-cell fusion.

SRT inhibits authentic SARS-CoV-2 infection in vitro.

We confirmed the antiviral activity of SRT against infectious SARS-CoV-2 in vitro. Inhibitory efficacies were measured in Vero E6 cells by quantification of viral copy numbers in the cellular supernatants by quantitative real-time PCR (qRT-PCR) and confirmed by visualizing the viral nucleoprotein (NP) expression via immunofluorescence microscopy at 24 h postinfection. SRT inhibited the replication of wild-type SARS-CoV-2 and Delta SARS-CoV-2 infection, with IC₅₀s of 1.638 ± 0.622 and 4.137 ± 0.930 μM, respectively (Fig. 2A). Similar results were obtained by immunofluorescence microscopy. SRT decreased the number of wild-type SARS-CoV-2-infected cells, and complete inhibition was observed at 6.7 μM, compared to that observed with CQ at 10 μM (Fig. 2B).

FIG 2.

Antiviral efficacy of SRT against live SARS-CoV-2 in vitro. (A) Vero E6 cells were infected with live SARS-CoV-2 at an MOI of 0.05 and treated with different concentrations of SRT for 24 h. Then the viral yield in the cell supernatant was quantified by qRT-PCR. Data are means and SD for triplicate replicates. (B) The inhibitory activity of SRT against SARS-CoV-2 infection (green) was detected by indirect immunofluorescence assay. Rabbit anti-NP polyclonal antibodies were used as the primary antibody and goat anti-rabbit IgG H&L (Alexa Fluor 488) (Abcam) as the secondary antibody. Cell nuclei were labeled with DAPI (blue). Bar = 200 μm.

SRT inhibits SARS-CoV-2 PsV entry at early stage.

A time-of-addition assay was performed to study the potential step of SARS-CoV-2 entry targeted by SRT. As shown in Fig. 3A, when SRT at 5 μM was added to 293T/ACE2 cells within 0.5 h after addition of SARS-CoV-2 PsV, the inhibitory activity was maintained at more than 70%. With the delay of the time of addition, the entry-inhibitory activity of SRT gradually decreased to 40% when SRT was added 4 h later. Additionally, when SRT was mixed with SARS-CoV-2 PsV for 0.5 h before addition to target cells, no significant difference in inhibitory activity was observed compared to that when drug was mixed with SARS-CoV-2 PsV and added to the cells at the same time (Fig. 3A). A virus binding and internalization assay was used to further confirm the targeted entry process of SARS-CoV-2 by SRT. SARS-CoV-2 PsV was preincubated with SRT before binding to cells on ice. Afterward, the cells were washed and shifted to 37°C for an additional hour to permit internalization. Lysates were analyzed by Western blotting. The presence of SRT resulted in a significant decrease in viral binding, compared to only a slight decrease in viral internalization (Fig. 3B and Fig. S3). These results indicate that SRT inhibits SARS-CoV-2 PsV entry at an early stage, possibly during the viral binding process.

FIG 3.

Effect of SRT on SARS-CoV-2 PsV entry. (A) The inhibitory activity of 5 μM SRT at 0, 0.5, 1, 2, 4, 6, and 8 h after SARS-CoV-2 PsV addition was compared with that of SRT premixed with SARS-CoV-2 PsV 0.5 h before PsV addition. (B) SRT inhibited the binding and internalization of SARS-CoV-2 PsV in Vero E6 cells. The inhibitory activity was determined by the level of p24. Representative data from three independent experiments are shown. (C) 293T/ACE2 cells were incubated with SRT at 25°C for 1 h and washed with DMEM or not before addition of the virus. Premixing SARS-CoV-2 PsV with inhibitor before addition to cells was done as a control.

Furthermore, we performed two kinds of binding assays directed toward host factors and viral proteins. When SARS-CoV-2 PsV was mixed with SRT before being added to the cells, SRT could fully inhibited SARS-CoV-2 PsV entry (Fig. 3C, red bars). However, when SRT was added to the cells before infection with SARS-CoV-2 PsV, it maintained about 50% inhibitory activity compared to the premixed condition (Fig. 3C, blue bars). In contrast, SRT lost its inhibitory ability when it was first incubated with target cells and then washed away before addition of SARS-CoV-2 PsV (Fig. 3C, black bars). These results show that the inhibitory of viral entry by SRT is dependent on the viral component but not a putative interaction with host cell factors.

SRT specifically binds to S protein.

Of note, we found that SRT showed a similar potency to inhibit the entry of SARS-CoV PsV, whose S protein shares about 76% amino acid identity with that of SARS-CoV-2 (22). SRT weakly inhibited Middle East respiratory syndrome coronavirus (MERS-CoV) PsV infection, but with no inhibitory activity against a control vesicular stomatitis virus (VSV) PsV (Fig. 4A), indicating SRT’s antiviral specificities. Thus, we speculated that SRT might act on the viral surface glycoprotein. The binding affinities of SRT with SARS-CoV-2 S protein and its various domains were determined with surface plasmon resonance (SPR) binding assay. The results showed that SRT could strongly bind to the S protein, with an equilibrium dissociation constant (K_D) of 142 nM (Fig. 4B). To determine the specific binding domain of S protein, we further evaluated the binding affinities of SRT with S1 subunit, S2 subunit, and S_RBD protein, respectively. We found that SRT showed the strongest binding affinity with S_RBD protein (K_D = 46.5 nM) (Fig. 4C), compared to those of SRT with S1 (K_D = 127 nM) (Fig. 4D) and S2 (K_D = 422 nM) (Fig. 4E). In contrast, SRT showed the least binding affinity to hACE2, with a K_D of 50,300 nM (Fig. 4F). It hardly bound to the NP (Fig. 4G). Computational molecular docking analysis showed that SRT binds to S protein and forms hydrophobic or cation-π interactions via Leu390, Leu517, and His519 in the RBD of the S protein. In addition, hydrophobic interactions also exist in the S1-SRT complex via Leu546 and Phe565 (Fig. 4H). These results suggest that SRT targets the SARS-CoV-2 S protein, especially the RBD domain, to inhibit viral entry.

FIG 4.

SRT specifically targeting SARS-CoV-2 S protein. (A) The viral-entry-inhibitory activities of SRT against infection by SARS-CoV-2 PsV, SARS-CoV PsV, and VSV PsV in 293T/ACE2 cells and MERS-CoV PsV in Huh-7 cells were tested. Experiments were performed in triplicate. Experiments were repeated twice, and similar results were obtained. (B to G) Purified SARS-CoV-2 S (B), S_RBD (C), S1 (D), S2 (E), ACE2 (F), and NP (G) proteins were immobilized on the sensor chip and SRT flowed over them for the SPR assay. SRT was diluted to different concentrations (from 0.03 to 0.5 mM) before injection. Five different concentrations were used to calculate the K_D values each time. Each experiment was repeated twice, and similar results were obtained. (H) Presumed binding sites of SRT to S protein analyzed by computational docking. SRT forms hydrophobic bonds with Leu390, Leu517, Leu546, and Phe565, and it binds to His519 by cation-π interactions.

We also investigated whether SRT might have other target sites involving in the SARS-CoV-2 entry process. First, because of its tertiary amine structure, SRT might increase the pH of an acidic environment by entering the endosome compartment (37), thus affecting the process of endosome acidification, which takes place at the late stage of viral fusion. However, a weak inhibition of acidification was observed with SRT at 5 μM, compared to that observed with NH₄Cl, a positive control (Fig. S4A). Moreover, the formation of a six-helical-bundle (6-HB) fusion core in the S2 subunit is a key step in SARS-CoV-2 S-mediated membrane fusion (36). We found that SRT did not interfere with the formation of the 6-HB fusion core (Fig. S4B). Last, SRT was found not to interfere with the activities of cathepsins, the endosomal proteases needed to prime S protein for cellular entry (Fig. S4C). Therefore, SRT specifically targets the S protein of SARS-CoV-2.

SRT inhibits S-hACE2 interaction and interferes with proteolysis of S protein.

Cell-based enzyme-linked immunosorbent assay (ELISA) showed that SRT could dose-dependently inhibit the binding of S and S1 protein to Vero E6 cells, which endogenously express hACE2 (Fig. 5A and B). Similarly, we found that SRT also effectively inhibited the binding of S and S1 to soluble hACE2 in a dose-dependent manner (Fig. 5C and D). In both assays, hACE2 was used as a positive control.

FIG 5.

Inhibition of S-hACE2 interaction and proteolysis of S protein by SRT. (A and B) Inhibition of the interaction between S (A) or S1 (B) protein and hACE2 protein on Vero E6 cells by SRT, as determined by a cell-based ELISA. (C and D) Inhibition of the interaction between soluble S (C) or S1 (D) and hACE2 protein by SRT, as determined by ELISA. Data were obtained from three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (E) SRT delayed SARS-CoV-2 S1 dissociation from S2. EGFP-tagged SARS-CoV-2 S protein from cell lysates of HEK-293T cells was immobilized on anti-GFP magnetic beads. With 0 or 10 μM SRT, the beads were exposed to pH 6 at 37°C for the indicated time. The extent of S1 dissociation was determined by Western blotting in the supernatant and bead-bound fractions. Results are means and SD from one experiment of triplicate measurements, which was representative of two independent experiments. (F) Analysis of CatL-mediated S-protein cleavage. Purified SARS-CoV-2 S protein (0.5 μg/mL) was incubated with or without CatL (5 μg/mL) or SRT (20 μM) in assay buffer (pH 5.5) at 37°C for 0, 5, 30, and 60 min. EST (20 μM) was used as a positive control. Proteins were subjected to Western blotting. Representative data from three independent experiments are shown.

The S protein is a trimeric class I fusion protein that exhibits substantial conformational flexibility. Binding to receptor hACE2 triggers the conformational changes of S protein that favor proteolytic cleavage, dissociation of the S1 subunit, and eventually conversion to the postfusion conformation (38, 39). We found that SRT prevented the dissociation of S1 subunit from S2 protein. The dissociation of S1 was not dependent on pH (Fig. S5), which was consistent with the previous results (40). Then, we treated the S samples at pH 6 and 37°C in the presence or absence of SRT. SRT appeared to introduce a 150-s delay of the shedding of S1 from S2 (Fig. 5E). Likewise, we further performed a cathepsin L (CatL) susceptibility assay to examine the proteolysis of the SARS-CoV-2 S protein with or without SRT. Treatment with SRT could prevent the cleavage of S protein by CatL, similar to the inhibition effects of aloxistatin (EST) (Fig. 5F). SRT alone did not inhibit the activities of cathepsins, as shown above (Fig. S4C).

Altogether, these results suggest that, by binding to the S protein, SRT inhibits the binding of S protein to hACE2. Moreover, SRT might stabilize the conformation of the S protein, resulting in the insusceptibility of S protein to dissociation or cleavage by proteases, which was necessary for SARS-CoV-2 entry into target cells.

SRT exhibits a synergistic effect in combination with other entry inhibitor-based antivirals against SARS-CoV-2 PsV infection.

Arbidol, HCQ, and CQ, which are entry inhibitors, have been shown to be clinically effective in COVID-19 patients (41, 42). Combinations of SRT with HCQ and with CQ exhibited strong synergistic effect against SARS-CoV-2 PsV (combination indexes [CIs], 0.12 and 0.15, respectively) with dose reductions of 426.07- and 94.99-fold for SRT based on IC₅₀s. SRT combined with arbidol showed very strong synergism against SARS-CoV-2 PsV (CI, 0.01) with dose reductions of 76.96-fold for SRT and 2435.77-fold for arbidol based on IC₅₀s (Table 1 and Fig. S6). These results suggest that SRT and clinically available entry inhibitors can be used in combination for the synergistic inhibition of SARS-CoV-2 infection.

TABLE 1.

CI and dose reduction values for inhibition of SARS-CoV-2 PsV infection by combining SRT with other antivirals

Drug combination (molar ratio) and % inhibitory concn	CI	SRT			Antiviral
		IC50 (nM)a		Dose reduction (fold)	IC50 (nM)a		Dose reduction (fold)
		Alone	Combined		Alone	Combined
SRT-HCQ (1:2.5)
50	0.12	294.68	0.69	426.07	14.40	1.74	7.28
90	0.44	7,042.88	586.03	11.02	4,055.29	1,465.07	1.77
SRT-CQ (1:3)
50	0.15	294.68	3.07	94.99	65.24	9.20	6.09
90	0.13	7,042.88	217.81	31.33	6,265.22	653.44	8.59
SRT-arbidol (3.2:1)
50	0.01	294.68	3.78	76.96	2875.39	1.18	2435.77
90	0.31	7,042.88	1,940.57	2.63	19091.00	606.43	30.48

^aData are the means of triplicate measurements. Experiments were repeated twice, and similar results were obtained.

SRT remains inhibitory against a panel of SARS-CoV-2 PsV variants.

To investigate whether SRT could inhibit various SARS-CoV-2 variants, we developed several PsV-expressing mutants with single, double, triple, or comprehensive amino acid substitutions in the pSpike-Env vector. A total of 12 key PsV mutants and Omicron variant were made, and each was used with similar levels of wild-type SARS-CoV-2 PsV for the inhibitory studies. In the absence of inhibitor, PsV with the D614G form exhibited increased luciferase expression, when all mutants were used at equivalent dose of p24, compared to wild-type virus (Fig. 6A). This is consistent with the previous finding that the D614G mutation increased SARS-CoV-2 infectivity (43). SRT exhibited broad and potent inhibitory activity against representative tested variants, with IC₅₀s ranging from 0.603 to 2.422 μM (Fig. 6B). Notably, the inhibitory activity of SRT against Omicron mutants was comparable to that against wild-type strains (Fig. 6B), which demonstrated that the antiviral activity of SRT was not affected by high numbers of mutations in Omicron S protein. The results indicate the broad-spectrum activity of SRT against multiple SARS-CoV-2 variants.

FIG 6.

Inhibitory activity of SRT against SARS-CoV-2 mutants. (A) Infectivity of SARS-CoV-2 PsV, as well as its S protein carrying single or multiple key mutations, in 293T/ACE2 cells. Data are means and SD (n = 3). RLU, relative light units. (B) Inhibitory activities of SRT against infection of SARS-CoV-2 PsV, as well as its S protein carrying single or multiple key mutations, in 293T/ACE2 cells. Data are means and SD (n = 3).

SRT inhibition of SARS-CoV-2 PsV infection is independent of serotonin transporter expression.

SRT is a serotonin transporter (SERT) inhibitor that blocks the reuptake of 5-HT in the synaptic cleft. Therefore, we decided to explore any possible role of SERT signaling in SARS-CoV-2 PsV infection. We first determined whether SERT was expressed in Vero E6 or 293T/ACE2 cells before or after SRT treatment. We found that SRT did not affect the expression of SERT in either cell type (Fig. 7A). Further, we used a set of cell lines with various combinations of SERT expression (Fig. 7A and B and Table S1), which were infected with SARS-CoV-2 PsV. Two cell lines, namely, the HeLa and TZM-bl cells, with high expression of SERT, were not readily infected with SARS-CoV-2 PsV. However, a cell line originating from human intestine in the panel, Caco-2, with low-level expression of SERT, was susceptible to SARS-CoV-2 PsV infection (Fig. 7B and C). After SRT treatment, we observed a similar dose-dependent inhibition of SARS-CoV-2 PsV infection in all SARS-CoV-2-susceptible cell lines regardless of the SERT status (Fig. 7D). This indicates that the inhibition by SRT of SARS-CoV-2 infection is independent of SERT expression. Moreover, our results suggested that SRT did not attenuate the viability of SARS-CoV-2 target cells, including 293T/ACE2, Vero E6, Caco-2, and Vero cells, with concentrations ranging from 0.39 to 25 μM (Fig. S7). SRT also displays limited hepatotoxicity and nephrotoxicity in vivo, with unaltered concentrations of serum alanine aminotransferase (ALT) and creatinine (Fig. S8A [females] and Fig. S8B [males]) in SRT-treated mice, compared with mice receiving vehicle only. SRT did not cause significant loss of neurons, apparent increase in astrocytes, or chaotic arrangement of granular cells in the hippocampal dentate gyrus (Fig. 7E). SRT is safe in vitro and in vivo.

FIG 7.

Neural assessment in vivo. (A) Western blotting of SERT expression in Vero E6 and 293T/ACE2 cells treated with 0 μM (NC) or 5 μM SRT. (B and C) Expression levels of SERT in different cell lines and their sensitivities to SARS-CoV-2 PsV infection, as determined by p24 (B) and luciferase (C). (D) Inhibitory effect of SRT on SARS-CoV-2 PsV-susceptible cell lines regardless of SERT status. (E) Representative immunohistochemistry for NeuN (marker for neurons) and glial fibrillary acidic protein (GFAP; marker for astrocytes [red arrow]) and comparison of pathological sections (HE) (magnifications: up, ×100; down, ×200) of granular cells in the dentate gyrus of the hippocampus in female ICR mice on day 15 after administration of SRT. Each image represents a group of 5 mice.

SRT inhibits SARS-CoV-2 PsV infection in Ad5-hACE2-sensitized mice.

The inhibitory effects of SRT on SARS-CoV-2 PsV infection in vivo were evaluated in Ad5-hACE2-transduced mice as previously described, with some modifications (44, 45). We first confirmed the establishment of mice sensitized to SARS-CoV-2 PsV infection. The expression of hACE2 could be observed in the lung, trachea, spleen, and intestinal tissues of BALB/c mice that were transduced intravenously with 2.5 × 10⁸ PFU Ad5-hACE2 (Fig. S9A and B). Control mice received the Ad5 empty vector. Five days later, transduced mice were challenged with SARS-CoV-2 PsV (Fig. 8A). Transduced mice infected with SARS-CoV-2 PsV lost about 15% of their body weight in the first 3 to 5 days of infection (Fig. S9C). Organ anatomy studies revealed that viruses, as determined by detection of p24, were mainly distributed in the lung, trachea, kidney, spleen, and intestine at 3 days postinfection (dpi) (Fig. S9B). As observed in the tissues of lung, spleen and intestine, the p24 signals gradually declined during the infection process from 3 dpi or by injecting PsV at 300 ng (p24) per mouse intravenously (Fig. S9D and E), similar to the expression of S gene determined by qRT-PCR in the lung (Fig. S9F and G). Also, SARS-CoV-2 PsV infection upregulated the expression of several cytokines and chemokines, including interleukin 10 (IL-10), IL-1β, tumor necrosis factor (TNF), and C-X-C motif chemokine ligand 10 (CXCL10), reflected in mRNA levels in the lung and protein levels in serum (Fig. S10A and B). Whether in lung or serum, the levels of these cytokines and chemokines peaked at 5 dpi (Fig. S10C and D), and the upward trend of these cytokines and chemokines was positively correlated with the exposure dose of virus (Fig. S10E and F). Therefore, this PsV infection mouse model, in which 300 ng (p24) per mouse was the inoculation dose and 5 dpi was the optimal time point for observation, was used for evaluation of SRT efficacy in vivo.

FIG 8.

Effect of SRT on SARS-CoV-2 PsV infection in vivo. (A) Scheme of the animal experiment. Mice were transduced with 2.5 × 10⁸ PFU of Ad5-hACE2 or Ad5-empty in 500 μL of DMEM intravenously. Five days later, transduced mice were challenged intravenously with 300 ng per mouse of p24 from SARS-CoV-2 PsV for 2 days. SRT was orally administered for two consecutive days prior to infection. (B) Tissues, including lung, spleen, and intestine, were harvested at 5 dpi for analysis. The levels of p24 were measured by Western blotting. (C) The levels of S were determined by qRT-PCR in lungs of each group (n = 5). (D and E) Levels of the indicated cytokines and chemokines from lung homogenates by qRT-PCR after normalization to GAPDH levels and comparison with normal mice (D) and protein levels of the indicated cytokines and chemokines from serum samples as determined by commercial ELISA at 5 dpi (n = 5) (E). (F) Pathological changes in lung tissues harvested at 5 dpi. Sections were stained with HE and imaged at ×50, ×100, and ×200 (bars, 1,000 μm, 200 μm, and 100 μm). Each image represents a group of 5 mice. (G) Histological scores for lung tissue from each group of mice. ***, P < 0.001; ns, not significant.

The in vivo study showed that using SRT treatment for two consecutive days prior to infection resulted in dose-dependent inhibition of pseudo-SARS-CoV-2 infection in Ad5-hACE2-transduced mice, as determined by p24 values in the tissues of the lung, spleen, and intestine (Fig. 8B) and the expression of S gene in the lung (Fig. 8C). Interestingly, mice treated with SRT also showed a dose-dependent attenuation of the inflammatory response in the lung (Fig. 8D) and serum (Fig. 8E). SRT protected mice from lung tissue histological changes (Fig. 8F and G). SRT possessed anti-inflammation activity in vitro. Of note, SRT could dose-dependently inhibit SARS-CoV-2 PsV- and lipopolysaccharide (LPS)-driven inflammation, manifesting as downregulation of the mRNA levels of IL-6, IL-1β, and TNF-α (Fig. S11). Thus, SRT treatment suppresses SARS-CoV-2 PsV infection, the subsequent inflammation process, and lung injury in mice.

Aerosol inhalation effectively delivers SRT to the lung.

SARS-CoV-2 initially attacks the epithelial cells in the respiratory tract, so drug delivery directly to the lungs may enhance the therapeutic effect and reduce side effects for other organs, such as stomach, liver, kidney, and brain. We studied the tissue distribution of SRT by inhalation, and compared the performance of inhalation adminstration (INH) with the conventional oral (p.o.) delivery of SRT in mice. We employed a previously described atomization device platform where the mice were placed in the whole-body exposure chamber and inhaled the aerosol via the mouth and nose under a constant SRT aerosol flow rate (46). Aerosol administration of SRT is effective in depositing the compound in lung, comparable to the p.o. dose (Fig. S12).

DISCUSSION

Vaccines have been regarded as an important weapon to end the COVID-19 pandemic. However, problems remain, such as their long-term safety and efficacy, as well as low awareness and acceptance by the public. There are still many ongoing attempts to develop pharmacological therapy, including small-molecule drugs, interferon therapies, oligonucleotides, peptides, and monoclonal antibodies (8, 47). Because of their high cost (such as peptides and monoclonal antibodies), rigorous storage conditions (such as interferon therapies, oligonucleotides, and monoclonal antibodies), difficulty of administration (such as interferon therapies, oligonucleotides, peptides, and monoclonal antibodies), and unclear pharmacological mechanisms (such as interferon therapies), the small-molecule regimen remains the most desirable treatment option. Notably, drug repurposing is an appealing alternative to achieve the implementation of safe antivirals with small molecules at a moderate cost in a shorter time (9).

In this study, we found that SRT potently blocks SARS-CoV-2 PsV entry at low-micromolar concentrations in various cell lines, as well as S protein-mediated cell-cell fusion. It also inhibits Vero E6 cell infection with live SARS-CoV-2.

The S protein is an important target for drug development. Although SRT has been reported to interfere with the process of endocytosis (48), the observed entry inhibition of SRT against SARS-CoV-2 PsV infection targeted the viral binding process in our study. Accordingly, we demonstrated that SRT directly binds to S protein and blocks the entry of SARS-CoV-2 PsV into cells. By comparing the binding affinities of SRT with different domains of S protein, we found that SRT shows the most potent ability to bind to the RBD of SARS-CoV-2 S1 protein. Although SRT targets epitopes distinct from the hACE2-binding site, it might inhibit RBD binding to hACE2 because of a steric hindrance rather than binding-site competition (49). The S protein of SARS-CoV-2 undergoes a series of structural rearrangements that cause fusion between the viral and cellular membranes. It is also reported that the RBDs of SARS-CoV-2 might exist in an equilibrium of up and down conformations in the context of the trimeric S on virions (50), and only the up conformation of the RBD is compatible with hACE2 binding. Then, the binding of hACE2 to S protein of SARS-CoV-2 provokes changes in the three-dimensional structure of S protein that might expose a cleavage site resulting in the protease (TMPRSS2 or CatL) digestion, subsequent shedding of the S1 subunit, and refolding of the S2 subunit required for viral fusion. The “down-to-up” conformational transition is a cooperative motion of the whole S protein involving subdomain 1 (SD1) and SD2. Therefore, the insertion of small-molecule compounds between SD1 and SD2 of S1 protein may impede the conformational motion of the S protein (51). Our molecular docking analysis reveals that SRT binds to Leu546 and Phe565, which are between SD1 and SD2. Hence, SRT might lock RBD in the “down” conformation and stabilize the S conformation, which leads to the inability of S protein to bind to hACE2 and the subsequent cleavage by proteases. The detailed binding mode of SRT-hACE2 and the mechanism of SRT blocking of S1 dissociation from the S2 subunit warrant further investigation.

The ongoing upsurge in the newly emerging COVID-19 variants raises serious concerns over the efficacy of current vaccines and therapies. Since the beginning of the pandemic, Alpha (B.1.1.7), Beta (B.1.351), Delta (B.1.617.2), Gamma (P.1), and Omicron, which is currently the global dominant circulating strain (B.1.1.529 and BA lineages), were named as SARS-CoV-2 variants of concern (VOC) (52). We construct several key mutations in the S protein of VOC (53), such as D614G, L452R, P681R, P681H, and N501Y, which are associated with higher virulence and increased viral infectivity (54–56), and E484Q and E484K, which can favor viral escape from antibodies and from convalescent-phase sera of recovered individuals (57, 58). SRT retains potent inhibitory activity against these SARS-CoV-2 variants. Interestingly, SRT displays strong inhibitory activity against the Omicron variant compared to the wild-type strain. These results reveal that SRT broadly inhibits SARS-CoV-2 variants. Furthermore, SRT shows strong synergistic effects with several small-molecule entry inhibitors of SARS-CoV-2 in vitro. These results indicate that SRT could be used in combination with compounds in current SARS-CoV-2 clinical pipelines to potentially prevent the emergence of drug resistance, augment efficacy, decrease toxicity, and provide a broader spectrum of activity than monotherapy regimens (59).

“Old drug, new use” has been an important field of clinical focus to treat COVID-19 and has addressed compounds such as CQ and arbidol. Compared with these drugs, SRT has several strengths. First, as an old drug targeting the central nervous system, SRT can penetrate the tissues of the central nervous system. It is reported that 80% of COVID-19 patients had at least one neurological complication and that neurological signs or syndromes significantly increased the risk of death during acute hospitalization (60). SRT might be effective in inhibiting infection of SARS-CoV-2 in the central nervous system, thus preventing potential long-term neuronal damage as well as decreasing COVID-19 mortality. Second, SRT has been confirmed as safe for clinical use in specific subjects, such as older patients (61, 62), pregnant women (63), and patients with cardiovascular complications (64, 65), who often have increased risk for SARS-CoV-2 infection and infection-associated morbidity and mortality (66, 67). They are also not eligible for vaccination. Therefore, SRT might be particularly useful for optimizing COVID-19 treatment and prevention strategies for older patients, pregnant women, and patients with cardiovascular complications. Third, coinfection with different microorganisms and SARS-CoV-2 is a serious problem in the COVID-19 pandemic (68). Because coinfection causes more serious damage to the immune system, COVID-19 patients with coinfection, including fungal, bacterial, and viral infections, might experience worse disease outcomes, more complicated treatment, and longer treatment cycles in general. SRT possesses strong intrinsic antibacterial and antifungal activities, and it could augment the antibacterial activities of antibiotics (69, 70). It might be beneficial to treat SARS-CoV-2 coinfections with SRT. Last, SRT could be used locally, such as in an inhalation formulation. As an alternative route, aerosol lung administration of SRT would diminish side effects and decrease the therapeutic dose, compared with systemic administration by the oral or intravenous route. Inhaled treatment is also suitable for older people who have difficulty in swallowing, compared with oral administration.

SARS-CoV-2 causes an extremely active inflammatory response, known as a cytokine storm, which is the major cause for the severity and mortality of COVID-19. Targeted approaches to curb inflammation hold the promise of improving prognoses and reducing COVID-19-associated mortality. SRT has been found to have anti-inflammatory effects in several clinical studies (71, 72). In our mouse model, inflammatory mediators, including IL-1β, IL-10, TNF-α, and CXCL10, were significantly upregulated after SARS-CoV-2 PsV infection, which was consistent with findings in the sera of SARS-CoV-2 patients (73). SRT has been shown to suppress the inflammatory process induced by SARS-CoV-2 PsV infection in vivo. It is reported that SRT can directly bind to TNF-α and TNF-α receptor 1 (TNFR1), thus exerting anti-inflammatory activities by suppressing TNF-α-induced NF-κB and the downstream signaling pathway (74). NF-κB signaling is associated with the ongoing SARS-CoV-2-mediated cytokine storm and tissue damage (75, 76). Therefore, SRT might target the NF-κB signaling pathway to inhibit SARS-CoV-2 PsV-mediated inflammation. However, the exact mechanism by which SRT inhibits the inflammation process induced by SARS-CoV-2 should be investigated.

In this study, we first employed PsV, which packages the S protein at the surface of replication-incompetent HIV, allowing us to identify SRT as a potent SARS-CoV-2 entry inhibitor. We then used a mouse model of preclinical PsV infection to further verify the efficacy of SRT in inhibiting SARS-CoV-2 entry in vivo (45). Due to the incompatibility between SARS-CoV-2 and mouse ACE2 receptor, wild-type mice are less susceptible to SARS-CoV-2. We employed a transduced-hACE2 mouse model, with the advantage of being able to carry out immediate studies using multiple mice compared to the time-consuming process of breeding hACE2-transgenic mice. After challenge with SARS-CoV-2 PsV, viruses were enriched in the lung, spleen, trachea, intestine, and kidney of transduced mice, which was analogous to the organ distribution of wild-type SARS-CoV-2 (77). In addition, the infection model also causes inflammation. Therefore, this model is a good choice for preliminary studies to explore drug and vaccine candidates that are based on SARS-CoV-2 S protein. Although we confirmed the anti-PsV efficacy of SRT in this model, further evaluation of the inhibitory effects of SRT against infectious SARS-CoV-2 in vivo is warranted.

We have identified SRT as a potent and safe entry inhibitor of SARS-CoV-2 that acts by binding to the S protein and blocking the viral attachment step. It can be further developed as a therapeutic medicine for the treatment of COVID-19 patients.

MATERIALS AND METHODS

Cell cultures, plasmids, and live virus.

Vero cells, HeLa cells, A549 cells, and Huh-7 cells were purchased from Guangzhou Saiku Biotechnology Co., Ltd. HEK-293T cells, Vero E6 cells and 293T/ACE2 cells (hACE2 is stably expressed by HEK-293T) were kindly provided by Lu Lu of Fudan University. Müllar cells (MIO-M1 cells) were kindly provided by Yi Wang of Wenzhou Medical University. TZM-bl cells were obtained from the NIH AIDS Research and Reference Reagent Program. All cells were grown in Dulbecco’s modified Eagle medium (DMEM) (Gibco, USA) supplemented with 10% fetal bovine serum (FBS) (Excell Bio, Shanghai) and 1% penicillin-streptomycin (PS) (Gibco, USA) at 37°C in a humidified atmosphere of 5% CO₂, except for A549, which was maintained in RPMI 1640 medium (Gibco, USA) supplemented with 10% FBS and 1% PS at 37°C in a humidified atmosphere of 5% CO₂.

The envelope-expressing plasmids of SARS-COV-2-S (pcDNA3.1-SARS-2-S), SARS-S (pcDNA3.1-SARS-S), MERS-CoV-S (pcDNA3.1-MERS-S), and the plasmids pAAV-IRES-EGFP and pAAV-IRES-SARS-CoV-2-S-EGFP were kindly provided by Lu Lu (Fudan University, China). The plasmid expressing full-length VSV glycoprotein (VSV-G) was maintained in our laboratory. The plasmid pNL4-3.Luc.R-E-, the luciferase reporter expressing the HIV-1 backbone, was obtained from the National Institutes of Health AIDS Research and Reference Reagent Program.

Authentic preserved SARS-CoV-2 (Beta-CoV/Wuhan/WIV04/2019, IVCAS6.7512; Delta, IVCAS6.7585) was obtained from the National Virus Resource, Wuhan Institute of Virology, Chinese Academy of Sciences, and handled in a biosafety level 3 (BSL-3) laboratory.

Antiviral compounds and reagents.

Benztropine mesylate (catalog no. T1336), clomiphene citrate (catalog no. T1193), toremifene citrate (catalog no. T1464), clemastine fumarate (catalog no. T0147), maprotiline hydrochloride (catalog no. T0172), clomipramine hydrochloride (catalog no. T0255), teicoplanin (catalog no. T0967), sertraline hydrochloride (catalog no. T0482), chloroquine diphosphate (catalog no. T0194), hydroxychloroquine sulfate (catalog no. T0951), SR-12813 (catalog no. T6994), arbidol hydrochloride (catalog no. T0104), and betaxolol hydrochloride (catalog no. T0226) were purchased from TargetMol (USA).

The purified proteins, including SARS-CoV-2 S protein (S1+S2 extracellular domain (ECD), amino acid residues Val16 to Pro1213; catalog no. 40589-V08H4), SARS-CoV-2 S2 recombinant protein (amino acid residues Ser686 to Pro1213; catalog no. 40590-V08B), SARS-CoV-2 S1 recombinant protein (amino acid residues Val16 to Arg685; catalog no. 40591-V08H), SARS-CoV-2 S RBD recombinant protein (amino acid residues Arg319 to Phe541; catalog no. 40592-V08H), hACE2 protein (catalog no. 10108-H05H), and the anti-SARS-CoV-2 S (catalog no. 40592-R190) and S1 (catalog no. 40150-R007-H) monoclonal antibodies were bought from Sino Biological (Beijing, China). The NP was a gift from VACURE Biotechnology Co., Ltd. (Chengdu, China).

HR1P (ANQFNSAIGKIQDSLSSTASALGKLQDVVNQNAQALNTLVKQ) and HR2P (DISGINASVVNIQKEIDRLNEVAKNLNESLIDLQEL) with purity of >95% were synthesized by GL Biochem Ltd. (Shanghai, China).

Hydroxypropyl-β-cyclodextrin (CAS no. 128446-35-5) aqueous solution was used as a solvent for SRT throughout the in vivo animal experiments.

PsV preparation.

The SARS-CoV-2 PsV were prepared as described elsewhere (36). In brief, HEK-293T cells were seeded at a density of 4 × 10⁵ cells/mL in 6-well plates. After reaching 80 to 85% confluence, HEK-293T cells were cotransfected with pNL4-3.luc.R-E- and pcDNA3.1-SARS-CoV-2-S (encoding SARS-CoV-2 S protein) using PolyJet reagent (SignaGen, USA). PsV, including VSV, MERS-CoV, and SARS-CoV, were prepared in an identical fashion, but with the substitution of a plasmid encoding VSV-G protein MERS-CoV or SARS-CoV S protein. The supernatant containing pseudotyped particles were harvested 48 h posttransfection, centrifuged at 3,000 × g for 10 min, and then frozen at −80°C.

Live SARS-CoV-2 inhibition assay.

The inhibition assay of live SARS-CoV-2 was performed as previously described (78). Briefly, SARS-CoV-2 was propagated and titrated with Vero E6 cells. The target Vero E6 cells were pretreated with gradient-diluted SRT for 1 h at 37°C, and SARS-CoV-2 were then added at a multiplicity of infection (MOI) of 0.05. The viruses and cells were incubated for 1 h at 37°C. After that, uninfected viruses were removed, and fresh medium with the corresponding concentration of SRT was added. Supernatants were lysed using a MiniBEST viral RNA/DNA extraction kit (TaKaRa, Japan) 24 h later, and the viral RNAs were reverse transcribed with a PrimeScript RT reagent kit and gDNA Eraser (TaKaRa, Japan). Viral copies in cell supernatants were quantified from viral cDNA by a standard curve method on an ABI 7500 system with a pair of primers targeting the S gene.

Indirect immunofluorescence assay.

Vero E6 cells collected from the live SARS-CoV-2 inhibition assay as described above were fixed with 4% paraformaldehyde (Bio-Rad, USA) and then permeabilized with 0.2% Triton X-100 (Sigma, USA) for 30 min. After blocking with 2% skim milk, cells were incubated with anti-NP poly-antibodies (Sino Biological, catalog no. 40143-T62). Subsequently, cells were incubated with goat-anti-rabbit IgG H&L (Alexa Fluor 488) (Abcam, UK) for 1 h followed by counterstaining with DAPI (4′,6-diamidino-2-phenylindole) staining solution (78). Fluorescence images were acquired using an Axio Observer microscope (Zeiss, Germany).

Time-of-addition assay.

293T/ACE2 cells seeded in 96-well plates were incubated with SARS-CoV-2 PsV (50 ng of p24), while SRT at a final concentration of 5 μM was added 0.5 h before or 0, 0.5, 1, 2, 4, 6, or 8 h after addition of the PsV. Cells were lysed 48 h postinfection to determine the entry inhibition efficacy.

Time-of-removal assay.

293T/ACE2 cells were plated in 96-well plates with a density of 2 × 10⁴ cells per well. After 24 h, cells were incubated with SRT (final concentration, 5 μM or 2.5 μM) at 25°C for 1 h and then added directly with SARS-CoV-2 PsV (50 ng of p24) or washed with DMEM 3 times before addition of SARS-CoV-2 PsV. An entry inhibition assay in which SRT was coincubated with SARS-CoV-2 PsV prior to addition to cells was also performed.

Cell-based ELISA.

Vero E6 cells, which express hACE2, were plated in 96-well plates at a density of 40,000 cells per well and incubated overnight at 37°C. Next, following washing with phosphate-buffered saline (PBS), cells were fixed in 10% formaldehyde buffer at 25°C for 30 min. Cells were then washed twice with PBS and blocked with 2% bovine serum albumin (BSA) at 25°C for 1 h. SARS-CoV-2 S or S1 protein (50 μL) was pretreated with inhibitor or hACE2 at 4°C for 30 min. The mixtures were then added to cells and incubated at 25°C for 1 h. Plates were washed three times and incubated with anti-SARS-CoV-2 S or S1 monoclonal antibody diluted 1:2,000 at 25°C for 1 h, followed by addition of 50 μL of 1:5,000-diluted horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibody (Abbkine, catalog no. A21020) at 25°C for another hour. The substrate 3,3,5,5-TMB (3,3′,5,5′-tetramethylbenzidine) was added, and optical density at 450 nm (OD₄₅₀) was determined with a multidetection microplate reader (22).

Inhibition of the binding of soluble hACE2 and S protein.

Wells of 96 half-area plates were coated with 10 μg/mL of hACE2 protein in 0.01 M Tris buffer (pH 8.8) at 4°C overnight. On the following day, the plates were washed three times with PBS containing 0.05% Tween 20 (PBS-T) and blocked with 3% skim milk prepared in PBS-T at 37°C for 1 h. Samples prepared as described above were added to plates at 37°C for 1 h. The plates were washed and incubated with monoclonal antibody against SARS-CoV-2 S or S1. Then, the supernatant was removed, and the plates were washed five times with PBS-T, followed by the addition of 1:5,000-diluted goat anti-rabbit IgG-HRP secondary antibody (catalog no. A21020; Abbkine, USA) at 37°C for 1 h. TMB and 1 N H₂SO₄ were added sequentially. The OD₄₅₀ was measured using the multidetection microplate reader.

Measurement of the ability to inhibit SARS-CoV-2 PsV entry of SRT individually and in combination with other antivirals.

The antiviral activity of SRT alone or in combination with other antiviral agents, was assessed in 293T/ACE2 cells as described above. Briefly, the precultured 293T/ACE2 cells were infected with SARS-CoV-2 PsV (50 ng of p24) in the presence or absence of an individual inhibitor or two inhibitors in combination at the graded concentration. Two days later, the cells were harvested and lysed with lysing reagent. Luciferase activity was measured using a luciferase kit.

The 50% inhibitory concentration (IC₅₀) and 90% inhibitory concentration (IC₉₀) were determined for each drug. The combination index (CI) was then calculated by using the CalcSyn program to assess the synergistic effect of combinations. CIs of >1, 1, and <1 indicate antagonism, additive effect, and synergism, respectively. The strength of synergism is shown by the CI values (<0.1, very strong synergism; 0.1 to 0.3, strong synergism; 0.3 to 0.7, synergism; 0.7 to 0.85, moderate synergism; and 0.85 to 0.90, slight synergism). Dose reduction (fold) was calculated as described elsewhere (79).

Antiviral experiments of SRT in vivo.

Female BALB/c mice (6 to 8 weeks old; weight, 18 to 22 g) were purchased from the Laboratory Animal Center of Southern Medical University (Guangzhou, China). Briefly, mice were first transduced by injecting 2.5 × 10⁸ PFU of Ad5-hACE2 or Ad5-empty (control) in 500 μL DMEM as before (44). For the day-optimizing test, mice challenged with 200 ng of SARS-CoV-2 PsV were monitored over a 10-day time course. For the viral challenge experiments, different doses of SARS-CoV-2 PsV (37.5, 75, 150, or 300 ng of p24/mouse) were injected via the tail vein (500 μL per mouse). The viral load was detected by Western blotting for p24 signal, and body weights were monitored at the indicated time points. For the compound activity experiments, at 4 days postransduction, mice were administered with SRT orally once daily on two consecutive days. At day 5 postransduction, mice were injected with SARS-CoV-2 PsV via the tail vein on two consecutive days. SARS-CoV-2 PsV for the in vivo experiments were purified and concentrated by sucrose gradient centrifugation to remove potentially immunogenic substances, such as cell culture media or cellular metabolites. The viral particles were diluted with DMEM and p24 were quantified by ELISA (80). Mice without any treatment were used as the untreated control group (NC). Mice transduced with the Ad5-hACE2 and inoculated with PsV were treated as the vehicle group. At the indicated time point, mice were sacrificed for determination of p24 levels by Western blotting and S protein and hACE2 levels by qRT-PCR. The levels of cytokine and chemokine in the serum and the lung tissue were determined by ELISA and qRT-PCR, respectively. Hematoxylin-eosin (HE) staining was used to monitor the lung histological changes.

qRT-PCR.

Total RNA was isolated from lung homogenates using the total RNA isolation kit (Foregene, China) and reverse transcribed with Prime-Script RT master mix (TaKaRa, Japan) according to the manufacturer’s instructions. Gene expression was quantified using Go Taq qPCR master mix (Promega, USA) with commercial primers/probe sets specific for IL-10, IL-1β, TNF-α, CXCL10, hACE2, and S protein. Results were normalized to GAPDH. The sequences used are shown in Table S2.

HE assay.

After fixation in 4% paraformaldehyde solution for 24 h, the lung tissues randomly selected in each group were cut into 5-μm-thick sections. The sections were stained with hematoxylin-eosin, and the histopathological changes were observed under an optical microscope (Zeiss, Germany). The severity of injury of the lung was monitoring using the scoring system reported previously (81), as follows: 0, normal; 1, extremely mild damage (<25% of the visual field); 2, mild damage (25% to 50% of the visual field); 3, moderate damage (50% to 75% of the visual field); and 4, severe damage (>75% of the visual field).

Study approvals.

All the procedures conducted on animals were complied with the ARRIVE guidelines and approved by the Ethics Committee of Southern Medical University.

Statistical analysis.

Significance analysis was performed using SPSS for Windows, version 22.0.0.0 (SPSS, Chicago, IL, USA) and analyzed using one-way analysis of variance (ANOVA). A P value of <0.05 was considered statistically significant. Data are presented as means and standard deviations (SD).