Sigma‐1 Receptor‐Targeting Compounds Show Antiviral Activity Across Multiple Coronavirus Models

2.1. Antiviral Activity of σ1R Antagonists Against SARS‐CoV‐2 In Vitro

To validate σ1R as a potential therapeutic target against SARS‐CoV‐2, a σ1R knockout Vero E6 cell line was established by employing gene editing technology to test viral replication. Viral replication was then compared between the σ1R‐knockout (σ1R‐KO) and wild‐type (WT) Vero E6 cells to assess the role of σ1R. Viral titration (TCID₅₀) was used to measure infectious viral titers in cell culture supernatants at 72 h post‐infection (hpi) to quantify viral load. As shown in Figure 1A, the viral titer in wild‐type Vero E6 (WT‐Vero E6) cells was significantly increased, whereas that in σ1R‐KO Vero E6 cells exhibited a marked decrease. These results suggest that σ1R is a key host factor involved in SARS‐CoV‐2 replication, demonstrating its potential value as a therapeutic target. To further validate σ1R as a target, two commercially available, highly specific σ1R antagonists BD‐1047 dihydrobromide (BD‐1047) and σ1 receptor antagonist‐1 (σ1RA1) were selected for evaluation of their antiviral activity and cytotoxicity in the Vero E6 cell model. BD‐1047 antagonists is widely used in neuroscience and virology research. σ1RA1 is a highly selective σ1 receptor antagonist. Cytotoxicity assay (Figure 1B) determined the median cytotoxic concentration (CC₅₀) of BD‐1047 was 125 μM, while that of σ1RA1 was 85.19 μM. The 90% inhibitory concentrations (IC₉₀) for SARS‐CoV‐2 replication are presented in Figure 1C–H, with specific values for each variant annotated within the figures. These data indicate that both σ1R antagonists exhibit potent antiviral effects at concentrations substantially below their cytotoxic thresholds, demonstrating a favorable safety profile alongside anti‐SARS‐CoV‐2 activity. Furthermore, the 50% inhibitory concentrations (IC50) and selectivity indices (SI, calculated as CC50/IC50) were determined to comprehensively evaluate the pharmacological profiles, providing robust 95% confidence intervals for each candidate compound. The SI value of BD‐1047 against SARS‐CoV‐2 is 108.7, and that of σ1RA1 is 45.38.

2.2. Antiviral Activity of σ1R Antagonists Against SARS‐CoV‐2 In Vivo

Research on the in vivo anti‐SARS‐CoV‐2 activity of σ1R antagonists remains limited. Previously, we have evaluated the antiviral efficacy of the σ1R antagonist PB28 in K18‐hACE2 male mice [19]. In this research, we established mouse model susceptible to SARS‐CoV‐2/BMA8 strain infection to assess the in vivo antiviral potency of σ1R antagonist BD‐1047 and σ1RA1 (Figure 2A) [20], with the Molnupiravir serving as a positive control. Our results indicated that treatment with BD‐1047 or σ1RA1 significantly improved the survival rate of mice infected with SARS‐CoV‐2/BMA8; notably, the survival rate reached 100% in the group treated with BD‐1047 at a dosage of 10 mg/kg (Figure 2B and Figure S1A). We continuously monitored the body weight changes of mice for 21 days. The results revealed that, within the first 5 days post‐viral challenge, the infected mice exhibited persistent weight loss attributed to SARS‐CoV‐2/BMA8 infection. During Days 6–7, body weight entered a plateau phase, followed by a gradual increase (Figure 2C and Supporting Information S1: Figure S1B). Comprehensive analysis incorporating survival rate data and body weight dynamics demonstrated that BD‐1047 exerted relatively prominent antiviral efficacy. Specifically, BD‐1047 at a dosage of 10 mg/kg provided superior antiviral activity compared to the positive control drug Molnupiravir, as well as BD‐1047 (3 mg/kg) and σ1RA1 (3 mg/kg) exhibited some protection against challenge with SARS‐CoV‐2/BMA8.

In addition, the SARS‐CoV‐2 N gene was detected in the nasal turbinate and lung tissues of treated mice at 4 days‐post‐infection (4dpi). The results showed N gene copies in drug‐treated groups were significantly lower than the infection‐only group (Figure 2D,E, Supporting Information S1: Figure S1C,D). It indicated that σ1R is an promising target for controlling coronaviruses by blocking SARS‐CoV‐2 proliferation in vivo. Pulmonary pathological damage is a major index of SARS‐CoV‐2 infection. Lungs of untreated mice exhibited typical severe inflammatory infiltration and alveolar edema, whereas pathological damage was significantly alleviated in all drug‐treated groups (Figure 2G,F). Pathology scores indicates σ1R antagonists effectively mitigate infection‐induced lung injury. In summary, all results demonstrate that BD‐1047 and σ1RA1 effectively suppress SARS‐CoV‐2 replication in vivo and mitigate disease severity, and further substantiate σ1R as a promising therapeutic target for infection with coronaviruses.

2.3. Compounds Screening In Silico Based on σ1R

To identify potential σ1R ligands with therapeutic potential against coronaviruses, molecular docking simulations were employed to predict the interactions between a large library of drug candidates (exceeding 10,000 compounds) and σ1R. The free energy of binding (in kcal/mol) was used as a critical metric to evaluate binding potency, where a value < −5 kcal/mol was defined as indicative of high‐affinity binding‐a threshold set to prioritize compounds with robust σ1R‐binding potential.

Subsequently, the top‐ranked lead compounds that met this high‐affinity criterion were selected and summarized in Table 1. As being depicted, top eighteen compounds all fall into the small‐molecule drug class, with binding affinities ranging from −12.7 to −16.9 kcal/mol. Each compound is associated with a unique CAS Number, enabling unambiguous chemical identification and facilitating subsequent experimental validation.

2.4. Anti‐SARS‐CoV‐2 Activity Evaluation of Drug Candidates In Vitro

To validate the antiviral potential of candidate compounds, the present study determined the in vitro anti‐SARS‐CoV‐2 activity of the top 18 prioritized compounds selected through preliminary screening. The safety and anti‐SARS‐CoV‐2 activity of 18 drugs were evaluated by measuring their CC₅₀ and IC₉₀. The results showed that Ocaperidone and Balaglitazone conferred significant antiviral activity (Figure 3A,B), while Pyrvinium displayed moderate antiviral efficacy but with relatively high cytotoxicity. Tasosartan exerted only modest antiviral effects (Supporting Information S1: Figure S2A,B). Ocaperidone is a benzisoxazole derivative that was previously developed as an atypical antipsychotic compound [21]. Balaglitazone is a partial agonist of peroxisome proliferator‐activated receptor γ (PPARγ) and was originally developed as an antidiabetic agent [22]. Our study results suggest both Ocaperidone and Balaglitazone possess anti‐SARS‐CoV‐2 activity without prior reports.

Given the ongoing evolution of SARS‐CoV‐2 variants and their enhanced immune escape capabilities, broad‐spectrum antiviral activity of candidate compounds is critical for clinical application. Therefore, we expanded the activity assessment to multiple clinically prevalent and highly pathogenic SARS‐CoV‐2 variants, including Beta, Delta, and the immune escape‐prone Omicron sublineages (BA.1, BA.2, and JN.1). Results (Figure 3C–G) show both compounds provided highly significant inhibitory activity against Beta variant, but Ocaperidone showed even greater efficacy against the original strain. Both compounds conferred potent inhibitory effect against the Delta variant and Omicron sublineages.

In summary, these findings hold significant biological and clinical translational prospection. First, the potent inhibitory effects of Ocaperidone and Balaglitazone against the wildtype SARS‐CoV‐2 and major variants suggest that target, σ1R, may be a conserved indispensable factor in viral replication. It implies that these compounds hold promise in treatment of viral immune escape variants of SARS‐CoV‐2, addressing the issue of drug resistance caused by rapid viral mutations.

2.5. Predicted Binding Modes of Ocaperidone and Balaglitazone With σ1R

To elucidate the molecular mechanisms by which Ocaperidone and Balaglitazone interact with the σ1R, molecular docking technique was employed to systematically analyze the binding profile of these two compounds with σ1R. Molecular docking experiments were performed for Balaglitazone with the σ1R yielding 3 conformational results with docking scores of −7.903, −7.773, and −7.028 kcal/mol, respectively. The interaction mode between Balaglitazone and σ1R is illustrated in Figure 4A. The primary interaction arises from the π‐π stacking between Balaglitazone and residues W89 and F107 of the target protein. Molecular docking of Ocaperidone with the σ1R generated 9 conformational results, with docking scores ranging from −10.611 to −3.564 kcal/mol. Among these, the highest docking score reached −10.611, which was the highest among all docking results, indicating that Ocaperidone had a stronger binding ability with the σ1R (Figure 4B). Overall, Ocaperidone showed better binding ability to the σ1R than Balaglitazone and exhibited certain binding stability in multiple docking experiments. However, there were relatively large differences among different docking results, which might be related to factors such as ligand conformational changes and the microenvironment of the receptor's active pocket. However, the main interactions were π‐π stacking between Ocaperidone and residue F133 of the target protein, and hydrogen bonding with residue E172. Overall, the docking analysis suggested that both compounds could be accommodated within the σ1R binding pocket, with Ocaperidone showing a more favorable predicted docking score than Balaglitazone.

2.6. In Vivo Anti‐SARS‐CoV‐2 Activity of the Drug Candidates

To comprehensively evaluate the anti‐SARS‐CoV‐2 activity of Ocaperidone and Balaglitazone, an pathogenic BALB/c mice infection model was established using the mouse‐adapted SARS‐CoV‐2/BMA8 strain. We comprehensively validate the in vivo protective effects of both compounds by measuring multiple parameters including body weight, survival rate, viral load, and histopathology.

We first performed statistical analysis of changes in body weight and survival rate of mice treated with different concentrations of the two candidate compounds. The results demonstrated that a relatively high concentration of Ocaperidone (2 mg/kg) induced toxicity in mice, whereas treatment with Ocaperidone at 0.5 mg/kg provided antiviral efficacy comparable to that of the positive control drug Molnupiravir. It is noteworthy that after the weight‐loss plateau phase [23], the body weight of Ocaperidone‐treated mice recovered at a faster rate than that of mice in Molnupiravir group. Given that SARS‐CoV‐2 infection is known to induce anxiety‐ and depression‐like disorders, the accelerated weight recovery observed herein may be attributed to the intrinsic psychotropic properties of Ocaperidone (Figure 5A,B). By contrast, Balaglitazone exhibited relatively weak antiviral efficacy in the mouse model, suggesting that it may not possess favorable potential as an anti‐SARS‐CoV‐2 agent (Supporting Information S1: Figure S1A,B).

To further clarify the in vivo antiviral effects of the Ocaperidone, we quantified SARS‐CoV‐2 N gene copy numbers in nasal turbinates and lung tissues of infected mice (Figure 5C) using qRT‐PCR. Results showed significantly elevated viral loads in respiratory tissues of infected mice without treatment, whereas viral loads were markedly reduced in mice treated with Ocaperidone (0.5 mg/kg). We evaluated the ameliorative effects on pathology via HE staining (Figure 5D). Lungs of infected mice without treatment exhibited severe pathological manifestations, including massive inflammatory cell infiltration (e.g., macrophages and neutrophils), alveolar structural destruction, and diffuse parenchymal lesions. In contrast, lung tissue damage was significantly reduced in groups treated with Ocaperidone, exhibiting relatively intact alveolar architecture and diminished inflammatory infiltration. Quantitative histopathology scoring further validated the difference. Pathology scores were markedly elevated in the infected group, while treatment groups exhibited substantial reductions. This indicates Ocaperidone effectively disrupt the “viral replication‐ inflammation amplification” vicious cycle, thereby mitigating SARS‐CoV‐2‐induced pulmonary inflammation and structural damage.

Taken together, these data indicate that Ocaperidone exerted protective effects in SARS‐CoV‐2/BMA8‐infected mice, as reflected by improved survival, reduced respiratory viral load, and alleviated pulmonary pathology. In contrast, Balaglitazone showed limited in vivo efficacy in this model. These findings support further evaluation of Ocaperidone as a candidate compound in host‐directed anti‐coronavirus studies.

Herein, this study confirms the potential of the σ1 receptor as a therapeutic target against SARS‐CoV‐2 and provides robust preclinical evidence supporting the clinical translation of Ocaperidone and Balaglitazone for COVID‐19 treatment.

2.7. In Vitro Broad‐Spectrum Antiviral Activity of Candidate Drugs Against Coronaviruses

To extend the antiviral evaluation beyond SARS‐CoV‐2, we assessed the activity of BD‐1047, σ1RA1, and Ocaperidone against representative coronaviruses from different genera, including HCoV‐OC43, PEDV, and IBV. We first determined the CC50 values of these compounds in HCT‐8 cells (Figure 6A) and then evaluated their antiviral activity using IC90 as the core efficacy metric.

The results showed that all candidate compounds exhibited potent inhibitory activity for HCoV‐OC43, with their IC₉₀ values falling within the effective concentration range established for preclinical studies. Among these compounds, BD‐1047 and Ocaperidone displayed significantly superior inhibitory efficacy against HCoV‐OC43 (Figure 6B). In efficacy evaluations against PEDV, BD‐1047 and Ocaperidone also exhibited remarkable inhibitory effects, but Balaglitazone with moderate activity (Figure 6C). For IBV, σ1RA1 achieved the second‐highest inhibitory efficacy, inferior only to BD‐1047 (Figure 6D).

Collectively, these findings indicate that BD‐1047, σ1RA1, and Ocaperidone exhibit antiviral activity across multiple coronavirus models, although the level of supporting evidence differs among viruses. Among the tested compounds, BD‐1047 and Ocaperidone showed the most consistent activity profiles and therefore warrant further investigation.

2.8. In Vivo Antiviral Activity Study of Compounds Against HCoV‐CO43

We further evaluated the in vivo antiviral efficacy of candidate compounds (Figure 7A). The two compounds BD‐1047 and Ocaperidone, that showed the most potent activity at the cellular level, against HCoV‐OC43 in a mouse model. Treatment with the positive control drug Arbidol hydrochloride (ARB) prolonged the survival time of infected mice. However, all mice succumbed to infection by day 6 post‐challenge. In contrast, BD‐1047 and Ocaperidone significantly improved the survival rate of infected mice, demonstrating superiority in vivo protective efficacy (Figure 7B, Supporting Information S1: Figure S4A). BD‐1047 (10 mg/kg) exerted a more potent anti‐HCoV‐OC43 effect than the positive control drug ARB. Additionally, Ocaperidone $(0.5\text{mg}/\text{kg})$ exhibited a prominent efficacy in ameliorating symptoms such as head shaking and convulsions in mice during the late phase of viral infection (Figure 7C, Supporting Information S1: Figure S4B). Importantly, both BD‐1047 and Ocaperidone significantly reduced the viral load in the lungs of infected mice, with BD‐1047 providing efficacy even superior to that of ARB (Figure 7D). Given the neuropsychiatric manifestations observed in mice post‐infection, we further assessed the viral load in the brain. The results revealed that Ocaperidone exerted the most pronounced inhibitory effect on cerebral viral replication (Figure 7E). It indicates that BD‐1047 and Ocaperidone can effectively suppress HCoV‐OC43 replication in both pulmonary and cerebral tissues. Specifically, BD‐1047 exhibited superior overall antiviral potency, while Ocaperidone showed an apparent advantage in reducing brain viral load and improving symptom‐related observations; however, whether these effects are related to its neuropharmacological properties requires further investigation.

4.1. Cell Lines and Viral Strains

Vero E6 (Cat. No. CL‐0491) and HCT‐8 (Cat. No. CL‐0098) were purchased from Wuhan nuosai life science and technology Ltd. σ1R Knockout cell line (Vero E6) was purchased from Guangzhou Yuanjing Biotechnology Co. Ltd. SARS‐CoV‐2 WT (IME‐BJ01 strain, GenBank No. MT291831), Beta (CSTR: 16698.06.NPRC2.062100001) isolated in a patient from South Africa, were kept in National Center for the Preservation of Pathogenic Microorganisms., Delta (CSTR.16698.06.NPRC6.CCPM‐B‐V‐049‐2105‐6) were kept in National Center for the Preservation of Pathogenic Microorganisms (NPRC), Omicron BA.1 Variant (SARS‐CoV‐2 strain Omicron CoV/human/CHN_CVRI‐01/2022), Omicron BA.2 Variant (SARS‐CoV‐2 strain Omicron CoV/human/CHN_CVRI‐12/2022), Omicron JN.1 Variant (SARS‐CoV‐2 strain Omicron CoV/human/CHN_CVRI‐05/2024),

SARS‐CoV‐2 BMA8, PEDV(AF353511.1), IBV(CKCHSDMH21) were reserved in our lab. All SARS‐CoV‐2 related procedure were conducted in biosafety 3 level laboratories.

4.2. Reagents

Molnupiravir (Cat. No. HY‐135853), BD‐1047 (Cat. No. HY‐16996), σ1 receptor antagonist‐1 (Cat. No. HY‐10815), Ocaperidone (Cat. No. HY‐101094), Balaglitazone (Cat. No. HY‐16086) were purchased from MCE Ltd. (USA). Dimethyl sulfoxide (DMSO, Cat. No. D4540) was purchased from Sigma‐aldrich (USA), Trizol (Cat. No. 15596018), UNlQ‐10 Column Trizol Total RNA Isolation Kit (Cat. No. B511321) were purchased from Sangon Biotech Ltd. Multiplex One‐Step qRT‐PCR Probe Kit (UNG) (Cat. No. A388) was purchased from Beijing Kangrun Chengye Biotechnology Co. Ltd. DMEM with high glucose (Cat. No. C11995500CP), Penicillin‐streptomycin (10000 U/mL, Cat. No. 15140‐122) and Fetal bovine serum (Cat. No. 2358184 P) were purchased from Gibco Ltd. 0.25% Trypsin with EDTA (Cat. No. T1300) was purchased from Beijing Solarbio Technology Co. Ltd. Phosphate‐buffered saline (PBS, Cat. No. PB180327) was purchased from Wuhan Procell system Ltd. CCK‐8 Cell Proliferation‐Toxicity Testing Kit (Cat. No. CK04) was purchased from Dongren Chemical Technology (Shanghai) Co. Ltd. Paraformaldehyde solution (4%, Batch No. P0099‐500 mL) was purchased from Beyotime biotech Ltd.

4.3. Cell Viability Assay

4.4. Detection of Effective Concentration of Antiviral Drugs

Vero E6 cells were seeded at a density of 8×10³ cells per well in a 96‐well cell culture plate. 100 μL of complete medium was added into each well and incubated overnight at 37°C with 5% CO₂. When cell confluence reached 90%, medium was removed and replaced with DMEM medium with 2.5% fetal bovine serum (FBS). Subsequently, diluted compounds were added into the pretreated wells. Concurrently, diluted SARS‐CoV‐2 viral suspension with MOI of 0.01 was added. The treated cells were incubated for 72 h at a 37°C with 5% CO₂ and supernatants were collected from each well. Viral genomic RNA was extracted using a commercial kit. Real‐time quantitative PCR (qRT‐PCR) was performed to quantify the viral gene copy number of SARS‐CoV‐2 nucleocapsid (N) protein gene. The viral inhibition was calculated as follows by using the data from each control group as a baseline. The antiviral efficacy of compound was determined as Inhibition rate (%) = [1 ‐ (Drug‐treated group N gene copy number ‐ Blank control group N gene copy number)/(Positive control group N gene copy number ‐ Blank control group N gene copy number)] × 100%.

4.5. qRT‐PCR

SARS‐CoV‐2 viral RNA in cell supernatants was extracted using the UNIQ‐10 Column Trizol Total RNA Extraction Kit following instructions. Test of SARS‐CoV‐2 nucleocapsid (N) gene number was performed using an ABI 7500 qRT‐PCR instrument and the Genstar Multiplex One‐Step qRT‐PCR Probe Kit (UNG). The forward primer sequence for the N gene was 5'‐gggaacttctcctgctagaat‐3'. The reverse primer sequence is 5'‐cagacattttgctctcaagct‐3'. The probe sequence is 5'‐FAM‐ttgctgctgcttgacagattta‐MRA‐3'. The qRT‐PCR reaction program was set as follows, reverse transcription at 50°C for 10 min, followed by 2 min of pre‐denaturation at 95°C. This was followed by 45 cycles (each cycle consisting of 10 s of denaturation at 95°C and 30 s of annealing and extension at 60°C). The absolute copy number of the viral N gene was calculated based on the CT values from each well in combination with the standard curve to quantify the level of viral replication.

4.6. Histopathology

Following PBS rinsing, lung tissues were fixed in 4% paraformaldehyde for 48 h, embedded in paraffin, sectioned at 4 μm, and stained with hematoxylin and eosin (H&E). The stained sections were observed under a light microscope. Histopathological damage in the nasal turbinates and lungs was scored using a modified grading system: 0, no lesion; 1, lesions affecting < 10% of the area; 2, 10%–50%; 3, 50%–90%; and 4, > 90% of the area affected.

The drug administration group received the indicated compounds once daily by gavage. The compounds were dissolved in 10% DMSO–40% PEG300–5% Tween‐80–45% saline. The control group and the naive group received an equal volume of vehicle. On Day 4 post‐infection, three mice from each group were randomly sacrificed for tissue collection, viral RNA quantification, and histopathological analysis. The remaining mice were monitored for body weight change and survival until day 21 post‐infection.

4.7. Computational Screening of Candidate Compounds Based on σ1R

Molecular docking was used to screen in silico with 10000 small chemical molecule candidate in the DrugBank database, in order to discover active chemical molecules targeting the σ1R (5HK1). The 3D structure of σ1R was retrieved from the PDB database (https://www.rcsb.org/) and preprocessed using AutoDock Vina 1.1.2, which involved removing water molecules, adding nonpolar hydrogens, calculating Gasteiger charges, and saving the structure in PDBQT format. Candidate drug molecules in SDF format were downloaded from the DrugBank database, converted to MOL2 format using Open Babel 2.4.1, and then preprocessed into PDBQT format. Virtual screening of the candidate molecules was performed using AutoDock Vina 1.1.2 to obtain docking scores and binding modes. Lead molecules and their interaction with σ1R were analyzed using PLIP to generate interaction mode diagrams and construct a chemical molecule‐amino acid residue interaction network. Finally, a list of lead molecules with high validation rates was recommended based on the molecular docking scores, providing a theoretical basis for subsequent tests to verify the screening results.

4.8. Molecular Docking Analysis of Balaglitazone and Ocaperidone

For the refined binding mode analysis of specific candidates, molecular docking was conducted in Glide‐XP mode of Schrödinger. Balaglitazone and Ocaperidone were processed via LigPrep, while the target σ1R was treated using Protein Preparation Wizard, with subsequent generation of files containing docking site information. During docking, ligands underwent conformational search and binding energy calculation within the designated active pocket. Finally, the docking scores of the two drugs with the σ1R were acquired.

4.9. The Protection of σ1R Inhibitor Against SARS‐CoV‐2 Infection in BALB/c Mice

Female BALB/c mice, 7 weeks old and specific pathogen‐free (SPF), were obtained from Beijing Sipeifu. Except the naive group, each mouse was infected with 50 μL SARS‐CoV‐2 mouse‐adapted strain BMA8 (10⁴ TCID50/mouse) under anesthesia (intraperitoneal injection of anaesthetization). The drug administration group was gavaged with the corresponding drugs on a daily basis, the drug was diluted in 10% DMSO–40% PEG300–5% Tween‐80–45% saline, while the control group and the naive group were injected with an equal amount of saline for 7 consecutive days. Based on preliminary tolerability assessments where higher concentrations (e.g., 2 mg/kg of Ocaperidone) induced observable toxicity, optimal therapeutic doses were selected. A total of eight mice were assigned to each experimental group (n = 8). On Day 4 post‐infection, three mice (n = 3) from each group were randomly sacrificed for tissue viral load and histopathological analysis. The remaining five mice (n = 5) were continuously monitored for body weight changes and survival until 21 days post‐infection to ensure sufficient statistical power for Kaplan‐Meier survival analysis.

4.10. Statistical Analysis

Data were analyzed using GraphPad Prism 6. Survival curves were analyzed using the log‐rank test. Comparisons among multiple groups were performed using one‐way ANOVA followed by Tukey's multiple‐comparison test when appropriate, and Fisher's exact test was used where applicable. A p < 0.05 was considered statistically significant.