TNF inhibits SARS-CoV-2 induced cell-cell fusion through activating the SDC4-RhoA signaling to promote actin bundles formation

TL;DR

TNF, a key immune cytokine, prevents SARS-CoV-2 from causing cell-cell fusion by activating a signaling pathway that forms actin barriers.

Contribution

Identifies TNF as a novel antiviral factor that inhibits SARS-CoV-2-induced cell fusion via the SDC4-RhoA signaling pathway.

Findings

TNF suppresses cell-cell fusion caused by various SARS-CoV-2 Spike variants across multiple cell types.

TNF activates the TNFR1-TRADD/TRAF2/RIPK1-MAPK-SDC4 axis to inhibit membrane fusion.

SDC4/RhoA/ROCK signaling promotes actin bundle formation, creating a mechanical barrier against syncytia.

Abstract

SARS-CoV-2 infection-induced syncytia formation accelerates cell-to-cell transmission of the virus and enhances viral evasion by neutralizing antibodies. Host innate immune response plays a key role in controlling viral infection. Our present work identifies tumor necrosis factor (TNF) as a key cytokine quickly released from activated innate immune cells that suppresses SARS-CoV-2 spike-mediated cell-cell fusion. Mechanistically, TNF signals through the TNFR1-TRADD/TRAF2/RIPK1-MAPK-SDC4 axis. SDC4 further activates the RhoA/ROCK signaling pathway, which promotes cytoskeletal reorganization, leading to the formation of actin bundles at the interface between infected cell and adjacent cell. Such remodeling of actin effectively blocks further propagation of syncytia and viral spreading. These findings provide critical insights into the dynamic interplay between host antiviral factors and syncytia formation, deepening our understanding of the innate immune control of SARS-CoV-2 infection.

Image 1

•TNF effectively suppresses cell-cell fusion induced by various SARS-CoV-2 Spike variants across multiple cell types.•As a primary inflammatory cytokine, TNF is released early during innate immune cell activation to restrict membrane fusion.•TNF signals through the TNFR1 receptor and the TRADD/TRAF2/RIPK1 complex, subsequently activating the MAPK and SDC4 pathways to inhibit cell-cell fusion.•Downstream signaling of SDC4/RhoA/ROCK triggered by TNF promotes actin bundle formation, creating a mechanical barrier against cell-cell fusion.

TNF effectively suppresses cell-cell fusion induced by various SARS-CoV-2 Spike variants across multiple cell types.

As a primary inflammatory cytokine, TNF is released early during innate immune cell activation to restrict membrane fusion.

TNF signals through the TNFR1 receptor and the TRADD/TRAF2/RIPK1 complex, subsequently activating the MAPK and SDC4 pathways to inhibit cell-cell fusion.

Downstream signaling of SDC4/RhoA/ROCK triggered by TNF promotes actin bundle formation, creating a mechanical barrier against cell-cell fusion.

Figures

Full text

Introduction

1

Since its emergence in late 2019, Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), the causative agent of Coronavirus Disease 2019 (COVID-19), has spread globally, resulting in over 770 million infections and more than 7 million deaths to date. SARS-CoV-2 exhibits high transmissibility and a high mutation rate, leading to the evolution of numerous variants since the onset of this pandemic. The most consequential variants include the Alpha, Beta, Gamma, Delta, and Omicron lineages, designated by the World Health Organization (WHO) as Variants of Concern (VOC) (Carabelli et al., 2023). The continuous evolution of these variants poses significant challenges to vaccine design and therapeutic development. Studies have reported that key mutations in Beta and Omicron variants, such as E484K and N501Y, alter the structure of the spike protein's receptor-binding domain (RBD), significantly impairing the binding capacity and thus the neutralization efficiency of antibodies induced by vaccination or prior infection (Cao et al., 2022; Yuan et al., 2021). Furthermore, multiple mutations (E484K + K417N + N501Y) in the Beta variant exhibit synergistic effects, further exacerbating immune evasion (Khan et al., 2021). Such immune evasion substantially increases the risk of reinfection for recovered individuals or vaccinated individuals, and hinders the clinical application of broadly neutralizing antibodies due to reduced efficacy (Du et al., 2022).

Clinically, SARS-CoV-2 infection often manifests with fever, cough, and dyspnea (Guan et al., 2020). Moreover, autopsies of severe fatal cases revealed the presence of extensive virus-induced multinucleated syncytia within lung tissue (Braga et al., 2021). Additionally, prolonged viral RNA persistence has been detected in the lung tissue of COVID-19 patients (Stein et al., 2022). SARS-CoV-2 infection induced cell-cell fusion is initiated by the binding of the Spike (S) protein to the host ACE2 receptor. Following proteolytic cleavage at the S1/S2 site, and subsequently at the S2′ site by host proteases (such as TMPRSS2 on the cell surface or cathepsin L in endosomes), the S2 subunit undergoes dramatic conformational changes. This leads to the insertion of the fusion peptide into the host cell membrane and the assembly of a "six-helix bundle" (6-HB) structure, which pulls the membranes together to complete fusion.

Syncytia formation represents a potential viral immune evasion strategy, facilitating direct cell-to-cell transmission that bypasses host defense mechanisms such as neutralizing antibodies and the mucociliary barrier (Lin et al., 2021; Zhang et al., 2021). Syncytia provide a physical barrier shielding viral particles, preventing immune molecules like interferons (IFNs) from effectively reaching the infected cell interior. Fusion-enhancing mutations (e.g., P681R) in the Delta variant significantly enhance syncytia formation and viral resistance to type I interferons (Saito et al., 2022). Of note, syncytia often undergo rapid apoptosis at later stages, and the viral particles released upon their disintegration may exacerbate infection (Dufloo and Sanjuán, 2024). Furthermore, syncytia have been found to induce lymphocyte depletion through entosis-like mechanisms, providing a potential explanation for the lymphopenia observed clinically (Zhang et al., 2021). Therefore, cell-cell fusion has been suggested to act as viral "sanctuaries", potentially contributing to persistent infections, including the Long COVID (Li et al., 2024).

Following SARS-CoV-2 infection, the host mounts an immune response involving various host factors, including upregulation of interferons and activation of numerous interferon-stimulated genes (ISGs), forming an antiviral defense network. ISGs and downstream effectors including IFITMs (Shi et al., 2021; Xu et al., 2022), cholesterol 25-hydroxylase (CH25H) (Wang et al., 2020), and LY6E (Pfaender et al., 2020) can inhibit SARS-CoV-2 infection by targeting membrane fusion processes. However, in the early phase of SARS-CoV-2 infection, the production of IFN is low, and impaired ISG responses may lead to inadequate immune responses in patients (Blanco-Melo et al., 2020). Moreover, although inhibitors of TMEM16 proteins suppress cell-cell fusion, their clinical application is severely limited due to low selectivity and significant toxicity issues (Braga et al., 2021). Thus, identifying novel host factors possessing antiviral activity or the ability to inhibit syncytia formation is critical for developing novel immunomodulatory therapeutics.

SARS-CoV-2 induced cell-cell fusion has been observed in various cell types, including lung Calu-3 cells, intestinal Caco-2 cells (Li et al., 2023), liver Huh-7 cells (Zhang et al., 2021), U2OS-ACE2 cells (Buchrieser et al., 2020), HEK293T-ACE2 cells, A549-ACE2 cells, and fused neuronal cells (Martinez-Marmol et al., 2023). Using such cell models, our recent work revealed that pro-inflammatory cytokines such as interleukin-1β (IL-1β) and IL-1α produced from activated innate immune cells inhibit syncytia formation induced by SARS-CoV-2 infection (Zheng et al., 2025). Intriguingly, activated innate immune cells that cannot produce IL-1 still generated soluble factor(s) that inhibited cell-cell fusion. Our current work discovered that Tumor Necrosis Factor (TNF) is produced from innate immune cells early after activation, which inhibits SARS-CoV-2-mediated syncytia formation. Mechanistically, through its receptor TNFR1, TNF triggers signaling cascades involving the TRADD/TRAF2/RIPK1 complex and SDC4, leading to RhoA/ROCK activation and actin bundles formation between virus-infected cells and adjacent cells, thus preventing cell-cell fusion and viral spreading. These findings elucidate the molecular mechanism by which a critical pro-inflammatory cytokine restricts viral dissemination via inhibiting cell-cell fusion, providing novel insights into the potential strategy for antiviral defense.

Results

2

Host factors produced by innate immune cells early after activation inhibit SARS-CoV-2 spike protein-induced cell-cell fusion

2.1

Respiratory viral infection is frequently accompanied by bacterial co-infections (Westblade et al., 2021; Zhou et al., 2020). To mimic the immune responses triggered by bacteria during viral infection, we employed TLR ligands as stimulants. Meanwhile, we established both quantitative and qualitative experimental models to assess the extent of cell-cell fusion induced by SARS-CoV-2 infection (Yu et al., 2022). In the quantitative assay, HEK293T donor cells are co-transfected with Spike (S) and Cre plasmids, while HEK293T acceptor cells are transfected with ACE2 and a Stop-luciferase (Stop-Luc) plasmid containing a stop codon. Donor and acceptor cells are co-cultured for 16 h, and luciferase activity is measured to quantify cell-cell fusion efficiency. In this system, we observed that luciferase luminescence is positively correlated with ACE2 expression levels. In parallel, protein lysates from cells after 16 h fusion are analyzed, and the extent of S2′ cleavage is positively correlated with ACE2 expression, serving as an additional biochemical readout for fusion strength. Finally, for qualitative visualization, donor HEK293T cells are transfected with a ZsGreen plasmid, co-cultured with acceptor cells for 16 h, and syncytia number and size are assessed under a fluorescence microscope (Yu et al., 2022, 2023; Zheng et al., 2025).

Our recent work demonstrated that IL-1β released from activated innate immune cells inhibits SARS-CoV-2 spike-mediated cell-cell fusion (syncytium formation) (Zheng et al., 2025). However, when IL-1β secretion was blocked either by treating the human monocytic THP-1 cells with the NLRP3 (NOD-like receptor family pyrin domain containing 3) inflammasome inhibitor MCC950, or by CRISPR-Cas9-mediated knockout of NLRP3, the supernatants from Pam3CSK4 (TLR1/2 ligand) stimulated THP-1 cells still markedly suppressed cell-cell fusion, as indicated by a significant reduction in luciferase activity (Fig. 1(A) and (C), Fig. S1A and C), decreased S2′ cleavage (Fig. 1(B) and (D), Fig. S1B and D), and a pronounced reduction in syncytia area observed by fluorescence microscopy (Fig. 1(E)–(H)). These data indicate that there are factors other than IL-1 in the supernatant of activated innate immune cells inhibiting cell-cell fusion.Fig. 1Host factors secreted by innate immune cells early after activation inhibit SARS-CoV-2 spike protein induced cell-cell fusion. (A) Luciferase assay showing the effect of supernatant from THP-1 cell cultures pretreated with the NLRP3 inhibitor MCC950 (10 μM) and then stimulated with the TLR1/2 ligand Pam3CSK4 (1 μg/mL) for 3 h on spike protein-induced HEK293T cell-cell fusion. Serum-free RPMI 1640 served as the medium control. Data points represent mean ± SEM from six independent experiments; P values are indicated. (B) Immunoblot analysis of S2′ cleavage in fused cells treated with supernatant from MCC950-pretreated THP-1 cultures stimulated with Pam3CSK4 for 3 h. Blots are representative of three independent experiments. (C) Luciferase assay showing the effect of supernatant from NLRP3-knockout THP-1 cell cultures stimulated with Pam3CSK4 for 3 h on spike-induced HEK293T cell-cell fusion. Serum-free RPMI 1640 served as the control. Data points represent mean ± SEM from six independent experiments; P values are indicated. (D) Immunoblot analysis of S2′ cleavage in fused cells treated with supernatant from NLRP3-knockout THP-1 cell cultures stimulated with Pam3CSK4 for 3 h. Blots are representative of three independent experiments. (E) Visualization of syncytium formation by ZsGreen fluorescence after treatment with supernatant from MCC950-pretreated THP-1 cultures stimulated with Pam3CSK4 for 3 h. Images are representative of three independent experiments; white arrows indicate syncytia. Scale bar, 50 μm. (F) Visualization of syncytium formation by ZsGreen fluorescence after treatment with supernatant from NLRP3-knockout THP-1 cell cultures stimulated with Pam3CSK4 for 3 h. Images are representative of three independent experiments; white arrows indicate syncytia. Scale bar, 50 μm. (G) Quantification of relative syncytium area (%) based on ZsGreen fluorescence images in Fig. 1(E). Data are presented as mean ± SEM from three independent experiments. P values are indicated above the bars. (H) Quantification of relative syncytium area (%) based on ZsGreen fluorescence images in Fig. 1(F). Data are presented as mean ± SEM from three independent experiments. P values are indicated.Fig. 1

TNF inhibits SARS-CoV-2 spike protein induced cell-cell fusion

2.2

To identify the specific host factors released from activated innate immune cells that suppress syncytium formation, we performed real-time PCR on THP-1 cells stimulated with TLR ligands for 4 h. In addition to IL-1, we detected significant upregulation of IL-6, IL-8, and TNF mRNAs (Fig. S2A and B) (Zheng et al., 2025). We then treated fusion assay cells with each of these cytokines, and luciferase assay revealed that besides IL-1, only TNF suppressed cell-cell fusion (Fig. 2(A)).Fig. 2TNF produced by innate immune cells early after activation inhibit SARS-CoV-2 spike induced cell-cell fusion. (A) Luciferase assay showing the effect of recombinant IL-6 (10 ng/mL), IL-8 (10 ng/mL), or TNF (10 ng/mL) on spike-induced cell-cell fusion. PBS was used as the vehicle control. Data points represent mean ± SEM from four independent experiments; P values are indicated. (B) Luciferase assay showing the effect of supernatant from THP-1 cultures stimulated with Pam3CSK4 for the indicated durations on cell-cell fusion in HEK293T cells pretreated with the IL-1 receptor antagonist (IL-1RA) (4 μg/mL). Data represent mean ± SEM from four independent experiments; P values are shown. (C) Immunoblot analysis of S2′ cleavage in IL-1RA-pretreated fused cells exposed to supernatant from THP-1 cultures stimulated with Pam3CSK4 for the indicated durations. Blots are representative of three independent experiments. (D) ELISA quantification of IL-1β and TNF levels in supernatant from THP-1 cell cultures after stimulation with the indicated TLR ligand at the specified time points. Data represent mean ± SEM from three independent experiments. (E) Luciferase assay showing the effect of supernatant from THP-1 cultures stimulated with Pam3CSK4 for the indicated durations on spike-induced fusion in HEK293T-sgcontrol and HEK293T-sgTNFR1 cells. Data points represent mean ± SEM from six independent experiments; P values are indicated. (F) Immunoblot analysis of S2′ cleavage in HEK293T-sgcontrol and HEK293T-sgTNFR1 fused cells treated with supernatant from THP-1 cell cultures stimulated with Pam3CSK4 for the indicated durations. Blots are representative of three independent experiments. (G) Luciferase assay showing the effect of supernatant from Pam3CSK4-stimulated THP-1 cell cultures on cell-cell fusion in IL-1RA-pretreated HEK293T-sgcontrol and HEK293T-sgTNFR1 cells. Data represent mean ± SEM from four independent experiments; P values are indicated. (H) Immunoblot analysis of S2′ cleavage in IL-1RA-pretreated HEK293T-sgcontrol and HEK293T-sgTNFR1 fused cells exposed to supernatant from Pam3CSK4-stimulated THP-1 cell cultures. Blots are representative of three independent experiments. (I–J) ELISA quantification of IL-1β and TNF levels in supernatant from THP-1 cell cultures under (I) MCC950 pharmacological inhibition of NLRP3, (J) NLRP3 knockout, after stimulation with the indicated TLR ligand at the specified time points. Data represent mean ± SEM from three independent experiments.Fig. 2

Upon TLR ligands stimulation, TNF can be produced by innate immune cells rather quickly (Marino et al., 1997; Skurski et al., 2021). To validate that TNF was indeed generated quickly after innate immune cell activation in the above experiments, we added the IL-1 receptor antagonist (IL-1RA) in our HEK293T cell-cell fusion assay, followed by adding supernatants from THP-1 cells stimulated with TLR ligands for various durations. Here, it was found that supernatants from short-term (1 h) stimulation still significantly inhibited cell-cell fusion regardless of IL-1RA treatment, whereas supernatants from long-term (24 h) stimulation showed markedly reduced inhibitory effects in the IL-1RA-treated group, as evidenced by decreased luciferase activity and S2′ cleavage (Fig. 2(B) and (C)). These data indicate that short-term (1 h) TLR stimulation of THP-1 cells primarily released TNF without IL-1, whereas long-term stimulation produced both IL-1 and TNF. Consistently, ELISA measurements of IL-1β and TNF in culture supernatants from THP-1 cells after different durations of TLR stimulation showed that 1 h supernatants contained TNF but no detectable IL-1β (Fig. 2(D)).

To validate the above findings, we generated TNFR1-knockout HEK293T cells via CRISPR-Cas9 method and repeated the fusion assay with supernatants from THP-1 cells stimulated with TLR ligands for different durations. Notably, supernatants from THP-1 cells stimulated for only 1 h strongly inhibited fusion in control cells but not in the TNFR1-knockout cells. In contrast, supernatants from THP-1 cells stimulated for 3 or 24 h still significantly inhibited fusion in both cell types (Fig. 2(E) and (F)). These findings confirmed that short-term (1 h) TLR stimulation of THP-1 cells released TNF but not IL-1.

Finally, to determine whether IL-1 and TNF were the only inhibitory host factors in TLR-stimulated innate immune cell supernatants, we performed fusion assays in both control and TNFR1-knockout HEK293T cells with IL-1RA treatment. In control cells, IL-1RA blocked IL-1 activity, yet the supernatants still suppressed cell-cell fusion. In contrast, in TNFR1-knockout cells treated with IL-1RA, the supernatants completely lost their inhibitory effect on cell-cell fusion (Fig. 2(G) and (H)). Consistently, ELISA measurements of IL-1β and TNF in culture supernatants from MCC950-treated, or NLRP3-knockout THP-1 cells after different durations of TLR stimulation showed that TNF but not IL-1β is released from such cells (Fig. 2(I) and (J)). These results indicate that IL-1 and TNF are the sole inhibitory factors in the supernatants that block SARS-CoV-2 Spike mediated cell-cell fusion.

TNF inhibits multiple SARS-CoV-2 variants induced cell-cell fusion in various cells

2.3

To clarify the mode of action by which TNF inhibits cell-cell fusion, we examined its dose dependency through a concentration gradient. When the TNF concentration reached 10 ng/mL, the inhibition of SARS-CoV-2 spike–induced HEK293T cell-cell fusion peaked, suggesting a saturation effect. At this concentration, the luciferase signal indicating cell-cell fusion dropped to its lowest level (Fig. 3(A)); similarly, S2′ cleavage was minimized (Fig. 3(B)), and the area of fused cells were markedly reduced (Fig. 3(C) and (D)).Fig. 3TNF inhibits various SARS-CoV-2 variants spike protein and authentic virus induced cell-cell fusion. (A) Luciferase assay showing the effects of different doses of recombinant TNF on luciferase activity resulting from spike protein induced HEK293T cell-cell fusion. PBS treatment served as the control. Data points represent four independent experiments, shown as mean ± SEM, with P values indicated. (B) Immunoblot analysis showing the effects of different TNF doses on S2′ cleavage. Blots are representative of three independent experiments. (C) ZsGreen fluorescence assay directly visualizing the effects of TNF on spike protein induced cell-cell fusion. Images are representative of three independent experiments; white arrows indicate syncytium formation. Scale bar, 50 μm. (D) Quantification of relative syncytium area (%) based on ZsGreen fluorescence images in Fig. 3(C). Data are presented as mean ± SEM from three independent experiments. P values are indicated. (E) Luciferase assay showing the effects of TNF on luciferase activity resulting from cell-cell fusion induced by different spike variants in HEK293T cells. PBS treatment served as the control. Data points represent four independent experiments, shown as mean ± SEM, with P values indicated. (F) Immunoblot analysis showing the effects of TNF on S2′ cleavage in fused cells induced by different spike protein variants. Blots are representative of three independent experiments. (G) ZsGreen fluorescence assay directly visualizing the effects of TNF on cell-cell fusion induced by different spike variants. Images are representative of three independent experiments; white arrows indicate syncytium formation. Scale bar, 50 μm. (H) Quantification of relative syncytium area (%) based on ZsGreen fluorescence images in Fig. 3(G). Data are presented as mean ± SEM from three independent experiments. P values are indicated. (I) Immunoblot analysis showing the effects of TNF pretreatment of donor or acceptor HEK293T cells on S2′ cleavage. Blots are representative of three independent experiments. (J) Immunoblot analysis showing the effects of TNF on S2′ cleavage in Calu-3 cells. Blots are representative of three independent experiments. (K) Immunoblot analysis showing the effects of TNF on S2′ cleavage in Caco-2 cells. Blots are representative of three independent experiments. (L) Immunoblot analysis showing the effects of TNF treatment on S2′ cleavage and N protein levels in Caco-2 cells harvested at 48 hpi with SARS-CoV-2 B.1.617.2 Variant. Blots are representative of three independent experiments.Fig. 3

Given that the SARS-CoV-2 spike protein continuously mutates during viral transmission—such as in Alpha, Beta, Delta, and Omicron variants—and that spike is critical for cell-cell fusion, we tested whether TNF's inhibitory effect is broad-spectrum. Of note, TNF treatment significantly reduced luciferase signals (Fig. 3(E)), decreased S2′ cleavage (Fig. 3(F)), and lowered the area of fused cells (Fig. 3(G) and (H)) across all tested spike variants, indicating its broad-spectrum inhibitory effect.

In our cell-cell fusion assay, TNF was applied to co-cultures of both the donor HEK293T cells expressing spike and the acceptor HEK293T cells expressing ACE2. To further determine TNF's cellular targets, we pretreated donor cells alone, acceptor cells alone, or both cell types with TNF (10 ng/mL, 6 h), washed with PBS, and then co-cultured them for 16 h. Pretreatment of either donor or acceptor cells partially inhibited cell-cell fusion, whereas treating both cell types yielded maximal inhibition, matching the effect at TNF's saturation concentration. Under these conditions, S2′ cleavage was reduced to the largest extent (Fig. 3(I)), indicating that TNF acts on both donor and acceptor cells to suppress cell-cell fusion.

Since SARS-CoV-2 can infect multiple cell types-including those from the lung and gut-and induce cell-cell fusion, we further evaluated TNF's inhibitory effect using Calu-3 cells (human lung adenocarcinoma cells) and Caco-2 cells (human colorectal adenocarcinoma cells). Immunoblot analysis revealed that TNF treatment markedly reduced S2′ cleavage in both cell types compared with controls (Fig. 3(J) and (K)), demonstrating that TNF significantly suppresses cell-cell fusion in various cell types. Moreover, in order to validate the effect of TNF on cell-cell fusion during authentic SARS-CoV-2 infection, we pre-treated Caco-2 cells with TNF before inoculating Delta authentic SARS-CoV-2. Cell lysates were used for the detection of SARS-CoV-2 spike and N protein 24 hpi (Fig. S3D) and 48 hpi (Fig. 3(L)), wherein TNF clearly inhibited the S2′ cleavage. These results collectively demonstrate that TNF effectively suppresses authentic SARS-CoV-2 infection-induced cell-cell fusion and restricts viral spreading.

Of note, previous work showed that during SARS-CoV-2 infection, excessive release of pro-inflammatory cytokines can cause acute lung injury. Specifically, TNF plus IFN-γ may trigger inflammatory cell death and tissue damage (Karki et al., 2021). However, our above data showed that TNF alone effectively blocks cell-cell fusion induced by the SARS-CoV-2 spike protein. To investigate whether TNF and/or IFN-γ affect cell death in our fusion system, we measured LDH release—a marker of cell damage in the culture medium. Results showed that neither TNF/IFN-γ alone, nor their combination caused significant LDH release compared to positive controls in the 16 h cell-cell fusion system (Fig. S3A). We also examined potential changes in PARP cleavage indicating cell death, and found that only the Raptinal positive control group showed clear PARP cleavage (Fig. S3B). However, when we treated the syncytia with TNF plus IFN-γ for 48 h, we observed obvious PARP cleavage (Fig. S3C). Together, these findings demonstrate that in our cell-cell fusion system, TNF does not induce cell death, thus its inhibitory effect is not connected with cell death.

TNF inhibits SARS-CoV-2 spike-induced cell-cell fusion through the TNFR1-TRADD/TRAF2/RIPK1-MAPK-SDC4 signaling cascade

2.4

To elucidate the molecular mechanism through which TNF suppresses cell-cell fusion, we employed both gene knockout and pharmacological inhibition strategies in HEK293T cells (Fig. 4(A)). TNF exerts its downstream effects through two receptors, TNFR1 and TNFR2. TNFR1 is constitutively expressed in nearly all nucleated cells, whereas TNFR2 expression is largely restricted to specific immune cell subsets (McCulloch et al., 2024). Given this distribution, we first generated TNFR1-deficient HEK293T cells via CRISPR/Cas9-mediated knockout and subsequently treated such cells with TNF. Both luciferase reporter assays (Fig. 4(B)) and immunoblot analysis (Fig. 4(C)) revealed that there are no significant changes in luciferase activity or S2′ cleavage in TNFR1-knockout cells upon TNF treatment, indicating that TNF-mediated suppression of cell-cell fusion requires TNFR1.Fig. 4TNF inhibits SARS-CoV-2 spike protein induced cell-cell fusion via the TNFR1-TRADD/TRAF2/RIPK1-MAPK-SDC4 signaling pathway. (A) Schematics of gene knockout or inhibitor treatment in the TNF receptor pathway. (B) Quantification of luciferase activity in TNFR1-knockout HEK293T cells following TNF treatment in the spike induced cell-cell fusion assay. Data points represent mean ± SEM from six independent experiments, with corresponding P values indicated. (C) Immunoblot analysis showing the effect of TNF on S2′ cleavage in TNFR1-deficient HEK293T cells. Blots are representative of three independent experiments. (D) Quantification of luciferase activity in TRADD/TRAF2/RIPK1 triple-knockout HEK293T cells following TNF treatment in the spike protein induced cell-cell fusion assay. Data points represent mean ± SEM from six independent experiments, with P values indicated. (E) Immunoblot analysis showing the effect of TNF on S2′ cleavage in TRADD/TRAF2/RIPK1 triple-knockout HEK293T cells. Blots are representative of three independent experiments. (F) Quantification of luciferase activity in HEK293T cells treated with MAPK inhibitor cocktail: SP600125 (JNK inhibitor), SB203580 (P38 inhibitor), PD98059 (ERK inhibitor) prior to TNF exposure in the spike protein induced cell-cell fusion assay. Data represent mean ± SEM from six independent experiments, with P values indicated. (G) Immunoblot analysis showing the effect of TNF on S2′ cleavage followed by MAPK inhibitor cocktail (SP600125, SB203580, PD98059) in HEK293T cells. Blots are representative of three independent experiments. (H) Heatmap depicting changes in mRNA expression in control HEK293T cells or TRADD/TRAF2/RIPK1 triple-knockout cells treated with TNF. Data are derived from three independent biological replicates, and color intensity represents normalized mean expression levels. (I) Quantification of luciferase activity in SDC4-knockout HEK293T cells following TNF treatment in the spike induced cell-cell fusion assay. Data points represent mean ± SEM from six independent experiments, with P values indicated. (J) Immunoblot analysis showing the effect of TNF on S2′ cleavage in SDC4 deficient HEK293T cells. Blots are representative of three independent experiments. (K) Quantification of luciferase activity in SDC4-overexpressing HEK293T cells following TNF treatment in the spike induced cell-cell fusion assay. Data points represent mean ± SEM from six independent experiments, with P values indicated. (L) Immunoblot analysis showing the effect of TNF on S2′ cleavage in SDC4-overexpressing HEK293T cells. Blots are representative of three independent experiments. (M) Immunoblot analysis showing the effect of TNF on S2′ cleavage in SDC4-overexpressing Calu-3 cells. Blots are representative of three independent experiments.Fig. 4

Downstream of TNFR1, the canonical TRADD/TRAF2/RIPK1 complex functions as a pivotal signaling hub. We found that individual knockout of TRADD, TRAF2, or RIPK1 alone did not abolish the inhibitory effect of TNF on cell-cell fusion (Fig. S4A). However, simultaneous deletion of all three components abolished TNF-induced suppression, as evidenced by unchanged luciferase signals (Fig. 4(D)) and S2′ cleavage (Fig. 4(E)). These results indicate that TNF-mediated inhibition of cell-cell fusion is dependent on the intact TRADD/TRAF2/RIPK1 complex.

The TRADD/TRAF2/RIPK1 complex can activate multiple downstream pathways, including the MAPK family (p38, JNK, ERK) and the NF-κB pathway (Chung et al., 2002; Devin et al., 2003). To delineate the relevant downstream effectors, we systematically blocked these pathways using pharmacological inhibitors or via gene knockout. Here, it was found that gene knockout of ERK1/2 and MEK1/2 complex (MAPK3, MAPK1, MAP2K1, MAP2K2) had no effect on basal cell-cell fusion, and did not prevent TNF from suppressing fusion (Fig. S4B), indicating that ERK is dispensable.

Similarly, inhibition of p38 (SB203580; Fig. S4C), JNK (SP600125; Fig. S4D), NF-κB (Bay 11-7082; Fig. S4E) alone, as well as knockout of NFKB1, NFKB2 and NFKBIA (Fig. S4F), did not alter TNF-mediated suppression of cell-cell fusion either. Interestingly, however, when we treated cells with a MAPK inhibitor cocktail targeting ERK (PD98059), p38 (SB203580) and JNK (SP600125), the cell-cell fusion was markedly enhanced. Meanwhile, it was found that when co-treated with the inhibitor cocktail and TNF, TNF failed to suppress cell-cell fusion, accompanied by recovery of luciferase activity and S2′ cleavage (Fig. 4(F) and (G)). These findings collectively suggest that TNF's inhibitory effect is dependent on MAPK but not NF-κB pathway.

To identify downstream effectors responsible for TNF's inhibitory effect, we performed RNA sequencing on control HEK293T cells and TRADD/TRAF2/RIPK1 triple-knockout cells treated with PBS or TNF (10 ng/mL) for 4 h. Genes significantly up- or down-regulated by TNF in control cells, but unaffected in the triple-knockout cells were considered potential effectors. Heatmap analysis (Fig. 4(H)) revealed that, except for GOPC, which was markedly downregulated, most candidate genes were upregulated following TNF treatment. Integrating our RNA-seq data with published literature, we identified Syndecan-4 (SDC4) as a strong candidate. SDC4 is a cell-surface heparan sulfate proteoglycan implicated in diverse biological processes, including cell adhesion, signal transduction, and growth regulation (Keller-Pinter et al., 2021; Woods and Couchman, 1994). Notably, SDC4 interacts with integrins to promote cytoskeletal reorganization and actin polymerization, which are processes closely linked to membrane fusion events. We therefore hypothesized that TNF may suppress cell-cell fusion via SDC4-dependent signaling.

To test this, we generated SDC4-knockout HEK293T cells. Functional studies showed that loss of SDC4 abolished TNF-mediated inhibition of cell-cell fusion, whereas control cells retained this response, as revealed from both the luciferase assay (Fig. 4(I)) and immunoblotting (Fig. 4(J)). Furthermore, SDC4 overexpression in HEK293T and Calu-3 cells directly suppressed cell-cell fusion and reduced S2′ cleavage, as shown by luciferase assays and immunoblotting in HEK293T cells (Fig. 4(K) and (L)) and immunoblotting in Calu-3 cells (Fig. 4(M)). These results establish SDC4 as a critical downstream effector mediating TNF's suppression of cell-cell fusion. Furthermore, Western blot analysis confirmed a significant increase in SDC4 protein expression in HEK293T cells following TNF treatment (Fig. S4G). As supplement, we confirmed the efficiency for the knockout of TNFR1, TRADD, TRAF2, RIPK1, and SDC4 used in the above experiments (Fig. S4H).

TNF activates the RhoA-ROCK pathway, inducing F-actin to form actin bundles, thereby inhibiting SARS-CoV-2 spike-induced cell-cell fusion

2.5

Previous studies have shown that TNF can activate the RhoA/ROCK signaling pathway (Mckenzie and Ridley, 2007). As a principal downstream effector of RhoA, ROCK regulates substrates associated with the actin cytoskeleton, cell adhesion, and cell motility through phosphorylation (Riento and Ridley, 2003). To examine potential changes in RhoA activity in our system, we employed a biosensor derived from the C-terminal region of anillin (GFP-AHPH) to directly visualize the distribution of endogenous GTP-bound RhoA (active RhoA) (Priya et al., 2015). Interestingly, TNF markedly enhanced the fluorescence intensity of GFP-AHPH in control HEK293T cells, whereas no such activation was observed in TRADD/TRAF2/RIPK1 triple-knockout or SDC4-knockout cells (Fig. 5(A) and (B)).Fig. 5TNF inhibits SARS-CoV-2 spike induced cell-cell fusion by inducing actin bundles formation at cell-cell junctions via the RhoA/ROCK signaling pathway. (A) Representative confocal images of GFP-AHPH in sgcontrol, sgTRADD/TRAF2/RIPK1 complex, and sgSDC4 HEK293T cells treated with 10 ng/mL TNF for 30 min. Images are representative of five independent experiments. Scale bars represent 10 μm. (B) Quantification of GFP-AHPH fluorescence intensity. Data points represent the mean ± SEM from five independent experiments, with P values shown in figure. (C) Quantification of luciferase luminescence showing the effect of ROCK inhibitor Y27632 and TNF treatment on spike induced cell-cell fusion in HEK293T cells. Data points represent the mean ± SEM from four independent experiments, with P values shown in figure. (D) Immunoblot analysis showing the effect of ROCK inhibitor Y27632 and TNF treatment on S2′ cleavage in spike induced HEK293T syncytia. Blots are representative of three independent experiments. (E) Representative confocal images of GFP-AHPH in syncytia formed by HEK293T-S-HA and HEK293T-ACE2-V5 cells treated with PBS or TNF for 16 h. Green indicates AHPH, red indicates Spike-expressing cells, and white arrows indicate regions of AHPH enrichment or disappearance. Images are representative of three independent experiments. Scale bars represent 10 μm. (F) White lines indicate SARS-CoV-2 spike protein induced cell-cell transmission and quantify with fluorescence intensity of GFP-AHPH in Fig. 5(E). (G) Representative confocal images of GFP-AHPH in syncytia formed by HEK293T-S-HA and HEK293T-ACE2-V5 cells transfected with Vector or SDC4 for 16 h. Green indicates AHPH, red indicates S-expressing cells, and white arrows indicate regions of AHPH enrichment or disappearance. Images are representative of three independent experiments. Scale bars represent 10 μm. (H) White lines indicate SARS-CoV-2 spike protein induced cell-cell transmission and quantify with fluorescence intensity of GFP-AHPH in Fig. 5(G). (I) Representative confocal images of F-actin stained with phalloidin-488 in syncytia formed by HEK293T-S-HA and HEK293T-ACE2-V5 cells treated or untreated with 10 ng/mL TNF for 16 h. Green indicates F-actin, red indicates Spike-expressing cells, magenta indicates ACE2-expressing cells, and white arrows indicate regions of F-actin enrichment or disappearance. Images are representative of three independent experiments. Scale bars represent 10 μm. (J) White lines indicate SARS-CoV-2 spike protein induced cell-cell transmission and quantify with fluorescence intensity of F-actin in Fig. 5(I). (K) Representative confocal images of F-actin stained with phalloidin-488 in syncytia formed by HEK293T-S-HA and HEK293T-ACE2-V5 cells transfected with Vector or SDC4 overexpression for 16 h. Green indicates F-actin, red indicates Spike-expressing cells, and white arrows indicate regions of F-actin enrichment or disappearance. Images are representative of three independent experiments. Scale bars represent 10 μm. (L) White lines indicate SARS-CoV-2 spike protein induced cell-cell transmission and quantify with fluorescence intensity of F-actin in Fig. 5(K). (M) Representative confocal images of F-actin stained with phalloidin-488 in syncytia formed by HEK293T-S-HA and HEK293T-ACE2-V5 cells treated with 10 ng/mL TNF or 40 μM Y27632 for 16 h. Green indicates F-actin, red indicates S-expressing cells, magenta indicates ACE2-expressing cells, and white arrows indicate regions of F-actin enrichment or disappearance. Images are representative of three independent experiments. Scale bars represent 10 μm. (N) White lines indicate SARS-CoV-2 spike protein induced cell-cell transmission and quantify with fluorescence intensity of F-actin in Fig. 5(M).Fig. 5

To further investigate whether TNF-mediated suppression of cell-cell fusion depends on the RhoA/ROCK pathway, we treated fusion-competent cells with the ROCK inhibitor Y-27632. Notably, Y-27632 enhanced luciferase activity and S2′ cleavage in a dose-dependent manner and promoted syncytium formation. At lower concentrations of Y-27632, TNF abolished the fusion-promoting effect of the inhibitor; however, at higher concentrations of Y-27632, TNF failed to suppress cell-cell fusion (Fig. 5(C) and (D)).

To investigate whether TNF or SDC4 inhibits SARS-CoV-2 spike-induced cell-cell fusion through the RhoA/ROCK pathway, we co-transfected GFP-AHPH in ACE2-expressing cells, then co-cultured them with Spike-expressing cells. In the process of syncytia formation, cell-cell contact is established between S-expressing cells and ACE2-expressing cells, GFP-AHPH initially localized distally from cell-cell junctions and was subsequently visualized at the periphery of the syncytium. However, in both the TNF-treated group (Fig. 5(E) and (F), and S5A) and the SDC4-overexpression group (Fig. 5(G) and (H), and S5B), GFP-AHPH foci were markedly enriched at the cell-cell junctions between S-expressing cells and ACE2-expressing cells, thereby preventing further cell-cell fusion.

Given that RhoA is known to initiate the formation of actin bundles (Zheng et al., 2025), we next assessed whether TNF modulates actin cytoskeleton organization, thus impacting SARS-CoV-2 spike induced cell-cell fusion. To this end, HEK293T fusion cells were treated with PBS or TNF for 16 h, followed by staining with Actin-Tracker-488 for actin, anti-HA for spike protein, and anti-V5 for the ACE2 receptor. Immunofluorescence analysis showed that, in PBS-treated controls, actin was distributed around the periphery of fused cells, which displayed large syncytia and big fusion areas. In contrast, TNF-treated cells exhibited fewer fusion events, with actin enriched at the interface between spike protein expressing cells and adjacent uninfected cells, thereby preventing further fusion (Fig. 5(I) and (J), and S5C). Next, we overexpressed SDC4 in HEK293T cells and observed that actin in such cells was concentrated at the boundary between infected and uninfected cells; whereas in the control group, actin was redistributed around the periphery of fused cells (Fig. 5(K) and (L), and S5D). Using the same approach, we treated HEK293T cells with TNF or a combination of TNF and Y-27632. Consistent with previous observations, actin in the TNF-treated group was concentrated at the boundary between infected and uninfected cells, whereas in the TNF plus ROCK inhibitor Y-27632 group, actin was redistributed around the periphery of fused cells (Fig. 5(M) and (N), and S5E). Collectively, these findings indicate that TNF prevents SARS-CoV-2 from spreading via cell-cell fusion by activating RhoA/ROCK signaling, which promotes actin bundle accumulation at cell-cell junctions and blocks infected cells from fusing with neighboring uninfected cells.

Discussion

3

The spike (S) protein of SARS-CoV-2 plays a pivotal role in viral entry by binding to the host angiotensin-converting enzyme 2 (ACE2) receptor, facilitating both cellular invasion and adjacent cell fusion. As a type II transmembrane glycoprotein, the S protein consists of S1 and S2 subunits: the S1 subunit mediates ACE2 recognition, while the S2 subunit drives membrane fusion. Upon ACE2 engagement, the S protein undergoes sequential proteolytic cleavage—first at the S1/S2 site by furin protease during viral assembly (maintaining S1/S2 in a non-covalent state) (Xing et al., 2025), and subsequently at the S2′ site by transmembrane serine protease 2 (TMPRSS2) or endosomal cathepsin B/L (Jackson et al., 2022). These cleavages trigger conformational changes in the S2 subunit, exposing the fusion peptide (FP) to mediate viral-host membrane fusion. Critically, S2′ cleavage is indispensable for syncytia formation and viral infectivity, positioning this process as a prime therapeutic target (Jackson et al., 2022; Xing et al., 2025).

TNF, a central mediator of innate immunity, exerts dual antiviral effects. Pre-infection administration of recombinant TNF significantly reduces viral load through non-hematopoietic cell-mediated mechanisms (Baker et al., 2024), while its activation of the NF-κB pathway enhances interferon-I (IFN-I) production. TNF amplifies immune responses by augmenting macrophage phagocytosis, bactericidal activity, and antigen presentation, while synergizing with cytokines (e.g., IL-1, IL-6, IFN-γ) to orchestrate adaptive immunity (Bancroft et al., 1989; Schroder et al., 2004). It further enhances neutrophil antiviral functions, including viral replication suppression and pyroptosis induction (Bordon et al., 2013). In general, limited local generation of TNF is beneficial for homeostasis, while excessive production of TNF can cause tissue damage. Of note, in COVID management, anti-TNF agents elevated SARS-CoV-2 infection risk by downregulating ACE2 via Notch-1/IL-6 signaling (Keewan et al., 2021), indicating the antiviral potential of TNF. Our study demonstrates that TNF potently inhibits spike protein-mediated cell-cell fusion across SARS-CoV-2 variants through TNFR1-dependent activation of the TRADD/TRAF2/RIPK1-MAPK-SDC4-RhoA/ROCK axis. This cascade drives actin bundle accumulation at intercellular junctions, thus physically blocking viral spread and syncytia formation. Therefore, TNFR1 ablation may lead to excessive cell-cell fusion and viral spreading. As shown in Fig. 5(A), TNF treatment causes activated RhoA (GFP-AHPH) to relocate from the peripheral edge of the cell to the cytoplasm, where it distributes uniformly. This may be linked to the formation of stress fibers induced by TNF, involving regulation of the RhoA/ROCK signaling pathway (Hanna et al., 2001). During cell-cell fusion, the cells likely experience a transitional state from resting that involves cytoskeletal reorganization. After this transition, actin redistributes from the periphery to cell-cell junctions, forming actin bundles that prevent infected cells from further fusing with normal cells.

SARS-CoV and MERS-CoV belong to the same coronavirus family as SARS-CoV-2, and their Spike(S) protein-mediated cell-cell fusion mechanisms are similar in many respects. First, regarding the fusion mechanism, their S proteins also require cleavage by host proteases for activation to mediate cell-cell fusion and form syncytia. Furthermore, we have previously demonstrated that IL-1β can inhibit cell-cell fusion induced by the S proteins of both SARS-CoV and MERS-CoV (Zheng et al., 2025). Therefore, it is reasonable to hypothesize that TNF may also inhibit cell-cell fusion induced by the S proteins of these two coronaviruses.

Of note, in our cell-cell fusion system, neither TNF alone nor TNF combined with IFN-γ stimulation induced cell death. This suggests that the functional outcomes of cytokines may be regulated by the microenvironment at different stages of infection. In the early phase, spatially and temporally controlled TNF activity may suppress viral spreading through inhibiting cell-cell fusion. In the late phase, however, persistently elevated TNF levels may amplify inflammatory signaling cascades and increase the risk of cell death. Therefore, temporally modulating TNF activity according to the course of infection may provide a novel therapeutic strategy to balance antiviral efficacy and the risk of immunopathology.

Current therapies for COVID-19 predominantly target viral components (e.g., small-molecule inhibitors, vaccines), yet viral evolution necessitates host-directed strategies. TNF exemplifies such an approach: its rapid, antigen-independent action inhibits early viral transmission via RhoA/ROCK-mediated cell-cell fusion blockade, while coordinating phagocyte activation and cytokine amplification. Future interventions may combine TNF pathway modulation with agents to eliminate infected syncytia, offering a robust framework against emerging variants.

In conclusion, this work elucidates TNF's role in suppressing SARS-CoV-2 induced syncytia formation, highlighting innate immune factors as critical barriers to coronavirus dissemination and promising therapeutic targets.

Materials and methods

4

Reagents and plasmids

4.1

The antibodies used for immunoblotting include: rabbit anti-SARS-CoV-2 S2 (Sino Biological, 40590- T62, 1:2 000), rabbit anti-ACE2 (Proteintech, 21115-1-AP, 1:2 000), mouse anti-Syndecan-4 (5G9, sc-12766), HRP-conjugated anti-β-tubulin (Abclonal, AC030, 1:5 000), anti-rabbit/anti-mouse IgG (Jackson Immuno Research, 111-035-003, 1:5 000). The antibodies and regents used for immunofluorescence include: rabbit anti-SARS-CoV-2 S2 (Sino Biological, 40590-T62, 1:200), mouse anti-Syndecan-4 (5G9, sc-12766), rabbit anti-ACE2 (Proteintech, 21115-1-AP, 1:200), mouse anti-HA-Tag (Abclonal, AE008, 1:200). Actin-Tracker Green-488 (Beyotime, C2201S, 1:100), goat anti-mouse IgG-555 (Invitrogen, A-21424, 1: 400) and goat anti-rabbit IgG-647 (Invitrogen, A-21236, 1: 400), DAPI (Abcam, ab228549, 1:2 000) and antifade mounting medium (vector labs, H-1400-10). Pam3CSK4 (Invivogen, tlrl-pms), TPCA1 (Selleck, S2824), Y-27632 (Selleck, S6390). Inhibitors were dissolved in dimethyl sulfoxide (DMSO, Sigma-Aldrich, D2650), and DMSO was added as solvent control. Recombinant human TNF (300-01A), human IL-1β (200-01B), human IL-1RA (200-01RA), human IL-6 (200-06) and human IL-8 (200-08M) were purchased from Peprotech. TNF and IL-1β concentrations in supernatants from THP-1 were determined using ELISA kit, according to the manufacturer's instructions (R&D Systems, DY201). SARS-CoV-2 spike (Wild type, GenBank: QHD43419.1) was homo sapiens codon-optimized and generated de novo into pVAX1 vector by recursive polymerase chain reaction (PCR). WT, Alpha, Beta and Delta variants containing point and deletion mutations were generated via stepwise mutagenesis using spike construct containing the truncated 19 amino acids at the C-terminal (CTΔ19). The latest human codon optimized Omicron was purchased from Genescripts, and subcloned into the pVAX1 backbone with CTΔ19 for comparison. Human ACE2 assembled in a pcDNA4.0 vector was used for transient expression of ACE2. GFP-AHPH (Addgene plasmid #71368; http://n2t.net/addgene:71368; RRID: Addgene_71368) was from Addgene (Zheng et al., 2025).

Cell culture

4.2

HEK293T cells (4201HUM-CCTCC00187) were purchased from the National Science & Technology Infrastructure (NSTI) Cell Bank (https://www.cellbank.org.cn/). Human colon epithelial Caco-2 cells (Catalog No. SCSP-5027) were from the Chinese Academy of Sciences Cell Bank. Human lung cancer Calu-3 cells and Vero E6-ACE2 cells were kindly provided by Prof. Dimitri Lavillette's laboratory at the Institut Pasteur Korea (Applied Molecular Virology Unit). The human monocytic THP-1 cell line (TIB-202; ATCC) was authenticated via short tandem repeat (STR) analysis by Suzhou Genetic Testing Biotech Co., Ltd, following the ANSI/ATCC ASN-0002-2012 standard (Chen et al., 2023; Gao et al., 2023). All cell lines were identity-validated by suppliers and routinely tested for mycoplasma contamination. HEK293T and Vero E6-ACE2 cells were cultured in Gibco DMEM medium (GE Healthcare) supplemented with 10% fetal bovine serum (FBS, Sigma) and 1% penicillin/streptomycin (P/S, Life Technologies), maintained at 37 °C in a humidified incubator with 5% CO_2_. Calu-3 cells were cultured in Minimum Essential Medium (MEM) supplemented with 10% FBS, 1% non-essential amino acids, and 1% P/S. THP-1 cells were maintained in RPMI 1640 medium containing 10% FBS, 1% P/S, and 50 μM β-mercaptoethanol (2-ME). Experiments utilized cells between passages 4 and 25.

Cell transfection and cell-cell fusion assays

4.3

HEK293T cells were seeded in 24-well plates at 5 × 10^5^ cells/mL (500 μL per well) and cultured overnight until reaching 70%–80% confluency. Cells were transfected using Lipofectamine 2000 (Life Technologies) with 200 ng of SARS-CoV-2 spike plasmid and 200 ng of Cre plasmid or 200 ng of ACE2 plasmid plus stop-luciferase plasmid. After 24 h of transfection, cell-cell fusion assays were initiated. For the luciferase-based fusion detection system, the Cre-loxP firefly luciferase (stop-luc) system was employed to monitor membrane fusion mediated by spike protein via site-specific DNA recombination. To visualize fusion, 100 ng of ZsGreen plasmid was co-transfected to label spike-expressing HEK293T cells for microscopic imaging. Post-transfection, cells were gently detached using ice-cold PBS, centrifuged at 600×g for 4 min at room temperature, and resuspended in DMEM. The HEK293T-S cell suspension was mixed at a 1:1 ratio with control HEK293T cells, HEK293T-ACE2 cells, Vero E6-ACE2 cells, or Calu-3 cells, then seeded into 48- or 96-well plates and incubated for 16 h.

For luciferase-based cell-cell fusion quantification, cells were lysed with NP40 lysis buffer (0.5% (vol/vol) NP40, 25 mM Tris (pH 7.3), 150 mM NaCl, 5% glycerol, and 1 × EDTA-free protease inhibitor cocktail (Roche)). Lysates were mixed with Bright-Glo luciferase substrate (E2610, Promega), and relative luminescence units (RLU, 1-min read) were measured using a Synergy H1 plate reader (Biotek). For microscopic imaging, syncytia were visualized using an Olympus IX73 inverted fluorescence microscope equipped with a 10 × objective and a 12-bit monochrome CMOS camera.

Western blotting

4.4

After 16 h of incubation, the supernatant of adherent cells was discarded, and the cells were placed on ice. Cells were lysed with 2 × reducing Laemmli sample buffer, followed by heating at 95 °C in a metal bath for 10 min. Protein samples were separated by standard Tris-glycine SDS-PAGE on 7.5% Tris-glycine polyacrylamide gels. Following electrophoresis, proteins were transferred to 0.45 μm PVDF membranes (Millipore) using Towbin transfer buffer under wet conditions. All membranes were blocked with 5% skimmed milk powder in PBST (PBS containing 0.1% Tween 20) at room temperature for 1 h, then incubated with primary antibodies overnight at 4 °C. After primary antibody incubation, nonspecifically bound antibodies were removed by washing three times with PBST (10 min per wash at room temperature). Horseradish peroxidase (HRP)-conjugated secondary antibodies (Jackson ImmunoResearch) were incubated at room temperature for 1 h, followed by three additional PBST washes. Protein bands were visualized using PicoLight Enhanced Chemiluminescence Substrate (Epizyme Scientific). Chemiluminescent signals were captured using a Tanon 5200 Chemiluminescence Imaging System, and molecular weight standards were labeled in the images.

Quantitative real-time PCR (qPCR)

4.5

THP-1 cells were seeded in 6-well plates at a density of 1 × 10^6^ cells/mL and stabilized in a 37 °C cell culture incubator for 30 min. After stimulation, cells and supernatants were collected by centrifugation at 2000×g for 5 min at room temperature. Cell pellets were washed three times with PBS, then lysed with 1 mL TRIzol reagent (Catalog No. 15596018, Thermo Fisher Scientific) for 5 min at room temperature. Chloroform (1/5 of the total volume) was added, and the mixture was gently inverted for 1 min. After centrifugation at 10 000 rpm for 10 min at 4 °C, the aqueous phase was transferred to a new 1.5 mL nuclease-free EP tube, avoiding contamination by the white DNA layer. RNA was precipitated by adding an equal volume of isopropanol, followed by incubation at −20 °C for ≥20 min or overnight. The RNA pellet was collected by centrifugation at 14 000 rpm for 20 min at 4 °C, washed twice with 75% ethanol (prepared with DEPC-treated water), and air-dried before dissolving in 30 μL–50 μL DEPC-treated water. cDNA was synthesized using the GoScript Reverse Transcription Kit (Promega). Quantitative real-time PCR was performed using SYBR Green Real-Time PCR Master Mix (TOYOBO) on an ABI QuantStudio 6 Flex Real-Time PCR System (Thermo Fisher Scientific). Gene-specific primer sequences are listed in Supplementary Table 1, with GAPDH as the internal reference gene. Relative gene expression levels (Relative Units, RU) were calculated using the ΔΔCt method (Zheng et al., 2025).

CRISPR-Cas9-mediated gene targeting

4.6

The principle of CRISPR-Cas9 is as follows: the gRNA guides the Cas9 enzyme to precisely locate the target DNA sequence, where Cas9 induces a double-strand break (DSB). The cell's natural repair mechanisms, such as Non-Homologous End Joining (NHEJ), are then exploited to achieve precise genome editing. To construct gene knockout THP-1 or HEK293T cells using CRISPR-Cas9 gene editing technology, the key steps are as follows: Specific single-guide RNA (sgRNA) targeting the gene of interest is designed based on the target gene sequence, then the sgRNA is cloned into the LentiCRISPRv2 lentiviral vector (Sanjana et al., 2014). To produce lentiviral particles, the LentiCRISPRv2-sgRNA plasmid, psPAX2 packaging plasmid, and VSV-G envelope plasmid are co-transfected into HEK293T cells at a ratio of 4:3:2 using Lipofectamine 2000; after 6-8 h, the medium is replaced. Viral supernatants are collected 48 h post-transfection, filtered through 0.22 μm PVDF membranes, and used to infect THP-1 or HEK293T cells in the presence of 10 μg/mL polybrene. 24 h after such infection, cells are cultured in medium containing 2 μg/mL puromycin for 48 h. Surviving cells are diluted to single-cell suspensions and seeded into 96-well plates via limiting dilution to obtain stable monoclonal knockout cell lines. Gene-specific primer sequences are listed in Supplementary Table 2.

Immunofluorescence staining and imaging

4.7

Cell preparation: HEK293T cells transfected with Spike or ACE2 are seeded onto sterile poly-D-lysine (100 μg/mL)-coated 12 mm coverslips in 24-well plates. Upon harvesting, cells are fixed with 4% paraformaldehyde (4% PFA) for 20 min at room temperature (RT), permeabilized with 0.1% Triton X-100 for 10 min, and blocked with immunostaining blocking buffer (Beyotime, P0102) for 1 h at RT or overnight at 4 °C. Then, primary antibodies are applied for 1 h at RT, followed by three PBS washes. Fluorescent secondary antibodies or Actin-Tracker Green-488 are incubated for 1 h in the dark. After DAPI staining (10 min) and PBS washes, coverslips are mounted with anti-fade reagent. Images are acquired using an Olympus SpinSR10 confocal microscope with a 100 × oil objective. DAPI (blue), Actin-Tracker Green-488 (green), and other fluorescent signals are captured separately and merged using ImageJ (NIH), with scale bars added.

Authentic SARS-CoV-2 infection of cells

4.8

All experiments involving authentic SARS-CoV-2 virus in vitro were conducted in the biosafety level 3 laboratory of the Shanghai Municipal Center for Disease Control and Prevention (CDC). The experiments and protocols in this study were approved by the Ethical Review Committee of the Shanghai CDC (Permit Number: 2022-51). Briefly, Caco-2 cells were seeded into 24-well plates at a density of 4 × 10^5^ cells/mL overnight, then pre-treated with different reagents for 1 h before infection with 0.5 multiplicity of infection (MOI) Delta authentic SARS-CoV-2 (B.1.617.2) for 24 h or 48 h. Cell lysates were collected for spike S2′ cleavage and N protein immunoblots.

Statistical analysis

4.9

Bar graph data are presented as mean ± standard error of the mean (SEM) with individual data points shown. All analyses are performed using GraphPad Prism v9.0. For multi-group comparisons, matched one-way ANOVA with Sidak's post hoc comparisons is applied. Statistical significance (P values) is annotated between comparison groups and displayed in figures.

CRediT authorship contribution statement

Dong Duan: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Xu Zheng: Writing – review & editing, Writing – original draft, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Yanqiu Zhou: Investigation. Mengmeng Cui: Investigation, Formal analysis. Yunyi Li: Investigation. Xiaoxian Cui: Investigation. Yuying Yang: Investigation. Min Chen: Investigation. Huanyu Wu: Investigation. Xin Chen: Investigation, Funding acquisition. Guangxun Meng: Writing – review & editing, Supervision, Resources, Methodology, Funding acquisition, Conceptualization.

Declaration of interest

All authors declare that we have no known financial conflicts of interest or personal relationships that could influence the results reported in this paper.