SARS-CoV-2 aberrantly elevates mitochondrial bioenergetics to induce robust virus propagation

Abstract

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a 'highly transmissible respiratory pathogen, leading to severe multi-organ damage. However, knowledge regarding SARS-CoV-2-induced cellular alterations is limited. In this study, we report that SARS-CoV-2 aberrantly elevates mitochondrial bioenergetics and activates the EGFR-mediated cell survival signal cascade during the early stage of viral infection. SARS-CoV-2 causes an increase in mitochondrial transmembrane potential via the SARS-CoV-2 RNA-nucleocapsid cluster, thereby abnormally promoting mitochondrial elongation and the OXPHOS process, followed by enhancing ATP production. Furthermore, SARS-CoV-2 activates the EGFR signal cascade and subsequently induces mitochondrial EGFR trafficking, contributing to abnormal OXPHOS process and viral propagation. Approved EGFR inhibitors remarkably reduce SARS-CoV-2 propagation, among which vandetanib exhibits the highest antiviral efficacy. Treatment of SARS-CoV-2-infected cells with vandetanib decreases SARS-CoV-2-induced EGFR trafficking to the mitochondria and restores SARS-CoV-2-induced aberrant elevation in OXPHOS process and ATP generation, thereby resulting in the reduction of SARS-CoV-2 propagation. Furthermore, oral administration of vandetanib to SARS-CoV-2-infected hACE2 transgenic mice reduces SARS-CoV-2 propagation in lung tissue and mitigates SARS-CoV-2-induced lung inflammation. Vandetanib also exhibits potent antiviral activity against various SARS-CoV-2 variants of concern, including alpha, beta, delta and omicron, in in vitro cell culture experiments. Taken together, our findings provide novel insight into SARS-CoV-2-induced alterations in mitochondrial dynamics and EGFR trafficking during the early stage of viral infection and their roles in robust SARS-CoV-2 propagation, suggesting that EGFR is an attractive host target for combating COVID-19.

Fig. 1

SCoV2 induces aberrant mitochondrial elongation and alterations in mitochondrial oxidative phosphorylation process. a Quantification of intracellular ATP level of SCoV2-infected HEK293T cells at the indicated time points (MOI of 1). Data shown are the average of two independent experiments (mean ± SD; n = 4; *p < 0.05). b Quantification of ΔΨ_m change in HEK293T (left) and Calu-3 (right) cells infected with SCoV2 and IAV (H1N1, PR8 strain), respectively (MOI of 1). ΔΨ_m in SCoV2 or IAV-infected cells were monitored by TMRE assay at 1 day post-infection. ΔΨ_m in FCCP-treated cells were monitored by TMRE assay at 2 h post-treatment. Data shown are the representative of three independent experiments (mean ± SD; n = 3; *p < 0.05, **p < 0.001, ***p < 0.0001). ΔΨ_m, mitochondrial membrane potential. FCCP, synthetic mitochondrial uncoupler, a control for ΔΨ_m loss. The accompanying diagram (right) represents the SCoV2-induced increase in ΔΨ_m. c Quantification of ΔΨ_m increase by SCoV2 RNA-N cluster. HEK293T cells were transfected with DNA plasmid encoding SCoV2 nucleocapsid (N) gene. ΔΨ_m in cells were monitored by TMRE assay at 24 h post-transfection (left panel). HEK293T cells transfected with DNA plasmid encoding SCoV2 nucleocapsid (N) gene for 24 h were transfected with SCoV2 RNA (1 μg/ml, middle panel), poly(I:C) (1 μg/ml, middle panel), human tRNA (1 μg/ml, right panel) and IAV RNA (1 μg/ml, right panel), respectively. ΔΨ_m in cells were monitored by TMRE assay at 4 h post-transfection. Data shown are the representative of three independent experiments (mean ± SD; n = 4 or n = 8; *p < 0.05, **p < 0.001). poly(I:C), synthetic analogue of double-stranded RNA, a control for ΔΨ_m loss; human tRNA, a control for endogenous sense RNA; TF ctrl., transfection reagent only; mock, pcDNA3.1 plasmid DNA. The accompanying diagram (right) represents the SCoV2 RNA-N cluster-induced increase in ΔΨ_m. d Confocal microscopy showing the elongated shape of mitochondria in SCoV2-infected cells. SCoV2-infected HEK293T cells were immunostained with SCoV2-S (red) and TOM20 (green) antibodies. Infected (+) and uninfected (-) cells are marked. The white arrows indicate the expression of SCoV2 spike (S) antigen as an infection marker. Nuclei are demarcated with white circles. Yellow scale bar, 10 μm. The zoomed images reveal elongated shape of mitochondria in SCoV2-infected cells (right) compared with typical mitochondrial tubular network in uninfected cells (left). The accompanying graph represents the quantification of mitochondrial length by MBF ImageJ. e Western blot analysis of MFN1/2 expression in SCoV2-infected cells. Mitochondrial fraction (Mito) isolated from HEK293T cells infected with SCoV2 at an MOI of 1 was analysed by immunoblotting with anti-MFN1/2 antibody. SCoV2 nucleocapsid (N) protein, infection control; TOM20, an internal loading control. f Western blot analysis of Drp1, phospho-Drp1 (S616) and mitochondrial fission factor (MFF) expression in SCoV2-infected cells. Whole cell lysates extracted from HEK293T cells infected with SCoV2 at an MOI of 1 was analysed by immunoblotting with antibodies specific to Drp1, phospho-Drp1 (Ser616), and MFF, respectively. SCoV2 nucleocapsid (N) protein, infection control; β-actin, an internal loading control. g Real-time qPCR data showing the increase in mitochondrial DNA of SCoV2-infected HEK293T cells. At 1 day post-infection, the expression level of mitochondrial ND2 and COX2 DNA was analysed by real-time qPCR. GAPDH was used to normalise changes in ND2 and COX2 expression. h Western blot analysis of mitochondrial respiratory chain complex enzyme expression in HEK293T cells infected with SCoV2 at an MOI of 1. At 1 day post-infection, the expression level of complex I, II, III, IV and V enzymes was analysed by immunoblotting with anti-Hu total OXPHOS complex antibody. C, complex; SCoV2 spike (S) protein, infection control; β-actin, an internal loading control. i Heat maps of relative mRNA of the indicated mitochondrial OXPHOS genes isolated from uninfected and SCoV2-infected HEK293T cells. Each box indicates an average of three independent experiments. Colour indicates log2 fold-change for uninfected vs. SCoV2-infected cells. j Western blot analysis of mitochondrial respiratory chain complex enzyme expression in lung tissue isolated hACE2 transgenic mouse infected with SCoV2. At 1 day post-infection, the expression level of complex I, II, III, IV and V enzymes was analysed by immunoblotting with anti-Rodent total OXPHOS complex antibody. C, complex; SCoV2 nucleocapsid (N) protein, infection control; β-actin, an internal loading control

Fig. 2

SCoV2 promotes EGFR-mediated cell survival signal cascade and mitochondrial EGFR translocation. a Quantification of SCoV2-infected cell viability. Total number of viable HEK293T (left) or Calu-3 (right) cells infected with SCoV2 at an MOI of 1 was measured at the indicated time points as described in Materials and Methods. Data shown are the representative of three independent experiments (mean ± SD; n = 2). b Quantitative analysis of EGFR gene expression in SCoV2-infected cells. At 1 day post-infection, the expression level of EGFR mRNA in HEK293T cells infected with SCoV2 at an MOI of 1 were analysed by real-time qRT-PCR (mean ± SD; n = 3; *p < 0.05). Data shown are the representative of two independent experiments. c Western blot analysis showing SCoV2-induced activation in EGFR-mediated signal cascade in SCoV2-infected cells. At 1 day post-infection, whole cell lysates (WCL) of HEK293T (left) or Calu-3 (right) cells infected with SCoV2 at an MOI of 1 were analysed by immunoblotting with antibodies specific to p-EGFR, EGFR, p-Akt (Thr308) and Akt. SCoV2 nucleocapsid (N) protein, infection control; β-actin, an internal loading control. d Western blot analysis showing SCoV2-induced activation in EGFR-mediated signal cascade in lung specimen of SCoV2-infected hACE2 transgenic mouse. At 1 day post-infection, whole tissue lysates (WTL) of lung specimens of SCoV2-infected hACE2 transgenic mouse were analysed by immunoblotting with antibodies specific to p-EGFR, EGFR, p-Akt (Thr308) and Akt. SCoV2 nucleocapsid (N) protein, infection control; β-actin, an internal loading control. e Western blot analysis of SCoV2-induced mitochondrial translocation of EGFR. HEK293T cells were infected with SCoV2 at an MOI of 1. At 2 days post-infection, cytosolic (Cyto) and mitochondrial (Mito) fractions isolated from uninfected and SCoV2-infected HEK293T cells were evaluated by immunoblotting with antibodies specific to p-EGFR and EGFR. Fractions: purified cytoplasm, Cyto; purified mitochondria, Mito. Organelle marker: TOM20, mitochondria; GAPDH, cytoplasm. Infection marker: SCoV2 nucleocapsid (N) antigen. The relative intensity of EGFR normalised to TOM20 and p-EGFR normalised to EGFR, respectively, was analysed by ImageJ. f Confocal microscopy showing mitochondrial translocation of EGFR in SCoV2-infected cells. Uninfected (upper) and SCoV2-infected (lower) HEK293T cells prestained with MitoTracker (red) were immunostained with antibodies specific to EGFR (green) and SCoV2-N (blue). Nuclei are demarcated with white circles. Infected (+) and uninfected (-) cells are marked. Yellow scale bar, 5 μm (upper), 10 μm (lower). In the zoomed images, the white arrow indicate endogenous EGFR recruited to the mitochondria in SCoV2-infected cells (yellow). g Western blot analysis of mitochondrial EGFR translocation in the lung tissues of hACE2 transgenic mouse infected with SCoV2. Mitochondrial fractions (Mito) isolated from lung tissue of uninfected and SCoV2-infected hACE2 transgenic mouse were analysed by immunoblotting with EGFR antibody. Organelle marker: TOM20, mitochondria. Infection marker: SCoV2 nucleocapsid (N) antigen. The relative intensity of EGFR expression was normalised to TOM20 expression

Fig. 3

Tyrosine kinase inhibitors targeting EGFR signalling pathway reduces SCoV2 propagation. a A scheme for primary screening of the antiviral effect of EGFR inhibitors against SCoV2 propagation. At 4 h post-infection, HEK293T cells infected with SCoV2 at an MOI of 1 were washed with fresh cell culture media 5 times and subsequently treated with the indicated EGFR inhibitors (10 μM) for 44 h. Cell culture supernatant and pellet were used for further analyses (b, c). b Extracellular RNA was isolated from culture supernatants of SCoV2-infected HEK293T cells and used for the quantification of SCoV2 RNA levels by real-time qRT-PCR. Data shown are the representative of two independent experiments (mean ± SD; n = 2). c Whole cell lysates of SCoV2-infected HEK293T cells were analysed by immunoblotting with antibody specific to SCoV2 N protein. β-actin, an internal loading control. d Rescue of SCoV2-induced mitochondrial translocation of EGFR by vandetanib treatment. Cytosolic (Cyto) and mitochondrial (Mito) fractions isolated from uninfected and SCoV2-infected HEK293T cells (MOI of 1) treated with vandetanib (10 μM) were analysed by immunoblotting with EGFR antibody. Fractions: purified cytoplasm, Cyto; purified mitochondria, Mito. Organelle marker: TOM20, mitochondria; GAPDH, cytoplasm. Infection marker: SCoV2 nucleocapsid (N). e Rescue of SCoV2-induced elevation in intracellular ATP level by vandetanib treatment. At 1 day post-treatment of vandetanib (10 μM), intracellular ATP level in SCoV2-infected HEK293T cells (MOI of 1) was analysed as described in Materials and Methods. Data shown are the average of two independent experiments (mean ± SD; n = 2; *p < 0.05). f, Western blot analysis of mitochondrial respiratory chain complex enzyme expression. At 1 day post-treatment of vandetanib (10 μM), the expression level of complex I, II, III, IV and V enzymes was analysed by immunoblotting with anti-Hu total OXPHOS complex antibody. C, complex; β-actin, an internal loading control; SCoV2 N, infection marker. g–h, HEK293T cells infected with SCoV2 for 4 h at an MOI of 1 were washed with fresh cell culture media 5 times and then further cultured for 20 h in the presence of vandetanib (10 μM) for RNAseq analysis. g Heat maps of relative mRNA expression of the indicated mitochondrial OXPHOS genes isolated from SCoV2-infected and vandetanib-treated SCoV2-infected cells. Each box indicates an average of three independent experiments. Colour indicates log2 fold-change for SCoV2-infected vs. uninfected cells and vandetanib-treated SCoV2-infected vs. SCoV2-infected cells, respectively. h Read coverage across the SCoV2 genome in the presence and absence of vandetanib. The graph represents the number of viral reads per position of the SCoV2 genome in HEK293T cells (SCoV2, dark blue; SCoV2/vandetanib, orange). A scaled model of the SCoV2 genome and its genes is portrayed below

Fig. 4

Vandetanib is a potent antiviral agent for SCoV2 propagation. a Apoptosis analysis of SCoV2-infected and uninfected cells treated with vandetanib. HEK293T cells infected with SCoV2 at an MOI of 1 for 4 h were further cultured in the presence of vandetanib (1 or 10 μM) for 44 h. Apoptotic cell death was analysed by flow cytometry as described in Materials and Methods. Staurosporine (200 nM, 44 h) was used as a positive control for inducing apoptotic cell death. Data shown are the representative of two independent experiments. b Dose-dependent antiviral effect of vandetanib against SCoV2 infection. At 2 days post-treatment, whole cell lysates of SCoV2-infected HEK293T cells (MOI of 1) treated with vandetanib at the indicated concentrations were analysed by immunoblotting with SCoV2 N antibody. β-actin, an internal loading control. c Anti-SCoV2 effect by post-treatment of vandetanib. HEK293T cells infected with SCoV2 at an MOI of 1 for 4 h were washed with fresh cell culture media and then treated with vandetanib (0.1, 1, or 10 μM) for 44 h. d Anti-SCoV2 effect by pre-treatment of vandetanib. HEK293T cells were treated with vandetanib (0.1, 1, or 10 μM) for 4 h and then infected with SCoV2 at an MOI of 1 for 44 h. Intracellular SCoV2 RNA level in SCoV2-infected cells was analysed by real-time qRT-PCR using PCR primers set specific to the SCoV2 N gene (c, d). Data shown are the representative of two independent experiments (mean ± SD; n = 2). e–h, In vivo anti-SCoV2 efficacy of vandetanib in hACE2 transgenic mouse model susceptible to SCoV2 infection. e A scheme for analysing clinical disease score and inflammation in SCoV-2-infected mice orally administrated with vandetanib (f, g). Eight-week-old hACE2 transgenic mice (n = 5 per group) were intranasally (IN) inoculated with SCoV2 (2 × 10³ pfu/head, 10 MLD₅₀, clade S). One hour later, they were orally administrated with vandetanib (25 mg/kg) daily. At 6 days post-infection, all mice were terminated for further analyses (f, g). f Clinical scores of SCoV2-uninfected hACE2 transgenic mice (open circle), SCoV2-infected hACE2 transgenic mice administrated with vehicle (black circle) and SCoV2-infected hACE2 transgenic mice administrated with vandetanib (red circle). Vehicle control, PBS with 1% Tween 80 (f–h). Clinical scores for all mice were monitored daily based on ruffled fur (1 point), reduced mobility (1 point), hunched posture (1 point) and death (4 points) as described in Materials and Methods. g Immunohistochemistry analysis showing the effect of vandetanib in SCoV2-induced severe lung injury in hACE2 transgenic mouse. At 6 days post-infection, immunohistochemistry analysis was performed as described in Materials and Methods. Black scale bar, 500 μm (upper), 100 μm (lower). H&E score (right panel, mean ± SD; n = 12; *p < 0.0001). h In vivo anti-SCoV2 efficacy of vandetanib in hACE2 transgenic mice. Eight-week-old hACE2 transgenic mice (n = 3 per group) were intranasally (IN) inoculated with SCoV2 (2 × 10³ pfu/head, clade S). One hour later, they were orally administrated with vandetanib (25 mg/kg) daily. At 3 days post-infection, all mice were terminated for further analyses (upper panel). Intracellular SCoV2 RNA levels of lung tissues isolated from SCoV2-infected hACE2 transgenic mice was analysed by real-time qRT-PCR. Each data point represents the average of two independent experiments (mean ± SD; n = 2, lower panel)

Fig. 5

Vandetanib reveals a potent antiviral effect against various SCoV2 variants. a A scheme for analysing the antiviral effect of vandetanib (EGFR inhibitor) against various SCoV2 variants. HEK293T cells were infected with SCoV2 (S, V, G, GH and GR clades) and their variants of concern (VOC) including alpha (B.1.1.7), beta (B.1.351), delta (B.1.617.2) and omicron (B.1.529) variants, respectively, at an MOI of 1. At 4 h post-infection, HEK293T cells were washed with fresh cull culture media 5 times and then further incubated in the presence of vandetanib for 44 h. Cell culture media was used for further analyses of real-time qRT-PCR (b) and FFU assay using Vero E6 cells (c). b Real-time qRT-PCR data showing reduced extracellular SCoV2 RNA levels following treatment with vandetanib. Total RNA was isolated from the culture media of HEK293T cells infected with SCoV2 in the presence of vandetanib (10 μM) and then used for analysis of extracellular SCoV2 RNA level by real-time qRT-PCR using PCR primers set specific to the SCoV2 N gene. Data shown are the representative of two independent experiments (mean ± SD; n = 2). DMSO was used as the negative control. c FFU assay data showing the reduction in SCoV2 infectivity following treatment with vandetanib. Cell culture media of SCoV2-infected HEK293T cells post-treated with vandetanib at the indicated concentrations (1, 5 and 10 μM) were transferred to fresh Vero E6 cells and then further incubated for 8 h for FFU assay as described in Materials and Methods. SCoV2, cell culture media of SCoV2-infected HEK293T cells in the presence of vandetanib; Uninfected, cell culture media of SCoV2-uninfected HEK293T cells. The accompanying graphs show the average of two independent experiments (right panel). DMSO was used as the negative control. N.D. not determined

Fig. 6

A schematic diagram showing the significant contribution of SCoV2-induced altered mitochondrial dynamics and mitochondrial EGFR translocation in sustaining viral propagation. First, SCoV2 RNA and nucleocapsid complex increase ΔΨm during the early stages of SCoV2 infection. This alteration subsequently promotes mitochondrial elongation. SCoV2 also activates the mitochondrial OXPHOS process, thereby promoting ATP production. Second, SCoV2 activates EGFR-mediated cell survival signalling and subsequently promotes mitochondrial EGFR internalisation, which contributes to the maintenance of abnormal mitochondrial bioenergetics. These alterations are physiologically relevant to the maintenance of homoeostasis of SCoV2-infected cells and robust SCoV2 propagation