Role of TOMM34 on NF-κB activation-related hyperinflammation in severely ill patients with COVID-19 and influenza

Abstract

Background: Highly pathogenic respiratory RNA viruses such as SARS-CoV-2 and its associated syndrome COVID-19 pose a tremendous threat to the global public health. Innate immune responses to SARS-CoV-2 depend mainly upon the NF-κB-mediated inflammation. Identifying unknown host factors driving the NF-κB activation and inflammation is crucial for the development of immune intervention strategies.

Methods: Published single-cell RNA sequencing (scRNA-seq) data was used to analyze the differential transcriptome profiles of bronchoalveolar lavage (BAL) cells between healthy individuals (n = 27) and patients with severe COVID-19 (n = 21), as well as the differential transcriptome profiles of peripheral blood mononuclear cells (PBMCs) between healthy individuals (n = 22) and severely ill patients with COVID-19 (n = 45) or influenza (n = 16). Loss-of-function and gain-of-function assays were performed in diverse viruses-infected cells and male mice models to identify the role of TOMM34 in antiviral innate immunity.

Findings: TOMM34, together with a list of genes encoding pro-inflammatory cytokines and antiviral immune proteins, was transcriptionally upregulated in circulating monocytes, lung epithelium and innate immune cells from individuals with severe COVID-19 or influenza. Deficiency of TOMM34/Tomm34 significantly impaired the type I interferon responses and NF-κB-mediated inflammation in various human/murine cell lines, murine bone marrow-derived macrophages (BMDMs) and in vivo. Mechanistically, TOMM34 recruits TRAF6 to facilitate the K63-linked polyubiquitination of NEMO upon viral infection, thus promoting the downstream NF-κB activation.

Interpretation: In this study, viral induction of TOMM34 is positively correlated with the hyperinflammation in severely ill patients with COVID-19 and influenza. Our findings also highlight the physiological role of TOMM34 in the innate antiviral signallings.

Funding: A full list of funding sources can be found in the acknowledgements section.

Keywords: COVID-19; Inflammation; Innate immunity; NF-κB; TOMM34.

Fig. 1

TOMM34 mRNA levels are elevated in severely ill patients with COVID-19 and influenza by viral stimulation. a) UMAP projection of BAL cell types from healthy controls (HC, n = 26) and severe ill patients with COIVD-19 (COVID-19, n = 21) using published scRNA-seq data. b) Heatmap illustrating the fold changes (COVID-19 group vs HC group) of TOMM genes expression in the indicated cell types. The sample size and related power analysis are indicated in Supplementary Table S3. c) Violin plot showing the mRNA levels of TOMM34 in the indicated cell types from HC group and COVID-19 group. The sample size and related power analysis are indicated in Supplementary Table S3. d) Gene expression correlation analysis of TOMM34 and the indicated genes encoding antiviral effectors and inflammatory agents in Mo-macrophages from HC group and COVID-19 group. The shape of the 'ovals' or ellipses is designed to represent the correlation. The upper right triangle displays the P values, and the lower left triangle represents the confidence intervals. A central horizontal dashed line within the left lower triangle represents the value of zero. The 95% confidence intervals (CI) for significance are marked by the upper and lower boundaries of the box plot, as shown in Supplementary Table S4. The sample size and related power analysis are indicated in Supplementary Table S3. e) Boxplots representing the TOMM34 protein levels in multiple organs from COVID-19 autopsy samples and non-COVID-19 control samples. The sample size and related power analysis are indicated in Supplementary Table S3. f) Boxplots showing the mRNA levels of TOMM34 in the indicated cell types of PBMCs from healthy controls and severely ill patients with COVID-19 or influenza. The UMAP projection of PBMCs myeolid cells are shown in Supplementary Fig. S2a. The sample size and related power analysis are indicated in Supplementary Table S3. g) A549-ACE2 (A549 stably expressing ACE2) cells were mock-infected (Mock) or infected with SARS-CoV-2 (multiplicity of infection, MOI = 0.4) for 24 h. qRT-PCR analysis was conducted to examine the relative TOMM34 mRNA levels. h) Caco-2-N cells (Caco-2 stably expressing SARS-CoV-2 nucleocapsid protein) were mock-infected or infected with SARS-CoV-2 trVLP (MOI = 0.1) for the indicated hours. qRT-PCR analysis was conducted to examine the relative TOMM34 mRNA levels. i) Primary murine BMDMs were mock-infected or infected with SeV (20 HA units/ml) and VSV (MOI = 0.1) for 12 h, or with PR8 (MOI = 1), SARS-CoV-2 trVLP (MOI = 0.1) and RSV (MOI = 1) for 24 h. qRT-PCR analysis was conducted to examine the relative Tomm34 mRNA levels. j) THP-1-derived macrophages were mock-infected or infected with SeV (20 HA units/ml) and VSV (MOI = 0.1) for 12 h, or with PR8 (MOI = 1) for 24 h. qRT-PCR analysis was conducted to examine the relative TOMM34 mRNA levels. Data are represented as mean ± SD calculated from at least three independent experiments (g–j). Wilcoxon rank-sum test (b, c, e, f), two-tailed Student's t test (g–j).

Fig. 2

TOMM34 potentiates antiviral cytokines expression upon virus infection. a) Effects of TOMM proteins overexpression on VSV-induced IFN-β promoter activation. HEK293T cells were co-transfected with IFN-β luciferase reporter plasmid (IFN-Beta_pGL3), renilla luciferase plasmid (pRL-TK) plus vectors encoding hemagglutinin (HA)-tagged TOMM proteins as indicated or empty vector (EV) for 24 h, and were then stimulated with VSV (MOI = 0.1) for 12 h. Cell lysates were obtained and subjected to dual luciferase assay. Activities of firefly luciferase were normalized to those of renilla luciferase and shown as fold induction. Protein expression levels are shown in Supplementary Fig. S3a. b) Effects of TOMM34 overexpression on IFN-β promoter activation induced by diverse stimuli. Similar to (a), HEK293T cells were transfected with IFN-Beta_pGL3, pRL-TK plus plasmid encoding HA-TOMM34 or EV for 24 h, and were then mock-infected (Mock) or stimulated with VSV, SeV (20 HA units/mL), PR8 (MOI = 1), RSV (MOI = 1) and poly(I:C) for 12 h. Cell lysates were obtained and subjected to dual luciferase assay. c) Effects of TOMM34 overexpression on antiviral cytokines induction. qRT-PCR analysis of IFNB1, IL6, and CCL2 mRNA levels in HEK293T cells transfected with HA-TOMM34 or EV for 24 h, followed by VSV infection for 12 h. The mRNA levels of TNF and ISG54 are shown in Supplementary Fig. S3b. d) Impacts on antiviral cytokines induction in the absence of TOMM34. qRT-PCR analysis of IFNB1, IL6, CCL2 and ISG54 mRNA levels in TOMM34^+/+ and TOMM34^−/− HEK293T cells mock-infected or infected with VSV for 12 h. Protein levels of TOMM34 in TOMM34^+/+ and TOMM34^−/− HEK293T cells, Supplementary Fig. S3c; mRNA levels of TNF in TOMM34^+/+ and TOMM34^−/− HEK293T cells, Supplementary Fig. S3d. e) Effects of TOMM34 recovery on antiviral cytokines induction in TOMM34^−/− HEK293T cells. qRT-PCR analysis of IFNB1 and CCL2 mRNA levels in TOMM34^+/+ and TOMM34^−/− HEK293T cells that were transfected with EV or vectors encoding HA-tagged TOMM34 and its mutant S93/160A for 24 h, followed by 12 h-VSV infection. Protein expression levels are shown in Supplementary Fig. S3e. f) Microscopy imaging of TOMM34^+/+ and TOMM34^−/− HEK293T cells that were mock-infected or infected with GFP-conjugated PR8 virus (PR8-GFP) for 12 h. Scale bars, 200 μm. g) Impacts on antiviral cytokines induction in the absence of TOMM34 in A549 cells. qRT-PCR analysis of IFNB1, IL6 and CCL2 mRNA levels in TOMM34^+/+ and TOMM34^−/− A549 cells mock-infected or infected with VSV, SeV (20 HA units/mL), PR8 (MOI = 1) and RSV (MOI = 1) for 12 h. Protein levels of TOMM34 in TOMM34^+/+ and TOMM34^−/− A549 cells, Supplementary Fig. S3f; mRNA levels of TNF in TOMM34^+/+ and TOMM34^−/− A549 cells, Supplementary Fig. S3g. h) Effects of murine Tomm34 overexpression on antiviral cytokines induction. qRT-PCR analysis of Ifnb1, Il6, Tnf, Ccl2 and Isg54 mRNA levels in mouse embryonic fibroblast NIH/3T3 cells that were transfected with plasmid encoding HA-Tomm34 or empty vector for 24 h, and were subsequently mock-stimulated or stimulated with poly(I:C) for 12 h. i) Impacts on antiviral cytokines induction in the absence of murine Tomm34. qRT-PCR analysis of Ifnb1, Il6, Ccl2 and Isg54 mRNA levels in Tomm34^+/+ and Tomm34^−/− NIH/3T3 cells that were mock-infected or infected with VSV (MOI = 0.1) for 12 h. Protein levels of Tomm34 in Tomm34^+/+ and Tomm34^−/− NIH/3T3 cells, Supplementary Fig. S3i; mRNA levels of Tnf in Tomm34^+/+ and Tomm34^−/− NIH/3T3 cells cells, Supplementary Fig. S3j. j) Effects of Tomm34 recovery on antiviral cytokines induction in Tomm34^−/− NIH/3T3 cells. qRT-PCR analysis of Ifnb1 and Il6 mRNA levels in Tomm34^+/+ and Tomm34^−/− NIH/3T3 cells that were transfected with HA-Tomm34 or EV for 24 h, followed by infection with VSV for 12 h. Protein expression levels are shown in Supplementary Figure S2k. Data are represented as mean ± SD calculated from three independent experiments, or are representative of three independent experiments. P values are reported with two significant digits, or shown as “P < 0.0001” (Unpaired two-tailed Student's t test).

Fig. 3

Tomm34 is pivotal for antiviral innate immunity in primary murine macrophages. a) ELISA of IFN-β, IFN-α and IL-6 in supernatants of Tomm34^fl/fl and Tomm34^fl/flLyz2-Cre BMDMs unstimulated (US) or stimulated for 12 h with SeV (20 HA units/mL), VSV (MOI = 0.1), PR8 (MOI = 1), SARS-CoV-2 trVLP (MOI = 0.1), RSV (MOI = 1), poly(I:C) or for 8 h with LPS. Protein levels of Tomm34 in BMDMs collected from Tomm34^fl/fl and Tomm34^fl/flLyz2-Cre mice, Supplementary Fig. S4a; numbers and percentages of BMDMs from Tomm34^fl/fl and Tomm34^fl/flLyz2-Cre mice, Supplementary Fig. S4b and c. b) qRT-PCR analysis of Ifnb1, Ifna4, Il6, Ccl2 and Isg54 in Tomm34^fl/fl and Tomm34^fl/flLyz2-Cre BMDMs infected for 8 h with VSV. c) Flow cytometry analysis (left graph) and microscopy imaging (right graph) for the replication of PR8-GFP in Tomm34^fl/fl and Tomm34^fl/flLyz2-Cre BMDMs infected with PR8-GFP for 24 h. Scale bars, 200 μm. d) ELISA of IFN-β in supernatants of Tomm34^fl/fl and Tomm34^fl/flLyz2-Cre PMs infected for indicated hours with VSV. e) qPCR analysis of VSV RNA in PMs that were treated as in (d). Data are represented as mean ± SD calculated from three independent experiments, or are representative of three independent experiments. P values are reported with two significant digits, or shown as “P < 0.0001” (Unpaired two-tailed Student's t test).

Fig. 4

Tomm34^fl/flLyz2-Cre mice exhibit increased susceptibility to RNA viruses infection. a) Survival (Kaplan–Meier curve) of 7-week-old male Tomm34^fl/fl and Tomm34^fl/flLyz2-Cre mice (n = 10) after intraperitoneal injection of VSV (5 × 10⁸ PFU per mouse). Time-specific numbers at risk for each group are indicated at the bottom. Mice were infected at day 0 and daily monitored until day 8 (experimnetal end-point as the body weight was decreased to 80%). Survival data showing the group, state (0 = survival; 1 = death) and end times of mice were displayed in Supplementary Table S5. The sample size and related power analysis are indicated in Supplementary Table S3. b) ELISA of cytokines in serum of mice intraperitoneally injected with VSV (5 × 10⁸ PFU per mouse) or PBS for 24 h (n = 6). c) qRT-PCR analysis of Ifnb1, Ifna4, Il6, Ccl2 and Isg54 mRNA in the lungs, spleens and livers of mice treated as in (b) (n = 6). d) Plaque assays analyzing VSV titers in the lungs, spleens and livers of mice treated as in (b) (n = 6). e) qRT-PCR analysis of VSV RNA in the lungs, spleens and livers of mice treated as in (b) (n = 6). f) Hematoxylin-and-eosin (H&E) staining of sections of lungs from mice treated as in B. Scale bars, 100 μm. g) Survival (Kaplan–Meier curve) of 7-week-old male Tomm34^fl/fl and Tomm34^fl/flLyz2-Cre mice after intranasal infection with 50 PFU PR8 (n = 10). Time-specific numbers at risk for each group are indicated at the bottom. Mice were infected at day 0 and daily monitored until day 14 (experimnetal end-point as the body weight was decreased to 80%). Survival data showing the group, state (0 = survival; 1 = death) and end times of mice were displayed in Supplementary Table S5. The sample size and related power analysis are indicated in Supplementary Table S3. h) Body weights of mice treated as in (g). Mice infected with PR8 were weighted at day 0 (Mock) and daily monitored until any one was dead (day 6, experimnetal end-point). Area under the curve (AUC) for weight change was calculated in Supplementary Fig. S4. i) IFN-β and IL-6 levels in BAL from mice at Day 3 after infection with 50 PFU PR8 or PBS (n = 6). j) IFN-β and IL-6 levels in lung homogenates from mice treated as in (i). k) Viral titer in lung homogenates from mice treated as in (i). l) H&E staining of sections of lungs from mice treated as in (i). Scale bars, 50 μm. Data are represented as mean ± SD calculated from three independent experiments, or are representative of three independent experiments. P values are reported with two significant digits, or shown as “P < 0.0001” (Unpaired two-tailed Student's t test).

Fig. 5

TOMM34 interacts with NEMO to facilitate the virally-induced NF-κB activation. a) Impacts on NF-κB and IRF3 promoter activation in the absence of TOMM34. Similar to Fig. 2A, TOMM34^+/+ and TOMM34^−/− HEK293T cells were transfected with the NF-κB or IRF3 luciferase reporter plasmid plus pRL-TK for 24 h, and were mock-infected or infected with VSV for another 12 h. Cell lysates were obtained and subjected to dual luciferase assay. Activities of firefly luciferase are normalized to those of renilla luciferase and shown as fold induction. b) Impacts of TOMM70-knocked down on antiviral cytokines induction in TOMM34^−/− HEK293T cells. qRT-PCR analysis of IFNB1, CCL2 and ISG54 mRNA levels in TOMM34^+/+ and TOMM34^−/− HEK293T cells that were treated with TOMM70 siRNA (siTOMM70) or scramble siRNA (Control) for 36 h, followed by infection with VSV for 12 h. c) Impacts of TOMM34 deficiency on activation of the IKKα/β/γ-NF-κB signalling. Immunoblotting (IB) analysis showing the protein levels of NF-κB p65 and its phosphorylated form (p-p65), IKKα/β and p-IKKα/β, IRF3 and p-IRF3, as well as TOMM34 in TOMM34^+/+ and TOMM34^−/− HEK293T cells infected with VSV for the indicated time. β-actin was immunoblotted as loading control. d) Immunofluorescence staining (with anti-p65 antibody) and miscroscopy imaging of Tomm34^+/+ and Tomm34^−/− NIH/3T3 cells that were mock-infected or infected with SeV for 12 h. Scale bars, 20 μm. e) Co-immunoprecipitation (Co-IP) determining the interaction between TOMM34 and the RIG-I signalling components. HEK293T cells were transfected with HA-Tomm34 and vectors encoding Flag-tagged RIG-I, MAVS, TBK1, IRF3, IKKβ or NEMO for 36 h, and were then subjected to Co-IP using anti-HA beads. Whole cell lysates (WCL) were analyzed by IB with anti-HA and anti-Flag antibodies. f) Fluorescent microscopic analysis of TOMM34 localization in HeLa cells transfected with HA-TOMM34. Nucleus and mitochondria were labeled with DAPI and Mito Tracker Red, respectively. Scale bars, 10 μm. The quantification data was shown in Supplementary Fig. S5a. g) Fluorescent microscopic analysis of TOMM34 and NEMO colocalization in HeLa cells co-transfected with HA-TOMM34 and Flag-NEMO for 24 h, and were then mock-infeted or infected with SeV for 12 h. Nucleus were labeled with DAPI. Scale bars, 10 μm. The quantification data was shown in Supplementary Fig. S5b. h) Immunoprecipitation (IP) examining the endogenous association between TOMM34 and NEMO. HEK293T cells were mock-infected or infected with VSV for 12 h, and were then subjected to IP using IgG or anti-TOMM34 antibodies (αTOMM34). IB analysis was conducted using indicated antibodies. i) Scheme for construction of TOMM34 and its deletion mutants. j) Co-IP determining the region of TOMM34 protein that is responsible for associating with NEMO. HEK293T cells were co-transfected with plasmid encoding Flag-tagged NEMO (Flag-NEMO) and plasmid encoding N-terminal HA- and C-terminal GFP-tagged TOMM34 (HA-TOMM34-GFP) or its mutants as indicated for 36 h, and were then subjected to Co-IP using anti-Flag beads. IB analysis was conducted using indicated antibodies. Asterisk indicates the nonspecific bands. Data are represented as mean ± SD calculated from three independent experiments, or are representative of three independent experiments. P values are reported with two significant digits or shown as “P < 0.0001” (Unpaired two-tailed Student's t test).

Fig. 6

TOMM34 promotes the K63-linked polyubiquitination of NEMO under viral stimulation. a) IP analysis showing the endogenous co-interaction among TOMM34, TRAF6 and NEMO upon viral infection. HEK293T cells were mock-infected or infected with SeV for 12 h, and were then subjected to IP analysis using IgG or anti-NEMO antibodies (αNEMO). IB analysis was conducted using indicated antibodies. b) Fluorescent microscopic analysis of TRAF6 and NEMO colocalization in Tomm34^+/+ and Tomm34^−/− NIH/3T3 cells co-transfected with plasmids encoding EGFP-tagged TRAF6 and mCherry-tagged NEMO for 24 h, and were then mock-infected or infected with SeV for 12 h. Nucleus was stained with DAPI. Scale bars, 10 μm. c) Effects of TOMM34 overexpression on the ubiquitination of NEMO under viral stimulation. HEK293T cells were transfected with indicated vectors for 24 h, followed by mock-infection or infection with VSV for 12 h. Cells were collected and subjected to IP using anti-Flag beads. IB analysis was conducted using indicated antibodies. d) Impacts of TOMM34 deficiency on the ubiquitination of NEMO under viral stimulation. TOMM34^+/+ and TOMM34^−/− HEK293T cells were transfected with HA-Ub for 24 h, and were mock-infected or infected with VSV for 12 h. Ubiquitination of endogenous NEMO were assessed by IP (using anti-NEMO antibodies and protein A/G beads) and IB (using indicated antibodies) analysis. e and f) Effects of TOMM34 deficiency on the K48- and K63-linked ubiquitination of NEMO under viral stimulation. TOMM34^+/+ and TOMM34^−/− HEK293T cells were transfected with plasmids expressing various HA-Ub (K48 only, K63 only, K48R, and K63R as indicated) and Flag-NEMO for 24 h, followed by infection with VSV for 12 h. Cells were subjected to IP (using anti-Flag beads) and IB analysis (using indicated antibodies) for determining the type of NEMO polyubiquitination. Data are representative of three independent experiments. g) A model for TOMM34 facilitates antiviral innate immunity. Upon RNA virus infection, cellular TOMM34 is induced to guide NEMO toward TARF6 in the MAVS signalsome, thus facilitating the K63-linked ubiquitination of NEMO, activation of NF-κB signalling and subsequent production of antiviral cytokines.