TRIM18 is a critical regulator of viral myocarditis and organ inflammation

TRIM18 downregulates type I IFN production in human macrophages in response to both RNA and DNA viruses

Previously, we screened all 70 members of the TRIM family proteins in the mouse alveolar macrophage cell line MH-S by a small-interfering-RNA approach and identified TRIM29 as a crucial negative regulator in antiviral innate immunity [38]. Meanwhile, we found another E3 ligase TRIM18 was a potential negative regulator in antiviral innate immunity. We next investigated the role of TRIM18 in antiviral innate immunity by knocking down TRIM18 via short hairpin RNA (shRNA) in human THP-1 macrophages. The TRIM18-targeting shRNA produced efficient knockdown of TRIM18 at the protein level (Fig. 1a). We then stimulated these cells with 5′-triphosphorylated RNA (5′pppRNA, the ligand of RIG-I-like receptors (RLRs)), dsRNA poly I:C (high molecular weight poly I:C, the ligand of RLRs) and dsDNA from HSV-1 virus (HSV60, the ligand of cytosolic DNA sensors), and measured type I IFN IFN-α and IFN-β by enzyme-linked immunosorbent assay (ELISA). As a positive control, knockdown of the key adaptor MAVS in RLRs signaling pathway via shRNA abrogated the production of IFN-β (Fig. 1b, c) and IFN-α (Additional file 2: Fig. S1a, b) in THP-1 macrophages stimulated by cytosolic 5’pppRNA and poly I:C. As a negative control, the production of IFN-β and IFN-α was not affected in STING-knockdown THP-1 macrophages (Fig. 1b, c) (Additional file 2: Fig. S1a, b), which confirmed a previous report showing that STING plays a critical role in DNA sensing but no role in RNA sensing [53]. In contrast, knockdown of TRIM18 markedly increased production of IFN-β (Fig. 1b, c) and IFN-α (Additional file 2: Fig. S1a, b) by THP-1 macrophages compared to cells treated with control shRNA (sh-Ctrl). We next determined whether TRIM18 controlled DNA sensing pathway in human THP-1 macrophages. As a result, STING knockdown led to significant reduction of IFN-β and IFN-α production in THP1 macrophages in response to dsDNA HSV60 (Fig. 1d and Additional file 2: Fig. S1c), while the knockdown of MAVS in THP1 macrophages had little effect on IFN-β and IFN-α production in response to dsDNA HSV60 (Fig. 1d and Additional file 2: Fig. S1c), which confirms a previous report showing that the RNA-sensing adaptor molecule MAVS is not required for cytokine production in response to cytosolic DNA [54]. However, knockdown of TRIM18 markedly increased production of IFN-β (Fig. 1d) and IFN-α (Additional file 2: Fig. S1c) by THP-1 macrophages compared to cells treated with control shRNA. Furthermore, knockdown of TRIM18 in THP1 macrophages had no effect of IFN-α and IFN-β production in response to LPS (Additional file 2: Fig. S1d, e). These data suggested that TRIM18 negatively regulates production of IFN-α and IFN-β in human THP-1 macrophages in response to cytosolic dsRNA and dsDNA.

Fig. 1figure 1

TRIM18 inhibits type I IFN production by human THP-1 macrophages in response to stimulations with dsRNA and dsDNA or infections with RNA and DNA viruses. a The immunoblot (IB) analysis of TRIM18, MAVS or STING expression in human THP-1 macrophages treated with shRNA to knockdown expression of TRIM18, MAVS or STING. A scrambled shRNA served as a control (sh-Ctrl) and the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as the loading control. The position of protein markers (shown in kDa) is indicated at right. b-h, ELISA of IFN-β production from human THP-1 macrophages treated with the indicated shRNA after a 12 h stimulations with dsRNA including 5’pppRNA (0.5 μg/ml) b and poly I:C (0.5 μg/ml) c or dsDNA from HSV-1 virus (HSV60, 2.5 μg/ml) d delivered by Lipofectamine 3000, or 12 h infections with RNA viruses including influenza A virus (influenza A virus PR8 strain, Flu PR8) e and Coxsackievirus B3 (CVB3) (f), or DNA viruses including HSV-1 g and adenovirus h at an MOI of 2 (n = 3 per group). Each circle represents an individual independent experiment and small solid black lines indicate the average of triplicates for results in (bh). Mock, scrambled shRNA-treated cells without stimulation. NS, not significant (p > 0.05), **p < 0.01, ***p < 0.001, and p value was calculated by unpaired two-tailed Student’s t test. Data are representative of three independent experiments

To further assess the role of TRIM18 in antiviral innate immunity against virus infection, we analyzed protein expression of TRIM18 in normal tissues from GeneCards [55]. TRIM18 had high expression in brain, heart and lung, but was less expressed in most of other tissues (Additional file 2: Fig. S2a). Therefore, we investigated the role of TRIM18 in antiviral immune response in human THP1 macrophages by infection with two RNA viruses coxsackievirus B3 (CVB3, a RNA virus targeting heart for inducing viral myocarditis) and influenza A virus PR8 strain (Flu PR8, a RNA virus targeting lung to induce viral pneumonia), and two DNA viruses including herpes simplex virus type I (HSV-1, a DNA virus targeting brain for inducing viral encephalitis) and human adenovirus (a DNA virus targeting lung for viral pneumonia). As expected, knockdown of the key adaptor MAVS abrogated the production of IFN-β and IFN-α in THP-1 macrophages in response to RNA viruses Flu PR8 and CVB3 (Fig. 1e, f, Additional file 2: Fig. S2b, c), but not to DNA viruses HSV-1 and adenovirus (Fig. 1g, h, Additional file 2: Fig. S2d, e), while STING knockdown led to significant reduction of IFN-β and IFN-α production in THP1 macrophages after infection with DNA viruses HSV-1 and adenovirus (Fig. 1g, h, Additional file 2: Fig. S2d, e), but not with RNA viruses Flu PR8 and CVB3 (Fig. 1e, f, Additional file 2: Fig. S2b, c). However, knockdown of TRIM18 significantly increased production of IFN-β (Fig. 1e–h) and IFN-α (Additional file 2: Fig. S2b–e) by THP-1 macrophages compared to cells treated with control shRNA after infection with RNA viruses Flu PR8 and CVB3 and DNA virus HSV-1 (Fig. 1e–h, Additional file 2: Fig. S2b–e). Taken together, these data indicate that TRIM18 negatively regulates type I IFN production in human THP1 macrophages in response to stimulation with dsRNA and dsDNA or infection with both RNA and DNA viruses.

TRIM18 negatively regulates innate immune response in mouse macrophages in response to RNA and DNA viruses

To further determine the role of TRIM18 in antiviral innate immunity in mice, we generated TRIM18 knockout (KO) mice. The Trim18 gene deletion was confirmed by PCR assisted genotyping analysis (Additional file 2: Fig. S3a). We first analyzed TRIM18 gene expression in different mouse immune cells using the Immunological Genome Project (ImmGen) and found TRIM18 was indeed highly expressed in mouse macrophages including peritoneal macrophages, splenic macrophages, alveolar macrophages and microglia macrophages (Additional file 2: Fig. S3b). Next, we isolated mouse peritoneal macrophages (MF PC) and splenic macrophages (MF Sp) and detected high expression of TRIM18 in macrophages from wild-type (WT) mice, while deletion of TRIM18 expression was confirmed by immunoblot analysis (Additional file 2: Fig. S3c). We also investigated if TRIM18 expression was affected by RNA virus or DNA virus infection in mouse bone marrow-derived macrophages (BMDM). We found TRIM18 was induced at both RNA (Additional file 2: Fig. S3d) and protein (Additional file 2: Fig. S3e) levels in mouse BMDM after RNA or DNA virus infection and the induction of TRIM18 was much stronger in mouse BMDM by DNA viruses HSV-1 and adenovirus than that by RNA viruses Flu PR8 and CVB3 (Additional file 2: Fig. S3d, e). Furthermore, TRIM18 had high expression in lung, brain and heart, and low expression in intestine, liver and kidney from WT mice after DNA virus HSV-1 infection (Additional file 2: Fig. S3f). Additionally, KO of TRIM18 did not change expression of surface markers CD11b and F4/80 by flow cytometry (Additional file 2: Fig. S4), indicating that TRIM18 does not affect differentiation markers of mouse macrophages.

To further investigate the role of TRIM18 in response to RNA viruses, we prepared mouse BMDM from WT and TRIM18 KO mice, and stimulated BMDM with dsRNA poly I:C and 5′-triphosphate RNA (5’pppRNA), and measured type I IFN proteins (IFN-α and IFN-β) by ELISA as well as mRNA levels of interferon stimulated gene 15 (ISG15) and ISG56 by qRT-PCR. The results showed that TRIM18 KO BMDM produced much more IFN-α and IFN-β proteins (Fig. 2a, b) and mRNAs of ISG15 and ISG56 (Additional file 2: Fig. S5a, b) than WT BMDM in response to 5’pppRNA and poly I:C stimulation. In addition, we employed two RNA viruses including Flu PR8 and CVB3 to investigate TRIM18 in response to RNA viruses in mouse BMDM. BMDM from WT and TRIM18 KO mice were isolated and infected with RNA viruses Flu PR8 and CVB3. Compared with WT BMDM, TRIM18 KO BMDM produced 2- to threefold more IFN-α and IFN-β proteins (Fig. 2c, d) and twofold more mRNAs of ISG15 and ISG56 (Additional file 2: Fig. S5c, d) post-infection by RNA viruses Flu PR8 and CVB3. Collectively, these data demonstrate a negative role for TRIM18 in regulating production of type I IFN and ISGs in mouse macrophages in response to dsRNA and RNA viruses.

Fig. 2figure 2

TRIM18 negatively regulates type I IFN production by BMDM upon stimulations of dsRNA and dsDNA or infections with RNA and DNA viruses. ad ELISA of IFN-α a, c and IFN-β b, d production by BMDM from Trim18+/+ (WT) and Trim18−/− (KO) mice after 12 h of stimulations with 5’pppRNA (0.5 μg/ml), poly I:C (0.5 μg/ml) delivered by Lipofectamine 3000 a, b or infections with RNA viruses including influenza A virus (influenza A virus PR8 strain, Flu PR8) and Coxsackievirus B3 (CVB3) c, d at an MOI of 2 (n = 3 per group). eh, ELISA of IFN-α e, g) and IFN-β f, h production by BMDM from Trim18+/+ (WT) and Trim18−/− (KO) mice after 12 h of stimulations with dsDNA from HSV-1 virus (HSV60, 2.5 μg/ml) and cGAMP (2.5 μg/ml) delivered by Lipofectamine 3000 e, f or infections with DNA viruses including adenovirus and HSV-1 g, h at an MOI of 2 (n = 3 per group). Each circle represents an individual independent experiment and small solid black lines indicate the average of triplicates for results. **p < 0.01, ***p < 0.001, and p value was calculated by unpaired two-tailed Student’s t test. Mock, wild-type BMDM without stimulation or infection. Data are representative of three independent experiments

To further determine the role of TRIM18 in response to DNA viruses, we prepared mouse BMDM from WT and TRIM18 KO mice, and stimulated BMDM with dsDNA HSV60 and cGAMP (STING stimulator in DNA sensing pathway), and measured type I IFN proteins by ELISA and mRNA levels of ISG15 and ISG56 by RT-qPCR. Consistent with the earlier results, TRIM18 KO BMDM produced much more IFN-α and IFN-β proteins (Fig. 2e, f) and mRNAs of ISG15 and ISG56 (Additional file 2: Fig. S5e, f) than WT BMDM in response to HSV60 and cGAMP stimulation. Furthermore, two DNA viruses including HSV-1 and human adenovirus were chosen to investigate role of TRIM18 in response to DNA viruses in mouse BMDM. BMDM from WT and TRIM18 KO mice were prepared and infected with DNA viruses adenovirus and HSV-1. The results showed TRIM18 KO BMDM produced significantly more IFN-α and IFN-β proteins (Fig. 2g, h) and mRNAs of ISG15 and ISG56 (Additional file 2: Fig. S5g, h) than WT BMDM after infection with DNA viruses. Taken together, these data suggest that TRIM18 is a negative regulator of type I IFN and ISGs productions in mouse macrophages in response to dsDNA and DNA viruses.

Deletion of TRIM18 protects mice from viral myocarditis

Viral myocarditis has been recognized as a cause of congestive heart failure and CVB3 infection is the main cause of viral myocarditis [20, 21]. Since TRIM18 is highly expressed in heart and TRIM18 negatively regulates innate immune response to RNA virus CVB3 in macrophages, we investigated the functional importance of TRIM18 in controlling CVB3 induced myocarditis in vivo. We first intraperitoneally infected both WT and TRIM18 KO mice with the RNA virus CVB3 and checked cardiac histology and functions. The heart histopathology revealed TRIM18 KO mice had significantly reduced cardiac inflammation and infiltration of inflammatory cells compared with WT mice following CVB3 infection (Fig. 3a, b). Additionally, TRIM18 expression was induced in hearts of WT mice with CVB3 infection (Fig. 3c), while TRIM18 induction was stronger in hearts of mice during CVB3 acute infection than that during CVB3 chronic infection (Fig. 3c). It’s reported that another TRIM family member TRIM21 could restrict CVB3 induced cardiac injury by positively regulate IRF3-mediated type I IFN production [68]. We then compared the expressions of TRIM18 and TRIM21 in hearts from WT and TRIM18 KO mice after CVB3 infection. The immunohistochemistry (IHC) data showed that there was high expression of TRIM21 in hearts from WT and TRIM18 KO mice after CVB3 infection (Fig. 3d). However, TRIM18 had higher expression than TRIM21 in heart from CVB3 infected WT mice, while TRIM18 expression was gone in heart from TRIM18 KO mice after CVB3 infection (Fig. 3d). Furthermore, IHC data showed that there were major macrophages and neutrophils, and minor NK cells and T cells in the infiltrated cells in hearts from mice after CVB3 infection for two days (Fig. 3e). In agreement, the echocardiography of WT mice revealed impaired cardiac function (Fig. 3f) as evidenced by decreased ejection fraction (EF) (Fig. 3g) and fractional shortening (FS) (Fig. 3h) when compared with TRIM18 KO mice. Compared to WT mice, TRIM18 KO mice had less heart weight increase during viral myocarditis (Fig. 3i), a marker of cardiac inflammatory edema. Additionally, brain natriuretic peptide (BNP), a marker of heart failure, was dramatically reduced in TRIM18 KO hearts compared with WT hearts (Fig. 3j). Importantly, we found that most of WT mice succumbed to CVB3 infection, while the survival of TRIM18 KO mice was significantly better than that of their WT littermates (Fig. 3k). These data suggested that deficiency of TRIM18 could protect mice from CVB3 induced myocarditis by reducing cardiac inflammation with improved function. To further investigate the mechanisms by which TRIM18 knockout mice reduced CVB3 induced myocarditis, we next checked viral replication and type I IFN protein levels in heart homogenates by plaque-forming assay and ELISA, respectively. We found that the CVB3 viral loads were significantly reduced in hearts from TRIM18 KO mice compared with WT mice at day 2 (D2), day 5 (D5) and day 14 (D14) after CVB3 infection (Fig. 3l). Furthermore, TRIM18 KO mice had higher concentrations of IFN-α (Fig. 3m) and IFN-β (Fig. 3n) in the hearts than did their WT littermates after infection with CVB3. However, WT mice had higher cardiac inflammatory cytokines IL-6 (Fig. 3o), TNF-α (Fig. 3p) and IL-1β (Fig. 3q) than did their TRIM18 KO littermates after CVB3 infection. These data indicate that deletion of TRIM18 protects mice from CVB3 induced myocarditis by improving cardiac function and promoting innate immune activation.

Fig. 3figure 3

Deletion of TRIM18 protects mice from myocarditis induced by RNA virus CVB3 in vivo. a Hematoxylin and eosin (H&E)-staining of heart sections from age-matched Trim18+/+ (WT) and Trim18−/− (KO) male mice after intraperitoneal infection with CVB3 (1 × 107 PFU per mouse) for 4 days. Scale bars represent 1000 μm for original images and 200 µm for enlarged images. b Histology score analysis of viral myocarditis in heart sections from mice as in (a). c Immunoblot (IB) analysis of TRIM18 expression in hearts from WT mice without or with intraperitoneal CVB3 acute infection (1 × 107 PFU per mouse) or chronic infection (1 × 103 PFU per mouse) for 2 days. The position of protein markers (shown in kDa) is indicated at right. d Immunohistochemistry (IHC) analysis of TRIM18 and TRIM21 expression in hearts from WT and KO male mice after CVB3 infection. e IHC analysis of the infiltrated cells in CVB3 infected hearts using anti-macrophage marker antibody, anti-neutrophil marker antibody, anti-NK cell marker antibody and anti-T cell marker antibody, respectively. Scale bars represent 100 μm in (d, e). f Representative M-mode images of hearts from WT and KO male mice at day 4 after CVB3 infection by echocardiography analysis. gh Cardiac function analysis of ejection fraction (EF) g and fractional shortening (FS) h of hearts from mice as in (f) (n = 5 per group). i The assessment of heart weight/baseline body weight from WT and KO male mice (n = 5 per group) at day 0 or day 6 after CVB3 infection. j The qRT-PCR analysis of brain natriuretic peptide (BNP) mRNA in the heart of from WT and KO male mice (n = 5 per group) at day 1, day 2 or day 4 after CVB3 infection.; results are presented relative to those of mock mice. k Survival of age-matched WT and KO male mice after intraperitoneal infection with CVB3 (1 × 107 PFU per mouse) (n = 10 per group). l Viral titers in homogenates of hearts from WT and KO male mice at day 2 (D2), day 5 (D5) and day 14 (D14) after CVB3 infection (n = 5 per group for D2 and D5, n = 3 per group for D14). mq, ELISA of IFN-α (m), IFN-β (n) IL-6 (o), TNF-α (p) and IL-1β q in hearts from mice as in k (n = 5 per group). Error bars indicate standard error of the mean for results in (b, gj, lq). NS, not significant (p > 0.05), **p < 0.01, ***p < 0.001, and p value was calculated by unpaired two-tailed Student’s t test and Gehan-Breslow-Wilcoxon test for survival analysis. Data are representative of three independent experiments

Knockout of TRIM18 protects mice from pneumonia and lung injury induced by viral infections

Viral pneumonia is an inflammation of the lungs caused by respiratory viruses, such as influenza virus, adenovirus and SARS-CoV-2 causing the ongoing pandemic of COVID-19 [23, 24]. Interestingly, the public GEO profile database show that patients with SARS-CoV-2 infection have higher expression of TRIM18 (Additional file 2: Fig. S6), we hypothesize that TRIM18 may play crucial roles in controlling pneumonia and lung injury induced by respiratory viruses including SARS-CoV-2. As shown above, TRIM18 was highly expressed in lung and TRIM18 downregulated the innate immune responses to respiratory viruses including RNA virus influenza virus and DNA virus adenovirus in both human and mouse macrophages. Therefore, we investigated if TRIM18 could control viral pneumonia induced by those respiratory viruses including influenza virus and adenovirus in vivo. First, we infected both WT and TRIM18 KO mice intranasally with respiratory RNA virus influenza virus and checked lung inflammation and injury by histology. Indeed, lung histopathology revealed edema, alveolar hemorrhaging, alveolar wall thickness and neutrophil infiltration in lungs from TRIM18 KO mice were less marked than those from WT after influenza virus infection (Fig. 4a, b). Importantly, we found that TRIM18 KO mice were more resistant to influenza virus infection than their WT littermates (Fig. 4c). These results suggested that knockout of TRIM18 protected mice from lung injury and inflammation induced by RNA virus influenza virus in vivo. We further investigated the mechanisms by which deletion of TRIM18 protected mice from pneumonia infected by influenza virus. We determined viral amplification in lungs at day 2 post-infection by plaque-forming assay. We detected significantly less influenza virus loads in TRIM18 KO mice than in their WT littermates (Fig. 4d). We then detected type I IFN production in lungs by ELISA. Compared with WT mice, TRIM18 KO mice produced significantly more IFN-α (Fig. 4e) and IFN-β (Fig. 4f) following influenza virus infection. Additionally, there were increased infiltration cells mainly containing macrophages, neutrophils, and lymphocytes in bronchoalveolar lavage fluid (BALF) of WT mice with Flu PR8 infection, which were dramatically reduced in TRIM18 KO mice (Fig. 4g). These data indicate that TRIM18 deficiency protects mice from pneumonia induced by RNA virus influenza virus through restricting viral replication and promoting innate immune activation in vivo.

Fig. 4figure 4

Knockout of TRIM18 protects mice from pneumonia and lung injury induced by RNA virus Flu PR8 or DNA virus adenovirus in vivo. a Hematoxylin and eosin (H&E)-staining of lung sections from age- and sex-matched Trim18+/+ (WT) and Trim18−/− (KO) mice left infected (Mock) or infected for 4 days by intranasal infection of Flu PR8 virus (1 × 105 PFU per mouse). b Histology score analysis of viral pneumonia in lung sections from mice as in (a). c Survival of age- and sex-matched WT and KO mice after intranasal infection with Flu PR8 virus (1 × 105 PFU per mouse) (n = 10 per group). d Plaque assay of Flu PR8 virus titers in the lung of WT and KO mice infected for 2 days by intranasal infection of Flu PR8 virus (n = 5 per group). ef, ELISA of and IFN-α e and IFN-β f in BALF samples from mice (n = 5 per group) as in (d). g Quantification of cell numbers in BALF samples from mice (n = 3 per group) as in (d). h Hematoxylin and eosin (H&E)-staining of lung sections from WT and KO mice left infected (Mock) or infected for 4 days by intranasal infection of adenovirus (1 × 108 PFU per mouse). i, Histology score analysis of viral pneumonia in lung sections from mice as in (h). j Viral titers in homogenates of lung from WT and KO mice (n = 5 per group) at day 2 of intranasal infection with adenovirus (1 × 108 PFU per mouse). k, l ELISA of IFN-α k and IFN-β l in BALF samples from mice as in j (n = 5 per group). m Quantification of cell numbers in BALF samples from mice (n = 3 per group) as in (j). Scale bars represent 400 μm for images in a and (h). Error bars indicate standard error of the mean for results in (b, d–g, i–m). NS, not significant (p > 0.05), **P < 0.01 and ***P < 0.001, and p value was calculated by unpaired two-tailed Student’s t test and Gehan-Breslow-Wilcoxon test for survival analysis. Data are representative of three independent experiments

Next, we evaluated the importance of TRIM18 in controlling pneumonia following infection by respiratory DNA virus adenovirus in vivo. Both WT and TRIM18 KO mice were infected intranasally with adenovirus, which is normally transmitted by the nasal route and targets the lungs for pneumonia. Lung histopathology revealed much-reduced edema, alveolar hemorrhage, alveolar wall thickness, and neutrophil infiltrations in TRIM18 KO mice compared to the lung pathology in WT mice (Fig. 4h, i), indicating the importance of TRIM18 in promoting adenovirus induced pneumonia and lung inflammation. To further investigate the underlying mechanisms, we measured adenovirus replication and type I IFN production in lung by plaque-forming assay and ELISA, respectively. We found that the adenovirus loads were dramatically reduced in TRIM18 KO mice compared with WT mice (Fig. 4j). Additionally, there was significantly more IFN-α (Fig. 4k) and IFN-β (Fig. 4l) in the BALF from TRIM18 KO mice than that from WT mice at day 2 post infection. Compared with TRIM18 KO mice, there were much more infiltration cells mainly consisted of macrophages, neutrophils, and lymphocytes in BALF of WT mice with adenovirus infection (Fig. 4m). Collectively, these findings demonstrate that knockout of TRIM18 protects mice from pneumonia and lung injury induced by viral infections through enhancing activation of innate immunity in vivo.

Deficiency of TRIM18 protects mice from encephalitis induced by HSV-1 infection

Herpes simplex encephalitis (HSE) is caused by HSV-1 infection of the brain and is the most common cause of sporadic fatal encephalitis worldwide[22]. Given that the high expression of TRIM18 in brain and the critical role of TRIM18 in regulating innate immune response to HSV-1 in macrophages, we further investigated if TRIM18 plays role in controlling brain damage and inflammation induced by DNA virus HSV-1 in vivo. We challenged WT and TRIM18 KO mice intravenously with HSV-1 virus and checked brain damage and inflammation by histology. The brain histopathology revealed much-reduced demyelination, necrosis, and inflammatory cell infiltration in TRIM18 KO mice as compared to the brain pathology in WT mice after HSV-1 infection (Fig. 5a, b). Compared with WT mice, TRIM18 KO mice had significantly higher survival rates (Fig. 5c). These results indicated that knockout of TRIM18 protected mice from brain damage and inflammation induced by HSV-1 virus in vivo. To further investigate the underlying mechanisms, we measured HSV-1 replication in brain and type I IFN production in serum by plaque-forming assay and ELISA, respectively. We detected significantly less HSV-1 virus in brain of TRIM18 KO mice than in WT mice on day 2 after infection (Fig. 5d). Moreover, TRIM18 KO mice had higher concentrations of IFN-α (Fig. 5e) and IFN-β (Fig. 5f) in the serum than did their WT littermates after HSV-1 infection. Collectively, these data demonstrate deficiency of TRIM18 protects mice from encephalitis induced by HSV-1 by enhancing innate immune activation.

Fig. 5figure 5

Deficiency of TRIM18 protects mice from brain damage induced by DNA virus HSV-1 in vivo. a Hematoxylin and eosin (H&E)-staining of brain sections from Trim18+/+ (WT) and Trim18−/− (KO) mice left infected (Mock) or infected for 4 days by intravenous infection of HSV-1 (1 × 107 PFU per mouse). Scale bars represent 1000 μm for original images and 200 µm for enlarged images. b Histology score analysis of viral encephalitis in brain sections from mice as in (a). c Survival of WT and KO mice (n = 10 per group) after intravenous injection of HSV-1 (1 × 107 PFU per mouse). d Viral titers in homogenates of brains from WT and KO mice (n = 5 per group) after intravenous injection of HSV-1 (1 × 107 PFU per mouse). e, f, ELISA of IFN-α e and IFN-β f in serum obtained from WT and KO mice (n = 5 per group) at 12 h after intravenous injection of HSV-1. Error bars indicate standard error of the mean for results in (b, df). NS, not significant (p > 0.05), **P < 0.01 and ***P < 0.001, and p value was calculated by unpaired two-tailed Student’s t test and Gehan-Breslow-Wilcoxon test for survival analysis. Data are representative of three independent experiments.

TRIM18 recruits PPM1A to inactivate TBK1 blocking TBK1 from interactions with its upstream adaptors during virus infection

To determine the molecular mechanisms by which the enhanced production of type I IFN and innate immune activation were achieved in BMDM from TRIM18 KO mice, we isolated BMDM from WT and TRIM18 KO mice and infected the cells without or with CVB3 and adenovirus for 1 h, then assessed activation of the transcription factors IRF3 by immunoblot analysis. We found that phosphorylation of IRF3 in TRIM18 KO BMDM was enhanced relative to WT BMDM after infection with CVB3 or adenovirus (Fig. 6a). Typically, cytosolic RNA or DNA is sensed by RIG-I-like receptors or DNA sensor such as cGAS, which then activate downstream MAVS and STING, respectively. MAVS or STING recruits downstream TBK1 to phosphorylate and activate IRF3 for inducing type I IFN. To further dissect the role of TRIM18 in these different IFN-I induction pathways, we examined the effect of TRIM18 on the IFN-β luciferase reporter activated by these components including MDA5, MAVS, TBK1, IKKi and cGAS/STING. Overexpression of TRIM18 reduced the IFN-β luciferase reporter activation by MDA5 and MAVS and to a greater extent by TBK1 and cGAS/STING (Additional file 2: Fig. S7a–d). However, TRIM18 did not inhibit downstream IKKi dependent IFN-β luciferase reporter activation (Additional file 2: Fig. S7e), indicating that TRIM18 targets the pathway at nodes between TBK1 and IKKi. To further investigate how TRIM18 regulates such signaling molecules, we used immunoprecipitation with an antibody specific to TRIM18 to identify TRIM18-interacting proteins in lysates of BMDM, followed by protein sequencing by liquid chromatography–mass spectrometry. Among a group of TRIM18-interacting proteins, we identified protein phosphatase, magnesium-dependent 1A (PPM1A; formerly called PP2C) (Additional file 1: Table S1). PPM1A has previously been reported to silence cytosolic RNA sensing and antiviral defense through direct dephosphorylation of TBK1 [56]. These collective data suggested that TRIM18 might recruit PPM1A to target and dephosphorylate TBK1 for dampening type I IFN production.

Fig. 6figure 6

TRIM18 recruits PPM1A to dephosphorylate TBK1 for its inactivation and black the interactions of TBK1 with its upstream adaptors. a Immunoblot (IB) analysis of total and phosphorylated (p-) IRF3 as well as β-actin in lysates of WT and TRIM18 KO BMDM infected for 1 h without (Mock) or with CVB3 or adenovirus at an MOI of 5. b Immunoblot analysis of endogenous proteins TRIM18 and PPM1A precipitated with anti-PPM1A, or immunoglobulin G (IgG) from whole-cell lysates of WT and TRIM18 KO BMDM. c Schematic diagram showing full-length PPM1A (Full) and serial truncations of PPM1A with deletion of various domain (left margin); numbers at ends indicate amino acid positions (top). M1, the catalytic domain; M2, the C-terminal no catalytic domain. d Immunoblot analysis of purified HA-tagged full-length PPM1A and serial truncations of PPM1A with deletion of various domains alone with anti-HA antibody (top blot) or after incubation with Myc-tagged TRIM18 and immunoprecipitation with anti-Myc antibody (second blot), and immunoblot analysis of purified Myc-tagged TRIM18 with anti-Myc antibody (third blot) or after incubation with Myc-tagged TRIM18 and immunoprecipitation with anti-Myc antibody (bottom blot). e Schematic diagram showing full-length TRIM18 (Full) and serial truncations of TRIM18 with deletion (Δ) of various domain (left margin); numbers at ends indicate amino acid positions (top). RING, the really interesting new gene domain; BBOX, the B-box zinc-finger domain; BBC, the B-box C-terminal domain; FN3, the fibronectin type 3 domain; SPRY, the Sp1A kinase and Ryanodine receptors domain. f Immunoblot analysis of purified HA-tagged full-length TRIM18 and serial truncations of TRIM18 with deletion of various domains alone with anti-HA antibody (top blot) or after incubation with Myc-tagged PPM1A and immunoprecipitation with anti-Myc antibody (second blot), and immunoblot analysis of purified Myc-tagged PPM1A with anti-Myc antibody (third blot) or after incubation with Myc-tagged PPM1A and immunoprecipitation with anti-Myc antibody (bottom blot). g Immunoblot analysis of total and phosphorylated (p-) TBK1, PPM1A, TRIM18 as well as β-actin in lysates of WT and TRIM18 KO BMDM infected for 1 h without (Mock) or with CVB3 or adenovirus at an MOI of 5. h Immunoblot analysis of endogenous proteins STING, MAVS, TBK1 and TRIM18 from whole-cell lysates (Input) or precipitated with anti-TBK1 antibody (IP: TBK1) from lysates of WT and TRIM18 KO BMDM left infected (Mock) or infected with CVB3 or adenovirus at MOI of 10 for 6 h. The position of protein markers (shown in kDa) is indicated on the right. Data are representative of three independent experiments

Next, we investigated if TRIM18 could interact with PPM1A in BMDM at the endogenous protein level. The anti-PPM1A antibody, but not control IgG, precipitated TRIM18 in WT BMDM, but not in TRIM18 KO BMDM (Fig. 6b), indicating real interaction between TRIM18 and PPM1A in resting BMDM. To further map the binding sites between TRIM18 and PPM1A, we analyzed interactions among Myc-tagged recombinant TRIM18 and HA-tagged recombinant full-length PPM1A, as well as truncation mutants of PPM1A (Fig. 6c). Both full-length PPM1A and the M1 catalytic domain of PPM1A bound to TRIM18 (Fig. 6d). Additionally, the mapping results for Myc-tagged recombinant PPM1A and HA-tagged full-length TRIM18 and their truncation mutants (Fig. 6e) showed that the C-terminal SPRY (Sp1A kinase and Ryanodine receptors) domain of TRIM18 bound to PPM1A (Fig. 6f). To further investigate whether TRIM18 recruits PPM1a to target and dephosphorylate TBK1, we isolated BMDM from WT and TRIM18 KO mice and infected the cells without or with CVB3 and adenovirus for 1 h, then assessed phosphorylation of TBK1 by immunoblot analysis. We found that phosphorylation of TBK1 in TRIM18 KO BMDM was significantly enhanced relative to that in WT BMDM after infection with CVB3 or adenovirus (Fig. 6g), suggesting that TRIM18 indeed recruited PPM1A to dephosphorylate TBK1 for TBK1 inactivation. Additionally, we found there was more PPM1A expression in WT BMDM than TRIM18 KO BMDM without and with CVB3 or adenovirus infection (Fig. 6b, g). To further investigate the outcomes of the recruitment and stability of PPM1A by TRIM18, we isolated BMDM from WT and TRIM18 KO mice and infected the cells without or with CVB3 and adenovirus for 5 h, then evaluated the interaction between TBK1 and its upstream adaptors MAVS or STING by co-immunoprecipitation and immunoblot analysis. We found that there was no interaction between TBK1 and MAVS or STING in BMDM from either WT or TRIM18 KO mice that had not been infected (Fig. 6h). However, after infection with CVB3 or adenovirus, interaction between TBK1 and MAVS or STING in both WT and TRIM18 KO BMDM was evident, but significantly enhanced in the TRIM18 KO BMDM relative to levels in WT BMDM (Fig. 6h), indicating that the recruitment of PPM1A by TRIM18 blocked interaction between TBK1 and its upstream adaptors MAVS or STING in macrophages after virus infection. Collectively, these data suggest that TRIM18 recruits PPM1A to dephosphorylate TBK1 for its inactivation and blocks interaction of TBK1 with its upstream adaptors MAVS and STING for signal transduction during virus infection.

TRIM18 stabilizes PPM1A by mediating K63-linked ubiquitination

Because TRIM18 is an E3 ubiquitin ligase and less PPM1A is seen in TRIM18 KO cells, we surmised that TRIM18 may mediate the stability of PPM1A. Consequently, we next determined whether TRIM18 was responsible for the ubiquitination of PPM1A ex vivo. We isolated BMDM from WT and TRIM18 KO mice and infected those cells with CVB3 or adenovirus for 6 h. Cell lysates were then prepared and analyzed for the ubiquitination of PPM1A. As results, PPM1A was modified via K63-mediated ubiquitination in BMDM from WT mice but not TRIM18 KO mice (Fig. 7a). Additionally, there was more PPM1A expression in WT BMDM that TRIM18 KO BMDM (Fig. 7a), which demonstrated the crucial role of K63-mediated ubiquitination in mediating stability of PPM1A. To investigate whether the ubiquitination of PPM1A was dependent on the binding site of TRIM18 with PPM1A, we transfected the HEK293T cells to co-express Myc-tagged PPM1A and HA-tagged full-length TRIM18, or truncated TRIM18 lacking the binding site of PPM1A (T18-∆SPRY). We then prepared cell lysates and incubated them for 5 min at 90 °C with 1% SDS (sodium dodecyl sulfate) to disrupt protein–protein interactions, followed by immunoprecipitation of Myc-tagged PPM1A. Immunoblot analysis of HA or ubiquitin demonstrated that the ubiquitination of PPM1A was strongly e

留言 (0)

沒有登入
gif