Disulfiram ameliorates STING/MITA-dependent inflammation and autoimmunity by targeting RNF115

Abstract

STING (also known as MITA) is an adaptor protein that mediates cytoplasmic DNA-triggered signaling, and aberrant activation of STING/MITA by cytosolic self-DNA or gain-of-function mutations causes severe inflammation. Here, we show that STING-mediated inflammation and autoimmunity are promoted by RNF115 and alleviated by the RNF115 inhibitor disulfiram (DSF). Knockout of RNF115 or treatment with DSF significantly inhibit systemic inflammation and autoimmune lethality and restore immune cell development in Trex1^–/– mice and STING^N153S/WT bone marrow chimeric mice. In addition, knockdown or pharmacological inhibition of RNF115 substantially downregulate the expression of IFN-α, IFN-γ and proinflammatory cytokines in PBMCs from patients with systemic lupus erythematosus (SLE) who exhibit high concentrations of dsDNA in peripheral blood. Mechanistically, knockout or inhibition of RNF115 impair the oligomerization and Golgi localization of STING in various types of cells transfected with cGAMP and in organs and cells from Trex1^–/– mice. Interestingly, knockout of RNF115 inhibits the activation and Golgi localization of STING^N153S as well as the expression of proinflammatory cytokines in myeloid cells but not in endothelial cells or fibroblasts. Taken together, these findings highlight the RNF115-mediated cell type-specific regulation of STING and STING^N153S and provide potential targeted intervention strategies for STING-related autoimmune diseases.

Introduction

The innate immune system employs pattern recognition receptors (PRRs) to recognize pathogen-associated molecular patterns (PAMPs) or damage-associated molecular patterns (DAMPs) and activate a series of signaling cascades for defense against invading pathogens and induction of inflammation [1–3]. Cyclic guanosine monophosphate-adenosine monophosphate (cGAMP) synthase (cGAS) is a major cytosolic DNA sensor that serves as a PRR to detect cytoplasmic dsDNA and catalyze the synthesis of cGAMP [4, 5]. cGAMP is a second messenger that binds to the adaptor protein Stimulator of interferon genes (STING; also known as MITA and ERIS) and induces the dimerization and oligomerization of STING [6–9]. Subsequently, STING is translocated from the ER to the Golgi apparatus, where it binds to sulfated glycosaminoglycans for further polymerization and recruits TBK1 and IRF3 to activate downstream signaling and induce the production of type I interferons (IFNs) and proinflammatory cytokines [10, 11]. Therefore, the cGAS-cGAMP-STING axis plays essential roles in cytosolic DNA-induced signaling and inflammation.

Host DNA is sequestered in the nucleus or mitochondria under homeostatic conditions. The leakage of nuclear DNA or mitochondrial DNA (mtDNA) into the cytosol and its cytosolic accumulation under stress conditions or as an effect of mutations in genes such as DNase II and TREX1 activates the cGAS-cGAMP-STING axis to induce severe inflammation [1, 2, 12–14]. For example, TREX1 is a DNA exonuclease that degrades self-DNA, and loss-of-function mutations in TREX1 are strongly correlated with human autoimmune diseases such as Aicardi-Goutieres syndrome (AGS), retinal vasculopathy with cerebral leukodystrophy (RVCL), and systemic lupus erythematosus (SLE) [15, 16]. Loss-of-function mutation or knockout of TREX1 results in the accumulation of self-DNA in the cytosol and continuous activation of the cGAS-cGAMP-STING pathway, thereby triggering autoimmune responses [17–19]. Accordingly, Trex1^–/– and TREX1^D18N mice develop severe systemic inflammation and exhibit lethal autoimmunity that can be genetically rescued by deletion of STING [20–22], suggesting an imperative role of STING in autoimmune diseases caused by self-DNA.

In addition to cytosolic DNA, gain-of-function mutations in STING (among which STING^N154S and STING^V155M are the most common) mediate the activation of downstream signaling pathways and result in an autoimmune disease termed STING-associated vasculopathy with onset in infancy (SAVI) [23, 24]. SAVI patients exhibit systemic inflammation, interstitial lung disease, and robust IFN production and ISG signature expression in peripheral blood mononuclear cells (PBMCs) [24]. In mouse models, heterozygous knock-in mice harboring the STING^N153S or STING^V154M mutation (orthologous to human STING^N154S or STING^V155M, respectively) exhibit myeloid cell expansion, T-cell cytopenia, pulmonary inflammation, and premature death [25–28]. Moreover, activation of the functional NF-κB pathway and inflammatory gene signatures, not activation of the IRF3 pathway, plays a dominant role in STING^N153S mice, as deletion of IRF3, IRF7, or IFNR1 in STING^N153S mice cannot rescue the lethal autoimmune phenotypes [28, 29]. Mechanistically, gain-of-function STING mutants are spontaneously and constitutively translocated from the ER to the Golgi apparatus in the absence of cGAMP, thereby triggering the induction of inflammatory cytokines [30]. However, whether gain-of-function mutants of STING require posttranslational modifications for their spontaneous ER-to-Golgi translocation and the related underlying mechanism are unclear. Therefore, deciphering the regulatory mechanisms involved in STING activation and targeting activated STING will provide insight for understanding autoimmunity and developing potential therapeutic strategies for autoimmune diseases.

SLE is a prototypical multisystem autoimmune disease characterized by autoantibody production and multiorgan damage [15, 31, 32]. Emerging evidence suggests that the cGAS-cGAMP-STING axis is extensively engaged in SLE progression [33, 34]. It has been shown that 1–2% of SLE patients carry TREX1 mutations that result in activation of the cGAS-cGAMP-STING axis [35, 36]. In addition, approximately 15% of patients with SLE exhibit detectable levels of cGAMP in peripheral blood cells, and these patients have higher SLE disease activity index (SLEDAI) scores than do patients without detectable cGAMP [37]; in addition, mtDNA derived from mitochondria-rich red blood cells induces the production of type I IFNs via the cGAS-cGAMP-STING axis in lupus monocytes [38]. Moreover, a high level of dsDNA is detected in the serum of a subset of SLE patients, and this dsDNA leads to the activation of STING-mediated signaling in PBMCs of these SLE patients [39]. In this context, it is critical to screen SLE patients for hyperactivation of STING, and downregulation of hyperactivated STING-mediated signaling might be beneficial for attenuating the symptoms of these patients.

RNF115 is involved in tumorigenesis and the migration of malignant cells in breast cancer, lung adenocarcinoma, and gastric cancer [40, 41]. We recently reported that RNF115 negatively mediates the post-ER and post-Golgi apparatus trafficking of TLRs by catalyzing the ubiquitination of RAB1A and RAB13 and inhibiting their interaction with guanosine diphosphate (GDP) dissociation inhibitor (GDI) for their reactivation [42]. We also demonstrated that RNF115 constitutively destabilizes MAVS by catalyzing its ubiquitination [43]. In contrast, RNF115 mediates the K63-linked ubiquitination of STING to promote its oligomerization, transport, and activation after HSV-1 infection. However, whether RNF115 regulates the activity of STING in the context of spontaneous inflammation, as well as the underlying mechanism, remains to be investigated.

In this study, we report that RNF115 is an E3 ligase that interacts with STING in Trex1^–/– cells and tissues. RNF115 promotes STING-mediated inflammation and autoimmunity by catalyzing the K63-linked ubiquitination of STING or STING^N153S, which is attenuated by the RNF115 inhibitor disulfiram (DSF). Deletion or pharmacological inhibition of RNF115 attenuates the autoimmune phenotypes of Trex1^–/– mice and STING^N153S/WT bone marrow chimeric mice. Moreover, knockdown of RNF115 or treatment with DSF substantially downregulates the expression of IFN-α, IFN-γ and proinflammatory cytokines in PBMCs from SLE patients with high peripheral blood concentrations of dsDNA. Mechanistically, knockout or inhibition of RNF115 impairs the K63-linked ubiquitination and oligomerization of STING in cells and organs from TREX1-deficient mice and suppresses the Golgi localization of STING^N153S in myeloid cells. Our findings highlight the process through which STING is regulated by RNF115 as a therapeutic intervention target for autoimmune diseases.

Results

Knockout of RNF115 attenuates the autoimmune phenotypes of Trex1^–/– mice

Loss-of-function mutations in TREX1 in humans and knockout of Trex1 in mice result in severe systemic inflammation and developmental abnormalities that are rescued by deletion of STING [19, 44, 45]. To identify potential regulators of STING involved in TREX1 deficiency-mediated autoimmunity, we performed two independent immunoprecipitation assays followed by mass spectrometry analysis of bone marrow-derived dendritic cells and macrophages (GM-DCs/Macs) and mouse lung fibroblasts (MLFs) from Trex1^−/− mice (Supplementary Fig. 1A), which led to the identification of RNF115 as an E3 ubiquitin ligase that interacts with STING (Supplementary Table 1). Interestingly, we previously demonstrated that RNF115 promotes the K63-linked ubiquitination and activation of STING after HSV-1 infection [43]. Subsequently, we found that RNF115 interacted with STING in the Trex1^–/– lungs or of GM-DCs/Macs but not in their WT counterparts (Supplementary Fig. 1B) and that knockdown of RNF115 significantly inhibited the mRNA expression of Ccl5, Cxcl10, and Isg15 in Trex1^–/– GM-DCs/Macs and MLFs (Supplementary Fig. 1C), indicating that RNF115 constitutively interacts with STING and promotes the expression of proinflammatory cytokines in the context of TREX1 deficiency.

To determine whether RNF115 is involved in the development of autoimmune phenotypes in TREX1-deficient mice, we generated Rnf115^–/–Trex1^–/– mice and investigated the survival and autoimmune phenotypes of these mice. As shown in Fig. 1A, B, deletion of RNF115 in Trex1^–/– mice partially but significantly rescued the developmental delay and autoimmune lethality in Trex1^–/– mice [44, 46], indicating that RNF115 facilitates autoimmunity caused by TREX1 deficiency. We then analyzed the autoimmune symptoms in multiple organs of wild-type, Trex1^–/–, and Rnf115^–/–Trex1^–/– mice. First, the serum levels of CXCL1, TNF-α, CCL5, and total IgG were significantly lower in Rnf115^–/–Trex1^–/– mice than in Trex1^–/– mice (Fig. 1C). Second, substantially fewer immune cells infiltrated into the lungs of Rnf115^–/–Trex1^–/– mice than into the lungs of Trex1^–/– mice, and the numbers of infiltrated cells in Rnf115^–/–Trex1^–/– mice were comparable to those in their wild-type counterparts (Fig. 1D). Third, the mRNA levels of Tnf, Cxcl9, Cxcl10, Ifna4, Ifit3, Il1b, Cxcl1, Ccl5, Il12p40, and Il10 in the kidneys and livers of Rnf115^–/–Trex1^–/– mice were significantly lower than those in the kidneys and livers of Trex1^–/– mice (Fig. 1E and Supplementary Fig. 2A). Finally, we observed that the spleens were substantially smaller and both the percentages and absolute numbers of splenic CD8⁺CD44⁺CD62L^- and CD4⁺CD44⁺CD62L^- T cells (effector T cells), GL7⁺FAS⁺ cells (germinal center B cells), CD8⁺IFN-γ⁺ and CD4⁺IFN-γ⁺ T cells, and NK1.1⁺CD69⁺ cells (activated NK cells) were significantly lower in Rnf115^–/–Trex1^–/– mice than in Trex1^–/– mice (Fig. 1F–H and Supplementary Fig. 2B). These results collectively suggest that RNF115 promotes self-DNA-triggered autoimmunity in Trex1^-/- mice.

Fig. 1.

Deletion of RNF115 attenuates the autoimmune phenotypes of Trex1^–/– mice. A Representative images (left) and body weights (right) of 5-week-old wild-type (WT) (n = 16), Trex1^–/– (n = 19), and Rnf115^–/–Trex1^–/– (n = 45) mice. B Survival (Kaplan–Meier curves) of Trex1^–/– (n = 72) and Rnf115^–/–Trex1^–/– (n = 49) mice. C ELISA of CXCL1, TNF-α, CCL5 and total IgG in the sera of 6-week-old WT (n = 6), Trex1^–/– (n = 8) and Rnf115^–/–Trex1^–/– (n = 8) mice. D Representative images of H&E-stained lung sections from 6-week-old Trex1^–/– (n = 3) and Rnf115^–/–Trex1^–/– (n = 3) mice. E qRT‒PCR analysis of Tnf, Cxcl9, and Cxcl10 mRNA in the kidneys and livers of 6-week-old WT (n = 6), Trex1^–/– (n = 9) and Rnf115^–/–Trex1^–/– (n = 8) mice. F Representative images (upper) of spleens and spleen weight/body weight ratios (lower) of 6-week-old WT (n = 9), Trex1^–/– (n = 15) and Rnf115^–/–Trex1^–/– (n = 10) mice. G, H Flow cytometric analysis of splenocytes from 6-week-old WT (n = 10), Trex1^–/– (n = 6), and Rnf115^–/–Trex1^–/– (n = 11) mice stained with fluorophore-conjugated antibodies against the indicated surface or intracellular molecules. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant (one-way ANOVA or the log-rank test). The graphs show the means ± SDs (A, C, E, F, H). The scale bars represent 200 μm (D). The data are representative of three independent experiments (C–H)

STING-mediated signaling induces IL-10 to attenuate inflammation in the gut [47]. Consistent with this finding, we observed increased levels of Il10 in the livers, kidneys and GM-DCs/Macs of Trex1^–/– mice compared to those of their wild-type counterparts (Supplementary Fig. 2A, C). Moreover, knockout of RNF115 or treatment with the JAK1 inhibitor ruxolitinib significantly decreased the levels of Il10 in Trex1^–/– GM-DCs/Macs but not in Rnf115^–/–Trex1^–/– GM-DCs/Macs (Supplementary Fig. 2C). These data suggest that RNF115 controls Il10 expression in a JAK1-dependent manner in the context of TREX1 deficiency.

Knockout of RNF115 barely affects the autoimmune phenotypes of STING^N153S/WT mice

Gain-of-function mutations in STING, such as STING^N154S and STING^V155M, which result in spontaneous and constitutive activation of STING, cause an autoimmune disease known as SAVI [24, 30]. To examine whether knockout of RNF115 inhibits autoimmunity caused by gain-of-function of STING, we generated STING^N153S/WT (designated NS) mice by CRISPR/Cas9-mediated genome editing, obtained Rnf115^–/–STING^N153S/WT (designated 115NS) mice and investigated the survival and autoimmune phenotypes of these mice (Supplementary Fig. 3A, B). NS mice started to die at 58 days after birth, and their median survival time was 65 days (Supplementary Fig. 3C). Unexpectedly, knockout of RNF115 in NS mice did not protect against autoimmune lethality (median survival time, 64 days) (Supplementary Fig. 3C). Moreover, the body weights and spleen sizes were comparable between NS and 115NS mice but were significantly decreased and increased, respectively, compared to those of wild-type mice (Supplementary Fig. 3D, E). It has been shown that NS mice exhibit T-cell cytopenia and myeloid cell expansion similar to those observed in SAVI patients [24, 26]. The results of flow cytometric analysis showed that the percentages and absolute numbers of CD11c⁺ and CD11b⁺ myeloid cells, GL7⁺FAS⁺ GCB cells, and CD8⁺GzmB⁺ T cells (cytotoxic T cells) in the spleens of 115NS mice were comparable to those in the spleens of NS mice but significantly greater than those in the spleens of wild-type mice (Supplementary Fig. 3F, G). In addition, the percentages and absolute numbers of CD4⁺, CD8⁺, and CD4⁺CD25⁺ T cells in the spleens of 115NS mice were comparable to those in the spleens of NS mice but lower than those in the spleens of wild-type mice (Supplementary Fig. 3F, G). Similarly, the percentages and absolute numbers of CD11c⁺ or CD11b⁺ myeloid cells and CD3⁺ T cells in the peripheral blood of 115NS mice were comparable to those in the peripheral blood of NS mice and were higher and lower, respectively, than those in the peripheral blood of wild-type mice (Supplementary Fig. 4A), indicating that deletion of RNF115 does not affect systemic autoimmunity, peripheral T-cell cytopenia or myeloid cell expansion in NS mice.

We further examined lymphoid and myeloid progenitor cells in central immune organs, including the thymus and bone marrow, from wild-type, NS, and 115NS mice. The results showed that the percentages and absolute numbers of double-negative (CD4^-CD8^-) thymocytes in 115NS mice were significantly increased compared to those in NS mice (Supplementary Fig. 4B). In addition, the percentages and absolute numbers of CD19⁺ cells (comprising B cells in all stages of development as well as immature B cells) and Lin^-Sca1^-c-kit⁺ myeloid progenitors in the bone marrow of 115NS mice were significantly higher and lower, respectively, than those in the bone marrow of NS mice but comparable to those in the bone marrow of wild-type mice (Supplementary Fig. 4C), suggesting that knockout of RNF115 restores immune cell progenitors in the central lymphoid organs of NS mice.

Rnf115^-/-STING^N153S/WT bone marrow chimeric mice exhibit alleviation of autoimmunity

To confirm the role of hematopoietic RNF115 in regulating immune cell development and STING^N153S-related autoimmunity in NS mice, we next generated bone marrow chimeric mice by transferring wild-type, NS, or 115NS bone marrow cells into lethally irradiated wild-type C57BL/6 mice and then conducted various analyses. As expected, the numbers of CD4^-CD8^- and CD3^-NK1.1⁺ thymocytes and the numbers of CD19⁺ cells and Lin^-Sca1^-c-kit⁺ myeloid progenitors in the bone marrow of 115NS bone marrow (115NS → WT) chimeric mice were comparable to those in the bone marrow of WT → WT chimeric mice (Supplementary Fig. 4D–F). Interestingly, we observed that 115NS → WT chimeric mice exhibited survival times and spleen sizes similar to those of WT → WT chimeric mice and significantly increased survival times and reduced spleen sizes compared to NS bone marrow (NS → WT) chimeric mice (Fig. 2A, B). Moreover, the levels of G-CSF, CXCL10, and TNF-α in serum, the mRNA levels of Cxcl10 and Cxcl9 in white blood cells, and the composition of immune cells (including CD3⁺ and CD19⁺ lymphocytes and CD11c⁺ and CD11b⁺ myeloid cells) in peripheral blood of 115NS → WT chimeric mice were comparable to those in WT → WT chimeric mice and were significantly increased compared to those of NS → WT chimeric mice (Fig. 2C, D and Supplementary Fig. 5A). In addition, the percentages and absolute numbers of CD11c⁺ and CD11b⁺ myeloid cells, GL7⁺FAS⁺ GCB cells, CD8⁺CD44⁺CD62L^- memory T cells, CD4⁺CD44⁺CD62L^- memory T cells, CD8⁺IFN-γ⁺ T cells, CD4⁺IFN-γ⁺ T cells, and NK1.1⁺CD69⁺ activated NK cells were significantly lower but the percentages of CD4⁺ and CD8⁺ T cells were significantly greater in the spleens of 115NS → WT chimeric mice than in the spleens of NS → WT chimeric mice but comparable to those in the spleens of WT → WT chimeric mice (Fig. 2E, F and Supplementary Fig. 5B). Collectively, these data suggest that NS bone marrow cells with RNF115 deficiency develop normally in the central and peripheral immune organs of irradiated wild-type mice and exhibit an impaired ability to cause autoimmunity in mice.

Fig. 2.

Rnf115^–/–STING^N153S/WT bone marrow chimeric mice exhibit attenuated autoimmunity. A Survival (Kaplan-Meier curves) of wild-type (WT) → WT (n = 8), STING^N153S/WT (NS) → WT (n = 20), and Rnf115^–/–STING^N153S/WT (115NS) → WT (n = 16) bone marrow chimeric mice. B Representative images of spleens (left) and the spleen weight/body weight ratios (right) of WT → WT (n = 5), NS → WT (n = 5), and 115NS → WT (n = 5) chimeric mice at the 7th week after bone marrow cell transfer. C ELISA of G-CSF, CXCL10, and TNF-α in the sera of WT → WT (n = 5), NS → WT (n = 5), and 115NS → WT (n = 5) chimeric mice as described in (B). D qRT‒PCR analysis of Cxcl10 and Cxcl9 mRNA expression in blood cells from WT → WT (n = 5), NS → WT (n = 5), and 115NS → WT (n = 5) chimeric mice as described in (B). E, F Flow cytometric analysis of splenocytes from WT → WT (n = 5), NS → WT (n = 5), and 115NS → WT (n = 5) chimeric mice as described in (B) after staining with fluorophore-conjugated antibodies against the indicated surface or intracellular molecules. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant (one-way ANOVA or the log-rank test). The graphs show the means ± SDs (B–F). The data are combined from two independent experiments (A) or are representative of two independent experiments (B–F)

To examine whether RNF115 in nonimmune cells is involved in the regulation of bone marrow cell development in 115NS mice, we transferred bone marrow cells from 115NS into Rnf115^+/+ (115NS → WT) and Rnf115^–/– (115NS → 115KO) mice and then performed various analyses. We observed that 115NS → WT and 115NS → 115KO chimeric mice exhibited comparable survival times that were significantly longer than those of NS → WT chimeric mice (Fig. 3A). The spleen size was comparable between 115NS → WT chimeric mice and 115NS → 115KO chimeric mice, as were the mRNA levels of Cxcl10, Cxcl9, Tnf, and Il6 in white blood cells and the serum levels of CXCL1, CCL5, and IL-6 (Fig. 3B–D). Moreover, the percentages and absolute numbers of various immune cells and immune progenitors, including CD11c⁺ and CD11b⁺ myeloid cells, GL7⁺FAS⁺ GCB cells, and CD4⁺, CD8⁺, and CD8⁺GzmB⁺ T cells, in the spleen and peripheral blood were comparable between 115NS → WT chimeric mice and 115NS → 115KO chimeric mice (Fig. 3E, F). These data suggest a bone marrow-intrinsic role of RNF115 in regulating STING^N153S immune cell development and STING^N153S-mediated autoimmunity.

Fig. 3.

RNF115 in hematopoietic cells contributes to autoimmunity in Rnf115^–/–STING^N153S/WT mice. A Survival (Kaplan-Meier curves) of STING^N153S/WT (NS)→wild-type (WT) (n = 10), Rnf115^-/-STING^N153S/WT (115NS) → WT (n = 10), and Rnf115^-/-STING^N153S/WT (115NS)→Rnf115^–/– (115KO) (n = 12) chimeric mice. B Representative images (left) of spleens and the spleen weight/body weight ratios (right) of NS → WT (n = 4), 115NS → WT (n = 6), and 115NS → 115KO (n = 6) chimeric mice at the 7th week after bone marrow cell transfer. C qRT‒PCR analysis of Cxcl10, Cxcl9, Tnf, and Il6 mRNA expression in blood cells from NS → WT (n = 4), 115NS → WT (n = 6), and 115NS → 115KO (n = 6) chimeric mice as described in (B). D ELISA of CXCL1, CCL5, and IL-6 in the sera of NS → WT (n = 4), 115NS → WT (n = 6), and 115NS → 115KO (n = 6) chimeric mice as described in (B). E, F Flow cytometric analysis of splenocytes and blood cells from NS → WT (n = 4), 115NS → WT (n = 6), and 115NS → 115KO (n = 6) chimeric mice as described in (B) after staining with fluorophore-conjugated antibodies against the indicated surface or intracellular molecules. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant (one-way ANOVA or the log-rank test). The graphs show the means ± SDs (B–D, F). The data are representative of two independent experiments (A–F)

RNF115 specifically promotes the Golgi localization of STING^N153S in myeloid cells

The observation that 115NS bone marrow cells developed normally in the periphery in wild-type and Rnf115^–/– mice but not in 115NS mice prompted us to hypothesize that the peripheral 115NS nonimmune cells might impart an unfavorable environment for the development of 115NS immune cells (Supplementary Fig. 3 and Figs. 2–3). Interestingly, we found that knockout of RNF115 significantly decreased the mRNA levels of Cxcl9, Tnf, and Cxcl10 in STING^N153S/WT GM-DCs/Macs but not in STING^N153S/WT MLFs and mouse pulmonary microvascular endothelial cells (MPMECs) (Fig. 4A), indicating that RNF115 promotes STING^N153S-mediated signaling in myeloid cells but not in peripheral nonimmune cells in vitro. Moreover, the mRNA expression levels of Cxcl9, Tnf, Cxcl10, Il1b, Cxcl1, Ccl5, Il12p40, and Il10 in the lungs and livers were comparable between 115NS mice and NS mice (Fig. 4B and Supplementary Fig. 6A, B) but were substantially lower in 115NS → WT chimeric mice than in NS → WT chimeric mice (Fig. 4C and Supplementary Fig. 6C, D). Notably, we found that the expression of Il10 was increased in NS GM-DCs/Macs compared to the wild-type controls and that knockout of RNF115 or treatment with ruxolitinib significantly downregulated the expression of Il10 in NS GM-DCs/Macs (Supplementary Fig. 6E), indicating that STING^N153S-induced expression of Il10 is controlled by RNF115 and JAK1. These results are consistent with a previous study showing that knockout of IFNAR1 inhibits the upregulation of Il10 in the lungs of NS mice [29].

Fig. 4.

RNF115 specifically facilitates the Golgi localization of STING^N153S in myeloid cells. A qRT‒PCR analysis of Cxcl9, Cxcl10, and Tnf mRNA expression in bone marrow-derived dendritic cells and macrophages (GM-DCs/Macs), mouse lung fibroblasts (MLFs), and mouse pulmonary microvascular endothelial cells (MPMECs) from 6-week-old wild-type (WT) (n = 4), STING^N153S/WT (NS) (n = 4), and Rnf115^–/–STING^N153S/WT (115NS) (n = 4) mice. B qRT‒PCR analysis of Cxcl9, Tnf, and Cxcl10 mRNA expression in lungs from 6-week-old WT (n = 4), NS (n = 4), and 115NS (n = 4) mice. C qRT‒PCR analysis of Cxcl10, Cxcl9, and Tnf mRNA expression in lungs from WT → WT (n = 5), NS → WT (n = 5), and 115NS → WT (n = 5) chimeric mice at the 7th week after bone marrow cell transfer. D Schematic diagram of the coculture assay of lung CD45^- nonimmune cells from WT, NS, or 115NS mice with lung CD45⁺ immune cells from WT mice. E qRT‒PCR analysis of Cxcl10, Ccl5, and Tnf mRNA expression in lung CD45^- nonimmune cells from WT (n = 4), NS (n = 4), and 115NS (n = 4) mice (upper graphs) and lung CD45⁺ immune cells from WT mice that were cocultured as described in (D) (lower graphs). F Immunofluorescence staining of Calnexin (an ER marker) (red), GM130 (a Golgi apparatus marker) (red), and FLAG (green) and confocal microscopy analysis of Rnf115^+/+ and Rnf115^–/– GM-DCs/Macs or MLFs reconstituted with STING^N153S-FLAG. Statistical analysis was based on two different colocalization images and was performed using ImageJ software. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant (one-way ANOVA). The orange and white scale bars represent 5 μm and 10 μm, respectively (D). The graphs show the means ± SDs (A–C, E). The data are representative of three (A, B) or two (C, E, F) independent experiments

To examine the potential effects of peripheral nonimmune 115NS cells on peripheral immune cells, we performed an in vitro co-culture assay (Fig. 4D). The CD45^- nonimmune cells from the lungs of wild-type, NS, and 115NS mice were subjected to qRT‒PCR analysis or cocultured with CD45⁺ immune cells from the lungs of wild-type mice for 24 h. Subsequently, the CD45⁺ cells were isolated for qRT‒PCR analysis. We found that the mRNA expression levels of Ccl5, Tnf, and Cxcl10 in the lung nonimmune cells of 115NS and NS mice were comparable and were higher than those in the lung nonimmune cells of wild-type mice (Fig. 4E). Importantly, the mRNA expression levels of Ccl5, Tnf, and Cxcl10 in CD45⁺ immune cells cocultured with CD45^- nonimmune cells from the lungs of NS or 115NS mice were higher than those in CD45^– nonimmune cells cocultured with CD45^– cells from the lungs of wild-type mice (Fig. 4E). These data support the idea that the peripheral nonimmune inflammatory environment in 115NS mice causes developmental abnormalities in peripheral 115NS immune cells and contributes to autoimmunity in these mice. Similarly, it has been shown that irradiated STING^N153S/WT mice transferred with wild-type bone marrow cells exhibit death rates and lung disease severity similar to those of STING^N153S/WT mice [29].

We previously demonstrated that knockout of RNF115 significantly inhibits HSV-1-induced ER-to-Golgi trafficking of STING in multiple types of cells [43]. Consistently, the cGAMP-induced translocation of STING from the ER to the Golgi apparatus was substantially inhibited in Rnf115^–/– GM-DCs/Macs and MLFs compared to their Rnf115^+/+ counterparts (Supplementary Fig. 7A, B). It has been reported that vesicles containing gain-of-function STING mutants, such as STING^N153S, spontaneously and constitutively bud from the ER and are translocated to the Golgi apparatus in the absence of cGAMP [30]. Interestingly, we found that knockout of RNF115 resulted in the retention of STING^N153S in the ER in GM-DCs/Macs but not in MLFs (Fig. 4F). Taken together, these data indicate that (i) the ER-to-Golgi translocation of STING induced by cGAMP and that induced by gain-of-function mutation are differentially regulated and that (ii) RNF115 promotes the ER-to-Golgi trafficking of STING^N153S and the expression of downstream inflammatory cytokines in myeloid immune cells but not in peripheral nonimmune cells, such as MLFs and MPMECs.

RNF115 promotes the ubiquitination and oligomerization of STING and STING^N153S

We previously showed that RNF115 catalyzes the K63-linked polyubiquitination of STING and promotes its phosphorylation during HSV-1 infection [43], leading to the oligomerization of STING and the activation of downstream signaling. To determine whether RNF115 regulates the ubiquitination and oligomerization and STING in the organs or cells of Trex1^–/– mice, we performed ubiquitination and oligomerization assays in various organs and cells from wild-type, Trex1^–/–, and Rnf115^–/–Trex1^–/– mice. As expected, knockout of RNF115 in Trex1^–/– mice substantially inhibited K63-linked polyubiquitination and oligomerization of STING as well as its interaction with TBK1 in the liver, kidney, GM-DCs/Macs and MLFs (Fig. 5A, B). Notably, ubiquitinated STING was phosphorylated in Trex1^–/– livers, and this phosphorylation was impaired by the additional deletion of RNF115 (Supplementary Fig. 7C). Moreover, knocking out RNF115 significantly inhibited the phosphorylation of STING and TBK1 in the livers of Trex1^–/– mice and in Trex1^–/– GM-DCs/Macs (Fig. 5C). Because both wild-type STING and STING^N153S are expressed in STING^N153S/WT mice and cells, we could not specifically examine the K63-linked ubiquitination or oligomerization of STING^N153S in the organs or cells of NS mice with an anti-STING antibody. Instead, we reconstituted Rnf115^+/+ and Rnf115^–/– GM-DCs/Macs and MLFs with STING^N153S-FLAG and subsequently analyzed the ubiquitination and oligomerization of STING^N153S. The K63-linked polyubiquitination and oligomerization of STING^N153S were substantially impaired in Rnf115^–/– GM-DCs/Macs but not in MLFs compared to their Rnf115^+/+ counterparts following reconstitution with STING^N153S-FLAG (Fig. 5D). Consistent with these observations, RNF115 constitutively interacted with STING^N153S in GM-DCs/Macs but not MLFs expressing STING^N153S-FLAG (Fig. 5E). These data indicate that RNF115 is essential for the K63-linked ubiquitination and oligomerization of STING in Trex1^–/– mice and of STING^N153S in GM-DCs/Macs.

Fig. 5.

RNF115 promotes the activation of wild-type STING and STING^N153S. Denaturing IP (IP with an anti-STING antibody and immunoblotting with an anti-K63-linked polyubiquitin antibody; upper two panels), immunoprecipitation (with an anti-STING antibody; middle two panels), immunoblot analysis (with an anti-STING, anti-RNF115, anti-TBK1, or anti-Tubulin antibody), and SDD-AGE analysis (with an anti-STING antibody; middle panels) of livers and kidneys (A), bone marrow-derived dendritic cells and macrophages (GM-DCs/Macs), and mouse lung fibroblasts (MLFs) (B) from 6-week-old WT, Trex1^–/–, and Rnf115^–/–Trex1^–/– mice. C Immunoblot analysis (with an anti-pTBK1, anti-pSTING, anti-TBK1, anti-STING, anti-RNF115, or anti-Tubulin antibody) of livers from 6-week-old WT, Trex1^–/–, and Rnf115^–/–Trex1^–/– mice (left panels) or from WT, Trex1^–/–, and Rnf115^–/–Trex1^–/– GM-DCs/Macs (right panels). D Denaturing IP (IP with an anti-FLAG antibody and immunoblotting with an anti-K63-linked polyubiquitin antibody; upper two panels), SDD-AGE analysis (with an anti-FLAG antibody; middle panel), and immunoblot analysis (with an anti-FLAG, anti-RNF115, or anti-Tubulin antibody; bottom three panels) of Rnf115^+/+ and Rnf115^–/– MLFs and GM-DCs/Macs reconstituted with STING^N153S-FLAG. E Immunoprecipitation (with control IgG or an anti-FLAG antibody) and immunoblot analysis (with an anti-FLAG, anti-RNF115, or anti-Tubulin antibody) of MLFs and GM-DCs/Macs reconstituted with STING^N153S-FLAG. The data are representative of two independent experiments

RNF115 catalyzes the K63-linked ubiquitination of human STING at Lys20/224/289 after HSV-1 infection [43]. Because mouse STING lacks the homologous Lys224 residue of human STING, we mutated Lys19/288 of mouse STING (homologous to Lys20/289 of hSTING) to Arg and examined the effects of these mutations on the activation of the resulting STING mutant (designated N153S^2KR). As shown in Supplementary Fig. 8A, N153S^2KR was localized in the ER in Rnf115^+/+ and Rnf115^–/– GM-DCs/Macs. In contrast, N153S^2KR was colocalized with the Golgi apparatus marker GM130 in both Rnf115^+/+ and Rnf115^–/– MLFs (Supplementary Fig. 8A). Consistent with these observations, compared to STING^N153S, N153S^2KR failed to upregulate the mRNA expression of Il6, Cxcl10, or Ccl5 in Rnf115^+/+ GM-DCs/Macs, and neither STING^N153S nor N153S^2KR upregulated the mRNA expression of Il6, Cxcl10, or Ccl5 in Rnf115^–/– GM-DCs/Macs (Supplementary Fig. 8B). However, STING^N153S and N153S^2KR induced Cxcl10 and Ccl5 mRNA expression to similar degrees in Rnf115^+/+ and Rnf115^–/– MLFs (Supplementary Fig. 7B). These data suggest that RNF115 targets Lys19/288 of STING^N153S to promote its ER-to-Golgi translocation and activation in myeloid cells.

Disulfiram inhibits STING-mediated antiviral immunity and inflammation

Disulfiram (DSF) is an FDA-approved drug for the treatment of chronic alcoholism that has been shown to inhibit the E3 ligase activity of RNF115 and to exhibit potent antitumor activity [48]. We next investigated whether DSF could inhibit HSV-1- or cGAMP-induced signaling by targeting RNF115. We found that DSF treatment dose-dependently and significantly inhibited HSV-1- or cGAMP-induced expression of Ifnb, Tnf, and Ccl5 in Rnf115^+/+ MLFs and GM-DCs/Macs but not in Rnf115^–/– MLFs and GM-DCs/Macs (Supplementary Fig. 9A, B). Moreover, DSF treatment inhibited RNF115-mediated K63-linked polyubiquitination of STING in HEK293 cells and suppressed cGAMP-induced K63-linked polyubiquitination and oligomerization of STING in Rnf115^+/+ GM-DCs/Macs but not in Rnf115^–/– GM-DCs/Macs (Supplementary Fig. 9C, D), suggesting that DSF targets RNF115 to inhibit the K63-linked ubiquitination, oligomerization, and activation of STING in GM-DCs/Macs. We next examined whether inhibition of RNF115 by DSF affects antiviral responses in mice. Rnf115^+/+ and Rnf115^–/– mice were injected intraperitoneally (i.p.) with DSF (20 mg/kg body weight) or DMSO one day before intravenous (i.v.) injection of HSV-1 (5 × 10⁶ PFU per mouse) followed by daily intraperitoneal injection of DSF or DMSO (Supplementary Fig. 9E). As shown in Supplementary Fig. 9F, G, injection of DSF significantly decreased the protein levels of IFN-β, IL-6, and CCL5 in serum and the mRNA levels of Cxcl10 and Il6 in the brain in Rnf115^+/+ mice but not in Rnf115^-/- mice after HSV-1 infection. In line with these findings, the mRNA level of the HSV-1 UL30 gene and the HSV-1 titer were significantly higher in the brains of Rnf115^+/+ mice than in those of Rnf115^–/– mice treated with DSF (Supplementary Fig. 9G), indicating that DSF targets RNF115 to inhibit HSV-1- and cGAMP-triggered STING-mediated antiviral immunity.

We next examined whether DSF affects STING-mediated inflammation. Trex1^–/– and Rnf115^–/–Trex1^–/– MLFs and GM-DCs/Macs were treated with or without DSF and subjected to qRT‒PCR analysis (Supplementary Fig. 10A). The results suggested that DSF treatment dose-dependently and significantly downregulated the expression of Ifnb, Isg15, and Ccl5 in Trex1^–/– MLFs and GM-DCs/Macs but not in Rnf115^–/–Trex1^–/– MLFs and GM-DCs/Macs (Supplementary Fig. 10B, C). Similarly, DSF treatment significantly inhibited the expression of Cxcl10, Il6, and Ccl5 in STING^N153S/WT GM-DCs/Macs but not in Rnf115^-/-STING^N153S/WT GM-DCs/Macs (Supplementary Fig. 10D). In contrast, DSF treatment barely affected the expression of Cxcl10, Il6, and Ccl5 in STING^N153S/WT or Rnf115^-/-STING^N153S/WT MLFs (Supplementary Fig. 10D). Taken together, these data suggest that DSF inhibits STING hyperactivation-mediated inflammation by targeting RNF115.

DSF attenuates the autoimmune phenotypes of Trex1^–/– mice and NS→WT chimeric mice

We next investigated whether DSF can attenuate the autoimmune phenotypes of Trex1^–/– mice. Four-week-old wild-type mice and Trex1^–/– mice were injected (i.p.) with DSF or DMSO every other day, and various analyses were then conducted. Injection of DSF significantly rescued the developmental delay and weight gain of Trex1^–/– mice (Fig. 6A, B). As expected, compared to DMSO treatment, DSF treatment significantly prolonged the survival of Trex1^–/– mice (Fig. 6C). H&E staining and qRT‒PCR analysis revealed that DSF injection suppressed the infiltration of immune cells into the lung; decreased the mRNA expression of inflammatory cytokines such as Tnf, Cxcl9 and Cxcl10 in the liver; and decreased the serum protein levels of CXCL1, CCL5, and IL-6 in Trex1^–/– mice (Fig. 6D–F). In addition, the spleen size and degree of immune cell development were restored in Trex1^–/– mice injected with DSF (Fig. 6G and Supplementary Fig. 11A, B). Consistent with the observation that knockout of RNF115 inhibited the K63-linked ubiquitination and oligomerization of STING in Trex1^–/– mice and in GM-DCs/Macs after cGAMP treatment (Fig. 5A, B and Supplementary Fig. 9B, C), DSF injection substantially impaired the K63-linked polyubiquitination and oligomerization of STING in the livers and kidneys of Trex1^–/– mice (Supplementary Fig. 11C). These data suggest that DSF attenuates the autoimmune phenotypes of Trex1^–/– mice by inhibiting the ubiquitination and oligomerization of STING.

Fig. 6.

DSF attenuates the autoimmune phenotypes of Trex1^-/- mice and NS → WT chimeric mice. A Representative images of 4-week-old wild-type C57BL/6 mice injected i.p. with DMSO (WT + DMSO), Trex1^–/– mice injected i.p. with DMSO (Trex1^–/– + DMSO), and Trex1^–/– mice injected i.p. with disulfiram (DSF) (50 mg/kg body weight) (Trex1^–/– + DSF) every other day for 4 weeks for survival analysis or for 2 weeks for other analyses. B Body weight changes in WT + DMSO (n = 6), Trex1^–/– + DMSO (n = 5), and Trex1^–/– + DSF (n = 8) mice treated as described in (A). C Survival (Kaplan-Meier curves) of Trex1^–/– + DMSO (n = 11) and Trex1^–/–+ DSF (n = 11) mice treated as described in (A). D Representative images of H&E-stained lung sections from WT + DMSO (n = 3), Trex1^–/– + DMSO (n = 3), and Trex1^–/– + DSF (n = 3) mice treated as described in (A). qRT‒PCR analysis of Tnf, Cxcl9, and Cxcl10 mRNA expression in the liver (E) and ELISA of CXCL1, CCL5, and IL-6 in the serum (F) of WT + DMSO (n = 6), Trex1^–/– + DMSO (n = 6), and Trex1^–/–+ DSF (n = 6) mice treated as described in (A). G Representative images (right) of spleens and spleen weight/body weight ratios (left) of WT + DMSO (n = 6), Trex1^–/– + DMSO (n = 6), and Trex1^–/– + DSF (n = 6) mice treated as described in (A). H Survival (Kaplan-Meier curves) of wild-type (WT) → WT chimeric mice injected i.p. with DMSO (WT → WT + DMSO) (n = 8), STING^N153S/WT (NS) → WT chimeric mice injected i.p. with DMSO (NS → WT + DMSO) (n = 11), and NS → WT chimeric mice injected i.p. with DSF (50 mg/kg body weight) (NS → WT + DSF) (n = 10) every other day starting from the 6th week after bone marrow cell transfer for 5 weeks for survival analysis or for 2 weeks for various other analyses. qRT‒PCR analysis of Tnf, Cxcl10, and Cxcl9 mRNA expression in the brain (I) and ELISA of CXCL1, CCL5, and IL-6 in the serum (J) of WT → WT + DMSO (n = 6), NS → WT + DMSO (n = 6), and NS → WT + DSF (n = 6) mice treated as described in (H). K Representative images (left) of spleens and spleen weight/body weight ratios (right) of DMSO (n = 6), Trex1^–/– + DMSO (n = 6), or Trex1^–/–+ DSF (n = 6) mice treated every other day as described in (H). *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant (one-way ANOVA or the log-rank test). The graphs show the means ± SDs (B, E–G, I–K). The scale bars represent 200 μm (D). The data are combined from two independent experiments (C, H) or are representative of two independent experiments (B, D–G, I–K)

We next injected (i.p.) NS → WT chimeric mice with DSF or DMSO every other day starting from the 6th week after bone marrow transfer and then conducted various analyses. We observed that compared with DMSO injection, DSF injection significantly prolonged the survival of NS → WT chimeric mice (Fig. 6H). In addition, in NS → WT chimeric mice, DSF injection inhibited the mRNA expression of Tnf, Cxcl10, and Cxcl9 in the brain; decreased the serum levels of CXCL1, CCL5, and IL-6; and restored the spleen size (Fig. 6I–K). Consistent with these findings, we observed that, compared with those treated with DMSO, the NS → WT chimeric mice treated with DSF exhibited restoration of immune cell development in the thymus, bone marrow, peripheral blood, and spleen (Supplementary Fig. 12A–F). These data collectively suggest that DSF attenuates the autoimmune phenotypes of NS → WT chimeric mice.

Inhibition of RNF115 downregulates the expression of proinflammatory cytokines in PBMCs from SLE patients

The peripheral blood of SLE patients contains high concentrations of dsDNA, which activates the STING-dependent signaling pathway [39], thereby inducing the expression of ISGs and proinflammatory cytokines. To examine whether RNF115 is involved in the expression of proinflammatory cytokines in the PBMCs of SLE patients, we collected peripheral blood, quantified the peripheral blood dsDNA concentration, and prepared PBMCs from 48 treatment-naive patients with SLE. We defined dsDNA^hi or dsDNA^low patients based on a cutoff peripheral blood dsDNA concentration of 0.1 μg/ml and classified 13 patients as dsDNA^hi and 35 patients as dsDNA^low [39] (Fig. 7A). In contrast, dsDNA was almost undetectable in the peripheral blood of healthy donors (HDs), and the peripheral blood dsDNA concentrations in HDs were significantly lower than those in the dsDNA^hi and dsDNA^low groups of SLE patients (Fig. 7A). In both the dsDNA^hi and dsDNA^low groups of patients, knockdown of RNF115 or treatment with DSF significantly downregulated the expression of Isg15, Cxcl10, and Ccl5 in PBMCs (Fig. 7B–C). Interestingly, the mRNA levels of STING and RNF115 were comparable between PBMCs from healthy donors and those from SLE patients, whereas the mRNA level of cGAS was significantly higher in PBMCs from SLE patients than in those from healthy donors (Supplementary Fig. 13A). These data are consistent with previous studies showing that cGAS is an interferon-stimulated gene and is highly expressed in PBMCs from SLE patients (Supplementary Fig. 13B) [37, 49, 50]. The fold decreases in the mRNA expression of Isg15, Cxcl10, and Ccl5 were positively correlated with the blood dsDNA concentration (Fig. 7D). The results from flow cytometric analysis showed that following siRNF115 transfection DSF treatment, peripheral blood PBMCs from dsDNA^hi patients but not those from dsDNA^low patients exhibited decreased expression of IFN-α (Fig. 7E). Moreover, knockdown or pharmacological inhibition of RNF115 significantly inhibited the expression of IFN-γ in CD4⁺ or CD8⁺ PBMCs (Supplementary Fig. 13C). These data suggest that knockdown or pharmacological inhibition of RNF115 downregulates the expression of type I and type II IFNs and proinflammatory cytokines in PBMCs from SLE patients and that targeting RNF115 with DSF is a potential therapeutic strategy for SLE.

Fig. 7.

Inhibition of RNF115 downregulates the expression of proinflammatory cytokines in PBMCs from SLE patients. A Quantitative analysis of the dsDNA concentration in plasma from SLE patients and healthy donors (HDs) (n = 15). High-dsDNA (n = 13) and low-dsDNA (n = 35) patients were defined based on a cutoff plasma dsDNA concentration of 0.1 μg/ml. qRT‒PCR analysis of ISG15, CXCL10, CCL5 (B), and RNF115 (C) mRNA expression in PBMCs obtained as described in (A) and transfected with siCon or siRNF115 in the presence of DMSO or disulfiram (DSF) (5 μM) for 24 h. D Correlation analysis between fold decreases in the expression levels of the genes listed in (B) and the dsDNA concentration in the SLE patients described in (A). E Flow cytometric analysis of intracellular IFN-α expression in PBMCs treated as described in (A) and then stained with an anti-IFN-α-PE antibody. *P < 0.05, **P < 0.01, ***P < 0.001 (one-way ANOVA). The graphs show the means ± SDs (A–C, E)

Discussion

The cGAS-cGAMP-STING axis plays a critical role in inflammation and autoimmunity. The concept of targeting this axis to treat autoimmune diseases is greatly appreciated, but doing so remains challenging. In this study, we demonstrated that the E3 ubiquitin ligase RNF115 promotes STING-mediated inflammation and autoimmunity, which are ameliorated by the RNF115 inhibitor DSF. In support of this idea, we found that (i) knockout of RNF115 or treatment with DSF alleviated inflammation and autoimmunity and prolonged the survival of Trex1^–/– mice and STING^N153S/WT→wild-type (NS → WT) bone marrow chimeric mice; (ii) knockdown of RNF115 or treatment with DSF downregulated the expression of IFN-α and proinflammatory cytokines in PBMCs from SLE patients with high peripheral blood concentrations of dsDNA; and (iii) knockout or inhibition of RNF115 suppressed the K63-linked ubiquitination and oligomerization of STING in cells and organs from TREX1-deficient mice and inhibited the Golgi localization of STING^N153S in myeloid cells. Collectively, these findings highlight the RNF115-mediated regulation of wild-type STING and STING^N153S under inflammatory conditions and suggest that DSF is a potential drug for the treatment of hyperactive STING-related autoimmune diseases.

Binding to cGAMP triggers the Golgi localization and oligomerization of STING to activate downstream STING signaling and the expression of proinflammatory cytokines. Molecular and biochemical evidence has shown that vesicles containing STING gain-of-function mutants, including N154S, V155M, and R284S, spontaneously bud from the ER and are translocated to the Golgi apparatus and induce the expression of proinflammatory cytokines in the absence of cGAMP [23, 24, 51, 52]. Moreover, the organs and cells of Trex1^–/– mice and embryos of Dnase II^–/– mice were found to exhibit enhanced signatures of type I IFNs and proinflammatory cytokines [13, 14, 19, 53, 54], while STING^N153S/WT and STING^V154M/WT knock-in mice exhibited robust activation of NF-κB but not the IRF3 pathway [26, 28, 29]. In line with this, deletion of IRF3 or IFNAR1 rescued autoimmune lethality in Trex1^–/– mice but not in STING^N153S/WT mice, indicating that the regulatory mechanisms of cGAMP-induced activation of STING and gain-of-function mutation-induced activation of STING are different. We found that deletion of RNF115 led to the retention of STING in the ER and inhibited the expression of proinflammatory cytokines in both Trex1^–/– and wild-type GM-DCs/Macs and MLFs after cGAMP treatment, whereas knockout of RNF115 resulted in the retention of STING^N153S in the ER in GM-DCs/Macs but not in MLFs. In addition, the expression levels of proinflammatory cytokines in GM-DCs/Macs but not in MLFs from Rnf115^–/–STING^N153S/WT mice were significantly lower than those in GM-DCs/Macs from STING^N153S/WT mice. Furthermore, knockout of RNF115 impaired the K63-linked ubiquitination of wild-type STING in Trex1^–/– MLFs, GM-DCs/Macs and organs, whereas knockout of RNF115 impaired the K63-linked ubiquitination and the ER-to-Golgi translocation of STING^N153S in GM-DCs/Macs but not in MLFs. These data indicate that RNF115 regulates the activation of wild-type STING in multiple cell types and that RNF115 and other currently unidentified factors regulate the activation of STING^N153S in immune cells and in peripheral nonimmune cells, respectively. Consistent with these observations, deletion of RNF115 in Trex1^–/– mice attenuated the autoimmune phenotypes, and irradiated wild-type mice transferred with Rnf115^–/–STING^N153S/WT bone marrow cells exhibited attenuated autoimmunity. Further structural evidence elucidating the molecular mechanism of STING^N153S activation and the RNF115-STING^N153S interaction awaits future investigations.

We observed that the levels of proinflammatory cytokines in the liver and lung were comparable between Rnf115^–/–STING^N153S/WT mice and their STING^N153S/WT counterparts. Because knockout of RNF115 did not inhibit the expression of proinflammatory cytokines in STING^N153S/WT MLFs or MPMECs, Rnf115^–/–STING^N153S/WT immune cells were immersed in a constant inflammatory microenvironment in peripheral organs such as the liver and lung in Rnf115^-/-STING^N153S/WT mice. Therefore, the expression of proinflammatory cytokines in these Rnf115^–/–STING^N153S/WT immune cells would be increased in the constantly inflamed organs of Rnf115^–/–STING^N153S/WT mice in vivo, although knockout of RNF115 impaired the expression of proinflammatory cytokines in STING^N153S/WT GM-DCs/Macs in vitro. From this perspective, we found that lung CD45⁺ immune cells from WT mice cocultured with lung CD45^- nonimmune cells from either Rnf115^–/–STING^N153S/WT mice or STING^N153S/WT mice produced higher levels of proinflammatory cytokines than those cocultured with lung CD45^- nonimmune cells from WT mice. It should be noted that immune cells and nonimmune cells cross-talk and communicate extensively within organs, which might mutually reinforce the expression of proinflammatory cytokines in both immune and nonimmune cells to cause hyperinflammation at the organ level in Rnf115^–/–STING^N153S/WT mice. In support of this idea, the levels of proinflammatory cytokines were significantly lower in the livers and lungs of Rnf115^–/–STING^N153S/WT → WT chimeric mice (in which the peripheral nonimmune cells were normal) than in those of STING^N153S/WT → WT chimeric mice, whereas WT → STING^N153S/WT chimeric mice and NS mice exhibit similar lethality and lung inflammation [29].

Notably, we and others have shown that the level of Il10 is higher in Trex1^–/– and NS cells and organs than in the corresponding wild-type controls [19, 26, 29]. Importantly, knockout of RNF115 impaired the expression of Il10 in the lungs of Trex1^–/– mice and 115NS → WT chimeric mice. Furthermore, inhibition of JAK1 or knockout of IFNAR1 significantly downregulated the expression of Il10 in Trex1^–/– and NS GM-DCs/Macs and organs [26, 29]. Taken together, these data suggest that RNF115 regulates the expression of Il10 in these autoimmune models through IFNAR signaling. From this perspective, type I IFNs upregulate factors such as IL-10 to resolve inflammation [47, 55, 56]. However, knockout of IFNAR1 neither results in more severe autoimmunity nor attenuates autoimmunity in NS mice [29] but completely rescues autoimmune phenotypes in Trex1^–/– mice [19], indicating the differential involvement of type I IFN signaling in these two autoimmune models. Notably, the levels of proinflammatory cytokines, such as IL-12, CCL5 and CCL3, are decreased in Ifnar1^–/–STING^N153S/WT lungs compared to their IFNAR1-sufficient counterparts but are still higher than those in WT or Ifnar1^–/– lungs; moreover, knockout of IFNAR1 partially downregulates type I IFNs and proinflammatory cytokines in STING^N153S/WT mice [19, 29]. Therefore, the balance between proinflammatory cytokines and IL-10 downstream of STING likely controls the progression of autoimmunity in these autoimmune models. However, a thorough understanding of the role of RNF115-mediated expression of IL-10 in the progression of autoimmunity in Trex1^–/– and NS mice requires further investigations with IL-10- or IL-10R-deficient mice.

Interestingly, although the abnormal immune cell development in the bone marrow and thymus was rescued in STING^N153S/WT mice by deletion of RNF115, Rnf115^–/–STING^N153S/WT mice exhibited autoimmune phenotypes similar to those of STING^N153S/WT mice. Further analysis suggested that the T-cell cytopenia and myeloid cell expansion in the blood and spleen of Rnf115^–/–STING^N153S/WT mice were similar to those in the blood and spleen of STING^N153S/WT mice. One explanation for these phenomena is that the constant inflammatory microenvironment in the periphery of Rnf115^–/–STING^N153S/WT mice, which alters the gene expression profile of peripheral Rnf115^–/–STING^N153S/WT immune cells, causes the abnormal development of peripheral Rnf115^–/–STING^N153S/WT immune cells. In this context, normal immune cell development in the peripheral blood and spleen was substantially restored in 115NS → WT chimeric mice compared to NS → WT chimeric mice.

STING is also involved in autophagy induction [57–59]. cGAMP binding results in STING localization to the ER-Golgi intermediate compartment (ERGIC), where STING promotes LC3 lipidation in a manner dependent on the autophagy-related proteins ATG5 and WIPI2 but not TBK1, thereby resulting in the formation of autophagosomes to defend against invading pathogens. Notably, RNF115 interacts with RAB7 to regulate late endocytic trafficking and autophagosome maturation [60–62]. K63-linked polyubiquitination of STING catalyzed by RNF115 is imperative for STING activation; therefore, investigating whether the regulatory effect of RNF115 on autophagy relies on STING is highly important. Moreover, STING was recently shown to activate the PERK-eIF2a pathway to coordinate the induction of a translational program preferentially promoting inflammatory and survival signaling upon cGAMP binding, which occurs in the ER and is independent of the TBK1-IRF3 cascade [63, 64]. We observed that knockout of RNF115 resulted in the retention of cGAMP-bound STING or STING^N153S in the ER. It is highly important to elucidate whether and how STING accumulation in the ER is involved in the activation of the PERK-eIF2a pathway to coordinate the attenuation of inflammation in Trex1^–/– mice and 115NS → WT chimeric mice.

Emerging evidence has shown that STING-mediated signaling is involved in multiple human diseases, such as AGS, SAVI, SLE, coatomer subunit alpha (COPA) syndrome, amyotrophic lateral sclerosis (ALS), and Parkinson’s disease (PD) [2, 65–70]. We found that RNF115 promoted STING activation in Trex1^–/– mice and in STING^N153S/WT myeloid cells and that knockdown or pharmacological inhibition of RNF115 significantly decreased the levels of IFN-α and proinflammatory cytokines in PBMCs from SLE patients with high peripheral blood concentrations of dsDNA, suggesting that targeting RNF115 to downregulate STING activation may be beneficial for SLE patients with hyperactivation of STING signaling. Patients with COPA syndrome have mutations in the COPA gene encoding the α-COP component of COPI complex vesicles. Such mutations (such as COPA^E241K) result in excessive Golgi localization and hyperactivation of STING independent of cGAMP binding [65, 68, 69]. The cytoplasmic DNA/RNA-binding protein TDP-43 has been shown to enter mitochondria and release DNA to activate STING signaling, leading to ALS [66, 67]. In addition, mutation or deficiency of the E3 ubiquitin ligase Parkin or the ubiquitin kinase PINK1 causes mtDNA release, which activates STING signaling cascades, resulting in neuroinflammation and PD [70]. Importantly, deletion or inhibition of STING in Copa^E241K/+, Prp-TDP-43^Tg/+ (human TDP-43-overexpressing), and Prkn^–/–/Pink1^–/– mice significantly inhibited the progression of the related diseases. Inhibiting STING activation by targeting RNF115 might be beneficial for patients with these autoimmune diseases. We and others have demonstrated that the FDA-approved drug DSF inhibits RNF115. Future clinical trials of DSF in patients with indications are needed to confirm its clinical utility. Taken together, these findings highlight the diverse activation mechanisms of wild-type MITA and MITA gain-of-function mutants mediated by RNF115 and indicate that DSF treatment is a potential intervention strategy for STING-related autoimmune diseases.

Methods

Mice

The Rnf115^–/– and Trex1^+/– mice were described previously [43, 71]. The STING^N153S/WT mice were generated by GemPharmatech Co., Ltd., through CRISPR/Cas9-mediated gene editing. In brief, guide RNA (5’-GTTAAATGTTGCCCACGGGC-3’) was synthesized through in vitro transcription and purification. The gRNA was incubated with purified Cas9 protein and injected into fertilized eggs (at the one-cell stage) together with the donor vector with a mutation in exon 5 of the Sting gene. The injected fertilized eggs were cultured to the two-cell stage and subsequently transplanted into pseudopregnant mice. The targeted genomes of F0 mice were amplified via PCR and sequenced. Wild-type C57BL/6 mice whose ovaries were removed and replaced with the ovaries of STING^N153S/WT F0 mice were crossed with wild-type C57BL/6 mice to obtain F1 STING^N153S/WT mice. Similarly, wild-type C57BL/6 mice bearing ovaries from F1 STING^N153S/WT mice were crossed with Rnf115^–/– mice to obtain Rnf115^+/-STING^N153S/WT mice. Finally, wild-type C57BL/6 mice bearing ovaries from Rnf115^+/-STING^N153S/WT mice were crossed with Rnf115^–/– mice to obtain Rnf115^–/–STING^N153S/WT mice. Trex1^+/- mice were crossed with Rnf115^–/– mice to obtain Rnf115^–/–Trex1^–/– mice. Age- and sex-matched wild-type, Trex1^–/– and Rnf115^–/–Trex1^–/– mice and wild-type, STING^N153S/WT, and Rnf115^–/–STING^N153S/WT mice were used for all the described experiments. All mice were housed in the specific pathogen-free animal facility at Wuhan University, and all animal experiments were carried out following protocols approved by the Institutional Animal Care and Use Committee of Wuhan University (Approval No. 21020 A). The mouse genotypes were determined by PCR analysis of tail DNA, and the genotyping primers used were as follows (wild-type Rnf115 allele, 940 bp; the Rnf115 knockout allele, 440 bp; positive Trex1 knockout allele, 200 bp; STING^N153S allele, 619 bp): Rnf115 forward, 5ʹ- GGGCATTAGGCACAAACAAAATACCTATAAACATAA-3ʹ and reverse, 5ʹ- ATTAATATTACAGCAGCAAGATGACAGTCTGACAAC-3ʹ; Trex1 forward, 5ʹ-AGGCAAATAAGTAGTGGA-3ʹ and reverse, 5ʹ-TCTCACTGGCCCCAGGGCTAC-3ʹ; the STING^N153S forward, 5ʹ-CTCAGTTGGATGTTTGGCCTTC-3ʹ and reverse, 5ʹ- GGTCACCCTCAAATAAATAGGGTG-3ʹ.

LC-MS Analysis

The experiments were performed as previously described [42]. GM-DCs/Macs from Trex1^–/– mice were prepared for two independent coimmunoprecipitation assays with an anti-STING antibody or IgG (as a control). The immunoprecipitates were washed three times with PBS and subjected to LC–MS analysis. The purified proteins obtained in the two independent experiments were digested with trypsin overnight. The resulting peptide mixtures were desalted on SDB-RPS stage tips and analyzed on an EASY-nLC 1200 system interfaced online with an Orbitrap Exploris 480 mass spectrometer (Thermo Scientific) in DDA mode. Peptides were dissolved in 0.1% formic acid, loaded onto a C18 trap column (100 μm × 20 mm, 3 μm particle size, 120 Å pore size) through an autosampler and eluted onto a C18 analytical column (75 μm × 250 mm, 2 μm particle size, 100 Å pore size). Mobile phase A (0.1% formic acid) and mobile phase B (90% ACN, 0.1% formic acid) were used to establish a 60 min separation gradient. The constant flow rate was set at 300 nL/min. Data were acquired using a spray voltage of 2 kV, an ion funnel RF setting of 40, and an ion transfer tube temperature of 320 °C. Each scan cycle consisted of one full-scan mass spectrum event (Res. 60 K, scan range 350–1500 m/z, AGC 300%, and IT 20 ms) followed by MS/MS events (Res. 15 K, AGC 100%, and IT auto). The cycle time was set to 2 s. The isolation window was set to 1.6 Da. The dynamic exclusion time was set to 35 s. The normalized collision energy was set to 30%. The MS data were analyzed by Proteome Discoverer (Thermo Scientific, version 2.4). The identified proteins are listed in Supplementary Table 1.

Clinical samples

The protocols for human participation and blood collection from volunteer healthy donors (HDs) and SLE patients were reviewed and approved by the Medical Ethics Committee of Zhongnan Hospital of Wuhan University (Approval No. 2021011-1). Written informed consent was obtained from all participants before blood collection. Ten healthy donors (HDs) (12 women and 3 men, with an average age of 42.5 years) and thirty-two hospitalized treatment-naive patients with SLE (42 women and 6 men, with an average age of 45.6 years) were recruited from the Department of Rheumatology and Immunology, Zhongnan Hospital of Wuhan University, between Jan 20th and May 20th, 2023 (Patients #1-#32), and between Oct 1st and Nov 30th, 2023 (Patients #33-#48). The SLE patients were diagnosed according to the European League Against Rheumatism (EULAR)/American College of Rheumatology (ACR) 2019 criteria. Peripheral blood (10 ml from each patient) was collected on the second morning after the patients were hospitalized, and PBMCs were immediately isolated for various analyses. The clinical characteristics of the SLE patients are presented in Supplementary Table 2.

Generation and treatment of PBMCs from SLE patients and HDs

Peripheral blood mononuclear cells (PBMCs) were isolated from freshly collected blood from healthy volunteer donors (HDs) or SLE patients with a Ficoll-PlaqueTM PLUS Kit (GE Healthcare, 17-1440-02) according to the manufacturer’s instructions. Approximately 5 × 10⁶ PBMCs were obtained from 9 ml of blood. The cells (1–2 × 10⁶) were transfected with control siRNA or siRNF115 with Ultra Fection 2.0 transfection reagent (4 A Biotech, FXP092-010) in the presence or absence of DSF (5 μM). 4 h later, the medium was changed to fresh RPMI 1640 medium containing 10% FBS, 1% streptomycin-penicillin, and DMSO or DSF (5 μM), after which the cells were cultured for an additional 20 h. The cells were harvested in either TRIzol (1 ml) or 90% FBS and 10% DMSO and subsequently cryopreserved at –80 °C until analysis. At the end of the study, total RNA was isolated from frozen TRIzol-treated cells prior to reverse transcription and quantitative real-time PCR analysis, and the cryopreserved cells were thawed for intracellular staining with an anti-IFN-α antibody and subsequent flow cytometric analysis.

siRNA

siRNAs targeting human or mouse RNF115 were synthesized by JTS Biotechnology. Control siRNA (siCon): 5′-UUCUCCGAACGUGUCACGUTT-3′. si-hRNF115: 5′-GCCGUGGCUA GAACUGCAUTT-3′. si-mRNF115: 5′-CCAAGAUAAUAGAGCCAAUTT-3′.

Extraction and quantification of dsDNA in blood

Blood (1 ml) from SLE patients or HDs was centrifuged at 800 × g for 3 min, and the supernatants were collected for dsDNA preparation and quantification. Extracellular dsDNA in plasma was extracted using a Serum/Plasma Circulating DNA Kit (Tiangen, DP336) and quantified with an ABQubit dsDNA HS Assay Kit (ABclonal, RK30140) following the manufacturer’s instructions.

Reagents, antibodies, and constructs

Ruxolitinib (TargetMol, T1829), DSF (TargetMol, #T0054), recombinant mouse GM-CSF (CHAMOT Biotechnology, #CM060-20MP), DMSO (Sigma, #D8418), and cGAMP (InvivoGen, #1441190-66-4) were purchased from the indicated manufacturers. HSV-1 and the plasmids used for STING, RNF115, and ubiquitin (K63O) expression were previously described [42, 43]. Mammalian expression plasmids for STING^N153S were constructed by standard molecular biology techniques. The following antibodies and control IgGs were purchased from the indicated manufacturers: mouse control IgG (Santa Cruz Biotechnology, sc-2025), rabbit control IgG (Millipore, 12-370), HRP-conjugated goat-anti mouse and rabbit IgG (Thermo Scientific, PA1-86717 and SA1-9510, respectively), mouse anti-FLAG (Sigma, F3165), anti-GFP (ABclonal, AE012), anti-Tubulin (ABclonal, A12289), anti-Ubiquitin (Santa Cruz Biotechnology, sc-8017), anti- K63-linked ubiquitin (Millipore, 05-1308), anti-TBK1 (Abcam, 96328-11), anti-STING (Cell Signaling Technology, #13647 and ABclonal, A3575), anti-RNF115 (Abcam, ab187642), anti-phosphorylated-STING (Ser365) (Cell Signaling Technology, #72971 S), and anti-phosphorylated-TBK1 (Ser172) (Abcam, ab109272). The staining antibodies for flow cytometric analysis were purchased from the indicated manufacturers: anti-mouse CD11b-PE (Biolegend,101208), anti-mouse CD11c-Percp (Biolegend,117328), anti-mouse Ly6G-Apc-CY7 (Biolegend,127624), anti-mouse F4/80-BV421 (Biolegend,123137), anti-mouse CD3-FITC (Biolegend,100306), anti-mouse CD19-APC (Biolegend,152409), anti-mouse GL7-Percp (Biolegend, 144610), anti-mouse FAS-PE (Biolegend, 152608), anti-mouse CD3-APC (Biolegend, 100312), anti-mouse CD4-FITC (Biolegend, 100406), anti-mouse CD8-BV510 (Biolegend, 100751), anti-mouse NK1.1-APC (Biolegend, 156506), anti-mouse CD69-FITC (Biolegend,104506), anti-mouse GzmB -PE (Biolegend, 372208), anti-mouse CD3-FITC (Biolegend, 100306), anti-mouse CD4-BV421 (Biolegend,100427), anti-mouse Sca-1-APC (Biolegend,108111), anti-mouse lineage-biotin (BD Biosciences, 559971), Streptavidin-APC-CY7 (BD Biosciences, 554063), anti-c-Kit-PE (Biolegend, 105807) (kindly provided by Dr. Haojian Zhang, Wuhan University) [72], anti-human CD3-PE (ABclonal, A22795), anti-human CD4-FITC (ABclonal, A22773), anti-human CD8-APC(ABclonal, A24799), anti-human IFN-γ-PE-CY7 (BD Biosciences, 557643), and anti-human-IFN-α-PE (BD Biosciences, 560097).

Flow cytometric analysis

Single-cell suspensions prepared from various tissues or PBMCs were resuspended in FACS buffer (PBS, 1% BSA) and blocked with anti-mouse or anti-human CD16/32 antibodies for 15 min prior to staining with antibodies against surface and intracellular markers (BioLegend). To analyze the expression of IFN-γ and GzmB in T cells, cells were stimulated with PMA (Sigma, P8139) and ionomycin (Sigma, I0634) in the presence of GolgiStop (BD Biosciences, 554724) for 4 h, and surface and intracellular staining was then performed with antibodies against CD3, CD4, CD8, IFN-γ, and GzmB. Flow cytometry data were acquired on a FACS Celesta or Fortessa flow cytometer and analyzed with FlowJo software 10.0 (TreeStar).

Bone marrow chimeric mice

C57BL/6 wild-type or Rnf115^–/– mice (10 weeks old) were lethally irradiated (9 Gy, delivered as two fractions of 4.5 Gy) prior to injection of bone marrow cells from wild-type mice (1 × 10⁶), NS (1 × 10⁶), or 115NS mice (1 × 10⁶) via the tail vein.

Isolation of mouse pulmonary microvascular endothelial cells (MPMECs)

Lungs from WT, STING^N153S/WT (NS), and Rnf115^–/–STING^N153S/WT (115NS) mice were perfused with PBS through cardiac puncture. The lungs were cut into small pieces (2–3 mm in diameter) and transferred into a gentleMACS C Tube with an enzyme mixture containing 2.35 ml of DMEM, 100 μl of Enzyme D, 50 μl of Enzyme R, and 12.5 μl of Enzyme A from a Lung Dissociation Kit (Miltenyi Biotech). The C Tube was tightly closed and attached to the sleeve of the gentleMACSTM Octo Dissociator (Miltenyi Biotech), and the lung single-cell isolation program was run. After termination of the program, the C tube was detached from the dissociator and incubated at 37 °C for 40 min with shaking. The lung single-cell isolation program was repeated twice, followed by a short spin with a force of up to 1500 × g to collect the sample at the bottom of the tube. The cells were resuspended in PBS containing 1% FBS and incubated with anti-CD45 microbeads for 30 min at 4 °C, after which the CD45⁺ cells were depleted by flowing the mixture through a magnetic column (Miltenyi Biotec). The residual cells were stained with an anti-CD31 prior to FACS sorting. The CD31⁺ lung single cells were collected as MPMECs for subsequent qRT‒PCR analysis.

Coculture assay

CD45⁺ immune cells were isolated from the lungs of wild-type C57BL/6 mice as described above. CD45^- nonimmune cells were isolated from lung single-cell suspensions of wild-type, NS, or 115NS mice. The CD45^- nonimmune cells and the CD45⁺ immune cells (2 × 10⁶ each) were cocultured in 6-well plates for 24 h. Subsequently, the CD45⁺ immune cells were isolated with anti-CD45 microbeads for qRT‒PCR analysis.

DSF administration

For HSV-1 infection, 8-week-old wild-type and Rnf115^–/– mice were injected (i.p.) with DMSO or DSF (20 mg/kg) one day prior to HSV-1 infection. The same volume of DMSO as DSF (in solution, 20 mg/kg) was diluted in corn oil (total volume of 200 μl) and injected i.p. into mice daily after HSV-1 infection. For the autoimmune models, 4-week-old wild-type and Trex1^–/– mice or WT → WT and NS → WT chimeric mice at the 6th week after bone marrow transfer were grouped for DSF treatment. These mice were injected intraperitoneally with DMSO or DSF (50 mg/kg) every other day for 4 weeks for survival analysis or for 2 weeks for other analyses.

Transfection of cGAMP

cGAMP (1 μg/μl) and 10 × digitonin were diluted in PBS (the volume of PBS was 50% that of cell culture medium). The cell culture supernatants were removed, and cGAMP-containing PBS was added to the cell cultures. Half an hour later, the cGAMP-containing PBS was removed, and complete medium was added to the cells for a 3-hour culture followed by various assays.

Coimmunoprecipitation and immunoblotting

GM-DCs/Macs and MLFs were collected and lysed for 10 min with 800 μl of Nonidet P-40 lysis buffer (20 mM Tris HCl, pH 7.4–7.5; 150 mM NaCl; 1 mM EDTA; and 1% Nonidet P-40) containing protease and phosphatase inhibitors (APExBIO, K1015). Cell lysates (700 μl) were incubated with control IgG or an anti-FLAG antibody and protein G agarose (20 μl) for 6 h. The immunoprecipitates were washed three times with 1 ml of prelysis buffer and subjected to immunoblot analysis. The remainder of each lysate (100 μl) was subjected to immunoblot analysis of RNF115, Tubulin, STING, and STING^N153S-FLAG expression.

Semidenaturing detergent agarose gel electrophoresis (SDD-AGE)

The experiments were performed as previously described [71]. Cells or tissues were collected and lysed in NP-40 lysis buffer, and the lysates were mixed with 1× sample buffer (0.5× TBE, 10% glycerol, 2% SDS, and 0.0025% bromophenol blue) and loaded onto a vertical 1.5% agarose gel (Bio-Rad). After electrophoresis in running buffer (1 × TBE and 0.1% SDS) for approximately 1 h at a constant voltage of 100 V and a temperature of 4 °C, the proteins were subjected to immunoblot analysis.

Denaturing IP and ubiquitination assays

The experiments were performed as previously described [42, 73, 74]. Cells or tissues were collected and lysed in standard lysis buffer (100 μl), and the cell lysates were denatured at 95 °C for 15 min in the presence of 1% SDS. A portion of each cell lysate (10 μl) was saved for immunoblot analysis of the expression of the target proteins. The remainder of each cell lysate (90 μl) was diluted with 1 ml of lysis buffer and immunoprecipitated (denaturing IP) with either anti-FLAG beads or protein G (20 μl) plus an anti-STING antibody (0.5 μg). The immunoprecipitates were washed three times and subjected to immunoblot analysis. Alternatively, cell lysates were pulled down by TUBE (LifeSensors, 50 μg/ml) prior to immunoblot analysis.

Immunofluorescence and confocal microscopy analysis

The experiments were performed as previously described [42]. MLFs or GM-DCs/Macs were cultured on coverslips, fixed in 4% paraformaldehyde for 10 min and washed with PBS three times. After that, the cells were permeabilized with 0.5% saponin in PBS for 5 min and washed with PBS three times. The cells were blocked in 1% BSA and 0.1% saponin in PBS for 30 min, stained in blocking buffer (1% BSA and 0.1% saponin) with primary antibodies for 2 h, and washed with PBS three times. The cells were further stained with Alexa Fluor 488- or 594-conjugated secondary antibodies for 1 h. Finally, the cells were stained with In Situ Microplate Nuclear Stain and Anti-Fade (Sigma, DUO82064-1KT), and the coverslips were mounted on slides. Images were acquired on an Olympus FV1000 fluorescence microscope.

Quantitative real-time PCR (qRT–PCR) and ELISA

The experiments were performed as previously described [74]. Total RNA was extracted from cells (2 × 10⁶) or tissues (0.5 mg of lung, liver or kidney tissue) using TRIzol (Life Technologies, 15596026). In brief, lungs, livers or kidneys (0.5 mg of tissue) were immersed in 1 ml of TRIzol reagent (Invitrogen) and cut into pieces (1–2 mm), which were subsequently homogenized with an IKA T10 basic homogenizer. The homogenates were mixed with 200 μl of chloroform and subsequently centrifuged (12,000 × g) for 10 min at 4 °C. The aqueous phase containing the RNA was carefully transferred to a tube containing 400 μl of isopropanol, and the tube was subsequently centrifuged (12,000 × g) for 10 min at 4 °C. The pellet was washed with 1 ml 75% ethanol and saved for reverse transcription and quantitative PCR analysis. For RNA extraction from cells, cells immersed in 1 ml of TRIzol reagent (Invitrogen) were mixed with 200 μl of chloroform prior to centrifugation (12,000 × g) for 10 min at 4 °C. The aqueous phase containing the RNA was carefully transferred to a tube containing 400 μl of isopropanol, and the tube was subsequently centrifuged (12,000 × g) for 10 min at 4 °C. The pellet was washed with 1 ml 75% ethanol and saved for reverse transcription with All-in-One cDNA Synthesis SuperMix (Aidlab Biotechnologies). Gene expression levels were measured with a Bio-Rad CFX Connect system by a fast two-step amplification program with Hieff® qPCR SYBR Green Master Mix (Yeasen Biotechnology, 11201ES08). The expression value obtained for each gene was normalized to that of the gene encoding β-actin. The sequences of the primers used for qRT‒PCR analysis are listed in Supplementary Table 3. The concentrations of G-CSF, CXCL1, CXCL10, CCL5, and TNF-α in serum samples were measured with ABplex Custom Panel ELISA kits from ABclonal Technology according to the manufacturer’s instructions.

Cell culture

The experiments were performed as previously described [71]. Bone marrow cells were isolated and cultured in DMEM supplemented with 10% (vol/vol) FBS and 1% streptomycin-penicillin. GM-CSF (20 ng/ml) was added to the cultured bone marrow cells for differentiation into GM-DCs/Macs. HEK293 cells were obtained from the American Type Culture Collection, authenticated by short tandem repeat (STR) profiling, and tested for mycoplasma contamination. Primary MLFs were isolated from ~8- to 10-week-old mice. Lungs were minced and digested in calcium and magnesium-free HBSS buffer supplemented with 10 mg/ml type I collagenase (Worthington) and 20 μg/ml DNase I (Sigma‒Aldrich) for 3 h at 37 °C with shaking. Cell suspensions were cultured in DMEM supplemented with 10% (vol/vol) FBS and 1% streptomycin-penicillin. Two days later, adherent fibroblasts were rinsed with PBS and subjected to various analyses or assays.

Lentivirus-mediated gene transfer

The experiments were performed as previously described [75]. HEK293 cells were transfected with phage-STING^N153S-FLAG along with the packaging plasmid psPAX2 and the envelope plasmid pMD2G. The medium was changed to fresh complete medium (15% FBS, 1% streptomycin-penicillin) 6 h after transfection with PEI reagent (Polysciences, Inc., 24765-1). 40 hs later, the supernatants were harvested for infection of primary Rnf115^+/+ or Rnf115^–/– MLFs or GM-DCs/Macs, after which various analyses were performed.

Viral infection

The experiments were performed as previously described [43]. For qRT‒PCR analysis, cells were seeded into 24-well plates (1 × 10⁶ cells per well) and infected with HSV-1 for 3 h. For viral replication assays, cells (1 × 10⁶) were infected with HSV-1. 1 h later, the supernatants were removed, and the cells were washed twice with 1 ml of prewarmed PBS and subsequently cultured in complete medium for 12 h. Viral replication was analyzed via qRT‒PCR or plaque assays. For intravenous infection of mice, age- and sex-matched control and RNF115-deficient mice were injected with HSV-1 (5 × 10⁶ PFU per mouse). Sera were collected 12 h after HSV-1 infection for ELISA. Brains and livers were collected 4 days after infection for qRT‒PCR analysis or plaque assays.

Plaque assay

The experiments were performed as previously described [43]. Homogenates (or serial dilutions) of livers and brains from infected mice were used to infect monolayers of Vero cells. 1 h later, the homogenates were removed, and the infected Vero cells were washed with prewarmed PBS twice, followed by incubation with DMEM containing 2% methylcellulose for 48 h. The cells were subsequently fixed with 4% paraformaldehyde for 10 min and stained with 1% crystal violet for 30 min before the plaques were counted.

Hematoxylin-eosin (H&E) staining

Lungs from mice were fixed in 4% paraformaldehyde for 4 h and embedded in paraffin blocks. The paraffin blocks were sectioned (5 μm) for H&E staining (Beyotime Biotech) followed by cover slipping. Images were acquired using an Aperio VERSA 8 (Leica) multifunctional scanner.

Statistical analysis

The significance of differences between the experimental and control groups was determined by Student’s t test (for comparisons of two groups of data) or by one-way ANOVA (for comparisons of more than two groups of data). For animal survival analysis, the Kaplan‒Meier method was used to generate survival curves, and differences in survival were analyzed with the log-rank test. P values of less than 0.05 were considered to indicate statistical significance.

Author contributions

BZ, DL, and XC designed and supervised the study; ZDZ performed the experiments; CRS, FXL, XSYLL and TCX helped with the mouse experiments; HG performed mass spectrometry analysis. YHW, QHZ, MC, and XC collected peripheral blood from the healthy donors and SLE patients; BZ, DL, and ZDZ wrote the paper. All the authors analyzed the data.

Data availability

All the other data supporting the findings of this study within the article and its Supplementary Information files are available from the corresponding author upon reasonable request. A reporting summary for this article is available as a Supplementary Information file.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41423-024-01131-3.