DUSP4 modulates RIG-I- and STING-mediated IRF3-type I IFN response

Abstract

Detection of cytosolic nucleic acids by pattern recognition receptors, including STING and RIG-I, leads to the activation of multiple signalling pathways that culminate in the production of type I interferons (IFNs) which are vital for host survival during virus infection. In addition to protective immune modulatory functions, type I IFNs are also associated with autoimmune diseases. Hence, it is important to elucidate the mechanisms that govern their expression. In this study, we identified a critical regulatory function of the DUSP4 phosphatase in innate immune signalling. We found that DUSP4 regulates the activation of TBK1 and ERK1/2 in a signalling complex containing DUSP4, TBK1, ERK1/2 and IRF3 to regulate the production of type I IFNs. Mice deficient in DUSP4 were more resistant to infections by both RNA and DNA viruses but more susceptible to malaria parasites. Therefore, our study establishes DUSP4 as a regulator of nucleic acid sensor signalling and sheds light on an important facet of the type I IFN regulatory system.

Fig. 1. Increased activation of ERK and TBK1-IRF3, and increased expression of cytokines in DUSP4 KO macrophages in response to RIG-I activation and influenza infection.

A Expression of DUSP4 in PBMCs from healthy individuals and patients with mild influenza (n = 16), and patients with severe influenza (n = 12) was analysed by quantitative real-time PCR (qPCR). Severe disease was defined as evidence of pulmonary involvement on chest radiography or need for supplemental oxygen during admission or both, while mild disease was absence of both criteria for severity. Mann-Whitney non-parametric test was use to perform statistical analysis. **P < 0.01. B Immunoblot analysis of DUSP4 protein expression in bone marrow-derived macrophages (BMDMs) from wildtype (WT) mice upon 5′-ppp dsRNA stimulation (0.5 μg/mL), or infected with PR8 influenza virus at a multiplicity of infection of 1. C Expression of IFNα, IFNβ, IL-6 and TNFα mRNA in WT and knockout (KO) BMDMs at 3 h after 5′-ppp dsRNA stimulation (0.5 μg/mL) was determined by qPCR. Cytokines in culture supernatants of WT and KO BMDMs at 6 h (for IFNα and IFNβ) or 24 h (for IL-6 and TNFα) after 5′-ppp dsRNA stimulation (0.5 μg/mL) were determined by ELISA. Activation of ERK, JNK and p38 (D), and TBK1, IKKε and IRF3 (E) in WT and KO BMDMs at various time-points after 5’-ppp dsRNA stimulation (0.5 μg/mL) was assessed by immunoblot analysis. The phosphorylation levels of TBK1 and IRF3 (n = 3) were quantified using ImageJ. F Cytokine expression in WT and KO BMDMs in response to influenza H1N1 PR8 virus infection (MOI of 1) at mRNA level at 6 h or at protein level at 6 h (for IFNα and IFNβ) or 24 h (for IL-6 and TNFα) post infection (PI) was determined by qPCR or ELISA respectively. G Increased activation of ERK, TBK1 and IRF3 in DUSP4 KO BMDMs at various time-points after PR8 infection at an MOI of 1. The phosphorylation levels of ERK, TBK1 and IRF3 (n = 3) were quantified using ImageJ. H Confocal microscopy of WT and KO BMDMs showing increased IRF3 nuclear accumulation compared to WT cells. Nuclei were stained with DAPI. Unpaired T-test was used for statistical analysis. *P < 0.05; **P < 0.01. Data are representative of at least three independent experiments with similar results.

Fig. 2. DUSP4 KO mice were more resistance to influenza infection compared to WT mice.

A WT and DUSP4 KO mice were infected with 50 plaque-forming units (PFU) of PR8 influenza virus intranasally. Viral titers in the lung of WT and DUSP4 KO mice (n = 3 for each time point) at day 2, 3 and 5 PI were analysed by plaque assay. B, C WT and KO mice (n = 5) were infected with 50 PFU of PR8 virus. Changes in body weight were monitored daily (B). Expression of influenza hemagglutinin (HA) and neuraminidase (NA) in WT and DUSP4 KO mice (n = 5) on day 11 PI was analyzed by qPCR (C). D, E On day 2 or 3 PI, lungs were harvested from WT and KO mice (n = 5). Cytokine mRNA expression and protein concentrations in lung homogenates were measured by qPCR and ELISA respectively. F WT and DUSP4 KO female mice (n = 5) were infected with a lethal dose (150 PFU) of H1N1 PR8 influenza viruses. The survival of the mice was monitored daily post infection (PI) (Log-rank test, P = 0.034). Unpaired T-test was used for statistical analysis. *P < 0.05; **P < 0.01. Data are representative of at least three independent experiments with similar results.

Fig. 3. Increased ERK and TBK1-IRF3 activation and cytokine expression in DUSP4 KO macrophages in response to STING stimulation and HSV-1 infection.

A WT BMDMs were transfected with 0.5 μg/mL of 2′3′-cGAMP, 3′3′-cGAMP or c-di-GMP. Cells were harvested at the indicated time-points to assess DUSP4 expression by immunoblot analysis. B, C WT and KO BMDMs were stimulated with c-di-GAP (0.5 μg/mL) for 3 h to examine the expression of IFNα, IFNβ, IL-6 and TNFα by qPCR (B), 6 h (for IFNα and IFNβ) or 24 h (for IL-6 and TNFα) to determine protein concentrations of cytokines in culture supernatants by ELISA (C). D WT and KO BMDMs were transfected with 0.5 μg/mL of c-di-GMP. Cells were harvested at the indicated time points to examine the activation of ERK, JNK and p38 (activation of JNK was undetectable), and TBK1, IKKε, IRF3 and NFκB by immunoblot analysis. The phosphorylation levels of ERK, TBK1 and IRF3 (n = 3) were quantified using ImageJ. E, F WT and DUSP4 KO BMDMs were infected with HSV-1 at an MOI of 0.01 for 6 h to examine cytokine mRNA expression by qPCR or protein expression by ELISA (E). Activation of TBK1, IKKε, IRF3, ERK, p38 and JNK at various time points PI was analysed by immunoblot analysis. The phosphorylation levels of ERK, TBK1 and IRF3 (n = 3) were quantified using ImageJ (F). G WT and DUSP4 KO mice (n = 5) were infected with 5 × 10⁴ HSV-1 viruses through intravenous injection. Mice were monitored daily for survival (Log-rank test, P < 0.01). Unpaired T-test was used for statistical analysis. *P < 0.05; **P < 0.01. Data are representative of three independent experiments with similar results.

Fig. 4. Increased susceptibility of DUSP4 deficient mice to experimental cerebral malaria.

A WT and KO BMDMs were stimulated with DNA (250 ng/mL) isolated from Plasmodium berghei ANKA (PbA) for 6 h to examine the expression of Ifnα, Ifnβ, Il6 and Tnfα by qPCR. 24 h after stimulation, the concentration of IFNα, IFNβ, IL6 or TNFα in the supernatants was determined by ELISA. Activation of ERK, p38, TBK1 and IRF3 was analyzed by immunoblot (B) and the phosphorylation levels of ERK, TBK1 and IRF3 (n = 3) were quantified using ImageJ (C). D-F WT and DUSP4 KO mice (n = 5) were infected with 1 × 10⁶ of PbA infected red blood cells. Parasitemia was determined by flow cytometry (D). Experimental cerebral malaria (ECM) progression was monitored and recorded using a previous validated clinical scoring algorithm [53] (E). Cumulative incidence of ECM of WT (n = 14) and KO (n = 14) on day 7 PI from three independent experiments (F). G Qualitative (left) and quantitative (right) brain capillary permeability was assessed after intracardica perfusion of Evans blue-injected mice. H On day 7 post PbA infection, H&E-stained brain sections (left) and semi-quantitative score of brain microvascular obstruction are shown. Data are representative of three independent experiments with similar results. I Percentage of death of WT (n = 14) and KO (n = 14) mice from three independent experiments. Unpaired T-test was used for statistical analysis. *P < 0.05; **P < 0.01.

Fig. 5. DUSP4 interacts with TBK1, IRF3 and ERK, and dephosphorylates TBK1.

A HEK293T cells were transfected with flag-DUSP4 and HA-IRF3 or HA-TBK1. Interaction between DUSP4 and IRF3 (left) or TBK1 (right) was examined by immunoprecipitation (IP). B DUSP4 inhibits TBK1-mediated IRF3 phosphorylation. HEK293T cells were transfected with IRF3-expressing plasmids with or without TBK1- or DUSP4-expressing plasmids. Phosphorylation of IRF3 and TBK1, and total protein of IRF3, TBK1 and DUSP4, were determined by Immunoblot analysis. C Recombinant DUSP4 or MKP5 protein was incubated with phospho-IRF3 (pIRF3) to perform in vitro phosphatase assay. The level of IRF3 phosphorylation was analysed by immunoblot. D Purified pTBK1 was incubated with recombinant flag-DUSP4 or flag-DUSP4 phosphatase-dead mutant (flag-DUSP4^mut) to perform in vitro phosphatase assay to determine the dephosphorylation of TBK1 by DUSP4. E HEK293T cells were transfected with HA-IRF3 and Flag-DUSP4 constructs in the combination indicated. Cell lysates were incubated with anti-Flag (left) or anti-HA (right) agarose beads to analyse the interaction between DUSP4, IRF3 and endogenous ERK by immunoblot. Data are representative of three independent experiments with similar results. Unpaired T-test was used for statistical analysis. *P < 0.05. WCL denotes whole cell lysate.

Fig. 6. MAPK-interaction domain of DUSP4 is critical for its regulation of IRF3-type I IFN response.

A Deletion of MAPK-interaction domain from DUSP4 abolished its interaction with ERK1/2 and p38. Flag-DUSP4, Flag-DUSP4^∆M or Flag-DUSP4^mut constructs were transfected to HEK293T cells to perform IP to examine their interaction with MAPKs including ERK1/2, p38 and JNK by immunoblot. B DUSP4 suppresses TBK1-mediated IRF3 activation, which is dependent on its MAPK-interaction domain. Immunoblot analysis of IRF3 phosphorylation in cells transfected with IRF3 together with TBK1 and DUSP4, DUSP4^ΔM or DUSP4^mut. C Deletion of MAPK-interaction domain from DUSP4 resulted in its increased interaction with TBK1. Flag-DUSP4 or Flag-DUSP4^ΔM constructs were transfected to HEK293T cells together with indicated constructs perform IP to assess their interaction with IRF3 or TBK1. D IFNβ promoter luciferase construct containing one AP-1, two IRF3/7 and one NFκB binding site, or its deletion mutants lacking AP-1 binding site, the two IRF3/7 binding sites, or both, were transfected into HEK293T cells to examine their ability to suppress TBK1-mediated IFNβ promoter activity by dual-luciferase assays. Unpaired T-test was used for statistical analysis. *P < 0.05; **P < 0.01. Data are representative of three independent experiments with similar results.

Fig. 7. ERK1/2 regulate IRF3 activation important for type I IFN expression in response to influenza virus infection.

A Recombinant IRF3 protein was incubated with purified pERK1, pERK2, ERK1, ERK2 or TBK1 for in vitro kinase assays. Phosphorylation of IRF3 was assessed by immunoblot and was quantified (n = 3) using ImageJ software. B BMDMs were pre-treated with vehicle or ERK inhibitor PD98059 (100 μM) for 1 h. Cells were then stimulated with 5’-ppp dsRNA (0.5 μg/mL) in the presence of vehicle or PD98059 respectively for the indicated period of time to assess the activation of ERK and IRF3 by immunoblot. C, D RAW264.7 cells were pre-treated with vehicle or ERK inhibitor PD98059 (100 μM) for 1 h followed by with PR8 at an MOI of 1. ERK activation at various time points PI was analysed by immunoblot (C). Cytokine expression with or without ERK inhibition was determined by qPCR. E, F WT and DUSP4 KO BMDMs were pre-treated with vehicle or PD98059 followed by infection with PR8 virus in the presence of vehicle or PD98059 for the indicated period of time to assess the activation of ERK and IRF3 by immunoblot (E). Expression of IFNα and IFNβ at 6 h PI was determined by qPCR (F). G, H. TBK1 deficient or TBK1 plus IKKε double deficient RAW264.7 cells were generated using CRISP/Cas9 technology. Cells were infected with PR8 virus with or without ERK inhibitor PD98059 to examine the phosphorylation of TBK1, IKKε, ERK, IRF3 by western blot analysis (G) or examine the expression of IFNβ by qPCR. Unpaired T-test was used for statistical analysis. *P < 0.05; **P < 0.01. Data are representative of two-three independent experiments with similar results.