We have previously shown that Mkp-1–deficient mice produce elevated TNF-α, IL-6, and IL-10 following systemic Escherichia coli infection, and they exhibited increased mortality, elevated bacterial burden, and profound metabolic alterations. To understand the function of Mkp-1 during bacterial infection, we performed RNA-sequencing analysis to compare the global gene expression between E. coli–infected wild-type and Mkp-1−/− mice. A large number of IFN-stimulated genes were more robustly expressed in E. coli–infected Mkp-1−/− mice than in wild-type mice. Multiplex analysis of the serum cytokine levels revealed profound increases in IFN-β, IFN-γ, TNF-α, IL-1α and β, IL-6, IL-10, IL-17A, IL-27, and GMSF levels in E. coli–infected Mkp-1−/− mice relative to wild-type mice. Administration of a neutralizing Ab against the receptor for type I IFN to Mkp-1−/− mice prior to E. coli infection augmented mortality and disease severity. Mkp-1−/− bone marrow–derived macrophages (BMDM) produced higher levels of IFN-β mRNA and protein than did wild-type BMDM upon treatment with LPS, E. coli, polyinosinic:polycytidylic acid, and herring sperm DNA. Augmented IFN-β induction in Mkp-1−/− BMDM was blocked by a p38 inhibitor but not by an JNK inhibitor. Enhanced Mkp-1 expression abolished IFN-β induction by both LPS and E. coli but had little effect on the IFN-β promoter activity in LPS-stimulated RAW264.7 cells. Mkp-1 deficiency did not have an overt effect on IRF3/7 phosphorylation or IKK activation but modestly enhanced IFN-β mRNA stability in LPS-stimulated BMDM. Our results suggest that Mkp-1 regulates IFN-β production primarily through a p38-mediated mechanism and that IFN-β plays a beneficial role in E. coli–induced sepsis.

The innate immune system acts as the first line of defense against invading bacterial pathogens (1). Mammals rely on germline-encoded pattern recognition receptors to detect bacterial components (2, 3). Recognition of the bacterial components activates a cascade of signaling events leading to activation of MAPK pathways and key transcription factors such as NF-κB (48). These transcription factors cooperate to initiate a transcriptional program by enhancing the expression of a variety of immune-related proteins, including proinflammatory cytokines, chemokines, and anti-inflammatory cytokines (7, 9). Some of the cytokines promote leukocyte recruitment and enhance cellular and humoral bactericidal activities (1014), whereas others restrain inflammation and limit the collateral damage to the host (15).

MAPK phosphatase-1 (Mkp-1), also referred to as DUSP1 (16), CL100 (17), 3CH134 (18), and Erp (19), is a dual-specificity protein phosphatase that preferentially acts on p38 and JNK MAPK subfamilies (20, 21). In innate immune cells, Mkp-1 is robustly induced in response to bacterial infection and serves as a negative regulator of the innate immune response (22, 23). We and others have shown that Mkp-1–deficient macrophages produce considerably greater amounts of cytokines, including TNF-α, IL-6, and IL-10 than do wild-type macrophages in vitro (2429). Manetsch et al. (30) found that knockdown of MKP-1 in TNF-α–stimulated human airway muscle cells enhanced both p38 and JNK activity and augmented IL-8 production, illustrating the importance of MKP-1 in the control of secondary inflammatory responses in stromal/parenchymal cells. In an Escherichia coli–induced sepsis model, Mkp-1−/− mice also produced markedly greater levels of cytokines, such as TNF-α, IL-6, and IL-10, and exhibited increased mortality (31). Increased bacterial burden, more severe organ damage, and metabolic abnormalities were also observed in Mkp-1−/− mice relative to wild-type mice after E. coli infection (31, 32). To understand the pathophysiology that Mkp-1−/− mice exhibit after systemic E. coli infection, we performed multiplex analysis of the serum cytokines in E. coli–infected wild-type and Mkp-1−/− mice. We observed enhanced cytokine production in Mkp-1−/− mice for 10 of the 13 cytokines tested. Among the cytokines enhanced by Mkp-1 deficiency in E. coli–infected mice is IFN-β, a type I IFN critical for host defense against viral infections (3335). Although regulation of IFN-β induction by TANK-binding kinase 1 (TBK1)–mediated IFN regulatory factors (IRFs) and NF-κB during viral infections has been well studied (3640), the regulation of IFN-β by Mkp-1 has not been fully understood, particularly in the context of bacterial infection. Type I IFN has been shown to play both beneficial and detrimental roles during bacterial infections, depending on the pathogens and mode of infections (4143). We found that, concurrent with a greater increase in IFN-β levels in the blood, many IFN-inducible genes were more robustly induced in Mkp-1−/− mice than in Mkp-1+/+ mice following E. coli infection. To address the role of Mkp-1 in the regulation of type I IFN, we studied the mechanism underlying Mkp-1–mediated regulation in macrophages using Mkp-1 knockout and overexpressing cells as well as pharmacological inhibitors. We also studied the physiological role of IFN-β during E. coli infection in Mkp-1−/− mice by using an IFN-α/β receptor 1 (IFNAR1) neutralizing Ab. Our studies indicate that Mkp-1 regulates IFN-β expression through controlling p38 but not JNK and that IFN-β induction is beneficial for the host during E. coli infection.

Mkp-1−/− mice have been described previously (25, 44) and have no obvious phenotype prior to experimental use. Mkp-1+/− mice on a C57/129 mixed background were generously provided by Bristol Myers Squibb Pharmaceutical Research Institute (Princeton, NJ). Mkp-1+/− mice were intercrossed to generate Mkp-1−/− and Mkp-1+/+ mice for E. coli infection experiments. Additionally, the Mkp-1+/− mice were backcrossed to C57BL/6J mice for eight generations to create Mkp-1−/− mice on a C57BL/6J background. Although Mkp-1−/− and Mkp-1+/+ mice on C57/129 background were used for all infection experiments, all macrophage studies in vitro were carried out using bone marrow isolated from the mice on C57BL/6J background. All mice were housed with a 12-h alternating light–dark cycle at 25°C with humidity between 30 and 70% and have access to food and water ad libitum. All experiments were performed according to National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committee at the Research Institute at Nationwide Children’s Hospital.

A wild-type (smooth) strain of E. coli (O55:B5, ATCC 12014) was purchased from American Type Culture Collection (Manassas, VA). Bacteria were grown in nutrient broth for 18 h at 37°C and refreshed by culturing in new medium for 2 h after a 1:5 dilution. Bacteria were washed three times with sterile PBS and adjusted to the appropriate final concentration. The bacterial suspension was injected into the tail vein of the mice at the volume of ∼250 μl per mouse, as previously described (31, 32). Mouse survival was monitored for 7 d. In the Ab neutralization experiments, mice were first given i.p. 100 μg of InVivoPlus murine monoclonal anti-mouse IFNAR1 Ab (catalog no. BP0241) or InVivoPlus mouse IgG1 isotype control Ab (catalog no. BP0083) purchased from Bio X Cell (Lebanon, NH). The mice were then i.v. infected with E. coli 1 h later. Mortality was monitored over 7 d. Disease severity was assessed using a sepsis morbidity scoring system (Table I), which was adopted and refined from the murine sepsis score system developed by Shrum et al. (45). This score system evaluates morbidity in seven categories: appearance, level of consciousness, activity, response to stimuli, eye state, respiration rate, and quality. Each of these categories was given a score between 0 and 4. The individual scores in all categories were added together to yield the total score for a specific animal at the time of examination. Mice that died or were in moribundity (euthanized) at the time of evaluation were given a maximal score of 28.

Bacterial burden was determined 24 h postinfection, as previously described (31). Bacterial colonies were counted separately for each sample. Spleen samples were normalized to organ weight, and blood samples were normalized to blood volume.

Mkp-1+/+ and Mkp-1−/− mice were infected i.v. with E. coli at the dose of 2.5 × 107 CFU/g body weight (b.w.). Livers were isolated 24 h postinfection, and total RNA was isolated from four animals in each treatment group for RNA-sequencing (RNA-seq) analysis (32). The RNA-seq data have been deposited in Gene Expression Omnibus (GSE122741). A comprehensive list of 71 known IFN-stimulated genes (Isg) was compiled, and the transcript copy numbers were used to calculate the fold changes and p values using a t test. The fold change of transcripts for each gene was calculated relative to the average expression in control Mkp-1+/+ mice (injected with PBS i.v.). The dataset was then sorted from the highest level to the lowest level. Values were log2 transformed to generate a heatmap, in which red indicates upregulation, white indicates no change, and blue indicates downregulation of gene expression.

IFN concentrations in blood and cell culture medium were measured by ELISA following a standard protocol (46) with minor modifications. Briefly, wells on 96-well plates were coated with an IFN-β capture Ab (catalog no. 519202; BioLegend, San Diego, CA) diluted in PBS overnight at 4°C. The wells were washed three times with PBS and then blocked with ELISA diluent (PBS containing 10% FBS) for 1 h at room temperature. Subsequently, the wells were washed three times with PBS, and the adequately diluted samples (serum or cell culture media) and Mouse IFN-β Standard (BioLegend) were added into the wells to allow incubation at room temperature for 2 h. The samples were aspirated, and the wells were washed five times with PBS. The detection Ab (catalog no. 32400-1; PBL Assay Science, Piscataway, NJ) was diluted to a concentration of 50 neutralization units per ml and was added to the wells and allowed to incubate at room temperature for 1 h. After five washes with PBS, HRP-conjugated goat anti-rabbit IgG (catalog no. 111-035-144; Jackson ImmunoReasearch Laboratories, West Grove, PA) was diluted by 5,000-fold in PBS containing 10% FBS and was added to the wells and allowed to incubate for 30 min at room temperature. After seven washes with PBS, color was developed using 1× 3,3′,5,5′-tetramethylbenzidine solution (Pierce Biotechnology, Rockford, IL) and the absorbance was measured at 450 nm using the SpectraMax M2 microplate reader (Molecular Devices, Sunnyvale, CA). The IFN-β concentration was calculated based on the standard curve using the SoftMax Pro program (Molecular Devices).

Multiplex cytokine assessment for mouse sera was carried out using a LEGENDplex multiplex kit (BioLegend), according to the manufacturer’s recommendations. We used a predefined mouse inflammation panel to quantify 13 mouse cytokines (GM-CSF, IFN-β, IFN-γ, IL-1α, IL-1β, IL-6, IL-10, IL-12 p70, IL-17A, IL-23, IL-27, MCP-1, and TNF-α).

Bone marrow was isolated from Mkp-1+/+ and Mkp-1−/− mice on the C57BL/6J background, and the RBCs were lysed by incubating in ACK Lysing Buffer (Invitrogen, Carlsbad, CA) for 2–3 min. The remaining bone marrow cells were cultured on petri dishes in DMEM supplemented with 10% FBS (Atlantic Biologicals, Flowery Branch, GA), 25% L929-conditioned medium, 10 mM HEPES buffer, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10 μg/ml gentamicin (Invitrogen). The cells were refed once with fresh medium after 4 d and cultured for three additional days to generate bone marrow–derived macrophages (BMDM).

BMDM were stimulated with LPS (O55:B5) (Calbiochem, San Diego, CA) or heat-killed E. coli for different times. Stimulation of BMDM with synthetic dsRNA polyinosinic:polycytidylic acid [poly(I:C)] (Invivogen, San Diego, CA) and sonicated herring sperm DNA (Sigma-Aldrich, St. Louis, MO) was carried out by transfection with polyethylenimine (Polysciences, Warrington, PA), as previously described (47). In some experiments, BMDM were pretreated with either vehicle (DMSO) or a pharmacological inhibitor of p38 [SB203580 (48); Calbiochem] or JNK inhibitor 8 (JNK-IN-8) (49) (Selleck Chemicals, Houston, TX) for 15 min prior to TLR ligand or E. coli stimulation. Medium was harvested for ELISA, and cells were lysed to harvest proteins for Western blot analysis or to harvest total RNA for quantitative RT-PCR (qRT-PCR) as previously described (50).

RAW264.7 cells were modified using the piggyBac expression system (51) (System Biosciences, Palo Alto, CA) to express rat Mkp-1 protein under a tetracycline-inducible (Tet-ON) promoter (52). We incorporated the one-vector tetracycline-inducible expression feature of the plasmid pCW57.1 MCS1-P2A-MCS2 (neo) (53) into the piggyBac vector to create a new vector, PB-SK2. The PB-SK2 vector carries two tandem expression cassettes sandwiched by two specific inverted terminal repeats (ITRs). The first cassette expresses reverse tetracycline-controlled transactivator (rtTA) and the neomycin phosphotransferase separated by a Thosea asigna virus 2A peptide bond-skipping sequence. The second expression cassette harbors a transgene under a promoter containing five tetracycline response elements. In the presence of doxycycline (Dox), the reverse tetracycline-controlled transactivator transcription factor generated by the first expression cassette will turn on the expression of the transgene in the second expression cassette. The neomycin phosphotransferase confers G418 resistance to the cells. When cotransfected with a hyperactive piggyBac transposase vector, the transposase can bind to the specific ITRs of the piggyBac vector and excise the ITR-flanked expression cassettes and insert into the genome at TTAA sites (51). We cloned the rat Mkp-1 cDNA into the EcoRI site downstream of a tetracycline response element–driven promoter. The authenticity of the constructs was confirmed by DNA sequencing. The empty piggyBac vector PB-SK2 or PB-SK2 containing the rat Mkp-1 cDNA was cotransfected with a hyperactive piggyBac transposase vector (51) into RAW264.7 cells, using Lipofectamine 3000 (Invitrogen). After transfection, the cells were selected for 2 wk in medium containing 500 μg/ml G418. Individual clones were isolated to test for the expression of Mkp-1 in the absence and presence of Dox. The leftover clones were pooled. Adequate RAW264.7 derivative clones and the pools were maintained in DMEM supplemented with 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 150 μg/ml G418 (Invitrogen). These cells were cultured in medium with or without Dox overnight and then stimulated with heat-killed E. coli or LPS to harvest cell lysates for Western blot analysis or collect culture medium for ELISA.

In the luciferase assay experiment, we transfected RAW264.7 cell clone (clone 22), which was stably integrated with a Tet-ON Mkp-1 expression cassette, with a hyperactive piggyBac transposase vector and a piggyBac vector carrying a luciferase reporter linked to the proximal mouse IFN-β promoter (nucleotides −53 to −195) (54). The cells were then selected with puromycin for 2 wk. The cell pool was then treated with Dox (100 ng/ml) overnight or left untreated. These cells were then stimulated with LPS or heat-killed E. coli for 6 h or left unstimulated. Cells were then washed with PBS and lysed to measure luciferase activity in the cell lysates using a Renilla luciferase assay system (Promega, Madison, WI), according to the manufacturer’s recommendations.

Western blot analysis was carried out as previously described (24, 55). The rabbit monoclonal Abs against phospho-IRF3 (Ser396), total IRF3, phospho-IRF7 (Ser437/538), phospho-TBK1 (Ser172), total TBK1, phospho-IκB kinase (IKK) α/β (Ser176/180), and Mkp-1 were purchased from Cell Signaling Technology (Danvers, MA). The mouse monoclonal Abs against radical S-adenosyl methionine domain-containing 2 (Rsad2), Isg15, IκBα, and fatty acid synthase (Fasn), as well as the polyclonal Ab against IKKα/β, were purchased from Santa Cruz Biotechnology (Dallas, Texas). The mouse mAb against β-actin was purchased from Sigma-Aldrich. The immunoblots were stripped and reprobed with an Ab against a housekeeping protein to control for loading. Western blots were developed using chemiluminescent reagent Immobilon ECL (Millipore, Billerica, MA). Western blot images were acquired using Epson Perfection 4990 PHOTO scanner (Epson, Long Beach, CA).

BMDM were stimulated with heat-killed E. coli (O55:B5, ATCC 12014) or various TLR ligands for different amounts of time. For E. coli stimulation, heat-killed E. coli were added to cell culture plates at a ratio of 10 bacteria per macrophage, then the plates were centrifuged in swinging buckets for 2 min at 2,000 rpm at 37°C. Total RNA samples were harvested from the cells using TRIzol. Genomic DNA was removed by digesting the total RNA with RQ1 RNase-Free DNase (Promega, Madison, WI). Liver RNA was then reverse transcribed on a PTC-200 DNA Engine Cycler (Bio-Rad Laboratories, Hercules, CA) with a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA). qRT-PCR was performed using PowerUp SYBR Green PCR Master Mix (Applied Biosystems) on a Realplex2 Mastercycler (Eppendorf, Hauppauge, NY).

For measuring IFN-β mRNA decay, BMDM were stimulated with heat-killed E. coli for 2 h. Actinomycin D was added into the culture medium at a concentration of 5 μg/ml. Total RNA was harvested from the cells using TRIzol (Invitrogen) after 0, 2, 4, and 8 h. IFN-β mRNA levels were quantified by qRT-PCR to assess IFN-β mRNA decay using primers 5′-GCCAGGAGCTTGAATAAAATG-3′ and 5′-GATGGTCCTTTCTGCCTCAG-3′, as previously described (46, 47). 18S rRNA was detected using 5′-GTAACCCGTTGAACCCCATT-3′ and 5′-CCATCCAATCGGTAGTAGCG-3′ and used as an internal control for normalization. The levels of IFN-β mRNA expression were calculated relative to 18S using the 2−ΔΔCT method (36).

Survival differences between groups were analyzed by Kaplan–Meier analysis with a log-rank test using the online statistics program (http://www.obg.cuhk.edu.hk/ResearchSupport/StatTools/Survival_Pgm.php) developed by Dr. A. Chang, Department of Obstetrics and Gynecology, Chinese University of Hong Kong. Differences in cytokine production or gene expression between groups were compared using a t test or two-way ANOVA with GraphPad Prism 8.2.0 (GraphPad Software, San Diego, CA). A p value <0.05 was considered statistically significant for all analyses.

Previously, we have found that Mkp-1−/− mice exhibited a profound defect in host defense against E. coli infection, indicated by substantial increases in mortality, bacterial burden, and organ damage associated with increased production of TNF-α, IL-6, and IL-10 compared with wild-type mice. To gain insight into the physiological function of Mkp-1 in sepsis after systemic E. coli infection, we analyzed the RNA-seq datasets (GSE122741) generated using livers of control and E. coli–infected Mkp-1+/+ and Mkp-1−/− mice (32). We noticed a profound enhancement of an IFN genetic response signature, although neither type I IFN nor type II IFN mRNA(s) were expressed in the livers of these mice. IFNs evoke a unique genetic program via inducing the expression of many Isgs. We compiled the mRNA transcript levels of all known Isgs that were expressed in the livers (Supplemental Table I), log-transformed the values, and generated a heatmap (Fig. 1). Nearly 60 of the 71 Isgs were upregulated in Mkp-1+/+ mice upon E. coli infection. Forty of the seventy-one Isgs, including myxovirus resistance (Mx) 1, Mx2, 2′-5′-oligoadenylate synthase-like protein 1 (Oasl1), Rsad2, nicotinamide phosphoribosyltransferase (Nampt), Isg15, Isg20, and ubiquitin-specific peptidase 18 (Usp18), were expressed at higher levels in Mkp-1−/− mice than in Mkp-1+/+ mice following E. coli infection. Western blot analysis confirmed the enhanced protein expression for Rsad2 (Fig. 2A, 2B) but not Isg15 (Fig. 2C), suggesting that the transcription of these IFN-regulated genes is not the only mechanism controlling their protein levels.

FIGURE 1.

Mkp-1 deficiency dramatically enhanced hepatic expression of IFN-responsive genes in E. coli–infected mice. Mkp-1−/− and Mkp-1+/+ mice on a C57/129 background were infected i.v. with live E. coli at a dose of 2.5 × 107 CFU/g b.w. or injected with PBS (controls). Mice were euthanized after 24 h, and total RNA was isolated from the livers using TRIzol for RNA-seq analyses. The copy numbers of RNA transcripts for each IFN-responsive gene were normalized to the average number in wild-type controls to calculate fold change. The IFN-responsive genes were ranked based on the fold change in RNA transcripts in wild-type mice following E. coli infection. These values were log2 transformed to generate the heat map. Log transformation allows for a greater scale in the heat map. When the transcript number is 0 for a given gene in a specific animal, we gave an arbitrary number that is lower than the lowest value in that group. Each column represents a distinct animal. *p < 0.05, comparing E. coli–infected Mkp-1−/− and E. coli–infected Mkp-1+/+ mice by t test (n = 4).

FIGURE 1.

Mkp-1 deficiency dramatically enhanced hepatic expression of IFN-responsive genes in E. coli–infected mice. Mkp-1−/− and Mkp-1+/+ mice on a C57/129 background were infected i.v. with live E. coli at a dose of 2.5 × 107 CFU/g b.w. or injected with PBS (controls). Mice were euthanized after 24 h, and total RNA was isolated from the livers using TRIzol for RNA-seq analyses. The copy numbers of RNA transcripts for each IFN-responsive gene were normalized to the average number in wild-type controls to calculate fold change. The IFN-responsive genes were ranked based on the fold change in RNA transcripts in wild-type mice following E. coli infection. These values were log2 transformed to generate the heat map. Log transformation allows for a greater scale in the heat map. When the transcript number is 0 for a given gene in a specific animal, we gave an arbitrary number that is lower than the lowest value in that group. Each column represents a distinct animal. *p < 0.05, comparing E. coli–infected Mkp-1−/− and E. coli–infected Mkp-1+/+ mice by t test (n = 4).

Close modal
FIGURE 2.

Increased Rsad2 protein expression and serum IFN-β production in Mkp-1−/− mice relative to Mkp-1+/+ mice following E. coli infection. Mkp-1−/− and Mkp-1+/+ mice on a C57/129 background were infected i.v. with live E. coli at a dose of 2.5 × 107 CFU/g b.w. or injected with PBS (controls). Mice were euthanized after 24 h to harvest blood and liver. Serum samples were used for measuring IFN-β by ELISA. Liver tissues were homogenized to extract protein for Western blot analysis using Rsad2 or Isg15 Ab. (A) Levels of Rsad2 proteins in the livers of control and E. coli–infected mice. Each lane represents a different animal. The same set of samples was also analyzed by Western blotting using a mouse mAb against a housekeeping protein, Fasn, to verify comparable loading (lower panel). Liver Rsad2 (B) and Isg15 (C) protein levels in control and E. coli–infected Mkp-1−/− and Mkp-1+/+ mice. Same membranes were stripped and reprobed with a β-actin Ab. Images shown in (B) and (C) are representative Western blotting results.

FIGURE 2.

Increased Rsad2 protein expression and serum IFN-β production in Mkp-1−/− mice relative to Mkp-1+/+ mice following E. coli infection. Mkp-1−/− and Mkp-1+/+ mice on a C57/129 background were infected i.v. with live E. coli at a dose of 2.5 × 107 CFU/g b.w. or injected with PBS (controls). Mice were euthanized after 24 h to harvest blood and liver. Serum samples were used for measuring IFN-β by ELISA. Liver tissues were homogenized to extract protein for Western blot analysis using Rsad2 or Isg15 Ab. (A) Levels of Rsad2 proteins in the livers of control and E. coli–infected mice. Each lane represents a different animal. The same set of samples was also analyzed by Western blotting using a mouse mAb against a housekeeping protein, Fasn, to verify comparable loading (lower panel). Liver Rsad2 (B) and Isg15 (C) protein levels in control and E. coli–infected Mkp-1−/− and Mkp-1+/+ mice. Same membranes were stripped and reprobed with a β-actin Ab. Images shown in (B) and (C) are representative Western blotting results.

Close modal

To understand the molecular basis for the enhanced IFN signature in E. coli–infected Mkp-1−/− mice, we measured the levels of 13 cytokines, including IFN-β, IFN-γ, TNF-α, IL-1α, IL-1β, IL-6, IL-10, IL-12p70, IL-17A, IL-23, IL-27, MCP-1, and GM-CSF, in the sera of Mkp-1+/+ and Mkp-1−/− mice prior and following E. coli infection. We infected Mkp-1+/+ and Mkp-1−/− mice i.v. with E. coli (2.5 × 107 CFU/g b.w.) and analyzed the cytokine levels after 3, 6, or 24 h using a multiplex cytokine assay kit. As previously described (31), Mkp-1−/− mice produced substantially more TNF-α, IL-6, and IL-10 than did Mkp-1+/+ mice following E. coli infection at all three timepoints (Fig. 3A–C). The differences in IL-6 and IL-10 at 24 h were particularly striking. Although IL-6 and IL-10 returned to close to basal levels in Mkp-1+/+ mice at 24 h, IL-6 and, particularly, IL-10 levels remained at very high levels in Mkp-1−/− mice at 24 h. At this point, the IL-6 and IL-10 levels in Mkp-1−/− mice were 3- and 31-fold higher than in Mkp-1+/+ mice, respectively. In addition, IL-17A and IL-27 production in Mkp-1−/− mice were also dramatically increased relative to Mkp-1+/+ mice following E. coli infection at both early and late timepoints (Fig. 3D, 3E). Interestingly, three cytokines, IL-1α, IL-1β, and GM-CSF, were similar in the two groups of mice at early timepoints (3 and 6 h) postinfection but diverged in different directions by 24 h (Fig. 3F–H). At that point, the levels of these cytokines were substantially decreased in Mkp-1+/+ mice but further increased (such as IL-1α) or persisted at the peak levels (GM-CSF and IL-1β) in Mkp-1−/− mice. Although serum MCP-1 levels were dramatically increased in both Mkp-1+/+ and Mkp-1−/− mice following E. coli infection, there was no significant difference between the two groups of E. coli–infected mice (data not shown). IL-23 levels did not change in either Mkp-1+/+ or Mkp-1−/− mice after E. coli infection (data not shown). IL-12p70 levels were only slightly increased in Mkp-1−/− mice after E. coli infection, and there were no differences between the two groups of E. coli–infected mice (data not shown). Interestingly, the levels of IFN-γ were very low in uninfected mice but dramatically increased 6 h after E. coli infection in both Mkp-1+/+ and Mkp-1−/− mice (Fig. 3I). Although IFN-γ levels in Mkp-1−/− mice were significantly higher than in Mkp-1+/+ mice 6 h after E. coli infection, IFN-γ levels in the Mkp-1+/+ mice plummeted to nearly basal levels by 24 h. In contrast, elevated levels of IFN-γ in E. coli–infected Mkp-1−/− mice persisted at 24 h postinfection. Although IFN-β levels in Mkp-1+/+ mice did not significantly change, IFN-β levels at both 3 h and 24 h after E. coli infection in Mkp-1−/− mice were significantly higher than in infected Mkp-1+/+ mice (Fig. 3J). We also developed an in-house sandwich ELISA to verify the observed differences in serum IFN-β levels between E. coli infected Mkp-1+/+ and Mkp-1−/− mice (Fig. 3K). The ELISA assays showed that IFN-β production was induced by E. coli infection in both Mkp-1+/+ and Mkp-1−/− mice. Although blood IFN-β levels in uninfected Mkp-1+/+ and Mkp-1−/− mice were comparable, serum IFN-β levels 24 h after E. coli infection were over 3-fold higher in Mkp-1−/− mice than in Mkp-1+/+ mice (Fig. 3K).

FIGURE 3.

Mkp-1 deficiency resulted in greater serum cytokine levels in E. coli–infected mice. Mkp-1−/− and Mkp-1+/+ mice on a C57/129 background were infected i.v. with live E. coli at a dose of 2.5 × 107 CFU/g b.w. or left uninfected. Mice were euthanized at 3, 6, and 24 h postinfection (n = 8 for all groups). Uninfected mice (n = 10 for both groups) were also euthanized and regarded as 0 timepoint. The blood was collected through cardiac puncture, and serum cytokine levels were measured using a LEGENDplex inflammation kit (AI) to quantify 13 predefined mouse cytokines. Cytokine levels are presented as mean ± S.E. in the graphs. Only cytokines that displayed significant differences between the two groups of mice are presented in the graphs. The filled circles represent Mkp-1+/+ group, and the filled squares represent Mkp-1−/− group. (A) TNF-α, (B) IL-6, (C) IL-10, (D) IL-17A, (E) IL-27, (F) IL-1α, (G) IL-1β, (H) GM-CSF, (I) IFN-γ, (J) IFN-β, (K) serum IFN-β concentration measured by ELISA. *p < 0.05, compared with cytokine levels in E. coli–infected Mkp-1+/+ mice at the same timepoint by t test (n = 5–10); #p < 0.05, compared with control of the same phenotype by t test (n = 3-10).

FIGURE 3.

Mkp-1 deficiency resulted in greater serum cytokine levels in E. coli–infected mice. Mkp-1−/− and Mkp-1+/+ mice on a C57/129 background were infected i.v. with live E. coli at a dose of 2.5 × 107 CFU/g b.w. or left uninfected. Mice were euthanized at 3, 6, and 24 h postinfection (n = 8 for all groups). Uninfected mice (n = 10 for both groups) were also euthanized and regarded as 0 timepoint. The blood was collected through cardiac puncture, and serum cytokine levels were measured using a LEGENDplex inflammation kit (AI) to quantify 13 predefined mouse cytokines. Cytokine levels are presented as mean ± S.E. in the graphs. Only cytokines that displayed significant differences between the two groups of mice are presented in the graphs. The filled circles represent Mkp-1+/+ group, and the filled squares represent Mkp-1−/− group. (A) TNF-α, (B) IL-6, (C) IL-10, (D) IL-17A, (E) IL-27, (F) IL-1α, (G) IL-1β, (H) GM-CSF, (I) IFN-γ, (J) IFN-β, (K) serum IFN-β concentration measured by ELISA. *p < 0.05, compared with cytokine levels in E. coli–infected Mkp-1+/+ mice at the same timepoint by t test (n = 5–10); #p < 0.05, compared with control of the same phenotype by t test (n = 3-10).

Close modal

To address the significance of elevated type I IFN, IFN-β, in the phenotype of Mkp-1−/− mice during E. coli infection, we blocked IFNAR1 with a neutralizing Ab and documented survival (Fig. 4). Although prophylactic neutralizing of IFNAR1 with a mAb appeared to increase mortality, the difference compared with the mortality after treatment with an isotype control Ab did not reach statistical significance. We then assessed the effect of IFNAR1 blockade on disease severity using a comprehensive murine sepsis scoring system (Table I) that evaluates appearance, consciousness, activity, response to stimuli, eyes, and respiratory rate and quality. We found that the IFNAR1 neutralizing Ab augmented disease severity in Mkp-1−/− mice following E. coli infection (Fig. 4B). Neutralization of IFNAR1 by the Ab almost completely abolished the increase in liver Rsad2 protein level in Mkp-1−/− mice triggered by E. coli infection (Fig. 4C), confirming the blockade of type I IFN signaling in mice that received the IFNAR1-neutralizing Ab. To our surprise, IFNAR1 neutralization had neither a significant effect on serum TNF-α or IL-6 levels (Fig. 4D) nor affected bacterial burdens in the blood or spleens (Fig. 4E). These results demonstrate that increased IFN-β is actually beneficial to the mice, suggesting that exacerbating IFN-β production is not responsible for the enhanced mortality and bacterial burden of Mkp-1−/− mice after E. coli infection.

FIGURE 4.

Neutralizing IFN-β exacerbated the severity of disease in E. coli–infected Mkp-1−/− mice. Mkp-1−/− mice on the C57/129 background were given i.p. 100 μg (per mouse) of either a monoclonal anti-mouse IFNAR1 or an isotype control (IgG1) Ab 1 h prior to E. coli infection. These mice were subsequently infected i.v. with live E. coli (O55:B5) at a dose of 3.2 × 106 CFU/g b.w. Mice were monitored for 7 d to evaluate mortality and morbidity. To assess IFNAR1 neutralization on liver protein levels, blood cytokines, and bacterial burdens, mice were sacrificed 24 h after E. coli infection. Livers, blood, and spleens were collected aseptically. (A) Survival curves of the Mkp-1−/− mice receiving either the isotype control or the anti-IFNAR1 Ab. (B) Morbidity scores for the Mkp-1−/− mice receiving either the isotype control or the anti-IFNAR1 Ab. The morbidity scores increased over time in both groups and were significantly greater in the group that got anti-IFNAR1 Ab than in the group that got isotype control. *p < 0.05, comparing the two groups over time by two-way ANOVA. In (A) and (B): nIsotype ctrl = 9; nIFNAR1 = 11. p < 0.05 (group) and p < 0.001 (time), by two-way ANOVA, indicating that the morbidity scores were significantly different for both, and there was a significant interaction between group and time, p < 0.001. (C) IFNAR-1 neutralizing Ab blocks Rsad2 induction in E. coli–infected mice. Livers were homogenized to extract soluble proteins for Western blot analysis using a mAb against Rsad2. The membrane was stripped and blotted with a mAb against Fasn. (D) Serum TNF-α and IL-6 levels. Serum TNF-α and IL-6 levels were measured by ELISA. Results represent the means ± SE (n = 6). (E) Bacterial burdens. Bacterial load in the blood and spleen homogenates were determined by culture. Each dot represents an individual animal. Horizontal line represents mean value of CFU.

FIGURE 4.

Neutralizing IFN-β exacerbated the severity of disease in E. coli–infected Mkp-1−/− mice. Mkp-1−/− mice on the C57/129 background were given i.p. 100 μg (per mouse) of either a monoclonal anti-mouse IFNAR1 or an isotype control (IgG1) Ab 1 h prior to E. coli infection. These mice were subsequently infected i.v. with live E. coli (O55:B5) at a dose of 3.2 × 106 CFU/g b.w. Mice were monitored for 7 d to evaluate mortality and morbidity. To assess IFNAR1 neutralization on liver protein levels, blood cytokines, and bacterial burdens, mice were sacrificed 24 h after E. coli infection. Livers, blood, and spleens were collected aseptically. (A) Survival curves of the Mkp-1−/− mice receiving either the isotype control or the anti-IFNAR1 Ab. (B) Morbidity scores for the Mkp-1−/− mice receiving either the isotype control or the anti-IFNAR1 Ab. The morbidity scores increased over time in both groups and were significantly greater in the group that got anti-IFNAR1 Ab than in the group that got isotype control. *p < 0.05, comparing the two groups over time by two-way ANOVA. In (A) and (B): nIsotype ctrl = 9; nIFNAR1 = 11. p < 0.05 (group) and p < 0.001 (time), by two-way ANOVA, indicating that the morbidity scores were significantly different for both, and there was a significant interaction between group and time, p < 0.001. (C) IFNAR-1 neutralizing Ab blocks Rsad2 induction in E. coli–infected mice. Livers were homogenized to extract soluble proteins for Western blot analysis using a mAb against Rsad2. The membrane was stripped and blotted with a mAb against Fasn. (D) Serum TNF-α and IL-6 levels. Serum TNF-α and IL-6 levels were measured by ELISA. Results represent the means ± SE (n = 6). (E) Bacterial burdens. Bacterial load in the blood and spleen homogenates were determined by culture. Each dot represents an individual animal. Horizontal line represents mean value of CFU.

Close modal
Table I.

Morbidity sepsis score to assess the severity of disease

VariableScoreDescription
Appearance Coat is smooth 
Patches of hair are piloerected 
Majority of back is piloerected 
Piloerection may or may not be present; mouse appears “puffy” 
Piloerection may or may not be present; mouse appears emaciated 
Level of consciousness Mouse is active 
Mouse is active but avoids standing upright 
Mouse activity is noticeably slowed. The mouse is still ambulant 
Activity is impaired. Mouse only moves when provoked; movements have a tremor 
Activity is severely impaired. Mouse remains stationary when provoked with possible tremor 
Activity Normal amount of activity, mouse is doing any of the following: eating, drinking, climbing, running, fighting 
Slightly suppressed activity, mouse is moving around bottom of cage 
Suppressed activity, mouse is stationary with occasional investigative movements 
No activity, mouse is stationary 
No activity, mouse is experiencing tremors, particularly in the hind legs 
Response to stimulus Mouse responds immediately to auditory stimulus or touch 
Slow or no response to auditory stimulus, mouse shows strong response to touch (moves to escape) 
No response to auditory stimulus, mouse shows moderate response to touch (moves a few steps) 
No response to auditory stimulus, mouse shows mild response to touch (no locomotion) 
No response to auditory stimulus, mouse shows little or no response to touch and cannot right itself if pushed over 
Eyes Open 
Eyes not fully open, possibly with secretions 
One eye at least half closed, possibly with secretions 
Both eyes half closed or more, possibly with secretions 
Eyes closed or milky 
Respiration rate Normal, rapid mouse respiration 
Slightly decreased respiration (rate not quantifiable by eye) 
Moderately reduced respiration (rate at the upper range of quantifying by eye) 
Severely reduced respiration (rate easily countable by eye, 0.5 s between breaths) 
Extremely reduced respiration (>1 s between breaths) 
Respiration quality Normal 
Brief periods of labored breathing 
Labored, no gasping 
Labored with intermittent gasps 
Gasping 
VariableScoreDescription
Appearance Coat is smooth 
Patches of hair are piloerected 
Majority of back is piloerected 
Piloerection may or may not be present; mouse appears “puffy” 
Piloerection may or may not be present; mouse appears emaciated 
Level of consciousness Mouse is active 
Mouse is active but avoids standing upright 
Mouse activity is noticeably slowed. The mouse is still ambulant 
Activity is impaired. Mouse only moves when provoked; movements have a tremor 
Activity is severely impaired. Mouse remains stationary when provoked with possible tremor 
Activity Normal amount of activity, mouse is doing any of the following: eating, drinking, climbing, running, fighting 
Slightly suppressed activity, mouse is moving around bottom of cage 
Suppressed activity, mouse is stationary with occasional investigative movements 
No activity, mouse is stationary 
No activity, mouse is experiencing tremors, particularly in the hind legs 
Response to stimulus Mouse responds immediately to auditory stimulus or touch 
Slow or no response to auditory stimulus, mouse shows strong response to touch (moves to escape) 
No response to auditory stimulus, mouse shows moderate response to touch (moves a few steps) 
No response to auditory stimulus, mouse shows mild response to touch (no locomotion) 
No response to auditory stimulus, mouse shows little or no response to touch and cannot right itself if pushed over 
Eyes Open 
Eyes not fully open, possibly with secretions 
One eye at least half closed, possibly with secretions 
Both eyes half closed or more, possibly with secretions 
Eyes closed or milky 
Respiration rate Normal, rapid mouse respiration 
Slightly decreased respiration (rate not quantifiable by eye) 
Moderately reduced respiration (rate at the upper range of quantifying by eye) 
Severely reduced respiration (rate easily countable by eye, 0.5 s between breaths) 
Extremely reduced respiration (>1 s between breaths) 
Respiration quality Normal 
Brief periods of labored breathing 
Labored, no gasping 
Labored with intermittent gasps 
Gasping 

To delineate the molecular mechanism by which Mkp-1 regulates IFN-expression, we analyzed the effects of Mkp-1 on IFN-β production in macrophages. First, we compared IFN-β production in Mkp-1+/+ and Mkp-1−/− BMDM following stimulation with heat-killed E. coli (Fig. 5A). Although E. coli stimulation resulted in IFN-β production in Mkp-1+/+ BMDM, substantially more IFN-β (>5-fold) was produced by Mkp-1−/− BMDM following E. coli stimulation than by Mkp-1+/+ BMDM. As LPS is an important pathogenic factor in Gram-negative bacteria and a potent stimulant for IFN-β production (56), we assessed the effect of Mkp-1 deficiency on LPS-stimulated IFN-β production (Fig. 5B). Similar to what was observed in E. coli–stimulated BMDM, Mkp-1 deficiency substantially enhanced IFN-β production in LPS-stimulated BMDM.

FIGURE 5.

Mkp-1–deficient BMDM produced significantly more IFN-β than did wild-type BMDM after LPS and E. coli stimulation. BMDM derived from Mkp-1−/− and Mkp-1+/+ mice on a C57BL/6J background were treated with heat-killed E. coli at a dose of 10 bacteria per macrophage or 100 ng/ml LPS (O55:B5). Medium was harvested at different timepoints to measure IFN-β concentration by ELISA. Total RNA was isolated from the cells to assess IFN-β mRNA levels by qRT-PCR. (A) IFN-β production by Mkp-1−/− and Mkp-1+/+ macrophages following E. coli stimulation. Values represent mean ± SE (n = 4). *p < 0.05, compared with Mkp-1+/+ macrophages at the same timepoint by t test. (B) IFN-β production by Mkp-1−/− and Mkp-1+/+ macrophages following LPS stimulation by t test. Values represent mean ± SE (n = 4). *p < 0.05, compared with Mkp-1+/+ macrophages at the same timepoint by t test. (C) Kinetics of IFN-β mRNA levels in macrophages stimulated with heat-killed E. coli. IFN-β mRNA expression was presented as fold change relative to the control cells. Values represent mean ± SE (n = 3). IFN-β mRNA significantly changed over time in both genotypes, and those changes were significantly greater in the Mkp-1−/− mice than in the Mkp-1+/+ mice. Both time and genotype were significantly different, *p < 0.001 (by two-way ANOVA), and there was an interaction between genotype and time, p < 0.001.

FIGURE 5.

Mkp-1–deficient BMDM produced significantly more IFN-β than did wild-type BMDM after LPS and E. coli stimulation. BMDM derived from Mkp-1−/− and Mkp-1+/+ mice on a C57BL/6J background were treated with heat-killed E. coli at a dose of 10 bacteria per macrophage or 100 ng/ml LPS (O55:B5). Medium was harvested at different timepoints to measure IFN-β concentration by ELISA. Total RNA was isolated from the cells to assess IFN-β mRNA levels by qRT-PCR. (A) IFN-β production by Mkp-1−/− and Mkp-1+/+ macrophages following E. coli stimulation. Values represent mean ± SE (n = 4). *p < 0.05, compared with Mkp-1+/+ macrophages at the same timepoint by t test. (B) IFN-β production by Mkp-1−/− and Mkp-1+/+ macrophages following LPS stimulation by t test. Values represent mean ± SE (n = 4). *p < 0.05, compared with Mkp-1+/+ macrophages at the same timepoint by t test. (C) Kinetics of IFN-β mRNA levels in macrophages stimulated with heat-killed E. coli. IFN-β mRNA expression was presented as fold change relative to the control cells. Values represent mean ± SE (n = 3). IFN-β mRNA significantly changed over time in both genotypes, and those changes were significantly greater in the Mkp-1−/− mice than in the Mkp-1+/+ mice. Both time and genotype were significantly different, *p < 0.001 (by two-way ANOVA), and there was an interaction between genotype and time, p < 0.001.

Close modal

We then assessed the effect on Mkp-1 deficiency on IFN-β mRNA levels in E. coli–stimulated BMDM by qRT-PCR (Fig. 5C). E. coli stimulation resulted in an ∼400-fold increase in IFN-β mRNA levels at 1 h in Mkp-1+/+ BMDM, followed by a gradual decrease such that by 6 h after stimulation, IFN-β levels were ∼40-fold above basal levels. The increase in IFN-β mRNA levels was dramatically enhanced in Mkp-1−/− BMDM. Following E. coli stimulation, IFN-β mRNA levels increased to >2,000-fold over the basal level in Mkp-1+/+ macrophages within 1 h. IFN-β mRNA levels maintained and slightly increased at 2 h and then rapidly declined.

We also assessed whether overexpression of Mkp-1 inhibits IFN-β production. We established stable RAW264.7 clones (clones 15, 13, 21, 22) that express the rat Mkp-1 protein in a Dox-inducible manner. RAW264.7 is a macrophage-like mouse cell line that can produce a variety of cytokines in response to pathogenic stimulation (57). Dox treatment dramatically increased Mkp-1 expression in these cells (Fig. 6A, upper panel). It is worth noting that in the absence of Dox these Mkp-1–inducible RAW264.7 clones exhibited an elevated basal Mkp-1 level that was comparable with the endogenous Mkp-1 level 1 h after LPS stimulation in the parental RAW264.7 cells (Fig. 6A, middle panel). Interestingly, LPS stimulation further increased the levels of Mkp-1 expression in these clones. Importantly, Dox treatment substantially inhibited IFN-β production in two of the tested Mkp-1–expressing clones (clones 15 and 22) following both LPS and E. coli stimulation (Fig. 6B). As there were substantial variability between the individual cell clones (Fig. 6B, note the differences in scales of the y-axis between clone 15 and clone 22), we assessed IFN-β production using pools of the stable Dox-inducible clones. In the stable RAW264.7 pool transfected with an empty vector, both E. coli and LPS triggered a substantial increase in IFN-β production, and Dox pretreatment had no effect on IFN-β production (Fig. 6C, left graph). In contrast, in the pool stably transfected with the Tet-ON Mkp-1 plasmid, Dox-induced Mkp-1 expression dramatically inhibited IFN-β production, although IFN-β production was potently stimulated by either E. coli or LPS in the absence of Dox (Fig. 6C, right graph).

FIGURE 6.

Overexpression of Mkp-1 potently inhibited IFN-β production in LPS-stimulated RAW264.7 macrophages. RAW264.7 cells were stably transfected with a Tet-ON expression cassette to express a rat Mkp-1 protein. Stable clones represent individual colonies after drug selection. (A) Expression of Mkp-1 in different clones after treatment with or without LPS in the absence or presence of Dox. RAW264.7 cells or individual clones were pretreated with or without 500 ng/ml Dox overnight and stimulated with or without LPS (100 ng/ml) for 1 h. Cells were harvested for Western blot analysis. Upper panel represents a short exposure, and the lower panel is a longer exposure to show induction of endogenous Mkp-1 by LPS. (B) Inhibition of IFN-β production by Dox-induced Mkp-1. Cells (1 × 106) of clone 15 and clone 22 were cultured on 24-well plates in 1 ml of medium containing 0 or 100 ng/ml Dox overnight and then stimulated with 100 ng/ml LPS or heat-killed E. coli at a dose of 10 bacteria per macrophage for 6 h. Media were harvested for ELISA. Values represent mean ± SE (n = 4). *p < 0.05, compared with untreated cells. *p < 0.05, compared with cells treated with LPS or E. coli in the absence of Dox. (C) The effect of Dox pretreatment on IFN-β production in control and Tet-ON Mkp-1–expressing pool. RAW264.7 pools stably transfected with a Tet-ON Mkp-1 expression cassette (Tet-ON Mkp-1 pool) or an empty vector (control pool) were first pretreated with 0 or 100 ng/ml Dox overnight and then stimulated with 100 ng/ml LPS or heat-killed E. coli at a dose of 10 bacteria per macrophage for 6 h. *p < 0.05, compared with cells that received no Dox pretreatment (−Dox) and were not treated by either LPS or E. coli. Values were compared between groups by t test.

FIGURE 6.

Overexpression of Mkp-1 potently inhibited IFN-β production in LPS-stimulated RAW264.7 macrophages. RAW264.7 cells were stably transfected with a Tet-ON expression cassette to express a rat Mkp-1 protein. Stable clones represent individual colonies after drug selection. (A) Expression of Mkp-1 in different clones after treatment with or without LPS in the absence or presence of Dox. RAW264.7 cells or individual clones were pretreated with or without 500 ng/ml Dox overnight and stimulated with or without LPS (100 ng/ml) for 1 h. Cells were harvested for Western blot analysis. Upper panel represents a short exposure, and the lower panel is a longer exposure to show induction of endogenous Mkp-1 by LPS. (B) Inhibition of IFN-β production by Dox-induced Mkp-1. Cells (1 × 106) of clone 15 and clone 22 were cultured on 24-well plates in 1 ml of medium containing 0 or 100 ng/ml Dox overnight and then stimulated with 100 ng/ml LPS or heat-killed E. coli at a dose of 10 bacteria per macrophage for 6 h. Media were harvested for ELISA. Values represent mean ± SE (n = 4). *p < 0.05, compared with untreated cells. *p < 0.05, compared with cells treated with LPS or E. coli in the absence of Dox. (C) The effect of Dox pretreatment on IFN-β production in control and Tet-ON Mkp-1–expressing pool. RAW264.7 pools stably transfected with a Tet-ON Mkp-1 expression cassette (Tet-ON Mkp-1 pool) or an empty vector (control pool) were first pretreated with 0 or 100 ng/ml Dox overnight and then stimulated with 100 ng/ml LPS or heat-killed E. coli at a dose of 10 bacteria per macrophage for 6 h. *p < 0.05, compared with cells that received no Dox pretreatment (−Dox) and were not treated by either LPS or E. coli. Values were compared between groups by t test.

Close modal

Mkp-1 prefers p38 and JNK as substrates (21), and Mkp-1 deficiency resulted in a considerable prolongation of p38 and JNK activity in BMDM following LPS stimulation (2629). We assessed whether p38 and JNK are involved in the production of IFN-β in BMDM stimulated with E. coli and other TLR ligands. First, we assessed the effects of p38 and JNK inhibition on the production of IFN-β in BMDM stimulated with E. coli and LPS. Pretreatment of BMDM with a pharmacological inhibitor of p38, SB203580, substantially attenuated IFN-β production in both Mkp-1+/+ and Mkp-1−/− BMDM stimulated with E. coli (Fig. 7A). Although the JNK inhibitor, JNK-IN-8, alone had little effect, combination of both SB203580 and JNK-IN-8 had a greater inhibitory effect on E. coli–induced IFN-β production than did SB203580 alone in Mkp-1−/− BMDM but not in Mkp-1+/+ BMDM (Fig. 7A). Although SB203580 substantially inhibited LPS-stimulated IFN-β production in both Mkp-1+/+ and Mkp-1−/− BMDM, JNK-IN-8 had little effect (Fig. 7B). The addition of both JNK-IN-8 and SB203580 did not significantly enhance the inhibition of SB203580 alone on LPS-induced IFN-β production in either Mkp-1−/− or Mkp-1+/+ BMDM. These results clearly demonstrate that p38 is primarily responsible for the dramatic increase in IFN-β in stimulated Mkp-1−/− BMDM.

FIGURE 7.

Enhanced IFN-β induction by E. coli and TLR ligands caused by Mkp-1 deficiency is substantially inhibited by the pharmacological inhibitor of p38 but not JNK. Mkp-1−/− and Mkp-1+/+ BMDM were pretreated with DMSO (vehicle) or 10 μM SB203580 or 3 μM JNK-IN-8 for 30 min, then stimulated with heat-killed E. coli at a dose of 10 bacteria per macrophages or with various TLR ligands. Media were harvested for ELISA to measure IFN-β concentration. Cells were harvested to isolate total RNA for qRT-PCR analysis on IFN-β expression. (A) The effects of p38 and JNK inhibition on IFN-β production by Mkp-1+/+ and Mkp-1−/− BMDM following 6 h of E. coli stimulation. (B) The effects of p38 and JNK inhibition on IFN-β production by Mkp-1+/+ and Mkp-1−/− BMDM following 6 h of LPS stimulation (100 ng/ml). (C) Kinetics of IFN-β production by E. coli–stimulated Mkp-1−/− and Mkp-1+/+ BMDM in the presence and absence of p38 inhibitor. BMDM (1 × 106 cells) were cultured in 1 ml of medium in 12-well plates. (D) Kinetics of IFN-β mRNA induction in Mkp-1−/− and Mkp-1+/+ BMDM by E. coli in the presence and absence of p38 inhibitor. (E) The effect of SB203580 on the IFN-β mRNA expression in Mkp-1−/− and Mkp-1+/+ BMDM stimulated with E. coli or LPS for 3 h. BMDM (9 × 105 cells) were cultured in 3 ml of medium in 60-mm plates. (F) The effect of SB203580 on the IFN-β mRNA expression in Mkp-1−/− and Mkp-1+/+ BMDM transfected with 1.8 μg of poly(I:C) or 1.8 μg of herring sperm DNA for 3 h. BMDM (9 × 105 cells) were cultured with 3 ml of medium in 60-mm plates were transfected with the indicated amounts of nucleic acids using polyethylenimine. Data in (A), (B), and (C) are presented as mean ± SE (n = 3) of the IFN-β concentrations. Data in (D) and (E) are presented as mean ± SE (n = 3) of fold change over the average levels in unstimulated Mkp-1+/+ cells. Data in (F) are presented as mean ± SE (n = 3) of fold change over the average levels in mock-transfected (with polyethylenimine) Mkp-1+/+ macrophages. *p < 0.05, compared with E. coli or TLR ligand-stimulated Mkp-1+/+ cells in their respective treatment group by t test; p < 0.05, compared with E. coli or TLR ligand-stimulated Mkp-1−/− cells in their respective treatment group by t test; #p < 0.05, compared with SB203580-pretreated, E. coli–stimulated Mkp-1−/− cells by t test.

FIGURE 7.

Enhanced IFN-β induction by E. coli and TLR ligands caused by Mkp-1 deficiency is substantially inhibited by the pharmacological inhibitor of p38 but not JNK. Mkp-1−/− and Mkp-1+/+ BMDM were pretreated with DMSO (vehicle) or 10 μM SB203580 or 3 μM JNK-IN-8 for 30 min, then stimulated with heat-killed E. coli at a dose of 10 bacteria per macrophages or with various TLR ligands. Media were harvested for ELISA to measure IFN-β concentration. Cells were harvested to isolate total RNA for qRT-PCR analysis on IFN-β expression. (A) The effects of p38 and JNK inhibition on IFN-β production by Mkp-1+/+ and Mkp-1−/− BMDM following 6 h of E. coli stimulation. (B) The effects of p38 and JNK inhibition on IFN-β production by Mkp-1+/+ and Mkp-1−/− BMDM following 6 h of LPS stimulation (100 ng/ml). (C) Kinetics of IFN-β production by E. coli–stimulated Mkp-1−/− and Mkp-1+/+ BMDM in the presence and absence of p38 inhibitor. BMDM (1 × 106 cells) were cultured in 1 ml of medium in 12-well plates. (D) Kinetics of IFN-β mRNA induction in Mkp-1−/− and Mkp-1+/+ BMDM by E. coli in the presence and absence of p38 inhibitor. (E) The effect of SB203580 on the IFN-β mRNA expression in Mkp-1−/− and Mkp-1+/+ BMDM stimulated with E. coli or LPS for 3 h. BMDM (9 × 105 cells) were cultured in 3 ml of medium in 60-mm plates. (F) The effect of SB203580 on the IFN-β mRNA expression in Mkp-1−/− and Mkp-1+/+ BMDM transfected with 1.8 μg of poly(I:C) or 1.8 μg of herring sperm DNA for 3 h. BMDM (9 × 105 cells) were cultured with 3 ml of medium in 60-mm plates were transfected with the indicated amounts of nucleic acids using polyethylenimine. Data in (A), (B), and (C) are presented as mean ± SE (n = 3) of the IFN-β concentrations. Data in (D) and (E) are presented as mean ± SE (n = 3) of fold change over the average levels in unstimulated Mkp-1+/+ cells. Data in (F) are presented as mean ± SE (n = 3) of fold change over the average levels in mock-transfected (with polyethylenimine) Mkp-1+/+ macrophages. *p < 0.05, compared with E. coli or TLR ligand-stimulated Mkp-1+/+ cells in their respective treatment group by t test; p < 0.05, compared with E. coli or TLR ligand-stimulated Mkp-1−/− cells in their respective treatment group by t test; #p < 0.05, compared with SB203580-pretreated, E. coli–stimulated Mkp-1−/− cells by t test.

Close modal

We then assessed the effect of p38 inhibition on kinetics of IFN-β production and IFN-β mRNA induction following E. coli stimulation. E. coli–induced IFN-β levels in the medium increased gradually over a 6-h period for both Mkp-1+/+ and Mkp-1−/− BMDM, although the increase was substantially greater for Mkp-1−/− BMDM (Fig. 7C). The increase in E. coli–induced IFN-β levels was detected within 2 h. IFN-β reached a peak level at ∼4 h in Mkp-1+/+ BMDM, whereas IFN-β levels continued to increase for 6 h in Mkp-1−/− BMDM. Pretreatment of both Mkp-1+/+ and Mkp-1−/− BMDM with SB203580 almost completely abolished E. coli–induced IFN-β production in both groups. IFN-β mRNA reached peak levels at 1–2 h in both Mkp-1+/+ and Mkp-1−/− BMDM following E. coli stimulation, then declined, although IFN-β mRNA reached substantially greater levels in Mkp-1−/− BMDM (Fig. 7D). SB203580 pretreatment decreased IFN-β mRNA levels substantially in both Mkp-1+/+ and Mkp-1−/− BMDM. We also assessed the effect of SB203580 on IFN-β mRNA induction in Mkp-1+/+ and Mkp-1−/− BMDM treated with LPS, poly(I:C), or herring sperm DNA. LPS and poly(I:C) activate TLR4 (58) and TLR3 (59), respectively. Poly(I:C) is also recognized by cytosolic pathogen sensor RIG-I (60). Herring sperm dsDNA has been shown to activate the cyclic GMP-AMP synthase (cGAS)–stimulator of IFN genes (STING) pathway to stimulate IFN-β expression (61). Similar to what was seen in cells stimulated with E. coli, LPS-stimulated IFN-β mRNA expression in Mkp-1+/+ and Mkp-1−/− BMDM was abolished by SB203580 (Fig. 7E). SB203580 also abolished IFN-β mRNA induction by transfection of either poly(I:C) or herring sperm DNA in both Mkp-1+/+ and Mkp-1−/− BMDM (Fig. 7F). These results further highlight the critical role of p38 in the regulation of IFN-β induction during pathogenic infections.

Because p38 controls the production of many cytokines via regulating the stability of cytokine mRNAs, we assessed whether IFN-β mRNA stability is affected by Mkp-1 deficiency. BMDM were stimulated with heat-killed E. coli for 2 h and then treated with actinomycin D to stop gene transcription. Cells were harvested at different times, and IFN-β mRNA levels in these samples were assessed by qRT-PCR (Fig. 8A). The half-life of IFN-β mRNA was calculated based on the rate of mRNA decay. The half-life of IFN-β mRNA in Mkp-1+/+ macrophages was ∼4.1 h, whereas the half-life in Mkp-1−/− macrophages was moderately longer, ∼5.9 h, suggesting that enhanced IFN-β mRNA stability contributes to the elevated IFN-β expression in Mkp-1−/− macrophages.

FIGURE 8.

Mkp-1 modestly shortened the stability of IFN-β mRNA but had little effect on IFN-β promoter activity or upstream signaling. (A) Decay of IFN-β mRNA in E. coli–stimulated Mkp-1+/+ and Mkp-1−/− BMDM after the addition of actinomycin. Mkp-1+/+ and Mkp-1−/− BMDM (5 × 106 cells on 60-mm plates in 3 ml of medium) were first stimulated with heat-killed E. coli for 2 h. Actinomycin D was then added into the culture medium (time 0) to a concentration of 2 μg/ml. Total RNA was harvested from the cells after 2, 4, and 8 h. IFN-β mRNA levels were quantified by qRT-PCR to assess IFN-β mRNA decay. IFN-β mRNA levels were normalized to 18S rRNA. The average levels of IFN-β mRNA at time 0 (no actinomycin D treatment) was set as 100%. The remaining mRNA levels (percentage) at other timepoints were calculated relative to the average level of the same genotype at time 0 and presented in the graph as mean ± SE (n = 3). Note, y-axis is set in log scale. The half-life (t1/2) of IFN-β mRNA was calculated using the formula N (t) = N0e−λt, where t1/2=ln (2)λ,N(t) represents level at time t, N0 represents level at time 0. (B) The effect of Mkp-1 overexpression on IFN-β promoter activity. RAW264.7 cells stably integrated with a Tet-ON Mkp-1 expression cassette (clone 22) were treated with 100 ng/ml overnight or left untreated. Cells were either stimulated with LPS (100 ng/ml) or with heat-killed E. coli at a dose of 10 bacteria per macrophage for 6 h or left unstimulated. Cells were harvested to measure luciferase activity. The activity was normalized to lysate protein contents. Values are expressed as mean ± SE (n = 4). *p < 0.05 by t test. (C) The effect of Mkp-1 deficiency on phosphorylation of IRF3 and IRF7 following LPS stimulation. Mkp-1+/+ and Mkp-1−/− BMDM were stimulated with 100 ng/ml LPS for the indicated times and harvested for Western blot analysis using Abs against phospho-IRF3, IRF3, phospho-IRF7, and Mkp-1. The membrane was stripped and blotted using an Ab against β-actin to control for sample loading. (D) The effect of Mkp-1 deficiency on TBK1 and IKKα/β activation following LPS stimulation. Mkp-1+/+ and Mkp-1−/− BMDM stimulated with 100 ng/ml LPS were analyzed by Western blotting using Abs against phospho-TBK1, TBK1, phospho-IKKα/β, IKKα/β, and IκB. The membrane was stripped and blotted using an Ab against β-actin to control for sample loading. The densities of the individual bands were quantitated by densitometry. The phosphorylated protein was normalized to total protein. IκB was normalized to β-actin. The fold change in protein phosphorylation (TBK1 or IKKα/β) or protein level (IκB) was calculated relative to the value in unstimulated Mkp-1+/+ cells and is marked underneath each lane. Representative results were presented in (C) and (D).

FIGURE 8.

Mkp-1 modestly shortened the stability of IFN-β mRNA but had little effect on IFN-β promoter activity or upstream signaling. (A) Decay of IFN-β mRNA in E. coli–stimulated Mkp-1+/+ and Mkp-1−/− BMDM after the addition of actinomycin. Mkp-1+/+ and Mkp-1−/− BMDM (5 × 106 cells on 60-mm plates in 3 ml of medium) were first stimulated with heat-killed E. coli for 2 h. Actinomycin D was then added into the culture medium (time 0) to a concentration of 2 μg/ml. Total RNA was harvested from the cells after 2, 4, and 8 h. IFN-β mRNA levels were quantified by qRT-PCR to assess IFN-β mRNA decay. IFN-β mRNA levels were normalized to 18S rRNA. The average levels of IFN-β mRNA at time 0 (no actinomycin D treatment) was set as 100%. The remaining mRNA levels (percentage) at other timepoints were calculated relative to the average level of the same genotype at time 0 and presented in the graph as mean ± SE (n = 3). Note, y-axis is set in log scale. The half-life (t1/2) of IFN-β mRNA was calculated using the formula N (t) = N0e−λt, where t1/2=ln (2)λ,N(t) represents level at time t, N0 represents level at time 0. (B) The effect of Mkp-1 overexpression on IFN-β promoter activity. RAW264.7 cells stably integrated with a Tet-ON Mkp-1 expression cassette (clone 22) were treated with 100 ng/ml overnight or left untreated. Cells were either stimulated with LPS (100 ng/ml) or with heat-killed E. coli at a dose of 10 bacteria per macrophage for 6 h or left unstimulated. Cells were harvested to measure luciferase activity. The activity was normalized to lysate protein contents. Values are expressed as mean ± SE (n = 4). *p < 0.05 by t test. (C) The effect of Mkp-1 deficiency on phosphorylation of IRF3 and IRF7 following LPS stimulation. Mkp-1+/+ and Mkp-1−/− BMDM were stimulated with 100 ng/ml LPS for the indicated times and harvested for Western blot analysis using Abs against phospho-IRF3, IRF3, phospho-IRF7, and Mkp-1. The membrane was stripped and blotted using an Ab against β-actin to control for sample loading. (D) The effect of Mkp-1 deficiency on TBK1 and IKKα/β activation following LPS stimulation. Mkp-1+/+ and Mkp-1−/− BMDM stimulated with 100 ng/ml LPS were analyzed by Western blotting using Abs against phospho-TBK1, TBK1, phospho-IKKα/β, IKKα/β, and IκB. The membrane was stripped and blotted using an Ab against β-actin to control for sample loading. The densities of the individual bands were quantitated by densitometry. The phosphorylated protein was normalized to total protein. IκB was normalized to β-actin. The fold change in protein phosphorylation (TBK1 or IKKα/β) or protein level (IκB) was calculated relative to the value in unstimulated Mkp-1+/+ cells and is marked underneath each lane. Representative results were presented in (C) and (D).

Close modal

To address whether Mkp-1 affects IFN-β gene transcription, we stably integrated an IFN-β–luciferase reporter into an RAW264.7 cell line harboring a Tet-ON Mkp-1 expression cassette. Cells were treated with Dox overnight or left untreated, then stimulated with LPS or heat-killed E. coli for 6 h prior to harvesting for luciferase activity assays (Fig. 8B). Dox treatment had no effect on the IFN-β–luciferase reporter in both unstimulated and LPS-stimulated cells, indicating that elevated Mkp-1 expression did not affect LPS-stimulated IFN-β promoter activity. However, Dox treatment slightly decreased IFN-β–luciferase reporter activity in E. coli–induced cells by ∼20%, indicating a slight inhibition on IFN-β promoter activity.

IFN-β transcription is regulated by multiple transcription factors, including IRF3, IRF7, and NF-κB, through phosphorylation mediated by protein kinases (62, 63). We compared IRF3 and IRF7 phosphorylation in Mkp-1+/+ and Mkp-1−/− BMDM after LPS stimulation by Western blotting using phospho-specific Abs (Fig. 8C). IRF3 was rapidly phosphorylated in both Mkp-1+/+ and Mkp-1−/− BMDM following LPS stimulation, and there was no overt difference in the kinetics or magnitude of IRF3 phosphorylation between these cells. Phosphorylation of IRF7 was similar. We then assessed the activities of upstream kinases using phospho-specific Abs (Fig. 8D). LPS stimulation led to a transient TBK1 phosphorylation/activation in both Mkp-1+/+ and Mkp-1−/− BMDM. TBK1 phosphorylation reached peak level at 30–90 min and substantially declined by 3 h post-LPS stimulation. There was no overt difference in TBK1 phosphorylation between the two groups. Activation of IKKα/β occurred within 15 min after LPS stimulation, then IKKα/β phosphorylation rapidly declined. No obvious difference in IKKα/β activity was observed between Mkp-1+/+ and Mkp-1−/− BMDM, although IκB levels appeared to recover faster in Mkp-1−/− BMDM than in Mkp-1+/+ BMDM.

Previously, we have shown that Mkp-1−/− mice exhibited a significantly greater mortality after E. coli infection than Mkp-1+/+ mice (31, 32). The increased mortality is associated with enhanced production of three cytokines (TNF-α, IL-6, and IL-10) and elevation of bacterial burden and greater organ damage. In this study, we showed that, in addition to these three cytokines, the production of at least another seven cytokines was enhanced in Mkp-1−/− mice following systemic E. coli infection, including IL-1α and β, IL-17A, IL-27, GM-CSF, IFN-β, and IFN-γ (Fig. 3), highlighting the pivotal role of Mkp-1 in the prevention of cytokine storms. Consistent with the significant increase in circulating IFNs, the expression of a large number of Isgs was substantially enhanced in the livers of E. coli–infected Mkp-1−/− mice compared with E. coli–infected Mkp-1+/+ mice (Fig. 1). Because the mRNA levels of type I IFNs were very low in livers (Supplemental Table II), they are unlikely a major source of type I IFN production during sepsis. We think that the augmented expression of these numerous Isgs in the livers of E. coli–infected Mkp-1−/− mice is likely a cellular reflection of elevated circulating type I IFNs. In this study, we focus on the function of type I IFN during E. coli infection and the regulation of IFN-β expression by Mkp-1. We showed that neutralizing IFNAR1 increased morbidity without affecting TNF-α and IL-6 or bacterial burden, supporting a beneficial role of type I IFNs in this sepsis model (Fig. 4). The expression of Rsad2 protein in E. coli–infected Mkp-1−/− mice was almost abolished by IFNAR1 neutralization, illustrating the importance of type I IFN signaling in Rsad2 induction and a nearly complete neutralization of type I IFN signaling in these mice (Fig. 4C). Supporting the critical role of Mkp-1 in the regulation of IFN-β production in phagocytes, we found that Mkp-1−/− BMDM produced dramatically more IFN-β protein than did Mkp-1+/+ cells following stimulation with E. coli, LPS, poly(I:C), or herring sperm DNA (Figs. 5, 7F). Moreover, enhanced IFN-β induction in response to these agents was almost completely blocked by a p38 inhibitor but not a JNK inhibitor (Fig. 7), suggesting that enhanced p38 activity is primarily responsible for the increased IFN-β production in Mkp-1−/− macrophages. We found that the half-life of E. coli–induced IFN-β mRNA was modestly longer in Mkp-1−/− BMDM than in Mkp-1+/+ BMDM (Fig. 8A). We also found that overexpression of Mkp-1 had little effect on the activity of the proximal IFN-β promoter (Fig. 8B). Mkp-1 deficiency had little effect on IRF3 phosphorylation (Fig. 8C) and TBK1 or IKKα/β activation (Fig. 8D). Taken together, our results clearly show that Mkp-1 constrains the exaggerated production of many cytokines that are both beneficial and harmful to the host during microbial infections, establishing Mkp-1 as a critical gate keeper of the cytokine storm in sepsis. Our studies indicate that Mkp-1 regulates IFN-β expression primarily through a p38-mediated mechanism.

We found that 10 of the 13 cytokines examined in this study exhibited enhanced production following E. coli infection in Mkp-1−/− mice relative to Mkp-1+/+ mice (Fig. 3). The exaggerated production of these cytokines in Mkp-1−/− mice following E. coli infection is not surprising. These are typical inflammatory cytokines and their mRNA transcripts contain adenylate-uridylate (AU)–rich elements in the 3′ untranslated regions (64). However, the regulation of IFN-β by Mkp-1 in response to bacterial infection has not been previously reported. The increased IFN-β production in E. coli–infected Mkp-1−/− mice and Mkp-1−/− BMDM following stimulation with E. coli, LPS, poly(I:C), and DNA clearly show an inhibitory role of Mkp-1 in IFN-β induction (Figs. 5, 7). The inhibitory action of Mkp-1 on IFN-β production is further supported by the almost complete suppression of IFN-β production by overexpression of Mkp-1 in E. coli- and LPS-stimulated RAW264.7 cells (Fig. 6).

Although it is clear that Mkp-1 inhibits IFN-β expression at both the mRNA and protein levels, the mechanism involved is unclear. In theory, Mkp-1 could regulate IFN-β induction through both transcriptional and posttranscriptional mechanisms (65). It has been well-established that IFN-β transcription can be regulated by the transcription factors IRF3, IRF7, NF-κB, and AP-1 in a cooperative manner via binding to the positive regulatory domain (PRD) I–IV on the IFN-β promoter (62, 63). However, our results suggest that, at least in response to LPS, enhanced IFN-β expression by Mkp-1 deficiency is unlikely mediated by these transcription factors. Mkp-1 deficiency did not overtly enhance IRF3 or IRF7 phosphorylation (Fig. 8C) or IKKα/β activity (Fig. 8D). The AP-1 transcription factor complex bound to PRD IV of the IFN-β promoter also does not appear to play a prominent role in the enhanced IFN-β expression in Mkp-1−/− macrophages, at least in response to LPS. PRD IV–bound AP-1 is composed of a c-Jun/activating transcription factor 2 (ATF2) heterodimer (63). JNK is known to enhance the transcriptional activity of AP-1 through phosphorylation of c-Jun and ATF2, and p38 also phosphorylates ATF2 (6669). Mkp-1 deficiency leads to enhanced JNK and p38 activity, which could, at least in theory, enhance AP-1 activity and enhanced IFN-β transcription. However, the following observations do not support this model. First, Mkp-1 overexpression in LPS-stimulated RAW264.7 cells had little effect on the activity of the proximal IFN-β promoter that contains PRD I–IV (Fig. 8B), although it almost completely inhibited IFN-β production from the endogenous gene after either LPS or E. coli stimulation (Fig. 6B, 6C). Second, a JNK-selective inhibitor had little effect on IFN-β expression following LPS stimulation (Fig. 7A, 7B), despite a substantial inhibition of c-Jun phosphorylation (data not shown). The role of AP-1 and JNK in IFN-β expression in response to E. coli stimulation appears to be more complicated (Figs. 7A, 8B). This is not surprising because bacteria are bound to stimulate more signaling pathways because of the increased complexity of the stimulant. JNK appears to play a minor role in E. coli–induced IFN-β expression in Mkp-1−/− BMDM because inhibition of both JNK and p38 led to a greater decrease in IFN-β expression than inhibition of p38 alone (Fig. 7A). This limited stimulatory effect of JNK could be mediated by AP-1 transcription factor via PRD IV of the proximal IFN-β promoter because E. coli–induced IFN-β–luciferase reporter activity was modestly inhibited by Mkp-1 overexpression (Fig. 8B). Unlike the limited role of JNK in IFN-β expression, p38 plays a critical role in IFN-β expression for all stimulations tested (Fig. 7). We postulate that if p38 positively regulates IFN-β transcription, the action is likely mediated by a transcription factor other than AP-1 through a distal element(s) on the IFN-β promoter.

We think that elevated p38 activity in Mkp-1−/− macrophages may also mediate IFN-β expression by stabilizing IFN-β mRNA and enhancing IFN-β translation (70). IFN-β mRNA contains several putative AU-rich elements (65, 71, 72). AU-rich elements have been shown to mediate mRNA decay through interaction with mRNA-binding proteins such as tristetraprolin (73). In the absence of Mkp-1, stronger p38 activity could lead to greater tristetraprolin phosphorylation, resulting in dissociation of tristetraprolin from the IFN-β mRNA decay machinery, leading to enhanced IFN-β mRNA stability and IFN-β translation. This is consistent with our observation that IFN-β mRNA half-life is longer in Mkp-1−/− BMDM than in Mkp-1+/+ BMDM (Fig. 8). Nevertheless, it is still puzzling whether a modest increase in IFN-β mRNA stability could explain the dramatic differences in IFN-β mRNA levels (Fig. 5).

Given the importance of IFN-β in host defense against viruses (41), we speculate that, as a negative regulator of IFN-β production, Mkp-1 could also be detrimental during certain viral infections, and inhibition of this phosphatase may represent a therapeutic treatment to contain viral spread. This is supported by the substantial enhancement in the expression of the large number of Isgs in the livers of Mkp-1−/− mice after E. coli infection (Fig. 1). It should be pointed out that IFN-γ likely also contributed to the induction of some of the genes, given the significant differences in IFN-γ levels between the E. coli–infected Mkp-1+/+ and Mkp-1−/− mice (Fig. 3).

Mkp-1−/− mice exhibit enhanced cytokine production, elevated bacterial burden, and increased mortality following systemic E. coli infection relative to Mkp-1+/+ mice (31). Because IFN-β has been shown to play both beneficial and detrimental roles during bacterial infection (33, 43), we blocked the receptor for IFN-β, IFNAR1, and examined the effects on both mortality and morbidity after E. coli infection. Complete blockade of type I IFN signaling by the IFNAR1-neutralizing Ab was supported by the absence of Rsad2 protein (Fig. 4C), a classic Isg (74). Neutralizing IFNAR1 not only led to significantly greater disease severity but also appeared to increase the mortality in Mkp-1−/− mice (Fig. 4A, 4B). These results suggest that elevated IFN-β is beneficial to the animals in this sepsis model, but how this occurs is unclear. Neutralizing IFN-β neither altered IL-6 production, an inflammatory index, nor changed bacterial burden (Fig. 4C, 4D). IFN-β has been shown to influence the immune system through a number of processes and systems, including tissue and cell integrity, and barrier functions (43, 75). Regardless of the exact mechanism via which IFN-β influences the host response in our E. coli–induced sepsis model, it is clear that elevated IFN-β production is not responsible for the elevated bacterial burden and increased mortality of Mkp-1−/− mice.

We are grateful to Bristol Myers Squibb Pharmaceutical Research Institute for proving Mkp-1 knockout mice. We thank Drs. Xianxi Wang, Jinhui Li, Xiaomei Meng, and Dimitrios Anastasakis for help with in vivo infection and RNA-seq experiments. We also gratefully acknowledge William M. White for technical support. We thank Dr. Yongliang Zhang for generously providing the IFN-β luciferase reporter and thank Dr. Stefanie Vogel for valuable suggestions.

This work was supported by grants from the National Institutes of Health, National Institute of Allergy and Infectious Diseases (AI124029 and AI142885 and to Y.L. and AI121196 to J.Z.).

The online version of this article contains supplemental material.

Abbreviations used in this article

ATF2

activating transcription factor 2

AU

adenylate-uridylate

BMDM

bone marrow–derived macrophage

b.w.

body weight

Dox

doxycycline

Fasn

fatty acid synthase

IFNAR1

IFN-α/β receptor 1

IKK

IκB kinase

IRF

IFN regulatory factor

Isg

IFN-stimulated gene

ITR

inverted terminal repeat

JNK-IN-8

JNK inhibitor 8

Mkp-1

MAP kinase phosphatase-1

poly(I:C)

polyinosinic:polycytidylic acid

PRD

positive regulatory domain

qRT-PCR

quantitative RT-PCR

RNA-seq

RNA-sequencing

Rsad2

radical S-adenosyl methionine domain-containing protein 2

TBK1

TANK-binding kinase 1

1.
Janeway
C. A.
 Jr.
,
P.
Travers
,
M.
Walport
,
M. J.
Shlomchik
.
2001
.
Immunobiology: The Immune System in Health and Disease.
Garland Publishing
,
New York
.
2.
Medzhitov
R.
,
C.
Janeway
Jr
.
2000
.
The Toll receptor family and microbial recognition.
Trends Microbiol.
8
:
452
456
.
3.
Akira
S.
2003
.
Mammalian Toll-like receptors.
Curr. Opin. Immunol.
15
:
5
11
.
4.
Cook
D. N.
,
D. S.
Pisetsky
,
D. A.
Schwartz
.
2004
.
Toll-like receptors in the pathogenesis of human disease.
Nat. Immunol.
5
:
975
979
.
5.
Kaisho
T.
,
S.
Akira
.
2000
.
Critical roles of Toll-like receptors in host defense.
Crit. Rev. Immunol.
20
:
393
405
.
6.
Takeuchi
O.
,
S.
Akira
.
2001
.
Toll-like receptors; their physiological role and signal transduction system.
Int. Immunopharmacol.
1
:
625
635
.
7.
Hawiger
J.
2001
.
Innate immunity and inflammation: a transcriptional paradigm.
Immunol. Res.
23
:
99
109
.
8.
Dong
C.
,
R. J.
Davis
,
R. A.
Flavell
.
2002
.
MAP kinases in the immune response.
Annu. Rev. Immunol.
20
:
55
72
.
9.
Hayden
M. S.
,
A. P.
West
,
S.
Ghosh
.
2006
.
NF-kappaB and the immune response.
Oncogene
25
:
6758
6780
.
10.
Kafka
D.
,
E.
Ling
,
G.
Feldman
,
D.
Benharroch
,
E.
Voronov
,
N.
Givon-Lavi
,
Y.
Iwakura
,
R.
Dagan
,
R. N.
Apte
,
Y.
Mizrachi-Nebenzahl
.
2008
.
Contribution of IL-1 to resistance to Streptococcus pneumoniae infection.
Int. Immunol.
20
:
1139
1146
.
11.
Prantner
D.
,
T.
Darville
,
J. D.
Sikes
,
C. W.
Andrews
Jr.
,
H.
Brade
,
R. G.
Rank
,
U. M.
Nagarajan
.
2009
.
Critical role for interleukin-1beta (IL-1beta) during Chlamydia muridarum genital infection and bacterial replication-independent secretion of IL-1beta in mouse macrophages.
Infect. Immun.
77
:
5334
5346
.
12.
Conti
H. R.
,
S. L.
Gaffen
.
2015
.
IL-17-mediated immunity to the opportunistic fungal pathogen Candida albicans.
J. Immunol.
195
:
780
788
.
13.
Dalrymple
S. A.
,
L. A.
Lucian
,
R.
Slattery
,
T.
McNeil
,
D. M.
Aud
,
S.
Fuchino
,
F.
Lee
,
R.
Murray
.
1995
.
Interleukin-6-deficient mice are highly susceptible to Listeria monocytogenes infection: correlation with inefficient neutrophilia.
Infect. Immun.
63
:
2262
2268
.
14.
Beutler
B.
1995
.
TNF, immunity and inflammatory disease: lessons of the past decade.
J. Investig. Med.
43
:
227
235
.
15.
Peñaloza
H. F.
,
P. A.
Nieto
,
N.
Muñoz-Durango
,
F. J.
Salazar-Echegarai
,
J.
Torres
,
M. J.
Parga
,
M.
Alvarez-Lobos
,
C. A.
Riedel
,
A. M.
Kalergis
,
S. M.
Bueno
.
2015
.
Interleukin-10 plays a key role in the modulation of neutrophils recruitment and lung inflammation during infection by Streptococcus pneumoniae.
Immunology
146
:
100
112
.
16.
Keyse
S. M.
1998
.
Protein phosphatases and the regulation of MAP kinase activity.
Semin. Cell Dev. Biol.
9
:
143
152
.
17.
Keyse
S. M.
,
E. A.
Emslie
.
1992
.
Oxidative stress and heat shock induce a human gene encoding a protein-tyrosine phosphatase.
Nature
359
:
644
647
.
18.
Charles
C. H.
,
H.
Sun
,
L. F.
Lau
,
N. K.
Tonks
.
1993
.
The growth factor-inducible immediate-early gene 3CH134 encodes a protein-tyrosine-phosphatase.
Proc. Natl. Acad. Sci. USA
90
:
5292
5296
.
19.
Noguchi
T.
,
R.
Metz
,
L.
Chen
,
M. G.
Mattéi
,
D.
Carrasco
,
R.
Bravo
.
1993
.
Structure, mapping, and expression of erp, a growth factor-inducible gene encoding a nontransmembrane protein tyrosine phosphatase, and effect of ERP on cell growth.
Mol. Cell. Biol.
13
:
5195
5205
.
20.
Keyse
S. M.
2000
.
Protein phosphatases and the regulation of mitogen-activated protein kinase signalling.
Curr. Opin. Cell Biol.
12
:
186
192
.
21.
Franklin
C. C.
,
A. S.
Kraft
.
1997
.
Conditional expression of the mitogen-activated protein kinase (MAPK) phosphatase MKP-1 preferentially inhibits p38 MAPK and stress-activated protein kinase in U937 cells.
J. Biol. Chem.
272
:
16917
16923
.
22.
Liu
Y.
,
E. G.
Shepherd
,
L. D.
Nelin
.
2007
.
MAPK phosphatases--regulating the immune response.
Nat. Rev. Immunol.
7
:
202
212
.
23.
Lang
R.
,
M.
Hammer
,
J.
Mages
.
2006
.
DUSP meet immunology: dual specificity MAPK phosphatases in control of the inflammatory response.
J. Immunol.
177
:
7497
7504
.
24.
Shepherd
E. G.
,
Q.
Zhao
,
S. E.
Welty
,
T. N.
Hansen
,
C. V.
Smith
,
Y.
Liu
.
2004
.
The function of mitogen-activated protein kinase phosphatase-1 in peptidoglycan-stimulated macrophages.
J. Biol. Chem.
279
:
54023
54031
.
25.
Zhao
Q.
,
E. G.
Shepherd
,
M. E.
Manson
,
L. D.
Nelin
,
A.
Sorokin
,
Y.
Liu
.
2005
.
The role of mitogen-activated protein kinase phosphatase-1 in the response of alveolar macrophages to lipopolysaccharide: attenuation of proinflammatory cytokine biosynthesis via feedback control of p38.
J. Biol. Chem.
280
:
8101
8108
.
26.
Zhao
Q.
,
X.
Wang
,
L. D.
Nelin
,
Y.
Yao
,
R.
Matta
,
M. E.
Manson
,
R. S.
Baliga
,
X.
Meng
,
C. V.
Smith
,
J. A.
Bauer
, et al
2006
.
MAP kinase phosphatase 1 controls innate immune responses and suppresses endotoxic shock.
J. Exp. Med.
203
:
131
140
.
27.
Hammer
M.
,
J.
Mages
,
H.
Dietrich
,
A.
Servatius
,
N.
Howells
,
A. C.
Cato
,
R.
Lang
.
2006
.
Dual specificity phosphatase 1 (DUSP1) regulates a subset of LPS-induced genes and protects mice from lethal endotoxin shock.
J. Exp. Med.
203
:
15
20
.
28.
Chi
H.
,
S. P.
Barry
,
R. J.
Roth
,
J. J.
Wu
,
E. A.
Jones
,
A. M.
Bennett
,
R. A.
Flavell
.
2006
.
Dynamic regulation of pro- and anti-inflammatory cytokines by MAPK phosphatase 1 (MKP-1) in innate immune responses.
Proc. Natl. Acad. Sci. USA
103
:
2274
2279
.
29.
Salojin
K. V.
,
I. B.
Owusu
,
K. A.
Millerchip
,
M.
Potter
,
K. A.
Platt
,
T.
Oravecz
.
2006
.
Essential role of MAPK phosphatase-1 in the negative control of innate immune responses.
J. Immunol.
176
:
1899
1907
.
30.
Manetsch
M.
,
W.
Che
,
P.
Seidel
,
Y.
Chen
,
A. J.
Ammit
.
2012
.
MKP-1: a negative feedback effector that represses MAPK-mediated pro-inflammatory signaling pathways and cytokine secretion in human airway smooth muscle cells.
Cell. Signal.
24
:
907
913
.
31.
Frazier
W. J.
,
X.
Wang
,
L. M.
Wancket
,
X. A.
Li
,
X.
Meng
,
L. D.
Nelin
,
A. C.
Cato
,
Y.
Liu
.
2009
.
Increased inflammation, impaired bacterial clearance, and metabolic disruption after gram-negative sepsis in Mkp-1-deficient mice.
J. Immunol.
183
:
7411
7419
.
32.
Li
J.
,
X.
Wang
,
W. E.
Ackerman
IV
,
A. J.
Batty
,
S. G.
Kirk
,
W. M.
White
,
X.
Wang
,
D.
Anastasakis
,
L.
Samavati
,
I.
Buhimschi
, et al
2018
.
Dysregulation of Lipid Metabolism in Mkp-1 Deficient Mice during Gram-Negative Sepsis.
Int. J. Mol. Sci.
19
:
3904
.
33.
Bogdan
C.
2000
.
The function of type I interferons in antimicrobial immunity.
Curr. Opin. Immunol.
12
:
419
424
.
34.
Patel
J. R.
,
A.
García-Sastre
.
2014
.
Activation and regulation of pathogen sensor RIG-I.
Cytokine Growth Factor Rev.
25
:
513
523
.
35.
Zuniga
E. I.
,
B.
Hahm
,
M. B.
Oldstone
.
2007
.
Type I interferon during viral infections: multiple triggers for a multifunctional mediator.
Curr. Top. Microbiol. Immunol.
316
:
337
357
.
36.
Hemmi
H.
,
O.
Takeuchi
,
S.
Sato
,
M.
Yamamoto
,
T.
Kaisho
,
H.
Sanjo
,
T.
Kawai
,
K.
Hoshino
,
K.
Takeda
,
S.
Akira
.
2004
.
The roles of two IkappaB kinase-related kinases in lipopolysaccharide and double stranded RNA signaling and viral infection.
J. Exp. Med.
199
:
1641
1650
.
37.
McWhirter
S. M.
,
K. A.
Fitzgerald
,
J.
Rosains
,
D. C.
Rowe
,
D. T.
Golenbock
,
T.
Maniatis
.
2004
.
IFN-regulatory factor 3-dependent gene expression is defective in Tbk1-deficient mouse embryonic fibroblasts.
Proc. Natl. Acad. Sci. USA
101
:
233
238
.
38.
Perry
A. K.
,
E. K.
Chow
,
J. B.
Goodnough
,
W. C.
Yeh
,
G.
Cheng
.
2004
.
Differential requirement for TANK-binding kinase-1 in type I interferon responses to toll-like receptor activation and viral infection.
J. Exp. Med.
199
:
1651
1658
.
39.
Scumpia
P. O.
,
G. A.
Botten
,
J. S.
Norman
,
K. M.
Kelly-Scumpia
,
R.
Spreafico
,
A. R.
Ruccia
,
P. K.
Purbey
,
B. J.
Thomas
,
R. L.
Modlin
,
S. T.
Smale
.
2017
.
Opposing roles of Toll-like receptor and cytosolic DNA-STING signaling pathways for Staphylococcus aureus cutaneous host defense.
PLoS Pathog.
13
:
e1006496
.
40.
Zevini
A.
,
D.
Olagnier
,
J.
Hiscott
.
2017
.
Crosstalk between cytoplasmic RIG-I and STING sensing pathways.
Trends Immunol.
38
:
194
205
.
41.
McNab
F.
,
K.
Mayer-Barber
,
A.
Sher
,
A.
Wack
,
A.
O’Garra
.
2015
.
Type I interferons in infectious disease.
Nat. Rev. Immunol.
15
:
87
103
.
42.
Trinchieri
G.
2010
.
Type I interferon: friend or foe?
J. Exp. Med.
207
:
2053
2063
.
43.
Boxx
G. M.
,
G.
Cheng
.
2016
.
The roles of type I interferon in bacterial infection.
Cell Host Microbe
19
:
760
769
.
44.
Dorfman
K.
,
D.
Carrasco
,
M.
Gruda
,
C.
Ryan
,
S. A.
Lira
,
R.
Bravo
.
1996
.
Disruption of the erp/mkp-1 gene does not affect mouse development: normal MAP kinase activity in ERP/MKP-1-deficient fibroblasts.
Oncogene
13
:
925
931
.
45.
Shrum
B.
,
R. V.
Anantha
,
S. X.
Xu
,
M.
Donnelly
,
S. M.
Haeryfar
,
J. K.
McCormick
,
T.
Mele
.
2014
.
A robust scoring system to evaluate sepsis severity in an animal model.
BMC Res. Notes
7
:
233
.
46.
Konstantinou
G. N.
2017
.
Enzyme-linked immunosorbent assay (ELISA).
Methods Mol. Biol.
1592
:
79
94
.
47.
Zhao
Q.
,
P.
Chen
,
M. E.
Manson
,
Y.
Liu
.
2006
.
Production of active recombinant mitogen-activated protein kinases through transient transfection of 293T cells.
Protein Expr. Purif.
46
:
468
474
.
48.
Lee
J. C.
,
J. T.
Laydon
,
P. C.
McDonnell
,
T. F.
Gallagher
,
S.
Kumar
,
D.
Green
,
D.
McNulty
,
M. J.
Blumenthal
,
J. R.
Heys
,
S. W.
Landvatter
, et al
1994
.
A protein kinase involved in the regulation of inflammatory cytokine biosynthesis.
Nature
372
:
739
746
.
49.
Zhang
T.
,
F.
Inesta-Vaquera
,
M.
Niepel
,
J.
Zhang
,
S. B.
Ficarro
,
T.
Machleidt
,
T.
Xie
,
J. A.
Marto
,
N.
Kim
,
T.
Sim
, et al
2012
.
Discovery of potent and selective covalent inhibitors of JNK.
Chem. Biol.
19
:
140
154
.
50.
Kim
V. Y.
,
A.
Batty
,
J.
Li
,
S. G.
Kirk
,
S. A.
Crowell
,
Y.
Jin
,
J.
Tang
,
J.
Zhang
,
L. K.
Rogers
,
H. X.
Deng
, et al
2019
.
Glutathione reductase promotes fungal clearance and suppresses inflammation during systemic Candida albicans infection in mice.
J. Immunol.
203
:
2239
2251
.
51.
Yusa
K.
,
L.
Zhou
,
M. A.
Li
,
A.
Bradley
,
N. L.
Craig
.
2011
.
A hyperactive piggyBac transposase for mammalian applications.
Proc. Natl. Acad. Sci. USA
108
:
1531
1536
.
52.
Cortazar
M. A.
,
R. M.
Sheridan
,
B.
Erickson
,
N.
Fong
,
K.
Glover-Cutter
,
K.
Brannan
,
D. L.
Bentley
.
2019
.
Control of RNA Pol II speed by PNUTS-PP1 and Spt5 dephosphorylation facilitates termination by a “sitting duck torpedo” mechanism.
Mol. Cell
76
:
896
908.e4
.
53.
Barger
C. J.
,
C.
Branick
,
L.
Chee
,
A. R.
Karpf
.
2019
.
Pan-cancer analyses reveal genomic features of FOXM1 overexpression in cancer.
Cancers (Basel)
11
:
251
.
54.
Rocha
B. C.
,
P. E.
Marques
,
F. M. S.
Leoratti
,
C.
Junqueira
,
D. B.
Pereira
,
L. R. D. V.
Antonelli
,
G. B.
Menezes
,
D. T.
Golenbock
,
R. T.
Gazzinelli
.
2015
.
Type I interferon transcriptional signature in neutrophils and low-density granulocytes are associated with tissue damage in malaria.
Cell Rep.
13
:
2829
2841
.
55.
Chen
P.
,
J.
Li
,
J.
Barnes
,
G. C.
Kokkonen
,
J. C.
Lee
,
Y.
Liu
.
2002
.
Restraint of proinflammatory cytokine biosynthesis by mitogen-activated protein kinase phosphatase-1 in lipopolysaccharide-stimulated macrophages.
J. Immunol.
169
:
6408
6416
.
56.
Yamamoto
M.
,
S.
Sato
,
H.
Hemmi
,
K.
Hoshino
,
T.
Kaisho
,
H.
Sanjo
,
O.
Takeuchi
,
M.
Sugiyama
,
M.
Okabe
,
K.
Takeda
,
S.
Akira
.
2003
.
Role of adaptor TRIF in the MyD88-independent toll-like receptor signaling pathway.
Science
301
:
640
643
.
57.
Schmitz
F.
,
J.
Mages
,
A.
Heit
,
R.
Lang
,
H.
Wagner
.
2004
.
Transcriptional activation induced in macrophages by Toll-like receptor (TLR) ligands: from expression profiling to a model of TLR signaling.
Eur. J. Immunol.
34
:
2863
2873
.
58.
Beutler
B.
2000
.
Tlr4: central component of the sole mammalian LPS sensor.
Curr. Opin. Immunol.
12
:
20
26
.
59.
Alexopoulou
L.
,
A. C.
Holt
,
R.
Medzhitov
,
R. A.
Flavell
.
2001
.
Recognition of double-stranded RNA and activation of NF-kappaB by Toll-like receptor 3.
Nature
413
:
732
738
.
60.
Kumar
H.
,
T.
Kawai
,
S.
Akira
.
2011
.
Pathogen recognition by the innate immune system.
Int. Rev. Immunol.
30
:
16
34
.
61.
Bode
C.
,
M.
Fox
,
P.
Tewary
,
A.
Steinhagen
,
R. K.
Ellerkmann
,
D.
Klinman
,
G.
Baumgarten
,
V.
Hornung
,
F.
Steinhagen
.
2016
.
Human plasmacytoid dentritic cells elicit a type I interferon response by sensing DNA via the cGAS-STING signaling pathway.
Eur. J. Immunol.
46
:
1615
1621
.
62.
Pålsson-McDermott
E. M.
,
L. A.
O’Neill
.
2004
.
Signal transduction by the lipopolysaccharide receptor, Toll-like receptor-4.
Immunology
113
:
153
162
.
63.
Honda
K.
,
A.
Takaoka
,
T.
Taniguchi
.
2006
.
Type I interferon [corrected] gene induction by the interferon regulatory factor family of transcription factors.
Immunity
25
:
349
360
.
64.
Clark
A.
,
J.
Dean
,
C.
Tudor
,
J.
Saklatvala
.
2009
.
Post-transcriptional gene regulation by MAP kinases via AU-rich elements.
Front. Biosci.
14
:
847
871
.
65.
McGuire
V. A.
,
D.
Rosner
,
O.
Ananieva
,
E. A.
Ross
,
S. E.
Elcombe
,
S.
Naqvi
,
M. M. W.
van den Bosch
,
C. E.
Monk
,
T.
Ruiz-Zorrilla Diez
,
A. R.
Clark
,
J. S. C.
Arthur
.
2016
.
Beta interferon production is regulated by p38 mitogen-activated protein kinase in macrophages via both MSK1/2- and tristetraprolin-dependent pathways.
Mol. Cell. Biol.
37
:
e00454-16
.
66.
Karin
M.
1995
.
The regulation of AP-1 activity by mitogen-activated protein kinases.
J. Biol. Chem.
270
:
16483
16486
.
67.
Davis
R. J.
1995
.
Transcriptional regulation by MAP kinases.
Mol. Reprod. Dev.
42
:
459
467
.
68.
Gupta
S.
,
D.
Campbell
,
B.
Dérijard
,
R. J.
Davis
.
1995
.
Transcription factor ATF2 regulation by the JNK signal transduction pathway.
Science
267
:
389
393
.
69.
Dérijard
B.
,
M.
Hibi
,
I. H.
Wu
,
T.
Barrett
,
B.
Su
,
T.
Deng
,
M.
Karin
,
R. J.
Davis
.
1994
.
JNK1: a protein kinase stimulated by UV light and Ha-Ras that binds and phosphorylates the c-Jun activation domain.
Cell
76
:
1025
1037
.
70.
O’Neil
J. D.
,
A. J.
Ammit
,
A. R.
Clark
.
2018
.
MAPK p38 regulates inflammatory gene expression via tristetraprolin: Doing good by stealth.
Int. J. Biochem. Cell Biol.
94
:
6
9
.
71.
Lin
W. J.
,
X.
Zheng
,
C. C.
Lin
,
J.
Tsao
,
X.
Zhu
,
J. J.
Cody
,
J. M.
Coleman
,
R.
Gherzi
,
M.
Luo
,
T. M.
Townes
, et al
2011
.
Posttranscriptional control of type I interferon genes by KSRP in the innate immune response against viral infection.
Mol. Cell. Biol.
31
:
3196
3207
.
72.
Pasté
M.
,
G.
Huez
,
V.
Kruys
.
2003
.
Deadenylation of interferon-beta mRNA is mediated by both the AU-rich element in the 3'-untranslated region and an instability sequence in the coding region.
Eur. J. Biochem.
270
:
1590
1597
.
73.
Rigby
W. F.
,
K.
Roy
,
J.
Collins
,
S.
Rigby
,
J. E.
Connolly
,
D. B.
Bloch
,
S. A.
Brooks
.
2005
.
Structure/function analysis of tristetraprolin (TTP): p38 stress-activated protein kinase and lipopolysaccharide stimulation do not alter TTP function.
J. Immunol.
174
:
7883
7893
.
74.
Nasr
N.
,
S.
Maddocks
,
S. G.
Turville
,
A. N.
Harman
,
N.
Woolger
,
K. J.
Helbig
,
J.
Wilkinson
,
C. R.
Bye
,
T. K.
Wright
,
D.
Rambukwelle
, et al
2012
.
HIV-1 infection of human macrophages directly induces viperin which inhibits viral production.
Blood
120
:
778
788
.
75.
Kovarik
P.
,
V.
Castiglia
,
M.
Ivin
,
F.
Ebner
.
2016
.
Type I interferons in bacterial infections: a balancing act.
Front. Immunol.
7
:
652
.

The authors have no financial conflicts of interest.

Supplementary data