IL-27 is an immunoregulatory cytokine consisting of p28 and EBI3. Its receptor also has two subunits, WSX1 and gp130. Although IL-27 promotes Th1 differentiation in naive T cells, it also induces IL-10 expression in effector Th1 cells to curtail excessive immune responses. By using p28-deficient mice and WSX1-deficient mice (collectively called IL-27–deficient mice), we examined the role of IL-27 in primary infection by murine γ-herpesvirus 68 (MHV68), a murine model of EBV. Upon airway infection with MHV68, IL-27–deficient mice had more aggravated lung inflammation than wild-type mice, although MHV68 infection per se was better controlled in IL-27–deficient mice. Although epithelial cells and alveolar macrophages were primarily infected by MHV68, interstitial macrophages and dendritic cells were the major producers of IL-27. The lung inflammation of IL-27–deficient mice was characterized by more IFN-γ–producing CD8+ T cells and fewer IL-10–producing CD8+ T cells than that of wild-type mice. An infectious mononucleosis–like disease was also aggravated in IL-27–deficient mice, with prominent splenomegaly and severe hepatitis. Infiltration of IFN-γ–producing effector cells and upregulation of the CXCR3 ligand chemokines CXCL9, CXCL10, and CXCL11 were noted in the liver of MHV68-infected mice. Oral neomycin effectively ameliorated hepatitis, with decreased production of these chemokines in the liver, suggesting that the intestinal microbiota plays a role in liver inflammation through upregulation of these chemokines. Collectively, IL-27 is essential for the generation of IL-10–producing effector cells in primary infection by MHV68. Our findings may also provide new insight into the mechanism of hepatitis associated with infectious mononucleosis.
Epstein–Barr virus is a ubiquitous human γ-herpesvirus infecting >90% of adult populations in the world (1, 2). EBV is usually transmitted via saliva and infects oropharyngeal epithelial cells. After initial replication, the virus infects B cells and spreads throughout the body using B cells as a vehicle. Then, the majority of EBV-infected B cells are eliminated by strong Th1-type immune responses that are characterized by vigorous proliferation of CD8+ T cells, which accounts for mononucleosis in the blood. Subsequently, EBV establishes a life-long latent infection in a small fraction of memory B cells and is intermittently released into the saliva through low-grade replication in oropharyngeal epithelial cells (1, 2). The primary EBV infection is mostly subclinical in early childhood but in later years often causes infectious mononucleosis (IM), a self-limiting lymphoproliferative disease that is characterized by high fever, sore throat, cervical lymph node enlargement, hepatosplenomegaly, and liver dysfunction (1, 2). Furthermore, EBV infection can be fatal in cases of chronic active EBV infection with life-threatening hemophagocytosis and in hosts with the X-linked lymphoproliferative disease trait (1, 2). EBV is also etiologically associated with various human malignancies, including Burkett’s lymphoma and nasopharyngeal carcinoma (1, 2). Thus, it is of great interest to elucidate the immunopathological processes associated with EBV infection.
Although the narrow host range of EBV prevents its study using small animal models, murine γ-herpesvirus 68 (MHV68) is a natural rodent pathogen that belongs to the same γ-herpesvirus subfamily as EBV (3, 4). Upon intranasal inoculation, MHV68 causes a self-limiting airway infection. Thereafter, the virus persists in lung epithelial cells and establishes a latent infection in macrophages, dendritic cells (DCs), and B cells (3–6). MHV68 also induces B cell lymphoma in immunocompromised mice as EBV does in humans (3, 7). Thus, MHV68 provides a useful small animal model for the study of immunopathological processes associated with EBV infection in humans (3, 8, 9).
IL-27, together with IL-12, IL-23, and IL-35, is a member of the IL-12 cytokine family (10, 11). IL-27 is a heterodimer of IL-27p28 and EBV-induced gene 3 (EBI3). Of note, p28 also functions as IL-30 as a homodimer (12) and yet as other cytokines by pairing with cytokine-like factor 1 (13) and with IL-12p40 (14). Similarly, EBI3 functions as IL-35 by pairing with IL-12p35 (10), and it may also pair with IL-23p19 to function as IL-39 (15). IL-27R is also composed of two subunits: WSX1 (also known as IL-27Rα or TCCR) and gp130 (10, 11). WSX1 is specific for IL-27, whereas gp130 is used by various IL-6 family cytokines (10, 11). Thus, the complex molecular relationships of the IL-12 cytokine family make the elucidation of their respective functions difficult.
It is now known that IL-27 is an immunoregulatory cytokine with pro- and anti-inflammatory functions (10, 11, 16–18). IL-27 promotes naive T cells to differentiate into Th1-type effector cells through STAT1 activation and T-bet induction, thus promoting responsiveness to IL-12 and production of IFN-γ (10, 11, 19). In contrast, IL-27 suppresses production of IL-2 (20, 21), IL-17 (22, 23), and GM-CSF (24). Furthermore, IL-27 induces IL-10 expression in IFN-γ–producing Th1-type effector CD4+ and CD8+ T cells via STAT1 and STAT3 activation (25), thereby promoting the generation of IFN-γ and IL-10 double-producing Th1 effector cells in late stages of immune responses (26). Because IL-10 is a potent anti-inflammatory cytokine (27–29), its induction by IL-27 is considered an important regulatory mechanism for preventing excessive Th1-type immune responses (10, 11). Of note, IFN-γ–producing CD4+ T cells are the major IL-10 producers in parasite infections (30–32), whereas the major IL-10 producers in viral infections are IFN-γ–producing CD8+ T cells (33–36).
APCs, such as DCs and macrophages, are the major producers of IL-27 (10, 11). Nelson et al. (37) demonstrated IL-27 production by macrophages and DCs upon infection with MHV68 in vitro; however, they found no significant increases in IL-27 in the serum or spleen of MHV68-infected mice, suggesting a minor role for IL-27 in MHV68 infection in vivo (37). Similarly, by using EBI3-deficient mice, Hu et al. (38) suggested that IL-27 may be dispensable for the generation of IL-10–producing CD8+ T cells in MHV68 infection. However, in addition to IL-27, EBI3 is a component of several other cytokines (10, 15). Thus, by using IL-27p28–deficient mice and WSX1-deficient mice, we re-examined the role of IL-27 in primary infection by MHV68. In this article, we report that airway infection with MHV68 caused severe lung inflammation in the absence of IL-27 that was characterized by increases in IFN-γ–producing CD8+ T cells and decreases in IL-10–producing CD8+ T cells. We confirmed that airway administration of anti–IL-27 aggravated lung inflammation of MHV68-infected wild-type (WT) mice, whereas that of rIL-10 ameliorated lung inflammation of MHV68-infected IL-27–deficient mice. Furthermore, IL-27–deficient mice had a more severe IM-like disease, with prominent splenomegaly and liver dysfunction, than did WT mice. We found that oral neomycin effectively attenuated liver inflammation in WT and IL-27–deficient mice, with downregulation of CXCL9, CXCL10, and CXCL11, suggesting that the intestinal microbiota plays a role in liver inflammation of MHV68-infected mice through upregulation of these chemokines, which attract Th1-type effector cells via CXCR3 (39, 40).
Materials and Methods
C57BL/6 mice were purchased from Japan SLC (Hamamatsu, Japan). IL-27–deficient mice (WSX1‒/‒ and p28‒/‒ mice) were described previously (11, 41, 42). Mice were housed and bred in a specific pathogen–free facility at Kindai University Faculty of Medicine and Tottori University Faculty of Medicine. Female mice (8–12 wk old) were used in the experiments. This study was approved by the Experimental Animal Care Committee of Kindai University and by the Institutional Animal Care and Use Committee of Tottori University; all animal experiments were performed in accordance with institutional guidelines.
Virus stock and plaque assay
BALB/3T12 cells and Vero cells were maintained in DMEM supplemented with 10% FBS, penicillin (100 U/ml), streptomycin (100 μg/ml), and 2 mM l-glutamine. The original stock of MHV68 (WUMS strain) was purchased from the American Type Culture Collection (Manassas, VA). Virus stocks were prepared in BALB/3T12 cells, as described previously (43). We routinely obtained virus titers > 1 × 107 PFU/ml. Virus titers were determined by plaque assays in Vero cells, as described previously (44). In brief, test samples were applied to Vero cell monolayers. After adsorption at 37°C for 1 h, Vero cells were washed with PBS and overlaid with 1% methyl cellulose in DMEM supplemented with 5% FBS. After 6–7 d at 37°C in 5% CO2, the monolayers were washed with PBS, fixed with 4% paraformaldehyde, and stained with methylene blue for counting plaques.
Airway infection by MHV68
Mice were anesthetized with an i.p. injection of pentobarbital sodium (Kyoritus Seiyaku, Tokyo, Japan) and inoculated intratracheally with 50 μl of PBS alone (mock infected) or with 50 μl of PBS containing MHV68 at 5 × 104 PFU (MHV68 infected). In some experiments, mice were also intratracheally administered 50 μl of PBS alone (mock treated) or 50 μl of PBS containing anti–IL-27 (500 ng/g mouse body weight) or recombinant mouse IL-10 (500 ng/g mouse body weight; both from R&D Systems, Minneapolis, MN). In some experiments, mice were divided into two groups; one group was treated with 1 mg/ml neomycin (Nacalai Tesque, Kyoto, Japan) in drinking water starting 2 d prior to infection. Then, mice were mock infected or intratracheally infected with MHV68. Neomycin in drinking water was continued until sacrifice. At the indicated days postinfection (dpi), mice were anesthetized with pentobarbital sodium and exsanguinated, and the organs were removed. For histological studies, right lungs and livers were fixed in 4% paraformaldehyde and embedded in paraffin. For viral plaque assays, left lungs and livers were homogenized in HBSS on ice to prepare tissue extracts. Total DNA and RNA were also extracted from cells and tissues and stored at −80°C until use.
Sections (2 μm thick) were made from paraffin-embedded tissues and stained with H&E for routine histological examination. Immunohistochemistry for CD8+ T cells was performed using polyclonal anti-CD8 Ab (Bioss, Woburn, MA) and a Histofine SAB-PO kit (Nichirei Biosciences, Tokyo, Japan).
Collection of serum and bronchoalveolar lavage fluid samples
To obtain serum samples, blood was allowed to clot for >30 min at room temperature and was centrifuged at 1000 × g for 10 min. Serum samples were stored at −20°C until use. To obtain bronchoalveolar lavage fluid (BALF), the trachea was exposed and intubated with 20G Surflo i.v. catheter (Terumo, Tokyo, Japan). One milliliter of 0.5% BSA-PBS was inoculated intratracheally and recovered. This procedure was repeated five times to obtain ∼5 ml of BALF from each mouse. Cells in BALF were pelleted by centrifugation at 700 × g for 5 min, washed, and resuspended in 2% FBS–RPMI 1640 for subsequent analyses. BALF supernatants were kept at −20°C until use.
Single-cell preparations from lung and liver
To prepare single cells from lungs, 5 ml of PBS was injected into the right ventricle to flush blood from lungs. To isolate lung epithelial cells, the trachea was intubated with 20G Surflo i.v. catheter (Terumo), and 1 ml of Dispase II solution (2.4 U/ml; Roche, Basel, Switzerland) was injected. Then, lungs were removed from the chest cavity, placed in 2 ml of Dispase II solution, incubated at 37°C for 10 min, finely minced, and filtered through a 40-μm nylon mesh (BD Biosciences, Sam Jose, CA). To isolate lung-infiltrating cells, lungs were removed from the chest cavity, finely minced, incubated in RPMI 1640 containing 4 mg/ml Collagenase D, 10 U/ml DNase I, and 1 U/ml heparin at 37°C for 40 min with constant shaking, and filtered through a 40-μm nylon mesh (BD Biosciences). Infiltrating cells were also isolated from liver. In brief, whole livers were flushed with PBS and homogenized on a 40-μm nylon mesh using a syringe plunger. Lymphoid cells were separated from hepatocytes and dead cells by centrifugation on 40% Percoll (GE Healthcare, Chicago, IL).
Cells were stained with various fluorochrome-labeled mAbs (all from BioLegend) and analyzed on an LSR Fortessa X-20 Cell Analyzer (BD) using FlowJo software (TreeStar, Ashland, OR). The cell fractions were defined as follows: Gr-1+ for neutrophils, CD11bloF4/80hi for resident macrophages, CD4+CD3+NK1.1− for CD4+ T cells, CD3−NK1.1+ for NK cells, CD11bhiF4/80lo for infiltrating macrophages, and CD8+CD3+NK1.1− for CD8+ T cells. To stain intracellular cytokines, cells were treated with a protein transport inhibitor (GolgiPlug; BD), stained for surface markers, fixed, and permeabilized in Cytofix/Cytoperm Buffer (BD). After washing in Perm/Wash Buffer (BD), cells were stained with PE-labeled anti-mouse IFN-γ (eBioscience, Santa Clara, CA) or PE-labeled anti-mouse IL-10 (BD).
Concentrations of cytokines and chemokines in BALF and serum samples were determined using ELISA. The following ELISA kits were used following the provided protocols: Quantikine ELISA Kits (R&D Systems) for mouse IFN-γ, IL-10, and CXCL9; Mini ELISA Development Kit (PeproTech, Rocky Hill, NJ) for mouse IL-4 and CXCL10; and a Mouse I-TAC ELISA Kit for mouse CXCL11 (RayBiotech, Norcross, GA).
Single cells were prepared from whole lungs, stained for various cell markers, and sorted into 11 cell fractions using a BD FACSAria II. The cell-gating strategies were as follows: EpCAM+CD45− for epithelial cells, CD11c+F4/80− for DCs, CD11c+F4/80+ for alveolar macrophages, CD11c−F4/80+ for interstitial macrophages, CD3+ for T cells, CD3−NK1.1+ for NK cells, CD3+NK1.1+ for NKT cells, Gr-1+ for neutrophils, B220+ for B cells, CD11c+PDCA-1+ for plasmacytoid DCs, and CD11bhiF4/80lo for infiltrating macrophages. For quantification of relative gene expression, we used the 2−ΔΔCT method with GAPDH as an internal control for normalization (45). In brief, total RNA was prepared from cells and homogenized tissues using an RNeasy Isolation Kit and on-column DNA digestion using RNase-Free DNase (both from QIAGEN), following the vendor’s protocols. cDNA was synthesized using a SuperScript First-Strand Synthesis System for RT-PCR (Life Technologies, Carlsbad, CA) with the oligo(dT) primer. RT-PCR was performed on a StepOnePlus (Life Technologies). We used THUNDERBIRD Probe qPCR Master Mix (Toyobo, Osaka, Japan) and the following TaqMan assay reagents (Thermo Fisher Scientific): Mm99999915_g1 for GAPDH, Mm00461162_m1 for IL-27p28, Mm00469294_m1 for EBI3, Mm00434946_m1 for Cxcl9, Mm00445235_m1 for Cxcl10, Mm00444662_m1 for Cxcl11, Mm01168134_m1 for IFN-γ, Mm00445259_m1 for IL-4, and Mm00439614_m1 for IL-10. The following oligonucleotides (made in-house) were also used: +5′-CCTGCACCACCAACTGCTTAG and −3′-GTGGATGCAGGGATGATGTTC for mouse GAPDH, +5′-GGCCGCAGACATTTAATGAC and −3′-GCCTCAACTTCTCTGGATATGCC for MHV68 open reading frame (ORF)50, and +5′-GTCAGGGCCCAGTCCGTA and −3′-TGGCCCTCTACCTTCTGTTGA for MHV68 ORF65. The MHV genome copy numbers were quantified as follows. Cells and tissues were lysed in a buffer containing 10 mM Tris-HCl (pH 8), 150 mM NaCl, 10 mM EDTA, 0.05% SDS, and 0.1 mg/ml proteinase K. DNA samples were obtained by phenol/chloroform extraction. Quantitative PCR for MHV68 DNA was performed using THUNDERBIRD SYBR qPCR Master Mix (Toyobo) and the MHV68 ORF50 primer pair. For the standard, the MHV68 ORF50 DNA fragment was subcloned in pGEM-T easy vector (Promega, Fitchburg, WI) and used.
Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were measured using Colorimetric/Fluorometric Assay Kits (BioVision, San Francisco, CA), following the vendor’s protocols.
Statistical analyses were performed using the two-tailed Student t test for two groups and one-way ANOVA with Fisher protected least significant difference test for three or more groups. We considered p < 0.05 statistically significant.
Aggravated lung inflammation in MHV68-infected IL-27–deficient mice
We used IL-27p28–deficient mice and WSX1/IL-27Rα–deficient mice to examine the role of IL-27 in primary infection by MHV68. Fig. 1A shows the average body weight of infected animals over time. Consistent with the self-limited nature of MHV68 infection (3, 4, 8), WT mice lost weight at 5–7 dpi but recovered to the original level by 12 dpi. In contrast, p28-deficient mice and WSX1-deficient mice continued to lose weight, even after 7 dpi, suggesting an aggravated disease. Histologically, p28-deficient mice and WSX1-deficient mice had more severe lung inflammation than WT mice (Fig. 1B).
To determine whether MHV68 infection per se was aggravated in p28-deficient mice and WSX1-deficient mice, we determined virus titers in lung homogenates. As shown in Fig. 1C, p28-deficient mice and WSX1-deficient mice had much lower virus titers at 6 dpi compared with WT mice. We also examined viral replication and gene expression in lung homogenates by PCR. The copy numbers of MHV68 DNA (Fig. 1D) and the transcripts of the MHV68 late genes ORF50 and ORF65 (Fig. 1E) were also much lower in p28-deficient mice and WSX1-deficient mice than in WT mice. These results suggested that the aggravated lung inflammation in p28-deficient mice and WSX1-deficient mice was not due to enhanced MHV68 infection but instead was due to lack of the immunoregulatory activity of IL-27 (11).
Identification of MHV68-infected cells and IL-27–producing cells in the lung
We next examined cells infected with MHV68 and cells producing IL-27 in the lung. Single cells were prepared from lungs of MHV68-infected WT mice 6 dpi, stained for various cell surface markers, and sorted into 11 cell fractions (Fig. 2). MHV68 DNA and transcripts of the p28 and EBI3 genes were quantified by PCR. As reported previously (3, 4), epithelial cells and alveolar macrophages were the main cell types that contained high copy numbers of MHV68 DNA and, thus, were primarily infected by MHV68 (Fig. 2A). Similarly, as reported previously (11), DCs and interstitial macrophages were the major cell types that expressed p28 and EBI3 mRNAs at high levels (Fig. 2B). Thus, the cells infected by MHV68 and those producing IL-27 were mostly different, suggesting that IL-27 expression in DCs and interstitial macrophages was induced by a mediator(s) released from MHV68-infected cells. In this context, previous studies have shown that type I IFNs are potent inducers of IL-27 in APCs (33, 46). Thus, it is likely that type I IFNs released from MHV68-infected lung epithelial cells and alveolar macrophages induce expression of p28 and EBI3 in DCs and interstitial macrophages.
Exudate cells and cytokines in BALF
We next examined cells in BALF. As shown in Fig. 3A, neutrophils and NK cells peaked in the early stage of infection (6 dpi), whereas lymphocytes and infiltrating macrophages were dominant in the late stage of infection (12 dpi). Furthermore, p28-deficient mice and WSX1-deficient mice had more neutrophils and NK cells in the early stage of infection and more CD8+ T cells and fewer CD4+ T cells in the late stage of infection compared with WT mice. No significant differences were seen in resident and infiltrating macrophages between WT mice and p28- or WSX1-deficient mice (Fig. 3A). We also quantitated IFN-γ, IL-4, and IL-10 in BALF obtained 14 dpi. As shown in Fig. 3B, BALF from p28-deficient mice and WSX1-deficient mice contained more IFN-γ and less IL-4 and IL-10 than did that of WT mice. Quantitative RT-PCR also revealed that lung tissues of p28-deficient mice and WSX1-deficient mice had much higher expression of IFN-γ and much lower expression of IL-4 and IL-10 compared with WT mice (Fig. 3C). These results suggested that IL-27–deficient mice had highly elevated Th1-dominant immune responses to MHV68 infection compared with WT mice.
Cells infiltrating lung tissues
We next examined cells infiltrating the lung tissues of MHV68-infected mice. As shown in Fig. 4, p28-deficient mice and WSX1-deficient mice had much greater numbers of total and IFN-γ–producing NK cells and CD8+ T cells than WT mice. In contrast, p28-deficient mice and WSX1-deficient mice had significantly fewer total CD4+ T cells than WT mice. No significant difference was seen in the number of IFN-γ–producing CD4+ T cells among these groups. Furthermore, p28-deficient mice and WSX1-deficient mice had much fewer IL-10–producing CD8+ T cells than WT mice. IL-10–producing CD4+ T cells were also significantly decreased in p28-deficient mice and WSX1-deficient mice compared with WT mice. No significant difference was seen in IL-10–producing NK cells among these groups. Of note, p28-deficient mice and WSX1-deficient mice had significantly more Foxp3-expressing CD4+ T cells than WT mice. This suggested an increase in Foxp3+ regulatory T cells to compensate for the decrease in IL-10–producing regulatory effector T cells. These results are consistent with the critical role of IL-27 in the induction of IL-10 expression in Th1-type effector T cells during the late stage of the immune response (25, 26, 47–49).
Effects of anti–IL-27 and rIL-10 on lung inflammation
To further support the role of IL-27 in IL-10 induction in lung inflammation by MHV68 infection, we next administered anti–IL-27 or rIL-10 intratracheally. By H&E staining, anti–IL-27 aggravated lung inflammation in MHV68-infected WT mice, whereas rIL-10 reduced lung inflammation in MHV68-infected WSX1-deficient mice (data not shown). As shown in Fig. 5A, anti–IL-27 significantly increased total and IFN-γ–producing CD8+ T cells in lung tissues of MHV68-infected WT mice, whereas rIL-10 significantly decreased total and IFN-γ–producing CD8+ T cells in lung tissues of MHV68-infected WSX1-deficient mice. In the case of NK cells, although anti–IL-27 did not significantly affect total or IFN-γ–producing NK cells in lung tissues of MHV68-infected WT mice, rIL-10 significantly decreased total and IFN-γ–producing NK cells in lung tissues of MHV68-infected WSX1-deficient mice (Fig. 5A). We also confirmed that anti–IL-27 treatment of MHV68-infected WT mice significantly decreased IL-10 mRNA in lung tissues (Fig. 5B) and IL-10 protein in BALF (Fig. 5C). These results supported that the reduction in IL-10–producing effector cells in the absence of IL-27 aggravated lung inflammation in MHV68-infected IL-27–deficient mice.
An aggravated IM-like disease in MHV68-infected IL-27–deficient mice
MHV68 infection is known to induce an IM-like disease in mice (3, 8). As shown in Fig. 6A, mice infected with MHV68 had profound splenomegaly. Furthermore, at least WSX1-deficient mice had significantly enhanced splenomegaly compared with WT mice. The serum liver enzymes AST and ALT were also highly elevated in MHV68-infected mice (Fig. 6B). Again, p28-deficient mice and WSX1-deficient mice, especially the latter, had much higher serum liver enzymes compared with WT mice. These results confirmed an IM-like disease in MHV68-infected mice. Furthermore, p28-deficient mice, and especially WSX1-deficient mice, had a much more aggravated IM-like disease than WT mice.
Aggravated liver inflammation in MHV68-infected IL-27–deficient mice
We further examined liver inflammation in MHV68-infected mice. H&E staining revealed infiltration of mononuclear cells in the liver, which was more pronounced in p28-deficient mice and WSX1-deficient mice than in WT mice (Fig. 7A). Immunohistochemistry also revealed increased infiltration of CD8+ T cells in p28-deficient mice and WSX1-deficient mice compared with WT mice (Fig. 7A). To further characterize liver-infiltrating cells, we prepared single cells from liver tissues and performed flow cytometric analysis. As shown in Fig. 7B, there were dramatic increases in lymphocytes, total and IFN-γ–producing CD8+ T cells, CD4+ T cells, and total and IFN-γ–producing NK cells in the livers of MHV68-infected mice. Furthermore, with the exception of CD4+ T cells, these cell fractions were much higher in p28-deficient mice and WSX1-deficient mice than in WT mice. We also quantified MHV68 DNA and gene expression in liver tissues by PCR. As shown in Fig. 7C, viral DNA copy numbers and ORF50 and ORF65 mRNAs were much lower in p28-deficient mice and WSX1-deficient mice than in WT mice. Thus, as in the lung, IL-27–deficient mice had much more aggravated liver inflammation than WT mice, although these mice controlled MHV68 infection in the liver better than WT mice. Separately, we confirmed that intratracheal administration of anti–IL-27 to MHV68-infected WT mice aggravated lung inflammation, as well as liver inflammation (Supplemental Fig. 1). These results strengthen the notion that liver infiltration of Th1-type effector cells in the IM-like disease in MHV68-infected mice is mostly independent of direct virus infection in the liver and is induced primarily by nonspecific mechanisms.
Expression of CXCR3 ligand chemokines in the liver
Because Th1-type effector cells are known to express CXCR3 and to be recruited into tissues by CXCR3 ligand chemokines (CXCL9, CXCL10, and CXCL11) (39, 40), we quantified the expression of CXCL9, CXCL10, and CXCL11 in livers of MHV68-infected mice by PCR. As shown in Fig. 7D, expression of CXCL9, CXCL10, and CXCL11 was highly elevated in MHV68-infected mice. Furthermore, expression of these chemokines was much more pronounced in p28-deficient mice and WSX1-deficient mice than in WT mice. Thus, the elevated expression of CXCR3 ligand chemokines in the liver of MHV68-infected mice might be responsible for the enhanced infiltration of Th1-type effector cells to the liver.
Therapeutic effect of oral neomycin on liver inflammation
Because the liver is infused by the portal vein, and the intestinal microbiota is known to promote liver diseases (50, 51), it is possible that the intestinal microbiota plays a role in liver inflammation in the IM-like disease seen in MHV68-infected mice. To test this hypothesis, we treated mice with neomycin, a gut-restricted aminoglycoside, in the drinking water to reduce the intestinal microbiota (52, 53). As shown in Fig. 8A, neomycin significantly ameliorated weight loss in MHV68-infected WT and WSX1-deficient mice. Consistently, neomycin significantly reduced serum AST and ATL levels in WT and WSX1-deficient mice (Fig. 8B). H&E staining also revealed that neomycin reduced the infiltration of inflammatory cells in livers of MHV68-infected WT and WSX1-deficient mice (data not shown). Consistently, neomycin significantly reduced liver infiltration of total and IFN-γ–producing CD8+ T cells and NK cells in WT and WSX1-deficient mice (Fig. 8C). Furthermore, neomycin significantly reduced the expression of CXCR3 ligand chemokines in livers of MHV68-infected WT and WSX1-deficient mice (Fig. 8D). Thus, toxic products from the intestinal microbiota, such as LPS, may promote liver inflammation in MHV68-infected mice by enhancing the production of the CXCR3 ligand chemokines in the liver, in synergy with IFN-γ. We confirmed that serum IFN-γ levels were highly elevated in MHV68-infected mice, especially in MHV68-infected p28-deficient mice and WSX1-deficient mice (Supplemental Fig. 2).
IL-27 is known to have proinflammatory and anti-inflammatory functions (11, 16–18). In the early stage of the immune response, IL-27 promotes Th1 immunity by inducing the expression of a transcription factor, T-bet, in naive T cells, which then upregulates the β2 subunit of IL-12R, thereby conferring responsiveness to IL-12 (10, 11, 19). Later, IL-27 induces the expression of IL-10 in IFN-γ–producing CD4+ and CD8+ effector Th1 cells, thereby curtailing excessive Th1 immune responses (26, 47). IL-10 is a well-known anti-inflammatory cytokine (27–29). IL-10–deficient mice develop a spontaneous enterocolitis due to uncontrolled T cell responses to normal bacterial flora (54). The lack of IL-10 also results in exaggerated immune responses during bacterial, protozoal, and viral infections (27–29). Similarly, IL-27–deficient mice have been shown to develop severe pathological inflammation in Th1 and Th2 responses (11). Thus, the axis of IL-27 and IL-10 is considered to have an important immune-regulatory role in the host to balance the benefits and deleterious effects of immune responses (48, 55).
By using IL-27p28–deficient mice and WSX1/IL-27Rα subunit–deficient mice, we have examined the role of IL-27 in primary infection by MHV68. Although IL-12 cytokine family members have highly complex molecular relationships with interchangeable subunits (10–15), we have obtained very similar results from mice with knockout of p28 (the ligand subunit) and WSX1 (the receptor subunit). Thus, we can confidently conclude that the present findings are primarily due to the absence of IL-27.
Upon airway infection by MHV68, IL-27–deficient mice had a highly aggravated lung inflammation (Fig. 1). This was not caused by enhanced lung infection by MHV68 in the absence of IL-27. On the contrary, IL-27–deficient mice suppressed MHV68 infection much more efficiently than WT mice (Fig. 1). CD8+ T cells and NK cells were markedly increased in BALF (Fig. 3) and lung tissues (Fig. 4) of IL-27–deficient mice. BALF and lung tissues of IL-27–deficient mice contained higher levels of IFN-γ and lower levels of IL-4 and IL-10 than those of WT mice (Fig. 3). Thus, IL-27–deficient mice had much stronger Th1-type immune responses to MHV68 infection than WT mice, accounting for the better suppression of infection per se. Previous studies have shown that IL-27 induces IL-10 in effector T cells, and CD8+ T cells are the major producers of IL-10 in viral infections (33–36). We indeed demonstrated that IL-10–producing CD8+ T cells were dramatically reduced in IL-27–deficient mice (Fig. 4). Thus, the reduced production of IL-10 by effector CD8+ T cells in the absence of IL-27 likely accounts for the enhanced Th1-type immune responses to MHV68 infection in IL-27–deficient mice. Indeed, intratracheal administration of anti–IL-27 to MHV68-infected WT mice aggravated the lung inflammation, with significant increases in lung infiltration of total and IFN-γ–producing CD8+ T cells (Fig. 5) and significant decreases in IL-10 mRNA in lung tissues and IL-10 protein in BALF (Fig. 5). Conversely, intratracheal administration of rIL-10 to MHV68-infected WSX1-deficient mice ameliorated lung inflammation, with significant decreases in lung infiltration of total and IFN-γ–producing CD8+ T cells (Fig. 5). Collectively, our results have demonstrated a critical role for IL-27 in the induction of IL-10–producing Th1-type effector cells during the course of primary airway infection by MHV68.
Recently, using mice intranasally infected with respiratory syncytial virus or influenza A virus, Pyle et al. (56) demonstrated that IL-6, the prototypic inflammatory cytokine, drives production of IL-27 by macrophages and monocytes; this, in turn, promotes local maturation of Foxp3+ regulatory T cells to suppress excessive lung inflammation. Similarly, Do et al. (57) reported that Foxp3+ regulatory T cells stimulated with IL-27 express Lag-3 and display an enhanced suppressive activity on T cell–mediated colitis. We observed significant increases in Foxp3+CD4+ T cells in MHV68-infected lungs of IL-27–deficient mice compared with those of WT mice (Fig. 4). We speculate that a compensatory increase in Foxp3+ regulatory T cells occurs upon the decrease in IL-10–producing regulatory effector T cells in the absence of IL-27. However, according to recent reports (56, 57), regulatory T cells in IL-27–deficient mice, although much greater in number than in WT mice, might have less suppressive activity than those in WT mice.
Primary infection by EBV often causes IM in young adults (1, 2). IM is known to be accompanied by splenomegaly and hepatitis, with the latter characterized by intrasinusoidal infiltration of CD8+ T cells (58, 59). EBV-specific CD8+ T cells recognize EBV-infected cells and destroy them with perforin and other cytotoxic effector molecules (1, 2); however, hepatocytes are not directly infected by EBV (58, 59). Thus, it is likely that infiltrating CD8+ T cells damage hepatocytes mostly through bystander mechanisms in IM hepatitis (60, 61), but it is not clear why CD8+ T cells infiltrate the liver during IM without EBV infection of hepatocytes. Because MHV68 infection provides a mouse model of EBV infection (3, 8), we investigated the IM-like pathology of MHV68-infected WT and IL-27–deficient mice. We observed prominent splenomegaly and hepatitis in MHV68-infected mice, and these were greatly aggravated in IL-27–deficient mice (Fig. 6). There were infiltrations of total and IFN-γ–producing CD8+ T cells and NK cells in the livers of MHV-infected mice and, again, these were strongly elevated in IL-27–deficient mice (Fig. 7). However, IL-27–deficient mice had much lower MHV68 DNA and gene expression in the liver than WT mice (Fig. 7). Thus, as in the lung, IL-27–deficient mice had much less MHV68 infection but more aggravated inflammation in the liver than WT mice. This led us to hypothesize that CD8+ T cells and NK cells infiltrate the liver of MHV68-infected mice primarily through virus-independent mechanisms. Th1-type effector cells are known to be recruited by the chemokines acting on CXCR3: CXCL9, CXCL10, and CXCL11 (39, 40). Thus, a significantly enhanced production of CXCR3 ligand chemokines in the liver might be responsible, in part, for the enhanced infiltration of Th1-type effector cells in the liver of MHV68-infected mice. We indeed demonstrated that the expression of CXCR3 ligand chemokines in the liver was highly elevated in MHV68-infected mice, especially in MHV-infected IL-27–deficient mice (Fig. 7).
The liver is infused by the portal vein that brings in nutrients, as well as toxic products of the intestinal microbiota, such as LPS (50, 51). In particular, LPS is known to synergize with IFN-γ to induce CXCR3 ligand chemokines in tissue cells, such as fibroblasts and endothelial cells (62, 63). We confirmed that MHV68-infected mice had highly elevated IFN-γ in their circulation (Supplemental Fig. 2). Thus, the production of CXCR3 ligand chemokines during strong Th1-type immune responses may be further enhanced in the liver under the influence of the intestinal microbiota, resulting in elevated recruitment of Th1-type effector cells to the liver via CXCR3. To test this hypothesis, we treated MHV68-infected mice with oral neomycin, a gut-restricted aminoglycoside, to reduce the intestinal bacterial load (52, 53). Indeed, we observed a significant therapeutic effect of oral neomycin on weight loss in MHV68-infected WT and WSX1-deficient mice (Fig. 8). Serum liver enzymes were also dramatically reduced by the neomycin treatment (Fig. 8). Total and IFN-γ–producing CD8+ T cells and NK cells infiltrating the liver were significantly reduced by the neomycin treatment (Fig. 8). Although the expression of CXCR3 ligand chemokines in the liver was highly elevated in MHV68-infected WT mice and further so in MHV68-infected IL-27–deficient mice, their expression levels were also significantly reduced by the neomycin treatment (Fig. 8). These results support our hypothesis that bacterial products, such as LPS derived from the portal vein, enhance the production of CXCR3 ligand chemokines in the liver of MHV68-infected mice. Th1-type effector cells are attracted by these chemokines via CXCR3 (39, 40) and cause liver damage, primarily through bystander mechanisms.
In conclusion, we have demonstrated a critical immunoregulatory role for IL-27 during primary infection by MHV68. IL-27 is needed to curtail excessive Th1-type immune responses through the induction of IL-10 in Th1-type effector cells. Furthermore, we have shown involvement of the intestinal microbiota in hepatitis associated with an IM-like disease caused by MHV68 infection. Because the liver is infused by the hepatic artery, as well as by the portal vein that drains the gastrointestinal tract, it is highly exposed to products of the intestinal microbiota, such as LPS. Thus, during strong Th1-type immune responses, intestinal microbial products, such as LPS, may synergize with IFN-γ to upregulate CXCR3 ligand chemokines in the liver (62, 63). Th1-type effector cells are then attracted by the CXCR3 ligand chemokines and cause liver damage, primarily through bystander mechanisms. Our results may provide new insights into the mechanism of hepatitis associated with human IM and other acute viral infections.
We thank Namie Sakiyama for excellent technical assistance.
This work was supported in part by a Core Research for Evolutional Science and Technology grant from the Japan Science and Technology Agency (to O.Y.).
The online version of this article contains supplemental material.
The authors have no financial conflicts of interest.