IFN-γ protects from lethal IL-17 mediated viral encephalomyelitis independent of neutrophils
© Savarin et al.; licensee BioMed Central Ltd. 2012
Received: 7 March 2012
Accepted: 10 April 2012
Published: 29 May 2012
Skip to main content
© Savarin et al.; licensee BioMed Central Ltd. 2012
Received: 7 March 2012
Accepted: 10 April 2012
Published: 29 May 2012
The interplay between IFN-γ, IL-17 and neutrophils during CNS inflammatory disease is complex due to cross-regulatory factors affecting both positive and negative feedback loops. These interactions have hindered the ability to distinguish the relative contributions of neutrophils, Th1 and Th17 cell-derived effector molecules from secondary mediators to tissue damage and morbidity.
Encephalitis induced by a gliatropic murine coronavirus was used as a model to assess the direct contributions of neutrophils, IFN-γ and IL-17 to virus-induced mortality. CNS inflammatory conditions were selectively manipulated by adoptive transfer of virus-primed wild-type (WT) or IFN-γ deficient (GKO) memory CD4+ T cells into infected SCID mice, coupled with antibody-mediated neutrophil depletion and cytokine blockade.
Transfer of GKO memory CD4+ T cells into infected SCID mice induced rapid mortality compared to recipients of WT memory CD4+ T cells, despite similar virus control and demyelination. In contrast to recipients of WT CD4+ T cells, extensive neutrophil infiltration and IL-17 expression within the CNS in recipients of GKO CD4+ T cells provided a model to directly assess their contribution(s) to disease. Recipients of WT CD4+ T cells depleted of IFN-γ did not express IL-17 and were spared from mortality despite abundant CNS neutrophil infiltration, indicating that mortality was not mediated by excessive CNS neutrophil accumulation. By contrast, IL-17 depletion rescued recipients of GKO CD4+ T cells from rapid mortality without diminishing neutrophils or reducing GM-CSF, associated with pathogenic Th17 cells in CNS autoimmune models. Furthermore, co-transfer of WT and GKO CD4+ T cells prolonged survival in an IFN-γ dependent manner, although IL-17 transcription was not reduced.
These data demonstrate that IL-17 mediates detrimental clinical consequences in an IFN-γ-deprived environment, independent of extensive neutrophil accumulation or GM-CSF upregulation. The results also suggest that IFN-γ overrides the detrimental IL-17 effector responses via a mechanism downstream of transcriptional regulation.
IL-17 and IFN-γ play diverse and often opposing functions during microbial infections, as well as autoimmune diseases. These interactions are partially attributed to their distinct regulation of the neutrophil response. Both IL-17A and IL-17 F signal through the IL-17R to induce granulocyte colony-stimulating factor and stem cell factor, thereby expanding neutrophil progenitors in the bone marrow and spleen as well as increasing mature neutrophils in the blood [1–3]. IL-17 also induces ELR+ CXC chemokines, which attract neutrophils [2, 3]. By contrast, IFN-γ opposes neutrophil recruitment by downregulating expression of neutrophil chemoattractants . Analysis of polarized T cell subsets and genetically deficient mice has provided insight into the distinct effector functions of IL-17 and IFN-γ; however, the interplay between IL-17 and IFN-γ in vivo remains complex [5, 6]. Moreover, downstream effector mechanisms mediating pathological consequences may be tissue- and pathogen-specific and are largely unresolved. For example, Th17 cell-mediated protection is critical during bacterial pneumonia . IL-17-mediated neutrophil recruitment to the infection site also indicates a protective role for Th17 cells during oropharyngeal candidiasis . By contrast, Th17-mediated inhibition of both protective Th1 responses and antimicrobial neutrophil functions increased tissue destruction following gastric candidiasis and pulmonary aspergillosis . These differences may reflect distinct infection sites, as indicated by the distinct immune responses to Candida albicans, which are dependent upon the anatomical site of infection .
Viral infections are often dominated by Th1 responses. However, the coemergence of Th17 and Th1 cells has recently been documented in several infections, including human immunodeficiency virus , simian immunodeficiency virus  and cytomegalovirus . A deleterious role of IL-17 is implied by acute lung injury associated with IL-17-mediated neutrophil recruitment during influenza virus infection . By contrast, Th17 responses are protective against lethal influenza virus infection in IL-10-deficient mice . Similarly, IFN-γ-mediated protection during herpes simplex virus-1 corneal infection correlated with reduced IL-17 production and subsequent neutrophil infiltration . However, the function of IL-17 during central nervous system (CNS) viral infections, including human immunodeficiency virus encephalitis, is unclear, although Th17 cells promote Theiler’s murine encephalomyelitis virus persistence and chronic demyelination by limiting the antiviral cytotoxic T-lymphocyte response .
In contrast to the limited information on IL-17 function during viral encephalitis, analysis of experimental autoimmune encephalitis (EAE) has revealed numerous insights into effector mechanisms as well as crosstalk between Th1 and Th17 cells . Although the inflammatory CNS disease multiple sclerosis and its animal model EAE were historically associated with a Th1 immune response [17, 18], a pro-inflammatory role of IFN-γ was contradicted by substantially increased disease severity and mortality in mice deficient in IFN-γ (GKO) or the IFN-γR [19, 20]. The correlation between increased EAE severity, enhanced Th17 responses and neutrophil infiltration into the CNS of GKO mice suggested that IFN-γ might be protective by inhibiting the Th17 response . Although IL-17−/− mice are susceptible to EAE , adoptive transfer of polarized encephalitogenic CD4+ T cells support Th17 cells as detrimental participants in EAE [23, 24]. However, the pathogenic mechanisms associated with Th17 cells remain an ongoing challenge and may involve multiple pathways. These include excessive CNS neutrophil infiltration and release of degrading enzymes, free radicals and pro-inflammatory cytokines, direct IL-17-mediated neuronal toxicity , and/or secretion of granulocyte macrophage colony-stimulating factor (GM-CSF) as the pathogenic effector molecule [26–28]. These data suggest that the balance between IFN-γ and IL-17 effector functions, as well as their regulation of neutrophils may dictate the outcome of non autoimmune-driven CNS inflammation, such as viral encephalitis.
During encephalomyelitis induced by the strain designated JHMV, CD4+ T cells not only contribute to antiviral effects by enhancing CD8+ T cell function within the CNS  but also mediate viral control in absence of CD8+ T cells . Nevertheless, they also contribute to both clinical disease and demyelination . To define the role of CD4+ relative to CD8+ T cells in viral encephalitis, memory CD4+ T cells from immunized donors were transferred into infected severe combined immunodeficiency (SCID) mice . This study revealed an early morbidity and mortality in infected recipients of CD4+ T cells lacking the ability to secrete IFN-γ compared to recipients of IFN-γ-sufficient CD4+ T cells or infected unreconstituted control mice . Notably, both memory populations were equally effective in controlling virus replication . The lethal outcome was specific for CD4+ T cells lacking IFN-γ , but not for a similar memory CD8+ T cell population deficient in IFN-γ . These data suggest that mortality was related to immune effector functions specific to CD4+ T cells and controlled by IFN-γ.
In this study, SCID recipients of GKO CD4+ T cells infected with JHMV were characterized by extensive neutrophil accumulation and IL-17 expression within the CNS. Neutrophil infiltration in the absence of IFN-γ correlated with significantly elevated levels of CXCL1, independent of IL-17. Moreover, comparison of infected recipients of wild-type (WT) CD4+ T cells depleted of IFN-γ and recipients of GKO CD4+ T cells depleted of IL-17 revealed mortality was due to IL-17, irrespective of abundant neutrophil accumulation. IFN-γ introduced by co-transfer of WT CD4+ T cells with IL-17-producing GKO CD4+ T cells abrogated the detrimental effects of IL-17 without affecting IL-17 transcription within the CNS. These data thus segregate the effects of toxic neutrophil components from IL-17-mediated pathogenesis.
Homozygous BALB/c Thy1.1 mice, provided by Dr. J. Harty (University of Iowa, Iowa City, IA, USA) and GKO BALB/c mice, provided by Dr. R. Coffman (DNAX Research, Palo Alto, CA, USA), were bred locally at the Cleveland Clinic. SCID mice were obtained from the National Cancer Institute (Frederick, MD, USA). Recipients and donors were maintained under sterile conditions and all procedures were performed in compliance with Cleveland Clinic Institutional Animal Care and Use Committee-approved protocols.
The gliatropic JHM strain of mouse hepatitis virus (JHMV)-neutralizing mAb variant designated 2.2v-1 was used for intracerebral infection . JHMV was propagated and plaque assayed on monolayers of DBT cells, a continuous murine astrocytoma cell line . SCID mice were injected in the left hemisphere with 30 μl volume containing 500 PFU of JHMV diluted in endotoxin-free Dulbecco’s modified PBS. The severity of the JHMV-induced clinical disease was graded as follows: 0, healthy; 1, ruffled fur and hunched back; 2, partial hind limb paralysis or inability to turn to the upright position; 3, complete hind limb paralysis; 4, moribund or dead. Virus titers were determined on plaque assay on monolayers of DBT cells as previously described [32, 33]. Briefly, brains were homogenized in ice-cold Dulbecco’s PBS using Ten Broeck tissue homogenizers (Kimble Chase, Vineland, NJ, USA). After clarification by centrifugation at 400 x g for 7 minutes at 4°C, supernatants were stored at −70°C whereas pellets containing CNS-derived cells were suspended in Percoll (GE Healthcare Bio-Sciences AB, Uppsala, Sweden) and used for flow cytometry analysis (see below).
BALB/c Thy1.1 and GKO donors were immunized by intraperitoneal (i.p.) injection with 2 × 106 PFU of JHMV. Donor splenocytes were prepared four to sixteen weeks post immunization. CD4+ T cells were purified by positive selection using anti-CD4-coated magnetic beads (Miltenyi Biotec Inc., Auburn, CA, USA). Purity of the purified population was assessed by flow cytometry using fluorescein isothiocyanate- (FITC) labeled anti-CD4 (clone GK1.5), phycoerythrin- (PE) labeled anti-CD8 (clone 53-6.7) and peridinin chlorophyll protein- (PerCP) labeled anti-CD19 (clone 1D3) mAbs (BD Pharmingen, San Diego, CA, USA). Recipients received 5 × 106 donor CD4+ T cells composed of 100% Thy1.1 (WT), 100% GKO or a 50/50% mixture of Thy1.1/GKO (WT/GKO) CD4+ T cells by intravenous (i.v.) injection coupled with a single i.p. injection of 250 μg of anti-CD8 mAb (clone TIB.210). Mice were challenged with virus two to three hours after adoptive transfer. For neutrophil depletion, mice received i.p. injections of either 500 μg of anti-Ly-6G (clone 1A8) or anti-Gr1 (clone RB6-8C5) mAb every other day until sacrifice, starting two days before infection. Depletion was confirmed in both cases by flow cytometric analysis using anti-Ly-6G (clone 1A8) mAb in addition to examination of hematoxylin and eosin- (H&E) stained sections of brain. Only data for the anti-Ly6G experiments are shown. No differences in survival relative to control-treated mice were observed following treatment with either neutrophil-depleting mAb. Similarly, for anti-IFN-γ treatment, mice received i.p. injections of 500 μg of anti-IFN-γ (clone XMG1.2) mAb every other day, starting two days before infection. For anti-IL17 treatment, mice received i.p. injections of 1 mg of anti-IL-17A (clone 1D10) mAb at day zero and six post infection (p.i.).
After brain homogenization and centrifugation to obtain supernatants for virus determination as described above, cell pellets were resuspended in RPMI containing 25 mM HEPES, pH 7.2 and adjusted to 30% Percoll (GE Healthcare Bio-Sciences BA). A 1 ml underlay of 70% Percoll was added prior to centrifugation at 800 x g for 30 minutes at 4°C. Cells were recovered from the 30% to 70% interface and washed with RPMI medium prior to analysis.
CNS mononuclear cell suspensions were blocked with anti-mouse CD16/CD32 (clone 2.4G2, BD Pharmingen) mAb on ice for 15 minutes prior to staining. Cells were then stained with FITC-, PE-, PerCP- or allophycocyanin-conjugated mAb for 30 minutes on ice in PBS containing 0.1% BSA. Expression of surface molecules was characterized using the following mAbs (all obtained from BD Pharmingen except when indicated): anti-CD45 (Clone Ly-5), anti-CD4 (clone GK1.5), anti-Thy1.1 (clone OX-7), anti-CD8 (clone 53-6.7), anti-CD11b (clone M1/70), anti-F4/80 (Serotec, Oxford, UK), anti-Ly6G (clone 1A8) and anti I-A/I-E (clone 2G9). Samples were analyzed on a FACSCalibur flow cytometer using CellQuest software (Becton Dickinson, Mountain View, CA, USA).
RNA was isolated from three or more individual brains per group using TRIzoL reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. cDNAs were prepared using SuperScript II Reverse Transcriptase (Invitrogen) and oligo (dT)12–18 primers (Invitrogen). Semi-quantitative RNA expression was assessed using LightCycler and SYBR Green kit (Roche, Basel, Switzerland) and the following primers; ubiquitin: F: 5’- TGGCTATTAATTATTCGGTCTGCAT-3’, R: 5’- GCAAGTGGCTAGAGTGCAGAGTAA -3’; IFN-γ: F: 5’- TGATGGCCTGATTGTCTTTCAA-3’, R: 5’- GGATATCTGGAGGAACTGGCAA-3’; IL-17: F: 5’-CTTCATCTGTGTCTCTGATGCTGTT-3’, R: 5’- TCGCTGCTGCCTTCACTGT-3’; IL-22: F: 5’- CATGCAGGAGGTGGTACCTT-3’, R: 5’- CAGACGCAAGCATTTCTCAG-3’; IL-21: F: 5’- GGACAGTATAGACGCTCACGAATG-3’, R: 5’- CGTATCGTACTTCTCCACTTGCA-3’; MHC class II: F: 5’- TCAACATCACATGGCTCAGAAATA-3’, R: 5’- AGACAGCTTGTGGAAGGAATGG-3’; GM-CSF: F: 5’- TTTCCTGGGCATTGTGGTCTA -3’, R: 5’- AAGGCCGGGTGACAGTGAT -3’; IL-6: F: 5’- ACACATGTTCTCTGGGAAATCGT -3’, R: 5’- AAGTGCATCATCGTTGTTCATACA-3’; IL-1β: F: 5’- GACGGCACACCCACCCT-3’, R: 5’- AAACCGTTTTTCCATCTTCTTCTTT-3’; CCL7: F: 5’-GGGAAGCTGTTATCTTCAAGACAAA-3’, R:5’-CTCCTCGACCCACTTCTGATG-3’; CCL20: F: 5’-GGTGGCAAGCGTCTGCTC-3’, R: 5’-GCCTGGCTGCAGAGGTGA-3’; CXCL2: F: 5’-CCTGCCAAGGGTTGACTTCA-3’, R: 5’-TTCTGTCTGGGCGCAGTG-3’; MMP9: F: 5’- CCATGCACTGGGCTTAGATCAT-3’, R: 5’- CAGATACTGGATGCCGTCTATGTC-3’; MMP3: F: 5’- TTTAAAGGAAATCAGTTCTGGGCTATA-3’, R: 5’-CGATCTTCTTCACGGTTGCA-3’; MMP12: F: 5’- GGAGCTCACGGAGACTTCAACT-3’, R: 5’-CCTTGAATACCAGGTCCAGGATA -3’. TaqMan primers and 2X TaqMan fast master mix (Applied Biosystems, Carlsbad, CA, USA) were used to assessed CXCL1 and CCL2 mRNA levels. Levels of mRNA expression were normalized to ubiquitin mRNA using ΔCt method as previously described .
After ice-cold PBS perfusion, brains in OCT were frozen in liquid nitrogen and stored at −80°C until 10 μm sections were prepared. Sections were fixed with methanol/acetone (1:1 ratio) for 15 minutes and then treated with blocking solution for 30 minutes at room temperature. Rat anti-mouse IL-17 (R&D systems, Minneapolis, MN, USA) and hamster anti-mouse CD3 primary mAbs (Serotec) were incubated overnight at 4°C. Alexa Fluor 488 goat anti-rat (Invitrogen) and Alexa Fluor 546 goat anti-hamster (Molecular Probes, Eugene, OR, USA) were added for 1 hour at room temperature. Sections were mounted with Vectashield mounting medium with 4’-6-Diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA, USA) and analyzed using a Leica DM4000B fluorescent microscope (Leica, Wetzlar, Germany).
Cytokine expression by CD4+ T cells derived from cervical lymph nodes of SCID recipients were analyzed directly at day eight p.i. without stimulation with viral antigen. For analysis of cytokine production by cells prior to transfer, JHMV was adsorbed to donor splenocytes for 60 minutes at 4°C and cells cultured for six days in RPMI complete, 10% FCS at 2.5 × 106 cells/ml. Cytokine production from both splenic cultures or ex vivo lymph node cells was measured following four hours stimulation with PMA (10 ng/ml) (Acros Organics, Geel, Belgium) and ionomycin (1 μM) (Calbiotech, Spring Valley, CA, USA). Monensin (2 μM) (Calbiotech) was added to the cultures for the last two hours. After stimulation, cells were harvested and stained for surface expression of CD4. Cells were then permeabilized using the cytofix/cytoperm kit (BD Pharmingen) according to the manufacturer’s instructions and stained for intracellular FITC-IFN-γ and PE-IL-17.
Statistical differences were calculated using the two-tailed unpaired Student’s t-test. P values <0.05 were considered significant. *p < 0.05, **p < 0.01, ***p < 0.001
IFN-γ and IL-17 are major effector molecules of tissue inflammation that play opposing roles in neutrophil recruitment/accumulation [4, 48, 49]. While their distinct influence on disease has been demonstrated during autoimmune-mediated neuroinflammatory responses, the interplay between IL-17 and IFN-γ, specifically the effects on downstream targets remain controversial. Furthermore, during microbial infections, protective and detrimental effects of IFN-γ and IL-17 depend on the pathogen and prominent cell types affecting microbial control [50–52]. The present study evaluated how the absence of IFN-γ secretion by CD4+ T cells contributes to a rapid lethal outcome during viral encephalomyelitis, without altering viral control. Early virus-induced mortality in SCID recipients of GKO virus-specific memory CD4+ T cells correlated with both IL-17 production and extensive neutrophil accumulation in the CNS. Selective blockade of either neutrophils or IL-17 demonstrated that early mortality did not correlate with CNS neutrophil recruitment, but rather with IL-17. This was confirmed by the prolonged survival of recipients of anti-IFN-γ mAb-treated WT recipients, which were characterized by extensive neutrophil inflammation, but an absence of IL-17.
Neutrophil-independent pathogenic effects of IL-17 in the JHMV model contrast with non-CNS viral infectious models including the influenza virus and herpes simplex virus-1 infections, which attribute Th17 cell-mediated pathogenesis to neutrophil attraction [12, 14]. However, neutrophil depletion following severe influenza virus infection also suggests that neutrophils play a protective, rather than a deleterious role . Our data also contrast with the deleterious role of neutrophils during EAE [49, 54]. Adoptive transfer of Th17 cells leads to excessive CNS neutrophil migration after EAE induction, while impaired neutrophil recruitment restrains leukocyte access into the CNS , indicating a prominent role of neutrophils in disrupting the blood-brain barrier. However, in contrast to EAE, neutrophils are not essential for the loss of blood-brain barrier integrity following sublethal JHMV infection . By contrast, JHMV-induced encephalomyelitis demonstrates that IFN-γ plays a more prominent role than IL-17 in regulating CNS neutrophil recruitment and/or retention by downregulating ELR+ neutrophil chemokine expression. Increased neutrophils correlated with high CXCL1 expression in the CNS of both IFN-γ-depleted WT recipients lacking IL-17, as well as in GKO recipients treated with anti-IL-17 Ab. Moreover, neutrophil infiltration was reduced by co-transfer of WT and GKO CD4+ T cells, despite sustained IL-17 expression in the CNS. These results are consistent with early studies identifying IFN-γ as a critical factor regulating CNS neutrophil infiltration , as well as recent observations implicating IFN-γ as a dominant molecule controlling CNS inflammation .
Despite evidence implicating IL-17 as a pathogenic mediator, independent of neutrophils, the mechanism(s) involved in IL-17-induced mortality of JHMV-infected mice remain unclear. Identical viral burden at day eight p.i. in all recipients  indicated that IL-17 does not alter control of virus replication, in contrast to its role in facilitating viral persistence following Theiler’s murine encephalomyelitis virus infection . Sustained Ag independent interaction between Th17 and neuronal cells during EAE correlated with increased neuronal damage due to IL-17-mediated neurotoxicity . Increased gray matter infection, especially in neuronal cells, is associated with premature death following JHMV infection of mice deficient in innate immune components . In addition, there is a preferential distribution of CD4+ T cells in the gray matter of GKO recipients compared to WT recipients , suggesting the possibility that in absence of IFN-γ, IL-17-secreting CD4+ T cells localize proximal to uninfected neurons, contributing to neuronal dysfunction and premature death. However, few neurons are infected early during JHMV pathogenesis in SCID mice and the types of infected cells were similar in all groups, suggesting no alteration in viral tropism . In addition, no differential neuronal loss was found comparing GKO and WT recipients . Similarly, increased expression of GM-CSF in GKO recipients compared to the WT counterparts suggested that GM-CSF might also contribute to disease outcome following JHMV infection. GM-CSF was implicated as a pathogenic effector molecule secreted by Th17 cells during EAE [27, 28]. However, the survival of GKO recipients treated with anti-IL17 did not correlate with a decrease in GM-CSF expression. Although GM-CSF expression is reduced by IFN-γ , the data do not support a pathogenic role of GM-CSF in early mortality of JHMV-infected GKO recipients.
IL-17 mRNA expression in GKO CD4+ T cell recipients suggested Th17 cells as the primary mediators of disease. Nevertheless, IL-17 can also be produced by neutrophils, γδ T cells, NK and CD8+ T cells [57, 58]. A deleterious contribution of neutrophil-derived IL-17, suggested during kidney ischemia-reperfusion , was ruled out by the inability of neutrophil-depletion to rescue mice from early death, as well as the absence of IL-17 mRNA in WT recipients treated with anti-IFN-γ, despite high CNS neutrophil infiltration. IL-17 production by CD4+ T cells derived from immunized GKO donors prior to transfer supports GKO CD4+ T cells as the primary source of IL-17. Moreover, stimulation of WT donor CD4+ T cells strongly induced IFN-γ, but not IL-17, indicating that virus-specific Th17 cells only differentiate in the absence of IFN-γ. These results support previous observations of a minor, if any, role of Th17 cells in the pathogenesis of JHMV-infected immunocompetent WT mice  and corroborate the inhibitory function of IFN-γ on Th17 differentiation during T cell priming . However, our data are novel in demonstrating that memory GKO CD4+ T cells are committed in their ability to produce IL-17 when restimulated in the recipient host, even in the presence of IFN-γ. Although unanticipated, this finding was confirmed by the inability of IFN-γ to downregulate IL-17 production in GKO donor cells in vitro, as well as on in vitro-differentiated mature Th17 cells . Similarly, the IL-27 suppressive function on Th17 differentiation from naïve CD4+ T cells could not be reproduced on memory Th17 cells , supporting the stability of committed Th17 cells. Importantly, the prolonged survival of co-transferred recipients, despite sustained CNS IL-17 expression, suggests that IFN-γ overcomes the deleterious effects of IL-17. However, the mechanisms by which IFN-γ overrides IL-17 function remain unclear. In EAE, IL-17 exerts detrimental effects via signaling in resident CNS cells, with astrocytes implicated as major targets . However, Th17 cell localization proximal to neurons also implicates potential dysregulation of neuronal function . Responsiveness of both cell types to IFN-γ [61, 62] suggests IFN-γ may counteract signaling molecules downstream of the IL-17R.
This study demonstrates that IL-17, in the absence of IFN-γ, can accelerate mortality during viral encephalomyelitis by a mechanism independent of the magnitude of CNS neutrophil infiltration and reversible by IFN-γ.
bovine serum albumin
central nervous system
experimental autoimmune encephalitis
granulocyte monocyte colony-stimulating factor
hematoxylin and eosin
gliatropic JHM strain of mouse hepatitis virus
major histocompatibility complex
polymerase chain reaction
peridinin chlorophyll protein
severe combined immunodeficiency
standard error of the mean
tumor necrosis factor
The authors thank Wenqiang Wei, Kate Stenson and Shabbir Hussain for technical assistance. This work was supported by National Institutes of Health grant NS18146 and National Multiple Sclerosis Society fellowship grant FG-1791-A-1 to CS.