MyD88 protects from lethal encephalitis during infection with vesicular stomatitis virus


  • Karl S. Lang,

    Corresponding author
    1. Institutes of Experimental Immunology, Neuropathology and Surgical Pathology, University Hospital of Zurich, Zurich, Switzerland
    • Institute of Experimental Immunology, University Hospital, Schmelzbergstrasse 12, 8091 Zurich, Switzerland, Fax: +41-1-255-4420
    Search for more papers by this author
    • The first three authors contributed equally to this work.

  • Alexander A. Navarini,

    1. Institutes of Experimental Immunology, Neuropathology and Surgical Pathology, University Hospital of Zurich, Zurich, Switzerland
    Search for more papers by this author
    • The first three authors contributed equally to this work.

  • Mike Recher,

    1. Institutes of Experimental Immunology, Neuropathology and Surgical Pathology, University Hospital of Zurich, Zurich, Switzerland
    Search for more papers by this author
    • The first three authors contributed equally to this work.

  • Philipp A. Lang,

    1. Institutes of Experimental Immunology, Neuropathology and Surgical Pathology, University Hospital of Zurich, Zurich, Switzerland
    Search for more papers by this author
  • Mathias Heikenwalder,

    1. Institutes of Experimental Immunology, Neuropathology and Surgical Pathology, University Hospital of Zurich, Zurich, Switzerland
    Search for more papers by this author
  • Barbara Stecher,

    1. Institute of Microbiology, ETH Zurich, Zurich, Switzerland
    Search for more papers by this author
  • Andreas Bergthaler,

    1. Institutes of Experimental Immunology, Neuropathology and Surgical Pathology, University Hospital of Zurich, Zurich, Switzerland
    Search for more papers by this author
  • Bernhard Odermatt,

    1. Institutes of Experimental Immunology, Neuropathology and Surgical Pathology, University Hospital of Zurich, Zurich, Switzerland
    Search for more papers by this author
  • Shizuo Akira,

    1. Department of Host Defense, Research Institute for Microbial Diseases, Osaka University, Osaka, Japan
    Search for more papers by this author
  • Kenya Honda,

    1. Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan
    Search for more papers by this author
  • Hans Hengartner,

    1. Institutes of Experimental Immunology, Neuropathology and Surgical Pathology, University Hospital of Zurich, Zurich, Switzerland
    Search for more papers by this author
  • Rolf M. Zinkernagel

    1. Institutes of Experimental Immunology, Neuropathology and Surgical Pathology, University Hospital of Zurich, Zurich, Switzerland
    Search for more papers by this author


MyD88 is a key adaptor molecule in innate resistance, engaged in most Toll-like receptor, as well as IL-1 and IL-18, signalling. Here, we analyzed the role of MyD88 in innate resistance during infection with vesicular stomatitis virus (VSV) using myd88–/– mice. We found an increased susceptibility to VSV in myd88–/– mice, which was not explained by reduced type I IFN or neutralizing antibody responses. Susceptibility of myd88–/– mice correlated with impaired recruitment of immune cells to the site of infection. In the absence of MyD88 signalling, VSV rapidly spread to the spinal cord and brain causing lethal encephalitis.


type I IFN


vesicular stomatitis virus


Control of pathogen replication initially depends on rapid activation of the innate immune response 1. Innate immune cells such as macrophages and dendritic cells sense foreign pathogens via so-called pattern recognition receptors. Toll-like receptors (TLR) are important sensors of such pathogens. In addition, a recently described family of receptors, the NOD/NALP receptors, are able to sense pathogen signals, which is followed by activation of a multiprotein complex, called the inflammasome 2, resulting in cleavage of pre-IL-1 and pre-IL-18 to its active form. MyD88 is one of the most central adapter molecules in activation of innate resistance. Beside the activation downstream of most TLR, additional sensing of inflammatory processes 3, 4 as well as apoptotic processes by the inflammasome requires MyD88-dependent signalling of IL-1 and IL-18 5.

Vesicular stomatitis virus (VSV) is a highly cytopathic rhabdovirus, closely related to rabies virus, causing rabies-like encephalitis in mice 6. Upon infection, type I interferons (IFN-I) and neutralizing antibodies are rapidly induced. Both are crucial for inhibiting local and systemic spreading of virus to the CNS and preventing death. Here we used mice deficient in MyD88 protein to study the role of innate signals in infection with VSV and found that MyD88 was essential for survival of the infection, but was neither required for production of IFN-I nor for neutralizing antibody formation. MyD88 signalling in the periphery led to accelerated recruitment of immune cells, resulting in phagocytosis of VSV-infected myocytes.

Results and discussion

MyD88 influences survival of VSV infection without influencing IFN-I and antibodies

To analyze the role of innate signals during VSV infection, we infected myd88–/– mice with 2 × 104 and 2 × 106 pfu VSV. Intravenous (i.v.) infection of myd88–/– mice with 2 × 106 or 2 × 104 pfu led to death of mice (Fig. 1A). This data are in accordance with a very recent report in which the LD50 of intranasal infection was between 5 × 102 pfu and 5 × 105 pfu compared with >5 × 105 for C57BL/6 mice 7. VSV replication in peripheral cells and organs is strongly inhibited by IFN-I 8. Mice deficient for the master transcription factor of IFN-I, IRF7, were not able to produce IFN upon VSV infection (Fig. 1B). Also, deficiency in IFN-I receptor (ifnar–/– mice) led to reduced production of IFN-α upon VSV infection, implying a positive feedback loop of IFN-I via its receptor upon VSV infection (Fig. 1B) 9, 10. Consistent with lack of IFN function in irf7–/– and ifnar–/– mice, replication of VSV was strongly enhanced in peripheral organs (Fig. 1C) and mice rapidly died (Table 1). In contrast, and rather unexpected, production of IFN-I was normal in myd88–/– mice and VSV did not detectably replicate in peripheral organs (Fig. 1B and C). Therefore, lack of IFN-I was obviously not responsible for the death of myd88–/– mice. Besides innate resistance mechanisms, neutralizing antibodies crucially control VSV replication in peripheral organs and inhibit lethal viremic spread of VSV to the CNS 11. Recently, it has been demonstrated that engagement of MyD88 signalling suppressed the activity of regulatory T cells, thereby favouring priming of the adaptive T cell responses 12. Furthermore, direct MyD88 signalling in B cells was reported to enhance antibody formation 13. Therefore, impairment of B cell responses might have contributed to death in VSV-infected myd88–/– mice. Although we found some differences in IgG isotype titers in myd88–/– compared to control mice measured by ELISA (Fig. 1D), the neutralizing pre-immune titer and the neutralizing protective antibody response during infection was comparable for myd88–/– and control C57BL/6 mice (Fig. 1E and F). The reduction of antibody titers in myd88–/– mice was unlikely to be responsible alone for lethal disease, as heterozygote TgH(KL25) mice in which 95% of B cells express a transgenic LCMV-specific Ig heavy chain, survived VSV infection, while VSV-neutralizing antibodies were much lower compared to VSV-infected myd88–/– mice (Supporting Information Fig. 1). While the role of MyD88 in antibody responses is controversial 1315, our data suggest that the MyD88-dependent innate immune response to VSV infection does not contribute importantly to the induction of the adaptive immune response, where B cell precursor frequencies are high against VSV. Nevertheless, innate resistance is of major importance to survive VSV infection.

Figure 1.

MyD88 signalling is crucial for survival of infection without affecting IFN-I or adaptive immune responses. (A) myd88–/– and corresponding control mice were infected i.v. with 2 × 104 or 2 × 106 pfu VSV-IND (n=4–5). Survival of mice was monitored. (B, C) myd88–/–, irf7–/–, ifnar–/– and control C57BL/6 mice were immunized with 2 × 106 pfu VSV-IND i.v. IFN-α was analyzed in the serum (B) and viral titers were measured in different organs after 3 days (n=2–6, ifnar–/– mice after 24 h, C). (D–F) myd88–/– and control C57BL/6 mice were infected with 2 × 106 pfu VSV-IND i.v. At day 6 after infection, sera were analyzed for VSV-specific IgG, IgG2b and IgG2c (n=3, D). Binding to the LCMV nucleoprotein served as a control. Sera obtained before virus infection (day 0, n=5, E) and at the indicated days after infection (n=6, F) were analyzed for neutralizing anti-VSV Ig and IgG.

Table 1. Survival of VSV infection in mice deficient in the innate immune system
Toll-like pathwaysAdditional innate pathways

MyD88 signalling is required for induction of IL-1 and chemokines

Pathophysiology of VSV infection in mice, similar to rabies in humans, suggests that after intramuscular infection, initial replication occurs in myocytes and the virus later reaches local neuronal axons. This start of infection is, therefore, more likely to unmask involved resistance mechanisms than a systemic intravenous infection. We found no obvious differences in viral replication in the presence or absence of MyD88 within the first 2 days after intramuscular VSV infection (Supporting Information Fig. 2). To screen the local innate immune response against VSV in muscle, we isolated RNA 10 h after infection and performed RT-PCR for genes known to be involved in innate immune responses. We found, in line with our results from i.v. infection, that VSV up-regulated IRF7 and IFN-I in the muscle independently of MyD88 signalling (Fig. 2A). A number of mRNA specific for genes involved in inhibition of viral replication were up-regulated comparably in myd88–/– and control mice (Supporting Information Fig. 3). IL-1 is an important molecule in triggering of inflammation via MyD88, known to regulate itself via a positive feedback loop 5, 1620. We found a strong up-regulation of IL-1β in muscles infected with VSV (Fig. 2B) but no up-regulation of IL-1α or IL-18. At the same time, no up-regulation of IL-1β was found in the spleen following intramuscular VSV infection, suggesting local activation (Fig. 2B). IL-1β injection into the muscle caused up-regulation of chemokines (Fig. 2C) in control mice. Local injection of VSV up-regulated chemokines in control mice but significantly less in the absence of MyD88 (Fig. 2D). Taken together, MyD88 deficiency, by impairing local IL-1β formation, led to reduced chemokine production upon VSV infection.

Figure 2.

MyD88 signalling is required for induction of IL-1 and proinflammatory chemokines. (A) C57BL/6 (black bars) and myd88–/– (white bars) mice were immunized with 2 × 106 pfu VSV intramuscularly (musculus gastrocnemius). After 10 h, muscles were removed and analyzed for mRNA up-regulation of IFN, IFN-regulating factors and mRNA of known IFN-dependent antiviral genes (Supporting Information Fig. 3). (B) C57BL/6 (black bars) and myd88–/– (white bars) mice were immunized with 2 × 106 pfu of VSV intramuscularly. After 10 h, muscles were analyzed for expression of IL-1α, IL-1β and IL-18. Spleens from the same mice at the same time point were additionally analyzed for expression of IL-1β (n=4–7, *p<0.05 in non-parametric t-test). (C) IL-1β (500 ng) was injected into muscles of C57BL/6 mice. Injection of PBS served as control. After 10 h, mice were analyzed for expression of the chemokines ccl2, cxcl2 and cxcl9 (n=3, * p<0.05 in parametric t-test). (D) C57BL/6 (black bars) and myd88–/– (white bars) mice were immunized with 2 × 106 pfu VSV intramuscularly. After 10 h, muscles were analyzed for up-regulation of the chemokines ccl2, cxcl2, cxcl9 and cxcl10 (n=4–7, * p<0.05 in non-parametric t-test).

Figure 3.

MyD88 is required to inhibit spread of VSV to the CNS. C57BL/6 and myd88–/– mice were immunized with 2 × 106 pfu VSV intramuscularly (musculus gastrocnemius). (A) After 5 days, muscle was removed and histologically analyzed in three to six steps (1 mm per step). Histological slides were stained for VSV, MHC class II (up-regulated on activated cells), CD4 (CD4+ T cells), CD8 (CD8+ T cells), B220 (B cells), F480 (Macrophages) and Gr1 (granulocytes). (B) VSV was quantified by plaque assay (Supporting Information Fig. 2). In addition, the number of VSV-positive myocytes were counted in all three to six cross sections, and the average of infected cells/cross section was calculated for every muscle 5 days after infection (n=9–11 muscles from six to seven animals). (C) The number of VSV-infected myocytes that were infiltrated by immune cells was counted in all three to six cross sections on day 5 after infection. The ratio between non-infiltrated infected myocytes to total infected myocytes was quantified for all muscles (n=9–11 muscles from six to seven animals, p<0.0001, t-test). (D) Spinal cords from myd88–/–and C57BL/6 mice were analyzed by histology 8 days after intramuscular infection with VSV. Sections were stained for VSV. Staining for CD4 and CD8 is shown in Supporting Information Fig. 4 (n=2). (E) For depletion of macrophages, C57BL/6 mice were treated with clodronate (100 μL). Control mice were treated with liposomes. At 1 h after depletion, mice were infected i.v. with VSV. Survival of mice was analyzed (n=5). Brain and spinal cord from three terminally ill clodronate-treated mice were analyzed for presence of replicating virus (n=3).

MyD88 requirement to terminate local VSV infection

Our results demonstrated that MyD88 enhanced peripheral inflammation upon VSV infection. When VSV-infected muscles were analyzed 5 days after infection, only one out of ten control mice, but four out of ten myd88–/– mice still demonstrated replicating virus in the muscle (Supporting Information Fig. 2). Immunohistology revealed similar, statistically not significantly different, amounts of VSV-infected myocytes per slice in myd88–/– mice and C57BL/6 mice (Fig. 3A and B). Nevertheless, VSV-infected myocytes in C57BL/6 mice were heavily infiltrated by mononuclear cells (Fig. 3A), whereas myd88–/– mice displayed statistically significantly reduced infiltrations (Fig. 3A and C). As myocytes are connected to local axons via the neuronal endplate, lysis of myocytes by cytopathic VSV would likely facilitate infection of neuronal axons and spread of virus to the spinal cord. Indeed, myd88–/– but not control mice exhibited VSV-infected neurons in the spinal cord (Fig. 3D, Supporting Information Fig. 4). Thus, death during VSV infection in myd88–/– mice correlated with reduced macrophage infiltration at the site of infection. Depletion of macrophages by clodronate treatment in C57BL/6 mice followed by VSV infection led to quick spread of VSV to the brain, followed by rapid death (Fig. 3E). This further emphasized the importance of macrophages during VSV infection. VSV has been described to activate the innate immune system via TLR7 21 and TLR4 22; however, neither of these two TLR, nor TLR3 and TLR9, were essential for mice to survive infection (Table 1). This could hint at redundancy between TLR7 and TLR4. The up-regulation of IL-1β upon VSV infection and the enhanced susceptibility of il1r–/– mice (Table 1) indicated a Toll-independent activation of the innate immune system, which might be crucial for clearance of VSV infection.

In conclusion, we found that, in addition to production of IFN-I and neutralizing antibodies, the recruitment of macrophages to the site of infection was important for control of the cytopathic rhabdovirus VSV.

Concluding remarks

Ever since the identification of the pathogen-sensing capacity of TLR and of MyD88 as a crucial adaptor protein involved in TLR signalling, both elements of innate resistance have commonly been viewed as inseparably linked. However, MyD88 is also an adaptor protein downstream of IL-1 and IL-18 signalling, being also part of the innate immune effector machinery but only indirectly linked to TLR signalling. As long as the infectious dose is low, activation of TLR requires initial pathogen replication within the host, so that the pathogen “pattern” concentration reaches a threshold to activate its distinct TLR. This early viral replication phase before an immune reaction is initiated may be fatal. Once replicating VSV reaches neuronal axons, the infected host usually gets terminally ill. However, the innate immune system may be activated through alternative and faster pathways. First, when entering a cell, VSV may activate IFN-I very rapidly, which subsequently inhibits viral propagation. Our experiments demonstrate that MyD88 signalling is not involved in rapid IFN formation following VSV infection. However, its activation was still essential for survival of infection. The inability to control infection was correlated with reduced production of IL-1β and reduced production of chemokines at the site of infection. The locally reduced MyD88 signalling resulted in impairment of inflammation in infected muscle tissue, which is in line with the recent finding of Zhou et al.7 that MyD88 signalling (probably in the mucosa) is important during intranasal infection. We suggest that during i.v. infection (with more that 2 × 104 pfu), VSV is not only trapped in the spleen, but is also distributed to peripheral sites. Probably some virus is captured at the site of injection, while some might reach lung and muscles. This could explain the death of myd88–/– mice after i.v. infection. Zhou et al.7 found that myd88–/– mice survived i.v. infection. This apparent difference of results is not explained by different virus doses but may reflect differences in hygienic status of mice influencing innate resistance.

In conclusion, signalling of MyD88 was crucial for survival of viral infection. While TLR/MyD88 signalling did not obviously play a role in induction of IFN or in activation of adaptive immune cells, we found that IL-1β induced in the infected solid tissue itself was crucial for recruitment of sufficient inflammatory cells to induce a myositis capable of controlling the virus and to inhibit spread to the nervous system.

Materials and methods

Mice and viruses

VSV, Indiana strain (VSV-IND, Mudd-Summers isolate) was originally obtained from Prof. D. Kolakofsky (University of Geneva, Switzerland). Virus was propagated on BHK-21 cells at an MOI of 0.01 and plaqued on Vero cells. All mouse experiments were performed in single ventilated cages. During survival experiments, health status of mice was checked twice daily. Upon appearance of clinical signs of VSV replication in the CNS, such as paralysis or scrubby pelt, mice were taken out of the experiment and counted as dead. All animals were on C57BL/6 background (at least nine times backcrossed), except for the tlr7–/– mice that were four times backcrossed. Animal experiments were carried out with authorization of the ‘Veterinäramt’ of the Kanton Zurich and in accordance with the Swiss law for animal protection.

Depletion of macrophages

For depletion of macrophages, mice were treated i.v. with 100 μL clodronate as described 23.

VSV-antibody measurement

Titers of VSV-neutralizing antibodies in sera were determined as described before 24. Sera were diluted in MEM-2% FCS, and the dilution that resulted in 50% reduction of virus plaques was taken as neutralizing titer. For determination of IgG titers, sera were incubated with equal volumes of 0.1 M 2-mercaptoethanol in PBS for 1 h at room temperature before dilution. All mouse sera were heated at 56°C for 30 min for complement inactivation. For VSV-ELISA, 96-well plates were coated with baculo VSV-GP 25. Sera, pre-diluted 1:40, were incubated for 90 min, washed with PBS containing 0.5% Tween 20, and incubated further with anti-mouse IgG-horseradish-peroxidase (HRP) (1:1000; Sigma-Aldrich) for detection of virus-specific IgG responses.


IFN-α ELISA was performed according to the manufacturers suggestions (Research diagnostics RDI, Flanders, NJ).

mRNA gene profiling by quantitative RT-PCR

Total RNA was extracted from 50 mg of muscle tissue using TRIzol reagent (Invitrogen, Paisley, Scotland, UK) following the manufacturers instructions. RNA (5 μg) were reverse transcribed using a ThermoScript RT-PCR System (Invitrogen) kit yielding the cDNA template. Pre-designed TaqMan low-density arrays (format 48 × 8; p/n 4342253), Micro Fluidic Cards (Applied Biosystems, Foster City, CA) were used in a two-step RT-PCR process using the ABI PRISM 7900 HT Sequence Detection System (Applied Biosystems) with a TaqMan Low Density Array Upgrade (Applied Biosystems). For each sample, 200 ng cDNA was loaded. Loading of the low-density arrays, standard thermal cycling conditions and data acquisition were done according to the manufacturer's suggestions. Quantitative real-time PCR amplification and data analysis were performed using an ABI Prism 7000 Sequence Detector System (Applied Biosystems). For analysis, all expression levels of target genes were normalized to the housekeeping gene 18s rRNA (▵Ct). Gene expression values were then calculated based on the ▵▵Ct method, using the mean of three untreated mice as calibrator to which all other samples were analyzed. Relative quantities (RQ) were determined using the equation: RQ=2^-▵▵Ct.


Histological analysis was performed on snap-frozen tissue. Sections were stained with antibodies against VSV, MHC class II, CD4, CD8, B220, F4/80 and Gr1. Staining was developed using a goat anti-rabbit (111-055-144, Jackon) or a goat anti-rat antibody (R40000 Caltag Laboratories, Burlingame, CA) and an alkaline phosphatase-coupled donkey anti-goat antibody (705–055–147 Jackson ImmunoResearch, West Grove, PA) with naphthol AS-BI (6-bromo-2-hydroxy-3-naphtholic acid 2-methoxy anilide) phosphate and new fuchsin as a substrate. The sections were counterstained with hemalaum.

Statistical analysis

Data are expressed as mean ± SEM. For statistical analysis student's t-test was used and p values <0.05 were considered as statistical significant.


We would like to thank Kathrin Tschannen and Iris Miescher for technical support with FACS staining and production of tetramers. We would like to thank Andrè Fitsche, Iris Markewitz, Anita Helminski and Silvia Behnke for histological analysis. This study was supported by the Swiss National Science Foundation grants to Hans Hengartner N°3100A0–100779 and Rolf Zinkernagel N°3100A0–100068, and the Deutsche Forschungsgemeinschaft (DFG) LA1419/1–1 to Karl Lang. The authors do not have any financial conflict of interest.


  1. 1


  2. 2


  3. 3