Role for MyD88 Signaling in Murine Gammaherpesvirus 68 Latency

Abstract

Toll-like receptors (TLRs) are known predominantly for their role in activating the innate immune response. Recently, TLR signaling via MyD88 has been reported to play an important function in development of a B-cell response. Since B cells are a major latency reservoir for murine gammaherpesvirus 68 (MHV68), we investigated the role of TLR signaling in the establishment and maintenance of MHV68 latency in vivo. Mice deficient in MyD88 (MyD88^−/−) or TLR3 (TLR3^−/−) were infected with MHV68. Analysis of splenocytes recovered at day 16 postinfection from MyD88^−/− mice compared to those from wild-type control mice revealed a lower frequency of (i) activated B cells, (ii) germinal-center B cells, and (iii) class-switched B cells. Accompanying this substantial defect in the B-cell response was an approximately 10-fold decrease in the establishment of splenic latency. In contrast, no defect in viral latency was observed in TLR3^−/− mice. Analysis of MHV68-specific antibody responses also demonstrated a substantial decrease in the kinetics of the response in MyD88^−/− mice. Analysis of wild-type × MyD88^−/− mixed-bone-marrow chimeric mice demonstrated that there is a selective failure of MyD88^−/− B cells to participate in germinal-center reactions as well as to become activated and undergo class switching. In addition, while MHV68 established latency efficiently in the MyD88-sufficient B cells, there was again a ca. 10-fold reduction in the frequency of MyD88^−/− B cells harboring latent MHV68. This phenotype indicates that MyD88 is important for the establishment of MHV68 latency and is directly related to the role of MyD88 in the generation of a B-cell response. Furthermore, the generation of a B-cell response to MHV68 was intrinsic to B cells and was independent of the interleukin-1 receptor, a cytokine receptor that also signals through MyD88. These data provide evidence for a unique role for MyD88 in the establishment of MHV68 latency.

Murine gammaherpesvirus 68 (MHV68) shares genomic colinearity with Epstein-Barr virus and Kaposi's sarcoma-associated herpesvirus, although it is more closely related to Kaposi's sarcoma-associated herpesvirus and herpesvirus saimiri (42, 44). It is capable of infecting inbred and outbred stains of laboratory mice and therefore provides a tractable small-animal model with which to study gammaherpesvirus pathogenesis (7, 15, 34, 39, 47-49). Hallmarks of MHV68 infection include acute viremia that is cleared approximately 2 weeks postinfection in wild-type mice and is accompanied by a massive expansion of immune cells (52). Acute infection leads to the establishment of long-term latency in the memory B-cell compartment, although other cell types, such as naïve B cells, macrophages, and dendritic cells (DCs), have been shown to harbor latent virus at early times postinfection (16, 17, 52). Like the other gammaherpesviruses, MHV68 has been shown to be associated with lymphoproliferative disease, and long-term infections can lead to the development of lymphomas (35, 47). MHV68 facilitates an understanding of viral and host determinants of gammaherpesvirus pathogenesis in vivo.

Toll-like receptors (TLRs), a type of pattern recognition receptor, are an important part of the innate immune system. TLRs recognize pathogens by detecting pathogen-associated molecular patterns (37, 51). There are 12 known mammalian TLRs, and with the exception of TLR3, engagement through their ligands activates the MyD88-interleukin-1 (IL-1)-associated receptor kinase-tumor necrosis factor receptor-associated factor 6 (MyD88-IRAK-TRAF6) signaling pathway, which then leads to activation of several transcription factors, such as NF-κB, mitogen-activated protein kinase, and interferon regulating factors. Engagement of TLRs expressed on antigen-presenting cells, including DCs and macrophages, with their ligand(s) results in chemokine and cytokine production, increased antigen presentation, and the expression of costimulatory molecules (2, 27). These events can initiate an inflammatory response through chemokine secretion and cellular recruitment (1, 2, 27). TLR engagement mediates the maturation and migration of DCs to lymph nodes, which facilitates interaction with T lymphocytes (27, 29). TLR ligands such as lipopolysaccharide and double-stranded RNA are known to act as adjuvants, enhancing the adaptive immune response (25). DC interactions with naïve T cells differentiate them into TH1 and TH2 or T regulatory lymphocytes (32).

There are several TLRs whose ligands are viral pathogen-associated molecular patterns. TLR3 recognizes double-stranded RNA (3), TLR7 and -8 recognize single-stranded RNA (13, 22, 23), and TLR9 recognizes CpG DNA (24). These TLRs induce an antiviral host defense response, especially secretion of alpha/beta interferon. TLRs have been shown to be important in activating the innate immune response to control virus replication during virus infections. MyD88 signaling is important to control lymphocytic choriomeningitis virus infection and the maturation/activation of virus-specific CD8⁺ T cells (58). TLR signaling has been shown to play a role in several herpesvirus infections. There is a requirement for TLR9 signaling in sensing murine cytomegalovirus to ensure a rapid antiviral response (11). Recently, it was demonstrated that certain laboratory and clinical strains of herpes simplex virus activate TLR2 and TLR9 (41). The TLR family plays an instructive role in innate immune responses against microbial pathogens as well as in the subsequent induction of adaptive immune responses.

There is also an emerging appreciation for the role that TLRs play in the activation of B-cell responses, including germinal-center reactions and antibody production (36). Signaling through TLRs has been shown to play a role in class switching (28, 36) as well as to be necessary for maintenance of neutralizing and long-term antibody responses against infections (20, 57). Signaling through the MyD88 pathway has been shown to be required for B cells to achieve an optimal response to foreign antigens (36). This optimization might be due to direct recognition of the pathogen through the TLR or, rather, as a tertiary signal following recognition of the antigen through the B-cell receptor in conjunction with B-cell receptor recognition and T-cell help.

In the present study, we explore the role of TLR signaling during MHV68 infection. We first examined the B-cell response during MHV68 infection in MyD88^−/− or TLR3^−/− compared to wild-type C57BL/6 mice. We demonstrate that MHV68 infection results in decreased B-cell activation, germinal-center formation, and class switching at early times postinfection in MyD88^−/− mice compared to wild-type mice. This decrease is not seen in TLR3^−/− mice, showing that the above effects are MyD88 specific. Since MHV68 requires germinal-center formation to establish latency (17, 43, 56), we next examined whether MyD88 deficiency had any effect on the ability of MHV68 to establish latency. Corresponding to the decreased B-cell response observed in the MyD88^−/− mice, there was also a decrease in the frequency of viral genome-positive cells. Importantly, this deficiency in B-cell response is intrinsic to MyD88^−/− B cells, as demonstrated by using mixed-bone-marrow chimeric mice. We also show that it is IL-1 receptor (IL-1R) signaling independent. This further illustrates that early establishment of MHV68 latency is dependent on the participation of germinal-center reactions to gain entry into the memory B-cell compartment.

MATERIALS AND METHODS

Cells, viruses, and virus culture.

MHV68 strain WUMS (ATCC VR1465) was used for all virus infections unless indicated otherwise. MHV68-IκBαM.1 and MHV68-IκBαM.MR were kindly donated by Laurie Krug. Virus passage, maintenance, and titration were performed as previously described (10). NIH 3T12 cells and mouse embryonic fibroblast (MEF) cells were maintained in Dulbecco's modified Eagle's medium supplemented with 100 U penicillin/ml, 100 mg streptomycin/ml, 10% fetal calf serum, and 2 mM l-glutamate (cMEM). Cells were maintained at 37°C in a 5% CO₂ environment. MEF cells were prepared from C57BL/6 mice as previously described (38).

Mice, infections, and organ harvests.

C57BL/6 (The Jackson Laboratory), MyD88^−/− (a generous gift from Andrew Gerwirtz), TLR3^−/− (The Jackson Laboratory), IL-1R^−/− (The Jackson Laboratory), and B6.SJL-PtprcaPepb/BoyJ wild-type (Ly5.1) (The Jackson Laboratory) mice were housed and bred in the Emory University Whitehead Building vivarium in accordance with all federal, university, and facility regulations. Mice were placed under isofluorane anesthesia prior to intranasal inoculation with 1,000 PFU of virus in 20 μl of cMEM or by intraperitoneal inoculation of 100 or 1,000 PFU of virus in 500 μl of cMEM. Mice were anesthetized with isofluorane prior to sacrifice by cervical dislocation. Spleens were harvested into cMEM, homogenized, and filtered through a 100-μm-pore-size nylon cell strainer (Becton Dickinson). Erythrocytes were removed with red blood cell lysis buffer (Sigma). Pooled splenocytes from 4 to 10 mice were used in all experiments.

Plaque assays.

Plaque assays were performed as previously described (10), with minor modifications. NIH 3T12 cells were plated onto six-well plates at 2.5 × 10⁵ cells/well 1 day prior to infection. Organs were subjected to four rounds of mechanical disruption of 1 min each, using 1.0-mm zirconia-silica beads (Biospec Products) in a Mini-Beadbeater-8 (Biospec Products). Serial 10-fold dilutions of organ homogenate were plated into NIH 3T12 monolayers in a 200-μl volume and allowed to absorb for 1 h at 37°C, with rocking every 15 min. Immediately after infection, plates were overlaid with 2% methylcellulose in cMEM. After 7 days, plates were stained with a neutral red overlay, and plaques were scored the next day. The limit of detection for this assay is 50 PFU per organ.

Antibodies for flow cytometry.

Cells were resuspended in phosphate-buffered saline (PBS) supplemented with 2% fetal calf serum (FCS) and stained for fluorescence-activated cell sorting (FACS) using a combination of the following antibodies: fluorescein isothiocyanate-conjugated antibodies to immunoglobulin A (IgA), IgG1, IgG2a/2b, IgG3, IgE, IgD, GL7, and CD45.1 (Ly5.1); phycoerythrin (PE)-conjugated antibodies to CD19; peridinin chlorophyll protein-conjugated antibodies to CD45.2 (Ly5.2) (all purchased from BD Pharmingen); and allophycocyanin-conjugated antibodies to CD45.1 (Ly5.1) and CD45.2 (Ly5.2) (purchased from eBioscience). When necessary, rat anti-mouse CD16/CD32 (Fc block) was used to block Fc receptors prior to staining.

Flow cytometry.

For flow cytometry analysis, cells were resuspended at 1 × 10⁶ cells/ml in PBS containing 2% FCS and, prior to being stained, were incubated with Fc block for 10 min. Cells were then stained with a 1:100 dilution of all antibodies except for PE-conjugated antibodies and fluorescein isothiocyanate-GL7, which were used at 1:200, for 20 min on ice in the dark. Cells were then washed with PBS-2% FCS and resuspended in a volume of 200 μl. Data were collected on either a FACSCaliber or LSR II flow cytometer (BD Biosciences) and analyzed using FloJo software (TreeStar).

For FACS, cells were resuspended at 2 × 10⁷ cells/ml and incubated for 10 min on ice in PBS-1% FCS and Fc block. Cells were then stained with PE-anti-CD19, -CD45.1 (Ly5.1), or -CD45.2 (Ly5.2) at 5 μl per 1 × 10⁷ cells for 20 min on ice in the dark. Cells were then washed twice with PBS-1% FCS and resuspended at 1 × 10⁸ cells/ml. Stained cell populations were then sorted by FACSVantage (BD Bioscience). Sorted populations were resuspended in cMEM supplemented with 10% dimethyl sulfoxide and stored at −80°C for limiting-dilution PCR analysis or resuspended in cMEM at 4°C for limiting-dilution ex vivo reactivation analyses as described below.

Magnetic cell separation.

Murine B cells were isolated by depletion of non-B cells by using a B-cell isolation kit (Miltenyi Biotec). Briefly, cells were resuspended at 2 × 10⁸ cells/ml in 1× PBS-0.5% FCS followed by staining with Fc block (0.125 μg/10⁶ cells) on ice for 15 min. Cells were then labeled with biotin-antibody cocktail (biotin-conjugated antibodies against CD43, CD4, and Ter-119) at 10 μl/10⁷ cells for 15 min on ice in the dark, followed by staining with anti-biotin microbeads at 20 μl/10⁷ cells for 15 min on ice. Cells were washed twice with PBS-0.5% FCS and subjected to magnetic separation using an autoMACS instrument (Miltenyi Biotec). Following separation, stained cell populations were collected as described earlier.

Limiting-dilution ex vivo reactivation analyses.

Limiting-dilution analysis to determine the frequency of cells containing virus capable of reactivating from latency was performed as described previously (52, 54). Briefly, bulk splenocytes or sorted cell populations were resuspended in cMEM and plated in serial twofold dilutions (starting with 10⁶ cells) onto MEF monolayers in 96-well tissue culture plates. Twelve dilutions were plated per sample, and 24 wells were plated per dilution. Wells were scored for cytopathic effect at 21 days postplating. To detect preformed infectious virus, parallel samples of mechanically disrupted cells were plated onto MEF monolayers. This process kills >99% of live cells, which allows preformed infectious virus to be discerned from virus reactivating from latently infected cells (52-54). The level of sensitivity of this assay is 0.2 PFU (53). Unless otherwise indicated, significant levels of preformed virus were not detected in these assays.

Limiting-dilution nested PCR detection of MHV68 genome-positive cells.

The frequency of cells harboring the MHV68 genome was determined by a single-copy-sensitivity nested PCR assay directed against the MHV68 ORF50 gene sequence as previously described (53, 54). Briefly, bulk splenocytes or sorted cells were thawed, counted, and resuspended in isotonic buffer. A series of six threefold serial dilutions, starting with 10⁴ cells/well, were plated in a background of 10⁴ uninfected NIH 3T12 cells in 96-well PCR plates. Cells were lysed prior to nested PCR by 6 h of treatment at 56°C in the presence of detergent and proteinase K. Next, 10 μl of round 1 PCR mix was added to each well. Following first-round PCR, 10 μl of round 2 PCR buffer was added to each well and samples were subjected to a second round of PCR. All cell lysis and PCRs were preformed on a PrimusHT thermal cycler (MWG Biotech). Products were resolved by ethidium bromide staining of 2% agarose gels. Twelve PCRs were performed for each sample dilution, and a total of six dilutions were performed for each sample. Every PCR plate contained control reaction mixtures (uninfected cells and 10 copies, 1 copy, and 0.1 copy of plasmid DNA in a background of 10⁴ cells). All of the assays demonstrated approximately single-copy sensitivity, with no false-positive results.

Antibody ELISA.

For MHV68-specific IgG assay, plates were first coated with a 1% final concentration of paraformaldehyde-fixed viral antigen for sample wells, with standard wells coated with a 2-μg/ml concentration of donkey anti-mouse IgG (Jackson ImmunoResearch) in coating buffer (0.1 M Na₂CO₃, 0.2% NaN₃, pH 9.6), and incubated at either 37°C for 2 h or 4°C overnight, and then plates were blocked with 3% bovine serum albumin in PBS and incubated for 2 h at 37°C. After washing of the plates with enzyme-linked immunosorbent assay (ELISA) wash buffer (200 ml 10× PBS, 1 ml Tween 20, with H₂O added to 200 ml), serum samples or mouse IgG at a concentration of 1 μg/ml (Jackson ImmunoResearch) was added to each well at a threefold dilution in ELISA diluent (BD Bioscience) and then incubated for 3 h at 37°C. Plates were then washed, and horseradish peroxidase-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch) was added at a 1:5,000 dilution and incubated for 2 h at 37°C. Plates were washed, and 100 μl of 1:1 developer was added (BD Bioscience). The reaction was stopped after 10 min with 50 μl stop solution (BD Bioscience), and plates were read on a microplate reader (Synergy HT; BioTek). The concentration of antibody in serum samples was determined by using a standard curve generated by threefold serial dilutions of standard in the same assay. Plates were read using the KC4 program.

Passive antibody transfer.

Blood from three to five naïve or infected C57BL/6 or MyD88^−/− mice was harvested by submandibular bleed into Microtainer serum separation tubes (Becton Dickinson) at the indicated times postinfection, pooled, and allowed to clot for 15 min at room temperature. Serum was aliquoted and stored at −80°C until used. One day prior to infection and 7 days after infection, 100 μl of serum was transferred intraperitoneally to each of five naïve C57BL/6 mice for each group. At 16 days postinfection, splenocytes were harvested and assayed as previously described.

Generation of MyD88-mixed-bone-marrow chimeric mice.

B6.SJL-PtprcaPepb/BoyJ wild-type (Ly5.1) mice (The Jackson Laboratory) were lethally irradiated (950 rad) with two doses of 475 rad in a 20-h interval and reconstituted with bone marrow from both MyD88^+/+ (Ly5.1) and MyD88^−/− (Ly5.2) mice. Bone marrow cells were flushed from femurs, depleted of red blood cells, and mixed. Recipient mice (Ly5.1) were anesthetized with 2,2,2-tribromoethanol, and a total of 2 × 10⁷ cells in 100 μl PBS were injected intraocularly. Mice were rested for 8 weeks to allow for reconstitution, which was confirmed by flow cytometry analysis of peripheral blood lymphocytes, before MHV68 infection. Chimeric mice received neomycin and polymyxin B (Sigma-Aldrich) in drinking water until reconstitution was confirmed.

Statistical analyses.

All data were analyzed by using GraphPad Prism software. Titer data were statistically analyzed with the Mann-Whitney nonparametric two-tailed t test. Based on Poisson distribution, the frequencies of reactivation and viral genome-positive cells were obtained from a nonlinear regression fit of the data where the regression line intersects at 63.2%. The frequencies of reactivation and genome-positive cells were statistically analyzed by an unpaired two-tailed t test of the log 63.2% effective concentration.

RESULTS

B cells in MyD88^−/− mice display a decrease in response.

The TLR family plays an instructive role in the innate immune response against microbial pathogens as well as in the subsequent induction of an adaptive immune response. Since the humoral response is a part of the adaptive immune response, we focused on changes in the B-cell response in MyD88^−/− and TLR3^−/− mice during the time when MHV68 is establishing latency, since any alteration in the B-cell response during this time might have impacted MHV68 latency. MyD88 has been reported to be required for B-cell activation, the germinal-center reaction, and class switching (36). To determine whether any of the same effects could be seen during MHV68 infection, MyD88^−/−, TLR3^−/−, and wild-type C57BL/6 mice were infected intranasally with 1,000 PFU of wild-type MHV68. Spleens were harvested at 16, 42, and 90 days postinfection (dpi), labeled with the indicated phenotypic markers, and analyzed by flow cytometry (Fig. 1). While total B-cell numbers did not differ between any of the mice (57.9% ± 4.6% for C57BL/6 mice, 58.5% ± 4.5% for MyD88^−/− mice, and 56.1% ± 4.9% for TLR3^−/− mice), the activation level, as determined by the percentage of cells positive for expression of the activation marker CD69, was significantly reduced in MyD88^−/− mice compared to those in C57BL/6 and TLR3^−/− mice at 16 dpi (Fig. 1A and B and Table 1).

FIG. 1.

Development and phenotype of B-cell response are attenuated in MyD88^−/− mice. Cells were prepared from spleens harvested at the indicated time points from four to six mice infected intranasally with 1,000 PFU as well as from two or three naïve uninfected mice. Representative flow cytometry plots for stained cells are shown. Values shown in the upper right quadrant are percentages of CD19⁺ cells that express the CD69 activation marker (A), the germinal-center markers CD95 (Fas) and GL7 (C), and class switching with a cocktail of antibodies (IgA, IgE, IgG1, IgG2a/b, and IgG3) (E). Bar graphs of absolute cell numbers of CD19⁺ CD69⁺ (B), CD19⁺ GL7⁺ Fas⁺ (D), and CD19⁺ isotype switch⁺ (F) cells from all mice for each time point are also shown. Error bars represent standard deviations between individual mice. The percentages of activated, germinal-center, and class-switched B cells from infected MyD88^−/− mice were significantly different from those of infected C57BL/6 mice at 16 dpi (P = 0.0324, P = 0.0005, and P = 0.0359, respectively).

TABLE 1.

Analysis of B-cell phenotypes in MHV68-infected mice

Mouse straina	dpi	Mean % (SD) of cells in subsetb
		CD19+	CD19+ CD69hi	CD19+ GL7+ CD95+	CD19+ isotype switch+
C57BL/6	16	57.9 (4.6)	22.7 (4.6)	11.7 (4.5)	13.2 (2.9)
MyD88−/−	16	58.5 (4.5)	7.3 (3.3)	4.4 (1.1)	4.7 (0.9)
TLR3−/−	16	56.1 (4.9)	18.8 (6.4)	9.6 (2.0)	11.0 (3.6)
C57BL/6	42	50.5 (2.1)	6.2 (1.8)	6.1 (1.2)	4.1 (1.7)
MyD88−/−	42	52.6 (6.8)	7.5 (2.7)	4.6 (0.74)	3.7 (0.30)
TLR3−/−	42	54.2 (6.0)	5.7 (1.4)	6.8 (0.65)	5.3 (0.86)
C57BL/6	90	52.7 (3.5)	4.1 (0.83)	4.0 (1.6)	0.59 (0.23)
MyD88−/−	90	55.9 (4.2)	3.4 (1.0)	3.2 (0.48)	0.56 (0.22)
TLR3−/−	90	58.9 (5.7)	4.0 (1.3)	3.7 (0.84)	0.62 (0.16)

^aC57BL/6, MyD88^−/−, and TLR3^−/− mice were infected intranasally with 1,000 PFU of MHV68.

^bCell subsets were derived from initial FACS gating on live lymphocyte populations of cell preparations from spleens. Cell subsets were derived from initial gating on CD19⁺ cells and then further fractionated into the indicated subsets by using additional surface markers, as indicated. The data shown are the mean percentages determined from FACS analysis of four to six individual mice in triplicate.

To ascertain if MyD88^−/− B cells were able to undergo a germinal-center reaction, we assessed the presence of the germinal-center markers GL7 and CD95 (Fas) on CD19⁺ B cells. At 16 dpi, MyD88^−/− mice had a decreased frequency of germinal-center B cells compared to the other mice (Fig. 1C and D and Table 1). Since immunoglobulin isotype switching is associated with the development of germinal centers and transition through germinal centers, we examined the frequency of class-switched B cells by using a cocktail of antibodies to assess surface expression of IgG, IgA, and IgE. Notably, we observed that at 16 dpi, MyD88^−/− mice had a statistically significant decrease in the ability to undergo class switching compared to MHV68-infected C57BL/6 and TLR ^−/− mice (Fig. 1E and F and Table 1). Taken together, these data demonstrate that MyD88 signaling is required for an optimal B-cell response to MHV68 infection. It should be noted that this phenotype is not due to the absence of receptor signaling for IL-1, a proinflammatory cytokine that has MyD88-dependent signaling. Analyses of MHV68-infected IL-1R^−/− mice at 16 dpi demonstrated that B-cell activation, germinal-center formation, and isotype switching were similar to those of infected C57BL/6 mice (data not shown).

Development of antibody response to MHV68 is delayed in MyD88^−/− mice.

The observed decrease in isotype-switched B cells in MHV68-infected MyD88^−/− mice suggested a defect in antibody production. To determine the impact on the MHV68-specific antibody response, the levels of MHV68 antibody-specific IgG in the sera of infected C57BL/6 and MyD88^−/− mice were determined by ELISA. In agreement with a published report (45), anti-MHV68 antibody was not detected during the first 20 days postinfection from either strain of mouse. However, by day 30, an MHV68-specific IgG response could be detected (Fig. 2A). As expected, at days 30 and 42 postinfection, there was a significant reduction in the amount of MHV68-specific IgG in the MyD88^−/− mice compared to that in wild-type C57BL/6 mice (Fig. 2A). However, by day 60 and later postinfection, both C57BL/6 and MyD88^−/− mice exhibited similar levels of MHV68-specific IgG (Fig. 2A), indicating that the defect in the humoral immune response is transient in MyD88^−/− mice. These data correlate with the decrease in isotype-switched B cells seen in the MyD88^−/− mice at early times postinfection (Fig. 1E and F).

FIG. 2.

MyD88^−/− mice have a decrease in MHV68-specific antibody response. (A) Blood was taken from three to five mice for each of the indicated groups. Sera from infected C57BL/6 and MyD88^−/− mice were analyzed using an MHV68-specific ELISA to detect the presence of antiviral antibodies. The differences in antibody titers determined at 30 and 42 dpi were statistically significant (P = 0.0012 and P = 0.0002, respectively). Naïve sera from both strains of mice had undetectable levels of ΜHV68-specific IgG (data not shown). Pooled serum from 42 to 45 dpi (B) or 180 dpi (C) was used to immunize C57BL/6 mice, followed by MHV68 infection and analysis of virus reactivation from latency at day 16 postinfection.

While ELISA showed a decrease in quantity of MHV68-specific IgG, we also wondered if there was a difference in quality of MHV68-specific antibody between C57BL/6 and MyD88^−/− mice. To investigate whether there was a difference in the abilities of C57BL/6 and MyD88^−/− mouse antibodies to give protection against a challenge infection (in an in vivo antibody neutralization assay), we injected naïve C57BL/6 mice with sera from either C57BL/6 or MyD88^−/− infected mice harvested at days 42 to 45 postinfection (the time at which we observed the greatest difference in the antibody response) (Fig. 2A) or from naïve mice, followed by MHV68 infection 1 day later. Mice were treated again with serum at day 7 postinfection. Immune sera from C57BL/6 mice reduced the level of latency to an undetectable level in the spleen following intraperitoneal infection, while immune sera from MyD88^−/− mice provided only partial protection (Fig. 2B). Naïve serum from either strain of mouse had no effect on the ability of MHV68 to establish latency (Fig. 2B). Thus, the decreased levels of MHV68-specific antibody observed in MyD88^−/− mice at days 42 to 45 postinfection directly correlated with a failure to provide full protection from MHV68 infection. In contrast, sera from MyD88^−/− and C57BL/6 mice harvested at day 180 postinfection displayed similar abilities to protect against a challenge MHV68 infection (Fig. 2C), which was consistent with ELISA data demonstrating comparable levels of MHV68-specific IgG at this time point (Fig. 2A).

TLR signaling is dispensable for acute MHV68 replication in vivo.

To evaluate the effects of TLR signaling on acute MHV68 replication, we infected MyD88^−/−, TLR3^−/−, and wild-type C57BL/6 mice intranasally with 1,000 PFU of wild-type MHV68. At days 4 and 9 postinfection, virus lung titers were determined by plaque assay. Titers in the lungs at both time points were comparable between all three types of mice (Fig. 3). Additionally, both MyD88^−/− and TLR3^−/− mice displayed no signs of illness after intranasal infection with 1 × 10⁶ PFU for >6 months (data not shown). Despite the precedence for TLRs to be used to control productive viral infections (11, 26, 41, 58), these findings indicate that TLR signaling is dispensable for control of acute MHV68 replication in vivo.

FIG. 3.

Acute replication in the lungs is unaffected in MyD88^−/− and TLR3^−/− mice. C57BL/6, MyD88^−/−, and TLR3^−/− mice were inoculated intranasally with 1,000 PFU of wild-type MHV68. Lungs were harvested at 4 and 9 dpi. The results were compiled from two independent experiments with four to six mice per group. Virus titers in lungs were determined by plaque assay on NIH 3T12 monolayers. Each point represents the virus titer from an individual mouse. The solid bars represent the geometric means, and the dashed line represents the limit of detection of the assay (50 PFU).

Loss of MyD88 signaling results in route-of-inoculation-specific defects in MHV68 latency.

Previous studies have shown that MHV68 latency is preferentially established in activated germinal-center B cells that allow access to memory B cells (17, 43, 56). Since a decrease in this B-cell population was seen in MyD88^−/− mice, but not in C57BL/6 or TLR3^−/− mice, we examined if this had any effect on the ability of MHV68 to establish latency. To evaluate this, we determined the frequency of splenocytes harboring the MHV68 genome by employing a limiting-dilution PCR assay. Spleens from infected mice were harvested at days 16, 42, and 90 postinfection. At 16 dpi, there was a ca. 10-fold reduction in the frequency of splenocytes harboring viral genomes (1 in 690 for MyD88^−/− mice compared to 1 in 68 for C57BL/6 mice and 1 in 110 for TLR3^−/− mice) (Fig. 4A). In addition, we also analyzed the frequency of latently infected cells that were capable of reactivating from latency ex vivo. In MyD88^−/− mice, there was also a ca. 10-fold decrease in this frequency (1 in 13,500) compared to those of C57BL/6 mice (1 in 1,800) and TLR3^−/− mice (1 in 2,600) (Fig. 4B). This decrease in reactivation reflects the decrease in the frequency of viral genome-positive cells, and thus, following intranasal inoculation, there does not appear to be any contribution of MyD88 to virus reactivation. Notably, MyD88^−/− mice exhibit lower levels of latent infection, despite a less effective humoral response against MHV68. Importantly, no defect in MHV68 latency was observed in IL-1R^−/− mice (data not shown).

FIG. 4.

Splenic latency is markedly reduced in MyD88^−/− mice. Spleens were harvested from mice that were infected intranasally with 1,000 PFU wild-type MHV68. The frequencies of bulk splenocytes harboring viral genomes at 16 dpi (A), 42 dpi (C), and 3 months postinfection (D) were determined by limiting dilution PCR. (B) Frequency of cells with reactivating virus at 16 dpi.

The defect in MHV68 latency in MyD88^−/− mice was maintained at day 42 postinfection, with the frequency of latently infected cells dropping in wild-type, TLR3^−/−, and MyD88^−/− mice (1 in 6,500 in MyD88^−/− mice versus 1 in 1,340 in C57BL/6 mice and 1 in 880 in TLR3^−/− mice) (Fig. 4C). However, by 3 months postinfection, all three strains of mice exhibited comparable frequencies of splenocytes harboring MHV68 genomes (1 in 5,000 for C57BL/6 mice, 1 in 4,050 for MyD88^−/− mice, and 1 in 3,300 for TLR3^−/− mice) (Fig. 4D). The latency defect in MyD88^−/− mice correlates with the loss in B-cell activation and germinal-center formation observed at the peak of MHV68 latency (Fig. 1).

To examine whether a more permissive route of infection would alter the phenotype of MHV68 in MyD88^−/− mice, we infected MyD88^−/− and C57BL/6 control mice with 1,000 PFU of virus via intraperitoneal inoculation. We assessed latency at day 18 postinfection in both peritoneal exudate cells (PECs) and splenocytes. Limiting-dilution PCR analyses to determine the frequency of cells harboring viral genomes revealed no defect in the establishment of MHV68 in either PECs or splenocytes (Fig. 5A and C). However, limiting-dilution ex vivo reactivation analyses did reveal distinct reactivation phenotypes in PECs and splenocytes recovered from MyD88^−/− mice. Notably, PECs from MyD88^−/− mice exhibited a mild hyperreactivation phenotype (Fig. 5B), suggesting that MyD88 may play a role in suppressing virus reactivation from latently infected peritoneal macrophages. In contrast, we observed a significant defect in virus reactivation from latently infected splenocytes recovered from MyD88^−/− mice (Fig. 5D). This is particularly notable since we did not observe an additional defect in reactivation following MHV68 infection via intranasal inoculation of MyD88^−/− mice but saw only a defect in establishment of latency (Fig. 4). This may reflect route-specific differences in the splenic B-cell populations harboring virus at day 16 and points out the potential importance of assessing multiple routes of infection in the characterization of MHV68 pathogenesis.

FIG. 5.

MHV68 reactivation from latently infected PECs is enhanced by loss of MyD88. PECs and spleens were harvested from mice infected with 1,000 PFU wild-type MHV68 via intraperitoneal inoculation. Frequencies of bulk PECs (A) and splenocytes (C) harboring viral genomes at 18 dpi were determined by limiting dilution PCR (see Materials and Methods). The frequencies of PECs (B) and splenocytes (D) with reactivating virus upon explantation at day 18 postinfection were determined using a limiting dilution ex vivo reactivation assay (see Materials and Methods).

Inhibition of NF-κB signaling in MHV68-infected MyD88^−/− mice leads to a further reduction in viral latency.

NF-κB is a key component in TLR signaling, and many genes induced by TLRs are NF-κB dependent. To discern if the defect in establishing latency in MyD88^−/− splenic B cells can be attributed solely to the absence of NF-κB signaling modulated by MyD88, we infected MyD88^−/− mice with a recombinant MHV68 expressing a mutant form of IκBα (IκBαM) that functions as a constitutive repressor of NF-κB activation (9, 31). As we previously observed in normal immunocompetent mice (31), upon infection with 1,000 PFU via intranasal inoculation we observed that acute virus replication in the lungs was unaffected by inhibiting NF-κB activation (Fig. 6A). However, when we assessed the frequency of latently infected splenocytes at day 16 postinfection, we observed that inhibiting NF-κB activation further exacerbated the latency phenotype in MyD88^−/− mice. MyD88^−/− mice infected with the control marker rescue virus (MHV68-IκBαM.MR) exhibited the expected frequency of latently infected cells (1 in 760), while MyD88^−/− mice infected with the recombinant MHV68 expressing the IκBαM superrepressor (MHV68-IκBαM.1) had a frequency of viral genome-positive cells of <1 in 10,000 (Fig. 6B). Thus, while we cannot rule out a role for NF-κB activation driven by MyD88 in the phenotype of MHV68 in B cells, there are clearly other pathways involved in activating NF-κB during the establishment of latency that are critical for MHV68 latency.

FIG. 6.

Decrease in B-cell response and MHV68 latency observed in MyD88^−/− mice cannot be attributed solely to a lack of NF-κB signaling. MyD88^−/− mice were infected with 1,000 PFU of MHV68-IκBαM.1 or MHV68-IκBαM.MR via intranasal inoculation. (A) At day 7 postinfection, lungs were harvested and titrated on NIH 3T12 fibroblasts. The data shown represent two independent experiments with three mice per group. (B) Frequencies of splenocytes harboring viral genomes were compiled from three independent experiments, with each containing four to six mice per group.

MyD88^+/+/MyD88^−/− mixed-bone-marrow chimeric mice are able to control acute MHV68 replication, but MyD88^−/− B cells in these mice retain defects in B-cell differentiation and establishment of MHV68 latency.

To determine if the defect in the B-cell response is intrinsic to MyD88 signaling in B cells or other cell types, we generated MyD88^+/+/MyD88^−/− mixed-bone-marrow chimeric mice. Lethally irradiated C57BL/6 (Ly5.1) mice were constituted with a mixture of wild-type (Ly5.1) and MyD88^−/− (Ly5.2) bone marrow cells. Eight weeks after generation of the chimeras, and prior to MHV68 infection, chimerism was verified by FACS analysis of peripheral blood (Fig. 7A). As expected, based on the replication of MHV68 in MyD88^−/− mice, the chimeric mice were able to control acute replication, with comparable kinetics to those of MyD88^+/+ wild-type and Ly5.2 MyD88^−/− mice, after a 1,000-PFU intranasal infection (Fig. 7B). Importantly, this demonstrated that the bone marrow transplants had appropriately reconstituted these chimeric mice.

FIG. 7.

MyD88^+/+/MyD88^−/− mixed-bone-marrow chimeric mice were successfully reconstituted and able to control acute lytic replication in the lungs. Mice were bled by submandibular bleed, and peripheral blood mononuclear cells were examined for reconstitution. (A) Wild-type marker, Ly5.1; MyD88^−/− marker, Ly5.2. Chimeras express both markers. Data shown are representational plots. After reconstitution was confirmed, mice were infected intranasally with 1,000 PFU wild-type MHV68, and lungs were harvested. (B) Viral titers were determined by plaque assay on NIH 3T12 cell monolayers as previously described.

We subsequently assessed whether the defective B-cell response observed in the MyD88^−/− mice also occurred in the MyD88^−/− B cells of the mixed-bone-marrow chimeras. The ability of the MyD88^+/+ B cells and MyD88^−/− B cells to respond to infection was analyzed by flow cytometry at days 16 and 42 postinfection. Notably, MyD88^−/− B cells from the chimeric mice exhibited the previously observed defect in the ability to respond to virus infection (Fig. 8). At 16 dpi, they displayed a decrease in the frequency of MyD88^−/− B cells that were activated compared to MyD88^+/+ B cells (Fig. 8A and Table 2). The ability to undergo a germinal-center reaction was also diminished for MyD88^−/− B cells (Fig. 8B and Table 2). Finally, the ability of MyD88^−/− B cells to class switch was also decreased compared to that of MyD88^+/+ B cells (Fig. 8C and Table 2). These results are entirely consistent with the B-cell responses observed following MHV68 infection of MyD88^−/− mice, giving support to the hypothesis that the observed defect is intrinsic to MyD88^−/− B cells and not the consequence of some other defect in the MyD88^−/− mice.

FIG. 8.

In MyD88^+/+/MyD88^−/− chimeric mice, there is a selective failure of MyD88^−/− B cells to differentiate. Cells were prepared from spleens harvested at the indicated time points from 5 to 10 mice infected intranasally with 1,000 PFU as well as from 5 naïve uninfected mice. Bar graphs show absolute cell numbers of CD19⁺ CD69⁺ (A), CD19⁺ GL7⁺ CD95⁺ (B), and CD19⁺ isotype switch⁺ (C) cells among MyD88^+/+ (Ly5.1) and MyD88^−/− (Ly5.2) cells from MyD88^+/+/MyD88^−/− chimeric mice for each time point. Error bars represent standard deviations between individual mice. Asterisks indicate that differences between MyD88^−/− and MyD88^+/+ B cells were statistically significant, for CD69 expression (P < 0.0001), germinal-center reactions (P < 0.0001), and class switching (P < 0.0001), at 16 dpi.

TABLE 2.

Analysis of B-cell phenotypes in infected MyD88^+/+/MyD88^−/− chimeric mice

Cell populationa	dpi	Mean % (SD) of cells in subsetb
		CD19+ CD69hi	CD19+ GL7+ CD95+	CD19+ isotype switch+
MyD88+/+	16	18.9 (4.8)	12.5 (1.5)	12.1 (3.6)
MyD88−/−	16	4.8 (3.2)	3.2 (1.3)	3.6 (1.6)
MyD88+/+	42	4.7 (1.4)	5.2 (1.3)	3.7 (1.5)
MyD88−/−	42	3.3 (1.9)	3.7 (1.2)	2.9 (1.5)

^aMice were infected with 1,000 PFU of wild-type virus by intranasal inoculation.

^bCell subsets were derived from initial FACS gating on live lymphocyte populations of cell preparations from spleens. Cell subsets were derived from initial gating on CD19⁺ cells and then further fractionated into the indicated subsets by using additional surface markers, as indicated. The data shown are the mean percentages determined from FACS analysis of 4 to 10 individual mice in triplicate.

The observed defect in the MyD88^−/− B-cell response in the MyD88^+/+/MyD88^−/− mixed-bone-marrow chimeric mice strongly suggested that there would also be a defect in MHV68 latency in the MyD88-null B cells. To assess the impact of MyD88 on B-cell latency in the chimeric mice, we separated splenocytes by magnetic separation into B-cell and non-B-cell populations (the purity of the B-cell population was >94% [data not shown]), followed by flow cytometry to further sort the purified B-cell population into Ly5.1 and Ly5.2 populations (>97% purity for each population [data not shown]). The MyD88^+/+ B-cell populations harbored the expected frequencies of latently infected cells at both days 16 and 42 postinfection (Fig. 9A and C). In contrast, the MyD88^−/− B cells exhibited a ca. 10-fold lower frequency of latently infected cells at day 16 and a ca. 5-fold lower frequency at day 42 postinfection (Fig. 9A and C). Finally, the expected decrease in the frequency of B cells reactivating virus was observed in the purified MyD88^−/− B cells, which could be accounted for by the decreased frequency of latently infected cells (Fig. 9B). These results very closely recapitulate the results obtained with MHV68-infected MyD88^−/− mice and, furthermore, are consistent with the analyses of B-cell responses in the MyD88^+/+/MyD88^−/− chimeric mice (Fig. 8). Therefore, these data show that MyD88 plays a role in enhancing the ability of MHV68 to establish latency in the memory B-cell compartment.

FIG. 9.

In MyD88^+/+/MyD88^−/− chimeric mice, MHV68 latency is established preferentially in MyD88^+/+ B cells. At the indicated times after intranasal infection of MyD88^+/+/MyD88^−/− chimeric mice, 4 to 10 spleens were pooled and sorted for Ly5.1-positive MyD88^+/+ B cells and Ly5.2-positive MyD88^−/− B cells. Frequencies of MHV68 latency were determined by limiting dilution PCR at 16 (A) and 42 (C) dpi. (B) MHV68 reactivation from latency, determined by scoring the cytopathic effect on MEF monolayers (see Materials and Methods).

DISCUSSION

Here we report several findings regarding a novel role for MyD88 signaling during MHV68 infection, including the following: (i) TLR signaling is dispensable during acute MHV68 replication; (ii) MyD88 signaling is required for the generation of an optimal B-cell response during MHV68 infection, which correlates with defects in B-cell latency; (iii) MyD88^−/− mice display a significant delay in the appearance of a robust MHV68-specific IgG response; (iv) the MHV68 latency phenotype observed in the MyD88^−/− mice does not appear to be due solely to the loss of NF-κB signaling, as the phenotypes of a recombinant MHV68 expressing a superrepressor of NF-κB were substantially different in MyD88-sufficient and MyD88-deficient mice; and (v) using mixed-bone-marrow chimeras, we have shown that the defective B-cell response to MHV68 infection, as well as the defects observed in MHV68 latency, are intrinsic to MyD88^−/− B cells. Taken together, these data demonstrate that MyD88 signaling in B cells plays an important role in the formation of a B-cell response during MHV68 infection as well as in the establishment of MHV68 latency, although both defects are transient, being ameliorated by 3 months postinfection.

The absence of a role for TLR signaling in controlling acute MHV68 replication contrasts with results obtained with some other viruses (e.g., murine cytomegalovirus and lymphocytic choriomeningitis virus), where TLR signaling has been reported to play an important role (11, 58). During MHV68 infection, other pathways, most notably type I interferons (4, 14), may be the primary mechanisms involved in controlling acute replication. This does not rule out an as yet undiscovered pattern recognition receptor or other innate immune mechanisms being involved in controlling MHV68 replication.

There is accumulating, albeit controversial, evidence that MyD88-dependent TLR signaling directly on B cells is critical for their full activation (19, 36). MyD88^−/− B cells during MHV68 infection display decreased activation and germinal-center participation and, as a consequence of this, a decrease in isotype switching. However, as we have noted, this impairment in the B-cell response appears to be transient and is a phenotype that, so far, is unique to MHV68. MyD88 signaling is required for the formation of a long-term humoral response to polyomavirus (20). The latter is not due to a lack of signaling of IL-1R and IL-18R, which signal through MyD88, as IL-1R^−/− and IL-18R^−/− mice did not display a defect (20). During an influenza virus infection, MyD88 fine-tunes the anti-influenza virus B-cell response via TLR7 signaling (21). These reports, in conjunction with our studies, highlight a role for MyD88 signaling in the formation of an optimal B-cell response to viral infections.

Early in MHV68 infection, latency is found primarily in proliferating B cells bearing markers characteristic of B cells participating in germinal-center reactions (17, 43, 56). Characterization of splenocytes recovered from MHV68-infected MyD88^−/− mice following intranasal inoculation revealed a decrease in latency at 16 and 42 dpi. This was not seen in TLR3^−/− mice, which signal through TRIF, not MyD88, illustrating that this is a MyD88-dependent phenomenon. Notably, the latency defect in MyD88^−/− mice was not absolute, indicating that other pathways that activate B cells, such as CD40-CD40 ligand interaction, direct activation through the B-cell receptor, or a direct infection of memory B cells (a pathway that may be independent of B-cell activation), may contribute to the pool of latently infected splenocytes. The reduction in splenic latency in MyD88^−/− mice could reflect, in part, a role for this host cell factor in mediating inflammatory cytokine production or up-regulation of homing molecules to the spleen by infected cells, which could lead to alterations in the recruitment, activation, and/or subsequent trafficking of infected cells to the spleen. MyD88 signaling may also be required for proliferation or survival signals upon entering the germinal center.

Following intraperitoneal inoculation, MHV68 did not require MyD88 signaling for the establishment of latency in either PECs or the spleen. However, unlike the intranasal route of inoculation, virus reactivation from splenocytes was significantly impaired in MyD88-deficient mice following intraperitoneal inoculation, suggesting route-specific differences in latently infected cell populations and/or their activation state. At this point, little is known about how latency is seeded to the spleen following intranasal inoculation, although studies with B-cell-deficient mice clearly demonstrate a critical role for B cells. With respect to latency in PECs, macrophages are the major cell type harboring latent MHV68 in PECs. As such, the observed hyperreactivation phenotype suggests that MyD88 signaling plays a role in suppressing MHV68 reactivation from latently infected macrophages.

MyD88^−/− mice also displayed slower kinetics in the development of an MHV68-specific IgG response. Notably, the slow appearance of a robust IgG response correlated with decreased MHV68 latency at days 16 and 42 postinfection after intranasal inoculation, suggesting that MyD88 signaling early in MHV68 infection is involved in activating B cells, which is beneficial to both the humoral immune response and the establishment of MHV68 latency. Consistent with this observation, sera recovered from MyD88^−/− mice at day 42 postinfection also displayed a decrease in the ability to protect against establishment of MHV68 latency following challenge with wild-type MHV68 compared to sera harvested from MHV68-infected C57BL/6 mice at the same time point. On the other hand, sera recovered at day 180 postinfection, a time point at which similar levels of MHV68-specific IgG were present in both strains of mice, were equally efficient in neutralizing establishment of MHV68 latency in vivo. This result underscores the transient nature of the defective B-cell response in MyD88^−/− mice, indicating that, ultimately, other activation pathways overcome the loss of MyD88 function. NF-κB signaling is one of the major activation pathways downstream of TLR signaling. Notably, we previously demonstrated that NF-κB activation is required for proper establishment of MHV68 latency in vivo (31). As such, we investigated whether the primary requirement for MyD88 involved NF-κB activation by using a recombinant MHV68 harboring a dominant inhibitor of NF-κB activation. Somewhat unexpectedly, this analysis revealed no amelioration of the phenotype observed with the MHV68-IκBαM recombinant virus (compared to the marker rescue virus) in MyD88^−/− mice. This suggests either that the required MyD88 activation of NF-κB in B cells occurs prior to MHV68 infection and expression of the NF-κB superrepressor or that an alternative activation pathway driven by MyD88 is involved in modulating MHV68 latency.

Experiments with mixed-bone-marrow chimeric mice provided a direct demonstration that the decrease in the response of MyD88^−/− B cells is intrinsic to these B cells. The defect in B-cell response was preserved in these mice in that there was a selective failure of MyD88^−/− B cells to become activated and enter the germinal centers. When presented with both MyD88^+/+ and MyD88^−/− B cells in the same mouse, MHV68 preferentially established latency in the MyD88^+/+ B cells. It is known that MHV68 infection results in polyclonal activation of B cells, characterized by an up-regulation of the CD69 activation marker and an increase in nonspecific antibody production, in a process that is dependent on CD4⁺ T cells (33, 40, 46). A reduction in CD69 activation has been correlated with both a decrease in splenomegaly and splenic latency (5, 8, 12, 31). This phenomenon was seen in both global MyD88 knockout mice and MyD88^+/+/MyD88^−/− mixed-bone-marrow chimeric mice. Nonspecific B-cell activation might represent a mechanism by which the virus drives B-cell participation in germinal-center reactions to increase the likelihood of gaining access to the long-lived memory B-cell reservoir, and MyD88 signaling may play a role in the activation and subsequent proliferation of B cells. In CD40^+/+/CD40^−/− mixed-bone-marrow chimeric mice, both CD40-deficient and CD40-sufficient B cells were found to harbor viral genomes at 14 dpi, at similar frequencies, although ultimately latency in the CD40-deficient B cells waned at later times postinfection (30). In contrast, the MyD88^+/+/MyD88^−/− mixed-bone-marrow chimeric mice exhibited a defect in the establishment of latency at 16 dpi, suggesting that activation of MyD88 signaling plays an important role in addition to, and perhaps prior to, the engagement of infected cells with CD40 ligand in germinal-center reactions.

The events involved in the activation of B cells, the transition of activated B cells into germinal-center cells, and the succeeding formation of memory B cells are not completely understood. This requirement of MyD88 for the establishment of MHV68 latency may define a unique differentiation step in the process of activation and transition of B cells into the germinal center and then to memory B cells. Alternatively, MyD88 may be required for the expansion, survival, or migration of B cells during an immune response. Ultimately, MHV68 may exploit MyD88 signaling to gain access to memory B cells, its long-term latency reservoir (55).

MyD88 is involved not only in TLR signaling but also in both IL-1R and IL-18R signaling. While IL-1R-mediated signaling is essential during Mycobacterium tuberculosis infection (18), as well as playing a role in controlling H5N1 influenza virus infection (50), neither IL-1R nor IL-18R signaling is necessary to maintain long-term antibody production in response to polyomavirus (20). We show here that IL-1R^−/− mice display no defect in B-cell response to MHV68 and no detectable changes in the frequency of latently MHV68-infected cells. Thus, it does not appear that IL-1R signaling is involved in the observed processes. However, it remains a formal possibility that several MyD88-dependent receptors are involved in the generation of the B-cell response to MHV68, and as such, a defect in only one may not impair the formation of the humoral response.

In summary, we have shown here that MyD88 signaling is critical in the formation of a B-cell response to MHV68 as well as in efficient establishment of viral latency. Our studies demonstrate that during MHV68 infection, B cells lacking MyD88 have a defective early B-cell response and this leads to a decrease in the early establishment of latency. Bone marrow chimeras illustrate that the role for MyD88 signaling is an intrinsic requirement of B cells. Finally, our study provides supporting evidence for the hypothesis that MHV68 benefits from a normal B-cell response to infection to gain access to the long-lived memory B-cell compartment and for the establishment of long-term latency.