Considering the potential of endosomal TLR signals to induce proinflammatory cytokine expression, UNC93B1 deficient 3d mice were first assessed for systemic IFN-α, IFN-γ, IL-12 and TNF-α production during early infection with a moderate (5×10⁴ PFU) dose of MCMV. C57BL/6 (WT) and 3d mice were uninfected or MCMV-infected for 40 h or 48 h. Serum was collected at indicated time points and IFN-α, IFN-γ, IL-12p70 and TNF-α were measured by enzyme-linked immunosorbent assay (ELISA). In WT mice, maximal production of IFN-α, IFN-γ and IL-12p70 was detected at 40 h post-MCMV infection before declining by 48 h post-infection (Fig. 1, A–C). In contrast, 3d mice exhibited lower serum levels of these cytokines in response to MCMV infection. Specifically, while serum IFN-α reached 1300±420 pg/mL at 40 h post-MCMV infection in WT mice, IFN-α production was reduced by three-fold at this infection time point in 3d mice (450±300 pg/mL), with comparable levels maintained at 48 h post-infection (Fig. 1A). Likewise, while average IFN-γ concentrations in WT mice reached maximal levels of 530±245 pg/mL at 40 h post-MCMV infection, 3d mice failed to induce detectable levels of this cytokine (Fig. 1B). IL-12p70 production similarly peaked at 40 h post-MCMV infection in WT mice, with levels reaching 1000±600 pg/mL. 3d mice, however, produced 12-fold less IL-12p70 at this infection time point (Fig. 1C). Serum TNF-α levels were elevated in response to MCMV infection at both 40 h and 48 h post-infection in WT mice (69±10 and 63±17 pg/mL, respectively, Fig. 1D). In 3d mice, average concentrations of TNF-α were 6-fold lower than WT at 40 h and 2.5-fold lower than WT at 48 h after infection. These results indicate a requirement for endosomal TLR signals for early systemic proinflammatory cytokine production, and concur with previous reports [35].

Having observed a reduction in the level of proinflammatory cytokines in the serum of 3d mice, cytokine responses in liver cells from 3d mice infected with MCMV were evaluated to address the impact of endosomal TLR signaling in a localized tissue site of infection. The best characterized liver cytokine responses are IFN-α and IFN-γ [20], [21], [36]. Therefore, to determine whether combined endosomal TLR signaling contributes to the production of these cytokines during MCMV infection, IFN-α and IFN-γ in liver homogenates and in individual cell populations were measured in WT and 3d mice uninfected or infected with MCMV for 40 h and 48 h. As shown in Fig. 2A, WT mice displayed four-fold higher levels of IFN-α than 3d mice at 40 h post-infection, with increased levels still evident at 48 h following infection. Since pDCs expressing the marker PDCA-1 have been shown to produce the majority of liver IFN-α at 40 h post-MCMV infection [20], this cell type was examined in WT and 3d livers during early infection. There was evidence of pDC accumulation in the livers of both WT and 3d mice (Fig. 2B). However, liver pDCs from 3d mice were impaired in their ability to express IFN-α. There were 4-fold fewer PDCA-1+ pDCs expressing intracellular IFN-α at 40 h, and 3-fold fewer at 48 h post-infection, in 3d mice as compared to WT (Fig. 2C). This trend was also reflected in the proportion of PDCA-1+ IFN-α+ pDCs at 40 h and 48 h after infection in 3d mice (0.8%±0.4% and 2%±1%) compared with WT (2%±1% and 6%±1%).

3d mice similarly demonstrated a defect in liver IFN-γ production in response to MCMV infection. In WT mice, IFN-γ reached maximal levels of 3500±1200 pg/g liver at 40 h before contracting by approximately half at 48 h post-infection. In contrast, at 40 h post-infection, 3d mice induced 5-fold less IFN-γ than WT (Fig. 2D). NK cells are an important early source of IFN-γ in the liver during MCMV infection [24], [25], and accumulated at this site in both WT and 3d mice (Fig. 2E). Using intracellular cytokine staining, results shown in Fig. 2F demonstrate a 7-fold reduction in the absolute numbers of NK1.1+ TCRβ- liver cells expressing IFN-γ in 3d mice at 40 h post-infection as compared to WT. There were also fewer IFN-γ-expressing liver NK cells in 3d mice by proportion (0.8%±0.3% and 1%±0.4%) when compared to WT (6%±2% and 4%±0.4%) at 40 h and 48 h, respectively. Together, these results demonstrate that, in addition to an effect on systemic cytokine production, combined endosomal TLR signals can affect the expression of critical proinflammatory cytokines in the liver during MCMV infection.

The innate immune response is important both in establishing early control of virus replication and in coordinating downstream adaptive responses. Following MCMV infection, virus-specific CD8+ T cells are recruited to the liver within 5 days and control viral replication at this site through release of cytotoxic molecules and production of cytokines such as IFN-γ and TNF-α [26], [27]. Given the abated liver cytokine responses observed in 3d mice, the effect of endosomal TLR signaling on liver CD8+ T cell responses was examined at late time points during acute MCMV infection. The results shown in Fig. 3A demonstrate comparable absolute numbers of CD8+ T cells in WT and 3d mice at days 5 and 7 post-MCMV infection, a trend that was also reflected in proportion (data not shown). To determine whether CD8+ T cells in 3d mice were properly activated against MCMV infection, intracellular expression of IFN-γ and TNF-α in CD8+ T cells was examined following ex vivo restimulation with H-2D^b M45 viral peptide, an immunodominant epitope of MCMV [29]. CD8+ T cells from 3d mice expressed these two cytokines at day 5 and day 7 post-MCMV infection by proportion and absolute numbers at levels that were comparable or slightly increased over WT (Fig. 3, B–E). As an indication of degranulation, surface expression of CD107a was also examined on liver CD8+ T cells from mice infected with MCMV; however, no differences in CD107a expression were detected between 3d and WT mice (data not shown). These results suggest that MCMV-specific CD8+ T cell responses in the liver are not compromised in the absence of endosomal TLR signals.

Previous studies have demonstrated resolution of virus-induced liver disease after 5 days of MCMV infection in WT mice [22], [24], [27], [37]. Therefore, given the impaired inflammatory responses observed in the absence of endosomal TLR signaling, liver sections prepared from 3d and WT mice that were uninfected or infected for 3, 5, or 7 days were hematoxylin and eosin (H&E) stained to evaluate pathology. The histological appearance of liver sections from uninfected 3d and WT mice appeared comparable (Fig. 4, A and B). By day 3 post-MCMV infection, clusters of infiltrating cells or inflammatory foci, which have been shown to coincide with sites of MCMV antigen expression [23], [37], were present in WT mice and persisted through day 5 before inflammation was resolved by day 7 post-infection (Table 1, Fig. 4, C, E and G). While the inflammatory foci per area of liver were equally apparent in liver sections from 3d mice infected for 3 days (Table 1), there was an increased presence of cytomegalic inclusion bodies characteristic of MCMV-infected cells that were not readily apparent in WT mice (Fig. 4D). By day 5 post-infection, livers from 3d mice were characterized by widespread areas of inflammation compared to the more punctate foci in WT livers (Fig. 4, E and F). Moreover, the inflammatory foci per area of liver in 3d mice at day 5 post-infection were significantly more numerous and contained a greater number of nucleated cells compared to WT (Table 1). However, by day 7 post-infection, inflammation in 3d mice showed signs of resolution that were similar to WT (Table 1, Fig. 4, G and H).

To further evaluate the effects of endosomal TLR responses on overall liver function, the liver enzyme alanine aminotransferase (ALT) was measured in serum samples from WT and 3d mice that were uninfected or infected with MCMV for 3, 5 or 7 days. Uninfected mice had comparable baseline levels of systemic ALT (Fig. 5). By day 3 of infection, similar elevations in ALT levels were detected in both groups. In contrast, by day 5 post-infection, the levels of systemic ALT in 3d mice reached values of 900±400 U/L and were three-fold higher than the ALT levels measured in WT mice (270±100 U/L; Fig. 5), indicating augmented liver disease. In contrast, by day 7 post-infection, 3d mice demonstrated a sharp decline in systemic ALT levels to values of 117±63 U/L, which were comparable to ALT levels exhibited in WT mice (108±85 U/L). Thus, at this time point of acute infection, both groups of mice exhibited ALT levels near baseline levels, consistent with the resolution of virus-induced liver pathology observed in 3d and WT mice (Fig. 4, G and H; Fig. 5). Taken together, these results suggest that in the absence of endosomal TLR signals, there is pronounced liver inflammation accompanied by a transient increase in liver damage during acute MCMV infection.

Given the diminished early cytokine responses and enhanced liver disease observed in 3d mice, the contribution of endosomal TLR responses to control of virus replication in the liver was assessed in WT and 3d mice infected with MCMV. Compared with those in WT mice, viral titers were elevated by 1 log on day 3 of infection in 3d mice, and remained significantly higher at day 5 post-infection when compared to WT (Fig. 6). Ultimately, while WT livers showed evidence of viral clearance at days 7 and 9 post-infection, 3d liver virus titers remained approximately 2 logs higher at these times of infection as compared to WT mice. Thus, endosomal TLR signaling contributes to the control of MCMV replication in the liver during acute infection.

Pathogen-free C57BL/6J mice were obtained from the Jackson Laboratory (Bar Harbor, ME). C57BL/6 Unc93b1^3d/3d mice were a kind gift from Dr. Bruce Beutler (The Scripps Research Institute, La Jolla, CA) and were generated as described [35]. C57BL/6J mice were housed and UNC93b1^3d/3d mice were bred in pathogen-free mouse facilities at Brown University. Male and female mice aged 8–10 weeks were used in experiments. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All animal work was approved by the Brown University Institutional Animal Care and Use Committee (Protocol Number: 0909082 and 0903035).

MCMV Smith strain was used in all experiments. This strain was prepared as a salivary gland-passaged stock from CD1 mice. Moderate dose infection (5×10⁴ PFU per mouse) was initiated on day 0 by intraperitoneal injection. In vivo responses were measured at indicated time points. For infectious viral titer quantification, organs were weighed, homogenized in cold supplemented DMEM (Invitrogen Life Technologies) and supernatants were collected following centrifugation. Serially diluted samples were used to inoculate confluent monolayers of bone marrow stromal cells (ATCC M2-10B4) in 24-well tissue culture plates and incubated for one hour at 37°C, 5% CO₂. Following incubation, inoculums were removed and monolayers were overlaid with a 1×DMEM/0.5% low-melt agarose solution. Cells were incubated for 7 days at 37°C, 5% CO₂, then fixed in 10% buffered formalin and stained with crystal violet. Plaques were counted to determine viral titer as previously described [22], [24], [26].

Liver leukocytes were prepared as previously described [23]–[25]. Briefly, following mechanical dissociation of tissue, red blood cells were removed by lysis with ammonium chloride and leukocytes separated by Percoll density gradient. To generate homogenates for cytokine analysis, the liver caudate lobe was homogenized in RPMI 1640 (Invitrogen Life Technologies) and supernatants collected following centrifugation. Serum was isolated from whole blood by centrifugation in the presence of heparin and stored at −80°C until further use in cytokine analyses or ALT assays.

Liver homogenates and serum were tested for cytokines by standard sandwich ELISA. IFN-γ, IL-12p70 and TNF-α were assayed by DuoSet (R&D Systems). IFN-α was measured by VeriKine mouse ELISA kit (R&D Systems). Limits of detection were 15 pg/mL for DuoSets and 12.5 pg/mL for VeriKine ELISA kits.

The following fluorochrome-conjugated mAbs were used in flow cytometric analyses: NK1.1-PE and TCRβ-APC to distinguish NK cells; CD8α-PECy7 and TCRβ-FITC to distinguish CD8+ T cells; and PDCA-1-APC (Miltenyi Biotec) to identify pDCs. Prior to surface staining, cells were incubated with anti-CD16/CD32 mAb to block nonspecific binding of Abs to Fcγ III/II receptors (clone 2.4G2). Unless otherwise noted, antibodies were obtained from BD Biosciences or eBioscience. For intracellular staining of cytokines, cells were treated with Brefeldin A (eBioscience) for 4 hours at 37°C, 5% CO₂ and permeabilized prior to labeling with IFN-α-FITC (R&D Systems), IFN-γ-PE, or TNF-α-APC (BD Biosciences). When indicated, leukocytes were stimulated for 5 hours with 100 ng/mL MCMV M45 peptide in addition to Brefeldin A treatment. Samples were acquired using a FACSCalibur and analyzed with BD Cell Quest software. For analysis, viable cells were gated by FSC and SSC. Isotype controls were used to set positive analysis gates.

Portions of the median liver lobes were isolated, fixed in 10% neutral buffered formalin, and paraffin embedded. Tissue sections (5 µm) were stained with H&E and analyzed microscopically. Images shown were photographed at the indicated magnification with a DP70 digital camera and software (Optical Analysis Corporation). Inflammatory foci, defined as discrete clusters containing between 6–60 nucleated cells, were quantitated as described previously [23]. In brief, inflammatory foci were identified, at a magnification of 200, as clusters of cells in totals of 8×50 µm² areas per representative tissue. In some cases, liver sections had larger areas of inflammation with >60 cells with less defined foci, and are indicated as such. Numbers of nucleated cells per inflammatory foci were counted in 20 individual foci per representative tissue at a magnification of 400. Liver alanine aminotransferase (ALT) was measured in serum samples by Marshfield Labs (Marshfield, WI).

Student's t test was used to determine statistical significance of experimental results when indicated (p≤0.05).