Cytomegalovirus exploits IL-10–mediated immune regulation in the salivary glands

MCMV replication can be detected for 1–2 mo in the submaxillary salivary glands, but not in the spleen, after a primary infection in C57BL/6 (B6) mice (6). Persistent MCMV replication in the salivary glands induced the accumulation of CD4 and CD8 T cells, B cells, NK cells, and antigen-presenting cells (Fig. 1 A). Cytokine expression analysis of salivary gland–derived leukocytes revealed that T cells (primarily CD4 T cells) were the predominant source of low spontaneous expression of antiviral cytokine IFN-γ, which was amplified by ex vivo polyclonal stimulation (Fig. 1 B).

Interestingly, we observed a considerable accumulation in the salivary glands of CD4 T cells expressing the immunosuppressive cytokine IL-10 during persistent MCMV replication, although these cells were not detected in visceral sites such as the spleen or in the lymph nodes draining the salivary glands (Fig. 1, C and E). We found a low percentage of CD4 T cells making detectable IL-10 (0.5–1%) without ex vivo restimulation (Fig. 1 C), suggesting that the results derived from stimulation with CD3/CD28 represented an amplification of cytokine production by these cells. Further reinforcing the notion that this IL-10 production reflected a physiological event in vivo, we found strong expression of IL-10 in salivary gland extracts from infected but not naive mice (Fig. 1 D). Importantly, CD8 T cells and non–T cells did not express notable levels of IL-10, suggesting CD4 T cells were the predominant source of this cytokine (Fig. 1 C). IL-10–producing CD4 T cells did not co-express IL-4, IL-5, IL-13, or FoxP3, implying these cells were not conventional T helper (Th)2 cells or naturally occurring T reg cells (Fig. 1 E). Moreover, the accumulation of IL-10–expressing CD4 T cells accompanied MCMV replication in the salivary glands (Fig. 1 F), demonstrating these T cells were actively induced by the virus infection.

IFN-γ–expressing CD4 T cells inhibit MCMV replication in the salivary glands (8). Although CD4⁺IFN-γ⁺ T cells were present (Fig. 1 B), the results suggested that IL-10– secreting cells might limit either antiviral activity of IFN-γ⁺ T cells or the IFN-γ itself. We hypothesized that shifting the ratio of CD4 T cells to favor IFN-γ production might limit virus replication. To accomplish this, we targeted the TNFR OX40 with an agonist antibody based on the ability of this receptor to enhance IFN-γ secretion (16, 17) and suppress IL-10 production by T cells (18, 19). OX40 was detected on CD4 but not CD8 T cells infiltrating the infected salivary glands (Fig. 2, A and B) and in draining lymph nodes (Fig. 2 B), indicating that ectopic targeting of OX40 with an agonist antibody was feasible. Mice treated with an agonist anti-OX40 antibody at the time of infection had no discernable effect on early (day 7 after infection) viral load in the salivary glands (Fig. 2 C). However, anti-OX40–treated mice had substantially reduced titers of replicating virus at 30 d after infection (Fig. 2 C). This was associated with increased infiltration of mononuclear cells compared with control mice (Fig. 2 D). Further analysis revealed that this cellular infiltrate consisted primarily of CD4 T cells (Fig. 2 E), and we found that the elevated CD4 response was critical for OX40-mediated inhibition of MCMV-persistent replication, as the protective effect was abolished in MHC class II–deficient mice (Fig. 2 F). In contrast, OX40 stimulation did not alter the accumulation of total or MCMV-specific CD8 T cells, or granzyme B–expressing CD8 T cells in the salivary glands (Fig. S1, A–C).

OX40 activation increased the percentage (Fig. 3, A and B) and total numbers (unpublished data) of CD4 T cells, but not CD8 T cells (Fig. S1 D), expressing IFN-γ. Moreover, anti-OX40 treatment also reduced the percentage of IL-10–expressing CD4 T cells (Fig. 3, C and D), altering the ratio of T cells expressing IFN-γ to IL-10 in the salivary glands (Fig. 3 E). This effect of OX40 activation was also reflected in a shift in the expression of IFN-γ and IL-10 in situ (Fig. 3 F). These data show that virus replication in the salivary glands is reduced when OX40 is activated, which is associated with an increased ratio of IFN-γ to IL-10–expressing CD4 T cells.

To determine whether the restricted virus replication and the increased ratio of IFN-γ to IL-10–secreting CD4 T cells involved IL-10, we used an antagonist IL-10R antibody (anti–IL-10R) to block signaling. Administration of this antibody enhanced the accumulation of total and IFN-γ–expressing CD4 T cells in the salivary glands (Fig. 4, A and B), but had no effect on the accumulation of CD8⁺ T cells or NK⁺CD3⁻ cells expressing IFN-γ (Fig. S2 A). The accumulation of CD4⁺IL-10⁺ T cells was not a result of IL-10R signaling in this organ (Fig. 4, C and D). Importantly, virus replication was dramatically inhibited by blocking IL-10R, with no virus detectable in two treated mice (n = 3) after 30 d of infection (Fig. 5 A). Moreover, when anti–IL-10R treatment was started at 7 d after infection to coincide with the appearance of IL-10⁺ cells (Fig. 1 E), virus levels in salivary glands remained undetectable in five out of eight mice (Fig. 5 B). Collectively, these data highlight the salivary glands as sites of strong IL-10–dependent immune regulation and demonstrate that IL-10–producing CD4 T cells suppress antiviral effector CD4 T cells. The results demonstrate that modulation of the OX40 and IL-10 pathways can block virus replication in this important excretory organ.

The salivary glands are a site of IL-10–dependent immune regulation. We demonstrated that IL-10–producing CD4 T cells suppressed effector CD4 T cells in a model of persistent virus replication at this mucosal site. The results establish that the IL-10 immunoregulatory pathway enables MCMV to persistently replicate in this mucosal organ. Counteracting this viral strategy by modulation of either the costimulatory OX40 pathway with an agonist antibody or targeting the immunoregulatory IL-10 pathway with an antagonist antibody, limited persistent virus replication in the salivary glands, attenuating shedding of virus, thus potentially reducing transmission to susceptible hosts.

Sustained activation of the TNFR OX40 with an agonist antibody favored the differentiation and accumulation of CD4 T cells that produced IFN-γ in the salivary glands to counteract the inhibitory effects of the IL-10–dominated immune response. OX40 was expressed on antigen-activated CD4 T cells, which were readily detected in the spleen, salivary glands, and draining lymph nodes from MCMV-infected mice. T cells express OX40 for short periods of time after antigen activation; thus, virus reactive T cells were the mostly likely targets of the antibody, and therefore, treatment subsequently altered virus-specific cytokine expression. Activation of OX40 can favor generation of effector Th1 or Th2 cells (16, 17, 20, 21), which are distinguished by their expression of IFN-γ or IL-4, IL-5, and IL-13, respectively. OX40 stimulation during MCMV infection enhanced the accumulation of CD4 T cells coexpressing IFN-γ and the activation-associated isoform of CD43 (unpublished data), suggesting that OX40 signaling directly impacted the differentiation of antiviral Th1 primary effector cells.

CD4 T cells in the salivary glands were identified as the predominant source of IL-10, representing ∼80–90% of all IL-10–expressing cells. In contrast to peripheral chronic arenavirus infection (22, 23), MCMV replication in the salivary glands did not induce expression of IL-10 by CD11c⁺ dendritic cells (unpublished data). IL-10–expressing CD4 T cells did not co-express the Th2-associated cytokines IL-4, IL-5, or IL-13, nor did they express the transcription factor FoxP3, which is a marker of naturally occurring CD4⁺CD25⁺ T reg cells. Collectively, the data indicated these T cells were IL-10– producing T reg cells, which have been described to mediate peripheral tolerance (24, 25) and suppresses mucosal inflammation in the gut (26). Interestingly, recent studies have reported that conventional Th1 cells can express IL-10 during chronic parasitic infections (27, 28). Whether MCMV-induced IL-10–expressing CD4 T cells represent T regulatory 1 or Th1 cells, or a combination of both, is currently unknown. Regardless of the exact lineage of these cells, however, this current study is the first description of a role for IL-10– expressing CD4 T cells in the salivary glands.

IL-10 suppresses inflammatory immune responses by inhibiting the expression of proinflammatory cytokines and chemokines, and decreasing expression of MHC class II and costimulatory molecules by APC, which subsequently limits differentiation or expansion of CD4 T cell effector responses (29). The dramatic reduction in persistent virus replication after IL-10R blockade was associated with increased IFN-γ–producing Th1 cells in the salivary glands, implying that IL-10 might act directly on effector T cells in the local microenvironment. Alternatively, IL-10 may suppress T cells indirectly by down-regulating one or more functions associated with APC. A previous study showed that MCMV- induced IL-10 in APC suppressed MHC class II expression by macrophages in vitro (30). However, we observed no change in MHC class II expression by cells infiltrating the salivary glands after IL-10R blockade (Fig. S2, B and C), suggesting this explanation cannot account for the action of IL-10 in vivo. The strong effect of anti-OX40 treatment in promoting IFN-γ–secreting CD4 T cell accumulation in the salivary glands suggests that IL-10 may predominantly act to suppress expression of costimulatory ligands such as OX40L and/or production of antiviral cytokines by APC, either in the salivary glands, or in the case of dendritic cells, after migration to draining lymph nodes.

In this mouse model of persistent virus replication, we have identified two distinct approaches to counteracting immune regulation. Regardless of the mechanism of IL-10–mediated suppression, the efficacy of IL-10R blockade in preventing viral replication was remarkable. Furthermore, delaying treatment until a week after infection, after CD4 and CD8 primary responses were established, dramatically altered virus replication in the salivary glands, showing the importance of IL-10 during this late stage of the antiviral response.

Targeting OX40 was also efficacious perhaps because of the expression of OX40 on antigen-reactive T cells, altering cytokine expression. This conclusion is guarded because MCMV peptides presented by class II MHC have not yet been defined to directly visualize antigen-specific T cells. The blockade of persistent viral replication by OX40 activation was associated with the accumulation of IFN-γ–secreting CD4 T cells previously described to afford protection from MCMV replication in the salivary glands (5, 8). Furthermore, the protective effect of OX40 stimulation was accompanied by a strong alteration in the ratio of IFN-γ to IL-10–expressing CD4 T cells. In a tissue culture model, the presence of OX40L on dendritic cells inhibited IL-10 expression in human CD4 T cells differentiating under both Th1 and Th2 priming conditions (19), suggesting that OX40 may directly antagonize IL-10 expression. However, the transcription factor GATA-3, which is implicated in the transcriptional control of IL-10 expression (31), can also be induced by OX40 stimulation in an IL-4R–dependent manner (21). This latter result suggests that, in certain conditions, OX40 could enhance IL-10 expression. Indeed, although we found an altered ratio of IFN-γ to IL-10–expressing T cells after OX40 stimulation, the actual numbers of IL-10–producing CD4 T cells did not change significantly. Furthermore, the level of IL-10 expression by CD4⁺IL-10⁺ T cells, as indicated by mean fluorescent intensity, was also unaltered by OX40 stimulation (unpublished data). Thus, the action of OX40 most likely reflected a preferential increase of Th1 effector cell differentiation and/or accumulation of Th1 cells in the salivary glands independent of IL-10. This conclusion is consistent with data where OX40 was shown to promote Th1 cell accumulation at the site of infection with influenza (32, 33) and Cryptococcus neoformans (20). Moreover, OX40 signaling can directly induce IFN-γ secretion and promote Th1 differentiation in tissue culture (16, 17). Anti-OX40 treatment reduced the viral load in the salivary glands by 1–2 log₁₀, whereas anti–IL-10R reduced titers by 3 log₁₀. The difference between the two pathways may reflect the finding that the IL-10–expressing CD4 T cells were not eliminated by treatment with anti-OX40. The finding that neither treatment altered the numbers of CD4⁺IL-10⁺ T cells indicates the difference in efficacy in reducing virus replication represented a fundamental distinction between the OX40 and IL-10R signaling pathways.

The ratio of T cells expressing IFN-γ to IL-10 provides a useful marker of immune regulation in the salivary glands and perhaps at other mucosal sites. A shift to IFN-γ expression pattern is critical for effective clearance of chronic parasitic infections (34), and correspondingly IFN-γ is necessary for the control of chronic MCMV infection (35). Elevated IFN-γ expression in the salivary glands may reduce viral persistence by directly inhibiting virus replication (35–37) or activation of additional effector mechanisms, such as NK, γδ T cells, and macrophages in the salivary glands (11, 38–40). IFN-γ also promotes CD8 T cell responses; however, in the context of the salivary glands, CD8 T cells do not limit virus replication (5), and neither anti-OX40 nor anti–IL-10R treatment had any effect on virus-specific CD8 T cell numbers, IFN-γ expression, or granzyme B expression, suggesting a minimal role for CD8 T cells in the protective effect of either treatment.

Induction of CD4⁺IL-10⁺ T cells after MCMV infection may represent a mucosal tolerance mechanism used by the salivary glands that is exploited by MCMV. It is likely that MCMV takes an active role in the exploitation of the mucosal tissue. Interestingly, MCMV alters the microarchitecture of the spleen by selective depression of CCL21 chemokine mRNA expression, which alters T cell trafficking within the spleen (41). Mutations in the viral genes M155 (42), M35 (43), or M133 (44, 45) abrogate replication in the salivary glands, although the role of these genes in promoting IL-10–expressing T cells is not understood. Of significance is the presence of viral IL-10 orthologs in primate CMV (46, 47), Epstein-Barr virus (48), and equine herpesvirus 2 (49). Although MCMV does not encode an obvious IL-10 ortholog, the observation that MCMV utilizes an IL-10–dependent mechanism to persist in salivary glands suggests a strong selective pressure is driving the conservation of targeting the IL-10 pathway in herpesviridae. The results raise the intriguing possibility that the primary function of cmvIL-10 in vivo may be to mimic IL-10–expressing T reg cells to block the action of the effector T cells secreting IFN-γ, allowing persistent virus replication specifically at mucosal sites. Thus, viral IL-10 orthologs may represent important targets to limit persistent virus replication in mucosal tissues such as the salivary glands, which in turn may block transmission of virus to new hosts.

Wild-type female C57BL/6 (B6) mice were purchased from The Jackson Laboratory, and MHC class II–deficient mice (50) were purchased from Taconic. All experiments were approved by institutional Animal Care and Use Committee.

MCMV Smith (American Type Culture Collection) was prepared and titered as described previously (51). Mice were infected with 5 × 10⁴ pfu MCMV by intraperitoneal route and, in some experiments, injected i.p. at the time of infection with 100 μg of either rat IgG (Chemicon) or anti-OX40 (clone OX86 [52]) purified by protein A affinity chromatography. In other experiments, mice were injected i.p. with 250 μg rat IgG1κ (clone R3-34; BD Biosciences) or anti-CD210 (anti–IL-10Rα clone 1B1.3a; BD Biosciences [53]).

Salivary glands were surgically excised and tissue leukocytes were isolated as previously described (11). In brief, salivary glands, free of associated lymph nodes, were cut into small pieces and incubated in RPMI 1640 supplemented with 5 mM CaCl₂, 5% fetal bovine serum, 1 mg/ml collagenase D (Roche Diagnostics), and 10 μg/ml DNase I (USB) at 37°C with mild agitation. After two consecutive incubations, dissociated cells and salivary gland tissue was passed through a 40-μm nylon cell strainer, and extruded cells were washed with PBS before analysis. The draining lymph nodes of the salivary glands were also excised, and cells were isolated by passage through a 40-μm nylon cell strainer.

Isolated leucocytes were then incubated at 5 × 10⁶ cells/ml for 5 h at 37°C in the presence of 2 μg/ml brefeldin A (Sigma-Aldrich) and stimulated with or without 1 μg/ml anti-CD3 (clone 145–2C11) and 1 μg/ml anti-CD28 (clone 37.51; BD Biosciences), or 2 μg/ml of MCMV-derived peptides (Genemed Synthesis Inc.). The peptides used were the H-2D^b–restricted M45-derived peptide HGIRNASFI and the H-2K^b–restricted peptides derived from the M38 (SSPPMFRVP), M57 (SCLEFWQRV), and M139 (TVYGFCLL) open reading frames, as previously described (54). Cells were then washed and incubated with Fc receptor blocking reagent and then stained with anti–CD8-PE-Cy7, anti–TCRβ-PE-cy5, and anti-CD4 antibody conjugated to FITC or APC-cy7 (BD Biosciences). Cells were then fixed with 4% formalin and permeabilized with saponin buffer (PBS, 2% FBS, 0.05% sodium azide, and 0.5% saponin) and stained with anti–IFN-γ-FITC or PE and, in some experiments, anti–IL-10-APC and PE-labeled anti–IL-5, anti–IL4, anti-FoxP3, or anti–IL-13 biotin (R&D Systems) followed by streptavidin-PE (BD Biosciences). In some experiments, unstimulated cells treated with brefeldin A were stained with anti-CD8 and anti–human granzyme B–PE (Catag/Invitrogen Life Technologies) or anti–NK1.1-APC and anti–IFN-γ. In other experiments, splenocytes were immediately stained with anti–MHC class II, anti-B220, anti-CD11c, anti–CD8-PE-cy7, anti–CD4-APC-cy7 (BD Biosciences), anti-F4/80 (eBioscience), and/or OX40-PE (Serotec). All cells were acquired on an LSR II flow cytometer (BD Bioscience).

Salivary glands were fixed in 10% formalin, paraffin-embedded, and 5-μm sections were stained with hematoxylin and eosin. Three sections from each gland were examined from four mice per group.

IFN-γ and IL-10 mRNA were assayed by quantitative RT-PCR using Stratagene Mx 4000 and SYBR green reagent (Bio-Rad Laboratories). Salivary glands from mock or MCMV-infected mice were snap frozen in liquid nitrogen before homogenization in TRIzol reagent (Invitrogen Life Technologies). Total cellular RNA was then isolated and quantified before performing DNase digestion and reverse transcription as described (55). The primer sequences for detection of mouse IFN-γ are GCCGTGGCAGTAACAGCC and AACGCTACACACTGCATCTGGG, and for IL-10 are CACTGCTATGCTGCCTTCTCT and TGTCTTCCCTGCTGTACTGT.

Fig. S1 shows time-course analysis of numbers of CD8 T cells (a), granzyme B expression by CD8 T cells (b), and accumulation of IFN-γ–expressing CD8 T cells (d), in the salivary glands of anti-OX40 or control IgG-treated mice. Panel c shows MCMV-specific CD8 T cell accumulation in the salivary glands at day 7 after infection in anti-OX40–treated or control IgG-treated mice.

Fig. S2 shows the numbers of CD8⁺/IFN-γ⁺ and NK⁺/CD3⁻/IFN-γ⁺ cells (a) and the percentage (b) and the mean fluorescent intensity (c) of MHC class II expression by leukocytes in the salivary glands of anti–IL-10R or control IgG-treated mice 14 d after infection.

The authors wish to thank Dr. Kirsten Schneider and Professor Brigitte Askonas for valuable discussions; Xiaohong Tang, Yan Fei Adams, and Antje Rhodes for excellent technical support; Dr. Ann Hill for MCMV peptide sequences recognized by CD8 T cells; and Dr. Tim Cheung for assistance with figure preparation.

This work was supported in part by National Institutes of Health (NIH) grants CA91837 and AI67341 (to M. Croft); NIH Predoctoral Fellowship T32 AI060536 (to A. Kinkade); and AI48073, AI057840, and R37AI33068 (to C.F. Ware). This is publication #793 from the La Jolla Institute for Allergy and Immunology.

The authors have no conflicting financial interests.