The mammalian immune system has evolved to encompass activating and inhibitory pathways. These counter-regulatory mechanisms enable the immune system to initiate protective coordinated responses against invading pathogens that are sufficiently limited in magnitude to avoid overt bystander damage to infected tissues. In addition, inhibitory immune pathways limit the development of harmful autoimmune reactions and over-exuberant immune responses to harmless antigens that are often encountered at mucosal surfaces.

Herpesviruses are large DNA viruses that are divided into the α-herpesviridae, β-herpesviridae and γ-herpesviridae families. Herpesviruses establish life-long infection in their hosts and are ubiquitous within the human population. These infections are often asymptomatic in healthy individuals. However, herpesvirus infections cause significant disease in immune compromised individuals, and the β-herpesvirus human cytomegalovirus (HCMV) is the leading infectious cause of congenital defects in the Western world. Moreover, the γ-herpesvirus Epstein Barr virus (EBV) causes infectious mononucleosis and, in addition to the γ-herpesvirus Kaposi's sarcoma-associated herpesvirus (KSHV), is associated with the development of several cancers [1].

Herpesviruses are thought to have emerged approximately 400 million years ago [2]. During co‑evolution with mammals, herpesviruses have acquired numerous genes encoding proteins that function primarily to evade or manipulate immune pathways, thus enabling the establishment of life‑long infection within their host. It is becoming increasingly apparent that host immune inhibitory pathways critically influence the outcome of herpesvirus infections. In this article, we summarize current understanding of the immune inhibitory mechanisms that modulate anti-herpesvirus immunity and highlight strategies adopted by herpesviruses to exploit such pathways to antagonize antiviral immunity.

The immune-modulatory cytokine interleukin-10 (IL-10) is a member of the same family of proteins as type I IFNs and IFNγ [3]. Mammalian IL-10 is produced predominantly by macrophages, dendritic cells (DCs), B cells and regulatory T cells (Tregs), although it can be produced by virtually all T cell subsets [4]. The IL-10 receptor (IL-10R) is composed of two subunits. The IL-10R1 (IL-10Rα) is the ligand-binding subunit expressed by hematopoietic cells whereas IL-10R2 (IL‑10Rβ) is an accessory subunit for signaling constitutively expressed by most hematopoietic and non-hematopoietic cells. The interaction between IL-10 and its receptor engages the Jak family tyrosine kinases Jak1 and Tyk2, which are constitutively associated with IL-10R1 and IL-10R2 respectively, leading to tyrosine phosphorylation and activation of STAT (signal transducers and activators of transcription) proteins, predominantly STAT3 and to a lesser extent STAT1 and STAT5.

IL-10 can induce a broad range of biological functions (reviewed in [5,6]). Although IL-10 displays certain immune-stimulatory activities, the majority of data to date demonstrates that IL-10 exerts suppressive effects on immune cells, particularly APCs and T cells (either directly or via suppression of APC function). Consequently, IL-10 profoundly influences a plethora of immune responses in vivo, including responses elicited in response to parasitic and bacterial infections [7].

During acute viral infections, IL-10 has paradoxical functions. For example, IL-10 can limit immunopathology induced by respiratory syncytial virus (RSV) [8,9], and influenza [10], suggesting the primary function of this pathway is to limit infection-induced immune-mediated tissue damage. Following high dose influenza infection however, IL-10-deficient mice exhibit accelerated clearance of virus which is associated with the induction of antiviral Th17 cells [11], suggesting that the biological outcome of IL-10R signaling during acute infection may vary depending on the virus pathogen and the infectious dose. Importantly, during chronic virus infection in vivo, as demonstrated in the LCMV model, IL-10 antagonizes antiviral immunity and promotes virus persistence [12,13].

The Programmed Death Receptor (PD-1) is a member of the CD28 family expressed on CD4⁺, CD8⁺ and NK T cells, B cells, monocytes and on some dendritic cell subsets upon activation. It is a type 1 transmembrane glycoprotein with an IgV-type extracellular domain and a cytoplasmic signaling domain containing two tyrosine residues [52]. It is a monomeric receptor which has two ligands, PD-L1 (B7-H1, CD274) [53] and PD-L2 (B7-DC, CD273) [54]. PD-L1 expression is detectable on resting B, T, myeloid and dendritic cells and can be up-regulated upon activation [55] as well as on non-lymphoid tissues such as heart, skeletal muscle, placenta and lung tissues [53]. PD-L2 has significant homology to PD-L1 (38% amino acid identity in mice [56]) and is expressed in an inducible manner on dendritic cells, macrophages [54] and activated T cells [57].

The binding of PD-1 to its ligands induces an inhibitory signal within the PD-1-expressing cell [56,58] thought to be triggered by immunoreceptor tyrosine-based inhibitory motif (ITIM)-mediated recruitment of Src homology region 2 domain-containing phosphatase(SHP)-1 and SHP-2. Engagement of this receptor with its ligand has an important regulatory role in the prevention of excessive immune responses against infections and the maintenance of peripheral tolerance against self-antigens (reviewed in [59,60]). However the biology of PD-1:PD-L1/L2 is perhaps more complex than a uni-directional co-inhibitory pathway (summarized in Figure 1). Indeed, signaling through PD‑L1 in tumor cells enhances resistance to apoptosis [61]. In addition, hyper activated immunity in response to influenza [62] and listeria monocytogenes [63] in the absence of PD-L1 mediated signals has been reported, with authors suggesting PD-L1 mediates conditioning of APCs and/or the presence of a stimulatory PD-L1-induced signal in these models. Interestingly, PD-1 deficient mice are extremely sensitive to Mycobacterium tuberculosis infection exhibiting heightened pro-inflammatory responses yet, paradoxically, reduced T and B cell responses and uncontrolled bacterial proliferation [64]. Thus, the role that PD-1 plays during infections in vivo can be complex.

In the context of acute viral infections, with the exception of influenza, PD-1/PD-L1 interactions inhibit antiviral T cell immunity [65]. Importantly, PD-1 signaling also drives T cell dysfunction and virus persistence in chronic LCMV infection [66]. Furthermore, blockade of PD-1:PD-L1 interactions improves the function of T cells reactive to human immunodeficiency virus (HIV) [67] and hepatitis B virus (HBV) [68]. Interestingly, PD-1 is up-regulated by monocytes during HIV infection, and ligation of this inhibitory receptor by PD-L1 induces IL-10 secretion and leads to CD4⁺ T cell dysfunction [69].

The PD-1:PD-L1/L2 pathway is a potentially critical regulator of anti-herpesvirus immunity. EBV‑specific CD8⁺ T cells up-regulate PD-1 upon activation [67], and a negative correlation between intensity of PD-1 expression and absolute numbers of circulating EBV-specific CD8⁺ T cells in the transition from acute infectious mononucleosis to convalescence has been reported [70]. Furthermore, PD-1 mediated exhaustion of EBV-specific (but not HCMV-specific) CD8⁺ T cells is detected in Systemic Lupus Erythematosus [71]. PD-1 ligation also suppresses HCMV-specific CD8⁺ T cell function [72] and has been implicated in driving attrition of CMV-specific CD8⁺ T cells in acute hepatitis B virus‑infected individuals [73].

These data suggest that PD-1 ligation on herpesvirus-specific T cells may antagonize control of infection in vivo; a hypothesis tested in murine models of herpesvirus infections. Although MCMV‑specific CD8⁺ T cells express PD-1 in a major site of persistence, the salivary glands, PD-L1 blockade failed to accelerate virus clearance [74]. Importantly, salivary gland-infiltrating antigen-specific CD8⁺ T cells are not exposed to cognate antigen due to a combination of efficient viral interference with MHC class I antigen processing in infected glandular epithelial cells and the inability of salivary gland APCs to cross-present antigen to CD8⁺ T cells [75]. Thus, any possible benefit of blockade of the PD-L1 pathway in this setting will be masked by impaired stimulation of virus-specific CD8⁺ T cells. Importantly, in a model of chronic murine γ-herpesvirus-68 (MHV-68) infection in MHC class II‑deficient mice in which PD-1 expression by virus-specific CD8⁺ T cells is observed, blockade of the PD-1:PD-L1 pathway reduces virus reactivation [76], underlining the potential importance of this inhibitory pathway in the suppression of protective anti-herpesvirus immunity.

The potential suppressive activity of this pathway represents an obvious target for herpesviruses to exploit. Although no herpesvirus proteins with homology to PD-1 ligands have been reported, MCMV infection of dendritic cells induces surface expression of both PD-L1 and PD-L2 [77,78]. In stark contrast to virus-induced down-regulation of stimulatory ligands (CD80, CD86, CD40), adhesion molecules (CD11b and CD11c) and MHC Class I and II, prolonged PD-L1 up-regulation is observed in infected cells [78]. Strikingly, poor activation of T cells by MCMV-infected DCs is reversed by PD-L1 blockade in vitro and in vivo, suggesting MCMV actively manipulates PD-L1 to inhibit T cell activation [78]. Interestingly, tumor cells in EBV-associated cancers such as classical Hodgkin’s Lymphoma (cHL) and post-transplant lymphoproliferative disorders also express significant levels of PD-L1 [79] suggesting possible active manipulation of this pathway by EBV.

Therefore, current data suggest it is beneficial for herpesviruses to exploit the PD-1:PD-L1/L2 inhibitory pathway. It is unclear whether these viruses specifically target the inhibition of antiviral T cell responses, or whether PD-1 mediated inhibition of other immune effector mechanisms may be exploited by these viruses. Furthermore, the potential virus exploitation of anti-apoptotic signaling events through PD-L1 in infected cells requires further investigation.

Like PD-1, BTLA (or CD272) is a CD28 family member. As with PD-1, BTLA contains an ITIM and an immunoreceptor tyrosine-based switch motif (ITSM) in the membrane proximal and distal regions of the cytoplasmic domain, respectively. Whereas the negative signal induced by PD-1 ligation is predominantly transmitted through ITSM recruitment of SHP-1 and SHP-2 [80], both ITIM and ITSM motifs appear critical for BTLA-mediated suppression of T cell activation [81]. In addition, BTLA contains a Grb-2 recognition consensus site that may contribute to negative signaling [81].

BTLA expression was first reported on Th1 cells but is known to be expressed on a broad array of hematopoietic cells [82]. Studies of BTLA-deficient mice have highlighted an important regulatory role for BTLA in limiting mucosal inflammation and autoimmunity (reviewed in [82,83]) and BTLA limits immune responsiveness to bacterial and parasitic infections [84,85], suggesting a classical inhibitory function for this CD28-family member. However, the biology of BTLA is complicated by positive signaling delivered through the cellular ligand of BTLA; the tumor necrosis factor receptor superfamily member (TNFRSF) herpesvirus entry mediator (HVEM, or TNFRSF14). BTLA:HVEM interactions induce bidirectional signaling resulting in HVEM-mediated NF-kappaB activation (Figure 1). This signal promotes survival of HVEM-expressing cells, thus demonstrating that BTLA can also promote T cell responses [86].

HCMV, however, has evolved to exploit the inhibitory properties of BTLA. UL144 is a virus‑encoded truncated TNFR member consisting of two cysteine-rich domains (CRDs) homologous to CRD1 and CRD2 of HVEM [87]. Despite sequence hypervariation in the ectodomain of UL144 variants expressed by clinical HCMV isolates [88], UL144 proteins from all groups bind BTLA and, critically, UL144 suppresses proliferation of polyclonally stimulated CD4⁺ T cells [87]. Furthermore, BTLA is highly expressed by HCMV-specific CD8⁺ T cells following activation and BTLA blockade enhances HCMV-specific CD8⁺ T cell proliferation in vitro [89]. Therefore, through the acquisition and retention of UL144, HCMV appears to induce inhibitory signals induced downstream of BTLA to suppress antiviral T cell immunity (Figure 1). Given the broad expression pattern of BTLA on hematopoietic cells, it is conceivable that HCMV has evolved to target this inhibitory pathway to antagonize multiple immune effector mechanisms.

CD200 is a member of the immunoglobulin superfamily (IgSF) expressed on membranes by a heterogeneous group of cells, including B cells, activated T cells, endothelial cells, epithelial cells, follicular dendritic cells and neurons [90]. The expression of CD200 in humans [91], mice [92] and rats [93] is highly conserved. CD200 contains two extracellular IgSF domains, a hydrophobic trans-membrane sequence, and a short cytoplasmic domain [94] that does not contain known signaling motifs or docking sites for adaptor proteins [91]. Thus, CD200 is thought to deliver a unidirectional signal to its cellular receptor, CD200R.

Expression of CD200R was originally detected on myeloid cells, including macrophages, granulocytes and dendritic cells [95], but has subsequently been detected on NK cells and T cells [96,97]. CD200R contains two IgSF domains and a cytoplasmic region containing two tyrosine based motifs that can be phosphorylated [98]. In myeloid cells the inhibitory signal induced by CD200R is facilitated by recruitment of DOK2 and RasGAP [99]. Upon interaction with CD200, CD200R delivers an immunosuppressive signal that antagonizes type 1 cytokine production by DCs [100], induces Treg development [101] and inhibits macrophage function [92,98]. Moreover, CD200:CD200R limits myeloid cell homeostasis in the periphery [92], lung [102] and, to a lesser extent, the intestinal mucosa [103]. CD200R is also a critical negative regulator of inflammation. CD200-deficient mice demonstrate rapid onset of experimental autoimmune encephalomyelitis [98] and increased susceptibility to acute inflammation induced by bacterial (Neisseria meningitidis) [104] and viral (influenza) [102] infections. Importantly, therapeutic ligation of CD200 during ocular HSV infection also reduced inflammatory lesions [105], demonstrating the anti-inflammatory nature of this pathway in an acute herpesvirus infection.

Immune inhibitory receptors represent important targets for herpesviruses to dampen antiviral immunity. Subsequently, these pathways or the virus-encoded proteins that exploit them represent possible targets of therapeutic strategies aiming to promote antiviral immunity.

The therapeutic success of specifically targeting viral homologues of immune-modulatory proteins will be partially determined by the sequence homology to their mammalian counterparts. Relatively low amino acid sequence similarity between HCMV UL111A and mammalian IL-10 [19] offers the potential of targeting UL111A in vaccine settings, as demonstrated with rhCMV UL111A [123,124]. Such an approach however is considerably more problematic for EBV due to the high sequence homology between BCRF-1 and mammalian IL-10 [20]. Moreover hypervariation in some viral proteins, for example in UL144 [88], represents further challenges for vaccination strategies targeting these molecules.

Importantly, in vivo data (predominantly from murine models of virus infections) highlight the important role of inhibitory pathways in limiting infection-induced pathology, particularly in acute infections. Therefore, herpesvirus persistence and dissemination that occurs as a consequence of signaling through these receptors is, from an evolutionary perspective, perhaps an acceptable price for mammals to pay to ensure effective control of virus-elicited immune-mediated damage. Thus, therapeutic treatment with antagonists of mammalian immune-inhibitory proteins during acute infections may have harmful consequences. Paradoxically, in certain conditions, ligation of inhibitory receptors may improve virus-induced pathology [46,105]. Importantly, to facilitate establishment of chronic infection and dissemination from their host, herpesviruses are also dependent upon survival and fitness of the mammal they infect. Therefore, an important function of viral proteins that target these pathways may be to contribute to the limitation of infection-induced pathology. In vivo infection models that utilize herpesviruses encoding such immune evasion proteins [51,116] will be particularly powerful tools to assess this possibility.

Much of the literature summarized herein implies that herpesviruses target immune inhibitory pathways to antagonize T cell responses. HCMV, however, induces the generation of a remarkably large virus-specific T cell response over time [125] suggesting that the primary function of these immune evasion strategies is not to antagonize memory T cell development. Importantly however, experiments using the MCMV model have demonstrated that memory T cell inflation is greatly increased in the absence of IL-10 [49], implying that inhibitory immune pathways could be exploited to further enhance the functional capabilities and/or in vivo survival of herpesvirus-specific memory T cells used, for example, in T cell-based therapies to treat herpesvirus-infected individuals.

It is likely that herpesviruses exploit immune inhibitory pathways to antagonize T cell responses primarily during acute infection, and during persistence in mucosal tissues, thus promoting virus survival and dissemination. Although targeting these pathways during acute infection may have harmful consequences (as discussed above), manipulation of these molecules to enhance virus-specific immunity induced by prophylactic vaccination strategies (including, possibly, attenuated herpesviruses) may greatly improve anti-herpesvirus immunity. Importantly, the observation that re‑infection of HCMV-seropositive women with different strains of HCMV can lead to intrauterine transmission of the virus [126] suggests that, in the case of this herpesvirus, boosting existing HCMV‑specific T cell responses, particularly in mucosal surfaces, may be crucial to reduce horizontal transmission and resulting congenital infections. Thus, understanding the role that immune inhibitory proteins play in regulating anti-herpesvirus immunity in mucosal (and non-mucosal) tissues of persistent/chronically infected hosts may inform strategies aiming to enhance heterosubtypic immunity afforded by existing herpesvirus-specific T cell responses. In vivo experiments will be crucial for the assessment of the efficacy and safety of targeting immune-inhibitory pathways to enhance herpesvirus-specific protective immunity in all of these settings.