Chronic viral infections, such as human immunodeficiency virus (HIV), hepatitis B virus (HBV) and hepatitis C virus (HCV), are great threats to human health, and developing effective therapies against those infections is a big challenge. During chronic viral infections, persistent viral load triggers continual stimulation signals via TCR to virus specific T cells, resulting in gradually loss of effector functions or even deletion of these T cells [1,2]. Recent studies have revealed that inhibitory receptors, programmed cell death receptor 1 (PD-1) and Cytotoxic T-Lymphocyte Antigen 4 (CTLA-4), are involved in the process of T cell exhaustion during chronic viral infections. PD-1, a 50-55 kDa membrane protein that belongs to the immunoglobulin superfamily, was first discovered in 1992 from a T cell hybridoma undergoing apoptosis [3]. PD-1 is expressed on a subset of thymocytes, activated T cells, B cells and myeloid cells [4-6], while CTLA-4, a typeone transmembrane glycoprotein, is only expressed on CD4⁺ and CD8⁺ T cells [7]. CTLA-4 is 30% homologous to CD28 and shares the same ligands CD80/86 with CD28 in antigen presenting cells (APCs) [8,9]. In the following sections, we will discuss PD-1 and CTLA-4 mediated signaling pathways, their regulatory roles in a variety of chronic viral infections, and their possible applications for antiviral therapy.



PD-1 deficiency results in development of progressive arthritis and lupus-like glomerulonephritis in aged C57BL/6 mice and autoimmune dilated cardiomyopathy in BALB/c mice [10,11]. Further in vitro studies show that signaling through PD-1 inhibits cell proliferation and cytokine secretion such as IFN-γ, TNF-α and IL-2, in both CD4⁺ and CD8⁺ T cells [12]. These reports suggest that PD-1 regulates peripheral T cell tolerance. Importantly, PD-1 deficient CD8⁺ TCR transgenic T cells show potent tumor rejection in vivo [13]. Furthermore, PD-1 plays a key role in cytotoxic T lymphocyte (CTL) exhaustion during chronic viral infections, indicating that PD-1 negatively regulates CD8⁺ T cell effector function.

PD-1 is not expressed on resting T cells, while it could be quickly induced upon T cell activation [4]. Previous reports have suggested that transcription factors including NFAT, T-bet and c-Fos, regulate PD-1 expression. NFATc1 or the activator protein 1 (AP-1) subunit c-Fos directly binds to regulatory element at the pdcd1 promoter to upregulate PD-1 expression upon stimulation [14,15]. By contrast, T-bet directly represses transcription of PD-1 [16]. Recently, two reports show that epigenetic modulation via DNA methylation might affect PD-1 expression. DNA demethylation at the pdcd1 promoter by 5-Zac contributes to PD-1 overexpression on lymphoid cell line- Molt-4 cells [17]. Consistently, lack of DNA remethylation in the pdcd1 regulatory region of exhausted CD8⁺ T cells from acute and chronic viral infections in human and mice may leave the pdcd1 locus poised for rapid expression [18] (Figure 1).

Figure 1: Different mechanisms regulate PD-1 expression in T cells. (1) NFATc1 or c-Fos induces PD-1 expression upon stimulation, while T-bet represses transcription of PD-1 (black lines). (2) HIV Nef induced PD-1 transcription is dependent on p38/MAPK activation. Hepatitis B Core Antigen (HBcAg) induces PD-1 expression on T cells, which depends on the phosphorylation of JNK, ERK and AKT (blue lines). (3) IFN-α promotes the induction and maintenance of PD-1 expression through an association of IFN-responsive factor 9 (IRF9) to the IFN stimulation response element on the pdcd1 promoter (green lines). (4) DNA demethylation at the pdcd1 promoter increases PD-1 expression (pink line).

Notably, virus-induced PD-1 upregulation has been reported. Nef is an HIV-1 accessory protein and plays an important role in the pathogenesis of both HIV-1 infected humans and primates. It has been reported that Nef induces PD-1 transcription, while specific inhibitor against p38/MAPK inhibits Nef activity and effectively blocks PD-1 upregulation. This suggests that Nef-mediated PD-1 expression is dependent on p38/MAPK activation [19]. In HBV infected patients, PD-1 expression is increased on CD4⁺ T cells from peripheral blood compared to those in healthy volunteers. In addition, there is a positive correlation between serum HBV DNA levels and the PD-1 expression levels on CD4⁺ T cells in the immune clearance phase. in vitro assay further shows that Hepatitis B Core Antigen (HBcAg) induces PD-1 expression on T cells, and inhibitors against JNK, ERK and PI3K/ AKT significantly decrease HBcAg-induced PD-1 expression on CD4⁺ T cells [20]. Interestingly, several reports demonstrate that anti-viral cytokine IFN-α promotes the induction and maintenance of PD-1 expression through an association of IFN-responsive factor 9 (IRF9) to the IFN stimulation response element on the pdcd1 promoter in T cells [21]. IFN-α also mediates upregulation of PD-1 in macrophages, which is dependent on interferon-sensitive responsive element (ISRE) and STAT1 and STAT2 [22]. Taken together, these observations suggest that both viral proteins and signaling proteins from host cells could regulate PD-1 expression (Figure 1).

Multiple studies have shown that signals induced by CTLA-4 dampen T cell responses. Engagement of CTLA-4 together with TCR and CD28 decreases T cell proliferation, delays T cell cycle transition from G0 to G1 phase and reduces IL-2 production as well as IL-2 receptor expression [23-25]. T cells from CTLA-4 deficient mice display activated phenotype (CD44^hi, CD69⁺) and undergo robust proliferation [26,27]. Autoimmune diseases developed in CTLA-4- deficient mice provide compelling evidence to demonstrate the crucial inhibitory role of CTLA-4, which results from accumulation of T cell blasts in spleen and LN, and depleting T cells in CTLA-4 deficient mice prevents these diseases. Consistently, CTLA-4 overexpression on cell surface impairs T-cell responses in vivo and in vitro as shown in CTLA- 4 transgenic mice [28].

In addition, recent findings also show that the biological function of CTLA-4 is tightly related to the elegant regulation of CTLA-4 trafficking and localization. Newly synthesized CTLA-4 is transported from Golgi to cell membrane that is dependent on ARF-ribosylation factor 1 and phospholipase D [29,30]. In resting T cells, the unphosphorylated YVKM motif in the cytoplasmic tail of CTLA-4 interacts with the clathrin adapter protein AP, leading to rapid internalization of CTLA- 4 in a clathrin-dependent way [31,32]. Upon TCR stimulation, both Y-165 and Y-182 in the cytoplasmic tail of CTLA-4 are phosphorylated by activated Src family kinases to abolish the association of CTLA-4 with AP2. This results in a large fraction of CTLA-4 retained on cell surface [33,34].

PD-1 and CTLA-4 have been reported to commit their biological function through ligation with individual ligands. PDL-1 (B7-H1, also termed CD274) and PDL-2 (B7-DC, also termed CD273), two ligands of PD-1, show different expression patterns on different cell types [35,36]. PDL-1 is ubiquitously expressed in multiple cells, including T, B cells, DCs, macrophages, bone marrow-derived mast cells, and many nonhematopoietic cells. By contrast, PDL-2 is mainly expressed on DCs, macrophages and bone marrow-derived mast cells [5]. Both PDL-1 and PDL-2 could engage PD-1 to deliver inhibitory signaling to PD-1 expressing cells, and also deliver reverse signaling to PDL-1 and PDL-2 expressing cells [37,38].

CTLA-4 and the costimulator CD28 share the same ligands B7-1 (CD80) and B7-2 (CD86) [39]. Contrast to inhibitory signals delivered by CTLA-4, engagement of CD28 favors T cell activation, proliferation and effective cytokines secretion. Without CD28 ligation, TCR signals alone lead to T cell anergy. The topology of CTLA-4 homodimer allows for its bivalent binding to B7, while only monovalent binding occurs in CD28-B7 interaction [40]. Therefore, CTLA-4 binds these ligands with greater affinity and avidity compared to CD28, suggesting that CTLA-4 might block CD28 signaling by competitive binding to B7s. It is reported that CTLA-4 engagement of B7-1 is functionally equivalent to engagement of B7-2 [41]. However, CTLA-4 dominantly binds B7-1 since B7-1 expression is rapidly enhanced to reach peak levels after activation [42], while B7-2 expression level is still very low at that period. Interestingly, B7-1 could also interact with PDL-1 through their IgV-like domain and deliver bidirectional inhibitory signals [43]. Therefore, there might be a crosstalk between PD-1 and CTLA-4 signaling in certain conditions.

While both PD-1 and CTLA-4 use same key downstream signaling molecules to restrain T cell activation, biased signal cascades are mediated by PD-1 or CTLA-4 engagement. PD-1 consists of a single IgVlike extracellular domain, a transmembrane domain, and a cytoplasmic domain. Its cytoplasmic domain contains an immunotyrosine-based inhibitory motif (ITIM) and an immunotyrosine-based switch motif (ITSM). ITSM plays an essential role for PD-1 inhibitory function and mutation of ITSM results in dysfunction of PD-1 [44,45]. Upon TCR stimulation and engagement of PD-1, tyrosines in the cytoplasmic domain of PD-1 might be possibly phosphorylated by Lck in T cells [46] or by Lyn in B cells [47], which in turn recruits SH2-domain containing tyrosine phosphatase 1 (SHP1) and SHP2 [44,48]. Due to stronger interaction of SHP2 with PD-1 than SHP1, PD-1 functions mainly by recruitment of SHP2 [46,49] (Figure 2). In the case of T cells interacting with antigen presenting cells (APCs), immunological synapse (IS) is formed at the contact site, which is also named supermolecular activation cluster (SMAC) [50,51]. Interestingly, PD-1 accumulates at the synapse extensively when T cells interact with dendritic cells (DCs) expressing high levels of PDL-2 [52]. Further, PD-1 forms microclusters with TCR at the center of SMAC in a ligand binding-dependent manner [53]. SHP2 is immediately but transiently recruited to the PD-1 microclusters, to decrease the phosphorylation of TCR proximal signaling molecules, such as CD3ξ, ZAP70, PKCθ and PI3K, to attenuate TCR signaling [45,46,53]. PD-1 also inhibits Erk activation, which could be rescued by cytokines IL-2, IL-7 and IL-15 via the activation of STAT5.

Figure 2: PD-1 and CTLA-4 target different molecules to inhibit T cell activation. Upon T cell conjugates with APC, PD-1 is located in the immune synapse at the T-APC interface and recruits SHP2 to inhibit TCR-induced activation of the PI3K-Akt and Ras-MEK/ERK pathways. PD-1 also suppresses transcription of SKP2 to result in accumulation of p27kip1, which is an inhibitor of cyclin-dependent kinases to block cell cycle and proliferation. Ligation of CTLA-4 dephosphorylates TCRζ chain and other signaling molecules including CD3ε, ZAP70 and Fyn. CTLA-4 inhibits Akt phosphorylation and activation by recruiting PP2A to its cytoplasmic tail. Ligation of CTLA-4 phosphorylates the pro-apoptotic factor BAD and enhances BcL-XL activity to prevent T cell apoptosis.

Recently, an inhibitory loop by PD-1 signaling to suppress T cell proliferation has been demonstrated in human CD4⁺ T cells (Figure 2, green colour) [54]. PD-1 engagement inhibits TCR-induced activation of the PI3K-Akt and Ras-MEK/ERK pathways, and suppresses SKP2 transcription, which results in accumulation of p27kip1, an inhibitor of cyclin-dependent kinases, and inhibition of CDK2 activity. Impaired CDK2 activity then acts on downstream effectors, which eventually suppresses SKP2 expression to inhibit cell cycle in a feed-back loop [54]. Alternatively, other findings have indicated that PD-1 could also inhibit T cell proliferation and effector function by upregulating the expression level of basic leucine transcription factor ATF-like (BATF) in exhausted CD8⁺ T cells from human and mice. BATF is a transcription factor in the AP-1 family. Overexpression of BATF markedly impairs T cell proliferation and cytokine secretion, while depletion of BATF reduces PD-1 mediated inhibition function. More importantly, silencing BATF rescues the function of exhausted HIVspecific T cells [55].

Except for key signaling molecules, miRNA has also been recently reported to play important roles in PD-1 signaling pathway. Silencing or deficiency of PD-1 upregulates miR-21 expression and enhances STAT5 binding to the promoter of miR-21 in T-cells. In addition, miR- 21 regulates the expression of programmed cell death 4 (PDCD4) [56]. Collectively, PD-1 deficiency activates a signaling cascade mediated by STAT5, miR-21, and PDCD4, which results in hyperproliferation of T cells and enhanced secretion of IFN-γ and IL-17.

Similar to PD-1, CTLA-4 is translocated to the cSMAC to stabilize its binding to the ligands CD80/CD86 upon TCR stimulation [57,58]. Ligation of CTLA-4 reduces the contact time between DC and T cells and prevents immunological synapse formation. This leads to a major reduction in Ca²⁺ influx/mobilization and an abrogation of ZAP70 microcluster formation [59]. It was also suggested that CTLA- 4 regulates T cell adhesion and promotes T cell mobility, possibly in a RAP1-dependent way [60]. Different from typical ITIM and ITSM motifs in PD-1, the YVKM motif in the cytoplasmic tail of CTLA-4 is phosphorylated by kinases Fyn, Lyn and Lck [61,62], and recruits the SH2 domain of PI3K to enhance PI3K activity upon CTLA-4 ligation [63,64]. Activated PI3K phosphorylates the pro-apoptotic factor BAD and enhances BcL-XL activity [65]. This decreases FasL expression to prevent T cell apoptosis and also sustains T cell in an anergy state through inhibiting phosphorylation of Forkhead transcription factor FKHRL1 [66]. Notably, the tail of CTLA-4 binds to the regulatory subunit of serine/threonine phosphatase PP2A (also called PP2AA) to inhibit Akt phosphorylation and activation (Figure 2, purple colour) [45,67]. Differently, PD-1 could also inhibit AKT activity via its ITSM motif in the cytoplasmic tail [45].

Ligation of CTLA-4 specifically dephosphorylates TCRζ chain [68]. Furthermore, other signaling molecules including CD3ε, ZAP70 and Fyn are hyperphosphorylated in CTLA-4 deficient mice [69,70]. These observations indicate that CTLA-4 may recruit phosphatases such as SHP2 to dephosphorylate these signaling proteins, resulting in blockade of TCR signals [71]. Although the typical binding motif I/VxYxxI/V/L for SHP2 is absent in CTLA-4, CTLA-4 could coimmunoprecipitate with SHP2 and CD3ε in vitro [72]. It indicates that SHP2 might be associated with CTLA-4 through unknown adaptor proteins. However, there is a debate about the inhibition role of the CTLA-4-SHP2 complex. In naive CD8⁺ T cells where CTLA-4 has no inhibitory function, CTLA-4 is still associated with SHP2 [73,74].

Recently, new evidences have been provided to explain the inhibitory ability of CTLA-4. CTLA-4 is found to constitively expressed on regulatory T cells (Tregs) and contributes to the suppressive function of Tregs. Tregs are crucial for maintaining peripheral tolerance, by suppressing proliferation and immune responses of effector T cells. Deficiency and functional alteration in Tregs cause autoimmune diseases. There are currently at least four mechanisms to explain the suppressive function of Tregs: 1) production of inhibitory cytokines such as IL10 and TGFβ to constrain effector T cells; 2) secretion of granzymes to cytolysis effector T cells; 3) down-modulation of DC function or maturation to inactivate effector T cells; 4) consumption of metabolite such as IL-2 and cAMP to deprive metabolite essential for effector T cell survival. CTLA-4 in Tregs could modulate CD80/CD86 expression on DCs in foxo3-dependent way [75]. Downregulation of CD80/CD86 expression decreases the potency of DCs to activate T cells. In addition, the production of indoleamine 2, 3-diooxygenase (IDO), an enzyme secreted by APC, could be upregulated by CTLA- 4. Soluble CTLA-4Ig protein induces IDO expression by binding to ligands CD80/CD86 [76,77]. Consistently, less IDO is secreted in CTLA-4 deficient Tregs than those from WT mice [78]. IDO has been shown to suppress T cells and simultaneously promote DC death due to tryptophan depletion [79]. Therefore, it is proposed that CTLA-4 might also use IDO to commit its inhibitory function.

To summary the above findings, distinct signaling transduction profiles between PD-1 and CTLA-4 are mainly caused by their different expression profile, different ligands, and distinct downstream signaling molecules. In addition, a possible crosstalk between PD-1 and CTLA-4 pathways support their synergizing inhibitory effect in some settings. Further studies are needed to demonstrate the precise signaling pathways transduced by the crosstalk of PD-1 and CTLA-4.

Increased levels of PD-1 and CTLA-4 were observed in various chronic viral infections, and their regulatory functions have been extensively studied on different T cell subsets during chronic viral infections. The behavior of CD8⁺ effector T cells is different in acute and chronic viral infections. During acute virus infection, CD8⁺ effector T cells are activated immediately and clear virus efficiently. Most of these virus-specific effector CD8⁺ T cells then undergo apoptosis, leaving behind a small number of long-lived memory cells against secondary infection. However, during chronic or persistent viral infection, longterm antigenic stimulation results in gradually lose of effector T cells, and generates exhausted T cells that unable to clear virus effectively. CD8⁺ T cell exhaustion was observed in persistent LCMV, HIV, HBV and HCV infections in human patients and mouse models [1,2].

PD-1 has been demonstrated to play a non-redundant role to induce CD8⁺ T cells exhaustion during viral infections. PD-1 was first reported to be selectively upregulated by exhausted T cells during chronic LCMV infection [80]. Further investigations have reported that PD-1 expression is upregulated in HIV, HBV, and HCV specific CD8⁺ T cells in patients, which is correlated with viral load, and blockade of PD-1 increases virus specific T cell functions [81-87]. in vivo blockade of PD-1 and PD-L1 interaction with antibody could enhance T cell responses with a great reduction of viral burden. Furthermore, even in persistently infected mice lacking CD4⁺ T cell help, blockade of the PD-1 pathway could restore the ability of ‘helpless’ CD8⁺ T cells to undergo proliferation, secrete cytokines, kill infected cells and decrease viral load. These studies identify importance of PD-1 to CD8⁺ T cell exhaustion and viral control, and purpose a potential effective strategy for the treatment of chronic viral infections.

By contrast, the role of CTLA-4 in CD8⁺ T cell responses against viral infection is controversial. Although CTLA-4 mRNA is increased on exhausted CD8⁺ T cell in chronic LCMV infection, blockade of the CTLA-4 inhibitory pathway shows no effect on either CD8⁺ T-cell function or viral control [80]. Consistently, CTLA-4Ig transgenic mice (blocking B7-CD28 interaction) infected with LCMV show no alternation of the function of virus specific cytotoxicity T lymphocytes (CTLs) [88]. In HIV infection, CTLA-4 expression is not altered on CD8⁺ T cells isolated from patients [89]. However, other reports suggest that CTLA-4 regulates CTL function during HBV or VSV infection. Upregulated CTLA-4 has been found on CD8⁺ T cells from HBV infected patients. HBV-specific CD8⁺ T cells with excessive CTLA-4 are prone to apoptosis due to enhanced expression of Bim, a proapoptotic protein. Abrogation of CTLA-4-mediated inhibition reduces Bim expression and simultaneously increases expansion of IFN-γ-producing HBV-specific CD8⁺ T cells [90]. Other supporting studies show that non-replicating or poorly replicating VSV infection dramatically impairs the proliferation of primed CTLs during acute phases in CTLA-4Ig transgenic mice [88]. Blockade of CTLA-4 in both the primary and secondary infection could increase memory CD8⁺ T cell expansion with enhanced effective function to clear foreign invaders without changing TCR repertories [91].

Although CD4⁺ T cells could produce IL-2 and IFN-γ and assist activation of CD8⁺ CTLs against viral infection, CD4⁺ T cells exhaustion also exists during chronic viral infections. CTLA-4 has a profound, non-redundant role in regulating virus specific-CD4⁺ T cells function. During HIV infection, the expression level of CTLA- 4 is positively correlated with virus load. Excessive CTLA-4 may contribute to impairment of CD4⁺ T cell proliferation and reduction of IL-2 expression [92]. CTLA-4 blockade restores HIV-specific CD4⁺ T cell proliferation and effective cytokine secretion, including IFN-γ and IL-2. Importantly, blockade of CTLA-4 decreases the production of TGF-β and IL-10, but increases IFN-γ production by HIV-specific CD8⁺ T cells [93]. Depletion of CD4⁺ T cells abrogates the impact of CTLA-4 on HIV-specific CD8⁺ T cells. This indicates that interestingly, CTLA-4 on HIV-specific CD4⁺ T cells confers negative roles to HIVspecific CD8⁺ T cells.

By contrast, the role of PD-1 for CD4⁺ T cell exhaustion is not consistent during different viral infections. In HIV infected patients, PD-1 expression is unregulated in CD4⁺ T cells, which is correlated with viral load and CD4⁺ T cell numbers [94]. However, some studies show controversial data that PD-1 may not play an important role for CD4⁺ T cell exhaustion. In an inducible transgenic mouse system in which antigen presentation is controlled, CD4⁺ T cell exhaustion does not require PD-1 expression. Further, successful tolerance is induced in PD-1 deficient CD4⁺ TCR transgenic cells, demonstrating that PD-1 signaling is not required for either the induction, or the maintenance of peptide-induced tolerance [95]. Therefore, further investigation is needed to answer if PD-1 involved tolerance-inducing mechanisms in CD4⁺ T cell exhaustion is probably dependent on different types of APCs or antigen doses [96].

Upon virus infection, Tregs are activated and proliferate at the sites of infection, which are double-edged sword in virus eradication. From the beneficial side, Tregs restrain excessive immune responses to minimize tissue damages. On the other hand, Tregs suppress the function of effector T cells, which leads to enhanced virus survival. CTLA-4 is found constitutively expressed on Tregs and its specific transcription factor Foxp3 activates CTLA-4 expression [97]. CTLA-4 has been confirmed to regulate the suppressive function of Tregs. CTLA- 4 deficiency impairs the suppressive activity of Tregs characterized by its incapability to downregulate the ligands CD80/86 on APC, leading to massive conventional T cells proliferation and excessive IL-2 and IFN-γ production [75].

Similar to the controversial role of PD-1 in CD4⁺ T-cell exhaustion, the role of PD-1 in Tregs is not consistent during different viral infections. Some studies report that PD-1-PDL-1 interaction has a pivotal role in the development, maintenance, and function of induced Tregs (iTregs). PDL-1 deficient APCs minimally convert naive CD4⁺ T cells into iTregs. In a naïve CD4⁺ T cell adoptive transfer model, PDL- 1 and PDL-2 double knock-out mice markedly reduce the conversion of iTregs and rapidly develop a fatal inflammatory phenotype. It has been suggested that PDL-1 downregulates phospho-Akt, mTOR, S6, ERK2 and upregulates PTEN to induce iTregs [98]. Consistently, PDL- 1 enhances foxp3 expression and increases the suppressive function of iTregs. However, other studies show that PD-1 dampers Treg function in chronic infection. In the livers of patients chronically infected with HCV, PD-1 expression is enhanced in Tregs and this is coincided with decreased expansion of Tregs than effector T cells. Blockade of PD-1 signaling enhances the expansion and function of Tregs from HCV patients via upregulation of STAT-5 phosphorylation ex vivo [99]. This indicates that PD-1 negatively regulates Tregs by limiting STAT-5 phosphorylation in patient infected with HCV.

It has been about 20 years since PD-1 and CTLA-4 were discovered. Extensive studies have been made to understand their transcriptional regulation mechanisms, detailed signaling pathways, and their regulatory roles in chronic viral infections. Taken together, it seems that PD-1 plays a more convincing role for CD8⁺ T cell exhaustion than CTLA-4. By contrast, CTLA-4 plays non-redundant role in regulating the suppressive function of Tregs. PD-1 may serve as both a marker of disease progression and a therapeutic target to reverse function of exhausted CD8⁺ CTLs. Many studies have shown that PD-1 blockade is a promising way to enhance effector T cell function in chronic infections [100]. PD-1 blockade using antibody in an SIV-macaque model was reported to enhance HIV-specific immune responses [101]. On the other hand, it is also important to consider CD4⁺ T-cell exhaustion and Treg function during antiviral therapy. Suppression of Treg function in short time may enhance anti-viral responses. Ipilimumab, a human monoclonal IgG1 antibody against CTLA-4, has been clinically used in the treatment of cancer and supports a possible usage of anti–CTLA-4 antibody in the context of infectious diseases. However, it is still unclear how CTLA-4 blockade influences Treg or CD4⁺ T-cell function in vivo. Further studies will elucidate the effect of anti- CTLA-4 therapy in vivo during chronic viral infections. It is also important and possible to consider the combination of PD-1 and CTLA-4 blockade for immunotherapy to enhance antiviral immune responses [102,103].

However, it is still a long way to go for the success clinical trial on PD-1 and CTLA-4 involved antiviral therapies. Since both PD-1 and CTLA-4 control the balance between an adequate protective immune response and suppression of immunopathology, a key issue is that blockade of PD-1/CTLA-4 may cause side effects or autoimmune diseases. Manipulation of PD-1 downstream signaling may be an alternative option to overcome the problem. Further investigations on PD-1 and CTLA-4 signaling pathways and the molecular mechanisms of T-cell exhaustion may provide new therapeutic opportunities to improve T cell mediated immune responses against chronic viral infections.

This work was supported by grants from the Ministry of Science and Technology of China (2011CB505005, 2011CB946100 and 2012CB910800), Natural Science Foundation of Shanghai (10ZR1435300) and National Natural Science Foundation of China (31070778).