Human CMV (HCMV) is a ubiquitous β-herpesvirus which has developed several mechanisms of escape from the immune system. IFN-γ-induced signaling relies on the integrity of the JAK/STAT pathway which is regulated by phosphorylation steps and leads to nuclear translocation of tyrosine-phosphorylated STAT1 (STAT1-P-Tyr), and its binding to IFN-γ activation site sequences of IFN-γ-inducible promoters. Activation of those promoters leads to the expression of genes involved in the immune response and in the antiviral effects of IFN-γ. Src homology region 2 domain-containing phosphatase 2 (SHP2) is a ubiquitous phosphatase involved in the regulation of IFN-γ-mediated tyrosine phosphorylation. Several mechanisms account for the inhibition IFN-γ signaling pathway by HCMV. In this study, we have identified a new mechanism that involved the inhibition of STAT1 tyrosine phosphorylation within 12–24 h postinfection. This defect was dependent on HCMV transcription. Consequences were impaired nuclear translocation of STAT1-P-Tyr, inhibition of IFN-γ activation site-STAT1 interaction, and inhibition of HLA-DR expression. Expression of indoleamine-2,3-dioxygenase which is involved in the antiviral effects of IFN-γ was also inhibited. Treatment of cells with sodium orthovanadate rescued STAT1 tyrosine phosphorylation, suggesting that a tyrosine phosphatase was involved in this inhibition. Coimmunoprecipitation of STAT1 and SHP2 was induced by HCMV infection, and SHP2 small interfering RNA restored the expression of STAT1-P-Tyr. Our data suggest that SHP2 activation induced by HCMV infection is responsible for the down-regulation of IFN-γ-induced STAT1 tyrosine phosphorylation.
Cytomegalovirus, a member of the β-herpesvirus subfamily, can persist in the host thanks to its capacity to enter into latency (1). A balance between the host’s strong immune response and numerous mechanisms of escape which target several aspects of the immune systems has developed during coevolution of human CMV (HCMV)3 with its specific host (2, 3). Cellular immune response is thought to be crucial in the fight against HCMV; besides γδ cells (4) and NK cells (5), CD8+ T cells, have been shown to play an important role (3, 6, 7). This was more directly demonstrated by transfers of CD8+ T cell clones (8) and, more recently, CD4+ and CD8+ T cell lines into immunodeficient patients (9). In addition, cytokines such as IFN-γ and TNF-α produced by CD4+ T cells have been demonstrated to block HCMV replication in vitro (10), and IFN-γ produced by CD4+ T lymphocytes has been shown to be important for the recovery of HCMV infection in vivo (11).
The IFN-γ signaling pathway involves the activation of STAT1 which, upon tyrosine phosphorylation by JAK1, homodimerizes and is translocated to the nucleus where it activates IFN-γ-responsive specific promoters such as those for MHC class I (MHC I) and MHC class II (MHC II) genes (12). IFN-γ up-regulates a number of other genes involved in Ag presentation, including peptidases (13, 14). Thus, although dendritic cells (DC) initiate immune responses, nonprofessional APCs can express MHC II upon IFN-γ stimulation and subsequently present Ag, including HCMV IE1 protein (15). In this respect, MHC II and proteolytic enzymes induced by IFN-γ may contribute to the antiviral defense. Besides inducing effector molecules of the adaptive immune response, IFN-γ also initiates antiviral effector mechanisms (16, 17, 18, 19).
Signaling by IFN-γ is regulated through inactivation of the JAK/STAT pathway mediated by suppressors of cytokine signaling (SOCS; Ref. 20) proteins and tyrosine phosphatases such as SHP1 and SHP2 (21). Whereas SHP1 is essentially expressed by cells of the hemopoietic lineage, SHP2 is a ubiquitous tyrosine phosphatase which has been found to be associated with and dephosphorylate STAT1 (22).
Escape mechanisms of HCMV have thus evolved to adapt and counteract control of infection by the immune system. Mechanisms are targeted to various steps within the same pathway, presumably leading to greater efficiency. For example, inhibition of constitutive MHC II expression, which results in decreased CD4+ T cell recognition, has been reported to result from altered localization of constitutive MHC II (23, 24, 25). Other mechanisms have involved either degradation or inhibition of assembly of MHC II by US2 and US3 proteins (26). Regarding inducible MHC II expression by IFN-γ, HCMV has targeted the JAK/STAT pathway through JAK1 degradation 72 h postinfection (p.i.; Ref. 27). In contrast, the mechanism we previously described involved inhibition of IFN-γ-induced CIITA mRNA expression 6 h postinfection with conserved STAT1 tyrosine phosphorylation (28). However, kinetics of IFN-γ exposure and HCMV infection is expected to be variable from one site of infection to another. We thus assessed the impact of infection on IFN-γ-induced STAT1 tyrpsine phosphorylation in settings that were as yet not fully investigated.
This allowed us to describe here a new mechanism through which HCMV impairs IFN-γ signaling by inhibiting STAT1 tyrosine phosphorylation. This inhibition requires viral replication and is mediated by the tyrosine phosphatase SHP2.
Materials and Methods
MRC5 human embryonic fibroblasts were maintained in DMEM (Life Technologies cell culture medium; Invitrogen) supplemented with 10% FBS (HyClone; Perbio Science), sodium pyruvate (1 mM), glutamine (2 mM), and penicillin (100 U/ml)-streptomycin (100 μg/ml). Human IFN-γ (Roche Diagnostics) treatment was performed using a 250-U/ml concentration for the times indicated.
Viral stocks of HCMV Towne strain were obtained by infection of MRC5 cells in 175-cm2 flasks, at a multiplicity of infection of 0.1. After 8–10 days, supernatants were collected, centifuged at 2000 × g for 30 min and stored at −80°C in 1-ml cryotubes. Viral titers, determined by the plaque lysis method, were generally comprised between 1 and 4 × 106 PFU/ml. Multiplicity of infection used throughout experiments was 5 PFU/cell. As indicated in the figures, the time of cell infection slightly varied from one experiment to another without impact on data.
Inactivation of HCMV was performed using either UV irradiation (Spectroline EF-140/F with a 10-cm exposure distance; Spectronics) or γ irradiation (4000 Gy; Biobeam 2000; Bebing).
Cells infected or not by HCMV were collected, and mRNA was extracted with an RNeasy kit (Qiagen France). cDNAs were prepared according to the standard protocol defined by the manufacturer (Finnzymes). Briefly, 2 μg of mRNA were incubated in reaction buffer (Finnzymes) supplemented by 250 nM concentrations of each dNTP, 100 ng of hexamer random primer (New England Biolabs), 5 U of RNaseOUT (Invitrogen), and 200 U of Moloney murine leukemia virus RNase H− reverse transcriptase for 1 h at 37°C.
PCR analysis were made using the MasterMix Eppendorf kit (Eppendorf).
Sequences of oligonucleotides were as follows: DRα forward 5′-GCTCTAGACCATGGCCATAAGTGGAGTCCCTGTGC-3′ and reverse 5′-CTGAATTCAGGTGATCGGAGTATAGTTGGAGCGC-3′; GAPDH forward 5′-ACCACAGTCCATGCCATCAC-3′ and reverse 5′-TCCACCACCCTGTTGCTGTA-3′ (29); indoleamine 2,3-dioxygenase (IDO) forward 5′-ACAGACCACAAGTCACAGCG-3′ and reverse 5′-AACTGAGCATGTCCTCC-3′.
β-Actin primers were from Invitrogen.
Gel shift assays
Preparation of cytoplasmic and nuclear extracts and oligonucleotide binding assays have been described elsewhere (30).
Oligonucleotide for IFN-γ activation site sequence (GAS) of the IFN regulatory factor promoter was 5′-GATCGATTTCCCCGAAATCATG-3′.
Immunoprecipitation and Western blot
Western blot was processed from samples of equivalent number of cells directly lysed in SDS loading buffer and sonicated. Final detection was made with the ECL Plus system (GE Healthcare).
For immunoprecipitation; cells (3 × 105) were lysed at 4°C in 250 μl of 20 mM Tris-HCl (pH 7.3), 0.15 M NaCl, 1% Nonidet P-40, 1 mM EDTA, and protease inhibitor mixture from Sigma-Aldrich at 1/50. After centrifugation, 40 μl of protein G (50% slurry; Sigma-Aldrich) and 0.2 μg of precipitating Ab were added. After a 4-h or overnight incubation at 4°C under agitation, protein G was washed once in lysis buffer and four times in TBS. Fixed proteins were eluted in 50 μl of SDS loading buffer and denatured at 95°C for PAGE and Western blot analysis. Semiquantitative analysis of relative expression of STAT1-P-Tyr, JAK1 and actin was performed using the Quantity One BioRad program. Significance was examined using a Mann-Whitney U test.
Abs used were STAT1 (9H2), phospho-STAT1 (Tyr701) (and phospho-SHP2 (Tyr580) from Cell Signaling Technology; JAK1 (polyclonal, rabbit) from Upstate Biotech; and SHP2 (B-1) from Santa Cruz Biotechnologies.
Small interfering RNA (siRNA) experiments
Stealth siRNA for ptpn11 (SHP2) and control (GFP) were supplied by Invitrogen. MRC5 cells (2 × 105 cells/well) in a six-well plate were transfected with the Xtreme GENE transfection reagent (Roche Diagnostics) using 2 μg of siRNA in 10 μl according to the suggested protocol by the manufacturer. After 72 h, cells were infected by HCMV for different times and analyzed for tyrosine-phosphorylated STAT1 (STAT1-P-Tyr) and SHP2 levels.
HCMV infection impaired IFN-γ-mediated tyrosine phosphorylation of STAT1
Tyrosine phosphorylation of STAT1 is an essential step in IFN-γ signaling (12). To investigate modulation of IFN-γ signaling in cells infected by HCMV, we analyzed STAT1 tyrosine phosphorylation at various times postinfection in MRC5 cells. As can be seen in Fig. 1, STAT1 tyrosine phosphorylation was impaired after 15 h of infection both in the cytoplasm and in the nucleus. Total STAT1 was also evaluated and, as expected, STAT1 present in the nucleus followed the profile of STAT1-P-Tyr because STAT1 migrates to the nucleus only when it is tyrosine phosphorylated (12). In the cytoplasm, the level of total STAT1 was constant, suggesting that the impaired phosphorylation observed was not due to degradation of the protein. Similar data were obtained using astrocytoma U373-MG cells (data not shown). Because degradation of JAK1 by the proteasome has been reported as a mechanism of inhibition of the IFN-γ pathway (27), we investigated this possibility. As can be seen in Fig. 2,A, in a representative experiment, the level of JAK1 was constant throughout infection. This was expected given that the duration of infection in our assay was much shorter than that required for JAK1 degradation which was reported to occur after 48 h of infection (27, 31). Fig. 2,B shows semiquantitative analysis of STAT1-P-Tyr vs total JAK1 expression, and Fig. 2 C shows total JAK1 vs actin expression from three independent experiments. STAT1-P-Tyr expression was diminished as compared with constant total JAK1, whereas expression of total JAK1 vs β-actin remained equivalent through the courses of experiments. Thus, the impaired tyrosine phosphorylation of STAT1 was not due to the degradation of JAK1.
To investigate whether HCMV replication was required for the phenomenon observed, HCMV was inactivated by UV irradiation (Fig. 3,A). Increasing time of irradiation led to complete inhibition of IE proteins expression and to restoration of STAT1-P-Tyr, suggesting that HCMV replication was required for inhibition of STAT1 tyrosine phosphorylation. This was further confirmed by γ-irradiation which also restored STAT1 tyrosine phosphorylation (Fig. 3,B). Ultracentrifugation of HCMV inoculum confirmed that HCMV particles were involved in the inhibition because pelleted HCMV, but not supernatant, was capable of inhibiting STAT1 tyrosine phosphorylation (Fig. 3 B). Requirement for replication of HCMV in the inhibition of STAT1 tyrosine phosphorylation suggests that binding of HCMV to its receptor(s) is not sufficient to inhibit STAT1 tyrosine phosphorylation.
Decreased IFN-γ signaling due to impaired STAT1 tyrosine phosphorylation
To investigate the consequences of impaired STAT1 tyrosine phosphorylation on cell signaling, steps involved in the activation of the immune response were investigated. Activation of promoters induced by IFN-γ is dependent on the binding of STAT1 to GAS (12). Thus, we first assessed the binding of STAT1 to GAS element using a consensus GAS oligonucleotide sequence. As shown in Fig. 4,A, binding of STAT1 to GAS induced by IFN-γ incubation was inhibited as the time of infection increased. HLA-DRα mRNA expression was severely impaired (Fig. 4 B).
The anti-HCMV activity of IFN-γ has been shown to rely on IDO (16). IDO expression, although constitutive in DCs (32), is inducible by IFN-γ in cells such as fibroblasts (33). Similar to MHC II, IDO mRNA expression was strongly diminished in MRC5 infected for 17 h (Fig. 4 B). Thus, HCMV infection impairs expression of genes involved in the immune response.
Involvement of the tyrosine phosphatase SHP2 in the impaired STAT1 tyrosine phosphorylation
To test whether a tyrosine phosphatase would be involved in the impaired phosphorylation observed, we inhibited tyrosine phosphatase activities using sodium orthovanadate. In the absence of infection, increasing concentrations of sodium orthovanadate potentiated the tyrosine phosphorylation of STAT1. This reflects the regulation of STAT1 tyrosine phosphorylation by tyrosine phosphatase (22, 34). After 16 h of infection, consistent with our original observation, a decrease of tyrosine phosphorylation of STAT1 was observed (Fig. 5,A). Increasing concentrations of sodium orthovanadate restored STAT1-P-Tyr up to levels observed in the absence of infection. Thus, a tyrosine phosphatase was shown to be involved in the impaired tyrosine phosphorylation of STAT1 during infection (Fig. 5 A). SHP2 is a tyrosine phosphatase which is expressed ubiquitously (35).
To investigate the involvement of SHP2 in the impaired activation of STAT1, we then assessed the interaction of SHP2 with STAT1. Fig. 5,B shows that HCMV induced coprecipitation of SHP2 and STAT1 which appeared as early as 6 h p.i. and was still visible 18 h p.i. In the presence of IFN-γ, coprecipitation was seen at the 18-h p.i. time point. Thus, HCMV infection was shown to induce coprecipitation of SHP2 and STAT1, suggesting interaction between those two proteins. Furthermore, phosphorylation of SHP2 was induced by HCMV, suggesting activation during infection (Fig. 5,C). Although basal levels of JAK1-P-Tyr were observed in MRC5 cells, tyrosine phosphorylation of JAK1 was also induced by HCMV, increased 6 h p.i., and was abolished at the last time point of the experiment (Fig. 5 C).
Implication of SHP2 in the impairment of STAT1 tyrosine phosphorylation was demonstrated using an siRNA approach: MRC5 cells were transfected with an SHP2 stealth siRNA. Cells were then assessed for STAT1 tyrosine phosphorylation after HCMV infection and IFN-γ stimulation. Fig. 6 shows that transfection of SHP2 siRNA, but not irrelevant siRNA, decreased the levels of endogenous SHP2. This decrease of SHP2 expression was correlated with the restoration of STAT1 Tyr phosphorylation at 24 h p.i. (Fig. 6, last lane). Therefore, this experiment demonstrated that SHP2 is involved in the decrease of STAT1 tyrosine phosphorylation induced by HCMV infection. Altogether, our data show that infection impairs STAT1 tyrosine phosphorylation through the activation of SHP2 and its interaction with STAT1. Whether other phosphatases are involved remains to be tested and is under investigation.
In this paper, we have described a new mechanism of inhibition of IFN-γ signaling by HCMV. This inhibition is due to impaired tyrosine phosphorylation of STAT1 as a consequence of SHP2 activity. Our observation defines a new mechanism of escape of HCMV from the immune response. Together with other mechanisms of escape from IFN-γ signaling (26, 36), this new mechanism may have a broad impact on the immune response.
A new time course of experiments allowed us to identify a new mechanism which neither involved JAK1 degradation (27) nor was related to our previously reported inhibition of CIITA expression which kept STAT1 tyrosine phosphorylation intact (28). Finding of similar inhibition of STAT1-P-Tyr in a different type of cell line (U373-MG cells; data not shown) emphasizes the potential relevance of our observation in HCMV infection. Fig. 7 summarizes our present data and gives an overview of timewise inhibition of the IFN-γ signaling pathway by HCMV.
HCMV has developed various strategies to block immune responses. By targeting different levels of the IFN-γ response at different time points of the infection, HCMV ensures an efficient blockade of the IFN-γ signaling pathway. Because IFN-γ is central in the development of the immune response such as Th cell polarization (37) and MHC gene expression (12), and in the combat against viruses (17, 18, 38, 39), inhibition of STAT1 tyrosine phosphorylation may have an impact on several aspects of immunity. In this respect, mechanisms such as inhibition of STAT1 tyrosine phosphorylation (this araticle) as well as induction of JAK1 degradation (27, 40), because they are upstream in the JAK/STAT pathway, are expected to have a broader impact on immunity than the previously described inhibition of CIITA in our laboratory (28).
Thus, both constitutive and IFN-γ-inducible expressions of MHC II proteins are down-regulated by HCMV (26). This contrasts with the high frequency of CD4+ T cells specific for HCMV Ags (41). The in vivo impact of the mechanisms of escape on the CD4+ T cell population thus remains to be evaluated. Surprisingly, there is no difference in the CD8+ T cell response of mice infected with wild-type CMV and CMV deleted from all MHC I evasion genes (42).
The role of STAT1 in the immune defense is illustrated by studies involving patients deficient in STAT1 (43) who showed susceptibility to major infectious agents. However, inhibition of the JAK/STAT pathway may not be as pervasive in HCMV infections and may only locally impair immune response at sites of infection. What are the potential consequences of STAT1 inhibition? IDO has been implicated in several compartments of the immune response. Inhibition of IDO, for example, may block an essential mechanism of anti-HCMV defense (16) and other microorganisms (44) and participate in global decrease of the immune response by reducing available tryptophan in the milieu. In contrast, IDO has been involved in the regulatory function of DCs (32). Because HCMV can infect DCs (45), regulatory functions of DCs may be perturbed by down-regulation of IFN-γ-inducible IDO. Constitutive IDO is essential in the maintenance of pregnancy (46). In addition, it has been reported that IFN-γ-inducible IDO may also play a role (47). More globally, inhibition of the IFN-γ pathway may block remodeling of the decidual vasculature, a crucial step in the development of pregnancy which is dependent on IFN-γ produced by uterine NK cells (48).
Several mechanisms account for impaired immune response in HCMV infection: HCMV encodes for a viral IL-10 with consequences on DC maturation and functionality (49). Impaired migration of DC (50), down-regulation of constitutive MHC I (40) and MHC II molecules (26) by HCMV may contribute to down-regulate the immune response. However, presumably due to its importance in the fight against pathogens, the JAK/STAT pathway seems to be a prominent target of viruses: Vaccinia (51), Sendai (52), Nipah (53), and Rinderpest viruses (54) have evolved their escape mechanisms toward this pathway to counteract the host’s immune response. However, no contribution of SHP2 has yet been described. Instead, vaccinia virus uses its own phosphatase, called VH1, to dephosphorylate STAT1 (51).
The receptor for epidermal growth factor (EGF) has been suggested to be a receptor for HCMV (55), although this is a matter of debate (56). Our data suggest that EGFR is not involved in the inhibition of STAT1 tyrosine phosphorylation by HCMV because this inhibition requires virus replication and is not induced by mere interaction of the virus with EGFR. Our conclusion has been further confirmed by demonstrating that rEGF does not modify IFN-γ-mediated STAT1 tyrosine phosphorylation (data not shown). This is of importance because signaling through EGFR has been shown to be controlled by the phosphatase SHP2 (57).
Kinetics of SHP2 phosphorylation vs STAT1 phosphorylation are crucial. JAK1 has been reported to interact with and activate SHP2 (58). We propose that SHP2 is phosphorylated and presumably activated by JAK1 in response to HCMV infection. Increase of STAT1-P-Tyr often observed at the 2- to 6-h p.i. time point (see Figs. 1 and 5 C) may be related to initial activation of the JAK1/STAT1 pathway by HCMV. There is then overlap between SHP2 activation and slowing down of ongoing STAT1 phosphorylation which leads to impaired detection in the 16- to 24-h time frame. At this point, SHP2 activation is declining. These hypotheses require further investigations.
Although other phosphatases may also be involved, SHP2 appears as a major component of the inhibition of STAT1 tyrosine phosphorylation by HCMV, as demonstrated by an almost complete restoration of STAT1-P-Tyr by SHP2 siRNA. In this regard, SHP1, although it may regulate the JAK/STAT pathway (21), was not tested because it is expressed exclusively in hemopoietic cells and some epithelial cells (35). Tyrosine phosphorylation of SHP2, as seen in our experiments in the presence of HCMV, may not always be indicative of its activation (21). However, restoration of STAT1-P-Tyr by both sodium pervanadate and SHP2 siRNA and STAT1/SHP2 association argue in favor of SHP2 activity in the impaired STAT1 tyrosine phosphorylation.
Other inhibitors such as SOCS proteins (59) have been also shown to play an important role in the regulation of the JAK/STAT pathway. However, the strong recovery of STAT1 tyrosine phosphorylation after treatment with SHP2 siRNA excludes a prominent role of SOCS proteins in the kinetics of infection used in our protocol. Regulation of gene expression downstream STAT1 is mediated by PIAS proteins (60). This mechanism is targeted at the level of regulation of gene expression and thus does not involve inhibition of STAT1 PIAS protein phosphorylation.
Our data are consistent with previous observation by Miller et al. (31) who showed that at earlier times points of HCMV infection, when JAK1 degradation is not yet detectable, there was inhibition of STAT1 homodimerization and functionality, although no mechanism was proposed. Our present study extends those observations and proposes a new mechanism which explains the lack of STAT1 functionality in the 12- to 24-h time frame of infection.
More studies are needed to identify viral protein(s) responsible for those mechanisms targeting the JAK/STAT pathway.
Inhibition of the JAK/STAT pathway by HCMV involves a series of events targeted to different steps of signaling, according to the time of infection (27, 28), and this article). The additive effect of all mechanisms used by HCMV to down-regulate the JAK/STAT pathway may have a broad impact on expression of proteins involved in the immune response and on the intensity of the immune response.
We thank Christian Davrinche for discussions and Delphine Nigon for statistical analysis.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work has been supported by Association pour la Recherche sur le Cancer and Sidaction.
Abbreviations used in this paper: HCMV, human CMV; MHC I, MHC class I; MHC II, MHC class II; IDO, indoleamine 2,3-dioxygenase; siRNA, small interfering RNA; DC, dendritic cell; SOCS, suppressors of cytokine signaling; SHP1, Src homology region 2 domain-containing phosphatase 1; SHP2, Src homology region 2 domain-containing phosphatase 2; GAS, IFN-γ activation site; STAT1-P-Tyr, tyrosine-phosphorylated STAT1; p.i., postinfection; EGF, epidermal growth factor.