The manipulation of signals downstream of the TCR can have profound consequences for T cell development, function, and homeostasis. Diacylglycerol (DAG) produced after TCR stimulation functions as a secondary messenger and mediates the signaling to Ras-MEK-Erk and NF-κB pathways in T cells. DAG kinases (DGKs) convert DAG into phosphatidic acid, resulting in termination of DAG signaling. In this study, we demonstrate that DAG metabolism by DGKs can serve a crucial function in viral clearance upon lymphocytic choriomeningitis virus infection. Ag-specific CD8+ T cells from DGKα−/− and DGKζ−/− mice show enhanced expansion and increased cytokine production after lymphocytic choriomeningitis virus infection, yet DGK-deficient memory CD8+ T cells exhibit impaired expansion after rechallenge. Thus, DGK activity plays opposing roles in the expansion of CD8+ T cells during the primary and memory phases of the immune response, whereas consistently inhibiting antiviral cytokine production.
Effective control of viral replication is dependent upon the expansion of Ag-specific CD8+ T cells that are poised to directly kill virally infected cells and are capable of producing cytokines that will help to shape the developing immune response. Postinfection, virally derived proteins that have been processed by the infected cells are presented in complex with MHC class I (MHC-I) molecules to the naive CD8+ T cell pool. Clones with the appropriate TCR will engage the peptide–MHC-I complex and become activated. Once activated, the reactive CD8+ T cell pool differentiates into effector CTLs and will migrate to the sites of infection. The effector phase is characterized by the aggressive expansion of the specific CD8+ T cell pool. Once the infection is resolved, the reacting T cell pool undergoes a contraction phase, where only 5–10% of the expanded effectors will persist. Finally, the formation and maintenance of the long-lived Ag-specific memory CD8+ T cell pool is formed, which will protect the host against subsequent infections by the same virus (1, 2).
The initiation of the T cell immune response begins with the recognition of foreign peptides, presented on MHC molecules, to the TCR. This immediate interaction, together with signals proximal to the TCR, has profound effects on the ensuing T cell response to bacterial and viral Ags (3–9). In addition, many extracellular determinants, together with intracellular signaling molecules and enzymes, have been identified as key regulators in the formation, function, and maintenance of the ensuing T cell response (10–15).
TCR signaling results in the production of secondary messengers, which serve to amplify and direct unique signaling pathways in activated T cells. The initiation of TCR signaling results in the activation of the Syk and Src family kinases. These proteins relay their signal by phosphorylating the adaptor proteins SLP-76 and LAT, which serve as docking sites for additional molecules in the TCR signaling cascade (16, 17). PLC-γ1 is then recruited to the SLP-76/LAT complex, where it becomes activated, and subsequently hydrolyzes phosphatidylinositol 4,5-bisphosphate into inositol 1,4,5-trisphosphate and diacylglycerol (DAG) (18, 19). Serving as a potent secondary messenger, inositol 1,4,5-trisphosphate initiates calcium release from the endoplasmic reticulum, activating the calcineurin pathway, and ultimately leading the nuclear translocation of NFAT (20, 21). DAG binds to and activates both RasGRP1 and protein kinase C (PKC)-θ via their cysteine-rich C1 domains. The result of DAG binding to RasGRP1 and PKC-θ is the activation of the Ras-ERK-AP1 and NF-κB pathways, respectively (22–27). The proximity of DAG production to the signaling events immediately downstream of the TCR, and its ability to activate multiple signaling pathways simultaneously, led investigators to hypothesize that the dysregulation of DAG signaling and metabolism might perturb normal T cell homeostasis and function.
As a potent positive regulator of T cell activation, the termination of DAG signaling is necessary to limit damage mediated by perpetually activated T cells or the development of an autoreactive T cell pool. A family of enzymes, the DAG kinases (DGKs), converts DAG to phosphatidic acid (PA), effectively terminating DAG-mediated signaling (28–30). Although 10 DGK isoforms have been identified in mammalian cells, DGKα and DGKζ are expressed in T cells (31–35). DGKα- and DGKζ-deficient mice were created to assess the functional relevance of DAG signaling in T cells in vivo. Analysis of DGK-deficient T cells revealed a variety of abnormalities. Enhanced activation of the Ras-MEK-Erk pathway and increased expression of CD25 and CD69 after anti-CD3 stimulation was observed in DGK-deficient T cells in comparison with their wild-type (WT) counterparts. In addition, DGKζ-deficient T cells were observed to be hyperproliferative. In an in vivo model, DGKζ-deficient mice infected with lymphocytic choriomeningitis virus (LCMV) exhibited an enhanced ability to control infection (9, 36). In CD4+ T cells, deficiency of DGKα and DGKζ results in resistance to the induction of T cell anergy both in vitro and in vivo (36). These observations clearly demonstrate a vital role for DAG metabolism in T cell-mediated immunity.
In this study, we examine the CD8+ T cell response to LCMV infection in DGK-deficient mice. The development of MHC-I tetramers (Tets) specific for LCMV-specific CD8+ T cell clones, as well as synthetic peptides that mimic virally derived Ags, have allowed us to enumerate the expansion of Ag-specific T cells on a single-cell scale in germline knockout (KO) models. Our data demonstrate that DGKα and ζ repress primary antiviral immune response by dampening CD8+ T cell expansion and cytokine production, but promote memory CD8+ T cell responses.
Materials and Methods
DGKα−/− (aKO) and DGKζ−/− (zKO) mice were generated as previously described (9, 36). TCR transgenic LCMV-specific P14 mice were purchased from Taconic and were bred to aKO or zKO to generate aKO-P14 and zKO-P14 mice in specific pathogen-free facilities at Duke University Medical Center. The experiments in this study were performed according to a protocol approved by the Institutional Animal Care and Usage Committee of Duke University. Eight- to 12-wk-old mice were used in this study.
Virus and infection
LCMV-Armstrong (Arm) stocks were propagated on BHK-21 cells and quantitated as described previously (37). A total of 2 × 105 PFU LCMV-Arm was administered to mice by i.p. injection for the acute viral infection.
To assess the primary immune response, we sorted 1 × 104 CD8+Vα2+CD44lo cells from the spleen (SP) of WT-P14, aKO-P14, and zKO-P14 mice, and transferred them i.v. into WT congenic hosts, which were infected the next day. For the memory response, 5 × 103 CD8+CD44hi cells were sorted from the SP and lymph nodes of WT, aKO, and zKO mice 4 mo postinfection with LCMV-Arm, and were transferred i.v. into WT congenic hosts. The next day, these hosts were infected with LCMV-Arm. Peripheral blood (PB) and splenocytes from infected mice were collected for analysis at the indicated times.
Cells from PB and SP were harvested, stimulated, and stained using appropriate Abs directly conjugated with fluorochrome. Tets of H-2Dbgp33–41 (TetG) and H-2DbNP396–404 (TetN) specific for LCMV were conjugated with allophycocyanin. The construction and purification of TetG and TetN have been described previously (38). For intracellular cytokine staining, splenocytes were stimulated with indicated peptides (10 μg/ml) in the presence of Golgi-Plug protein transport inhibitor (1:1000; BD Biosciences). After 5-h stimulation, the cells were harvested and stained using Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences). The peptides used for stimulation were gp33–41 (KAVYNFATM), gp276–286 (SGVENPGGYCL), and NP396–404 (FQPQNGQFI). To measure the mammalian target of rapamycin (mTOR) activity, we stained cells with a rabbit anti–phospho-S6 (Ser240/244) followed by FITC-conjugated anti-rabbit secondary Abs (Cell Signaling Technology). All of the data were collected using BD FACSCanto II and were analyzed with FlowJo software (Tree Star). Geometric mean fluorescence intensity (gMFI) is used to represent the amount of cytokine produced per cell (39).
Quantitative real-time PCR
Two-tailed unpaired Student t tests were performed for determining p values. Absolute numbers of gated groups such as CD8+ and CD4+ splenocytes were calculated with the percentages multiplied by total splenocyte number. All the graphs represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001.
Increased CD8+ T cell expansion in DGKα- and DGKζ-deficient mice infected with LCMV
To assess the physiological role for DAG metabolism during T cell expansion in vivo, we infected WT, aKO, and zKO mice with LCMV-Arm. Seven days postinfection, PB and SPs were harvested for analysis. Similar to WT mice, splenic cellularities were increased in aKO and zKO mice after LCMV infection (Fig. 1A). The expansion of CD4+ T cells was not obviously different between WT and aKO or zKO mice (Fig. 1B, 1C). Although the percentages of CD8+ T cells in the PB were similarly increased between WT and DGK-deficient mice, the increase of the percentages and absolute numbers of CD8+ cells in aKO or zKO SPs were 2-fold greater than the WT control (Fig. 1B, 1D). Thus, deficiency of either DGKα or ζ results in enhanced expansion of CD8+ T cells after LCMV infection.
One feature of CD8+ T cell-mediated responses to viral infection during the effector phase is the expansion of viral-specific T cell pool. Using MHC-I Tets, we were able to determine the frequency of LCMV-specific T cell clones within the bulk CD8+ T cell population. We observed a small increase in the frequency of CD8+ T cells reactive to the NP396–404 epitope and equivalent frequencies for the gp33–41 epitope in DGK-deficient mice as compared with WT mice (Fig. 1E). However, because of the increase in total CD8+ T cells in the SP, the number of Tet+ CD8+ T cells is higher in the SP of DGK-deficient mice infected with LCMV (Fig. 1F). Collectively, these data reveal that deficiency of either DGKα or ζ promotes the expansion of LCMV-specific CD8+ T cell clones in vivo.
Enhanced functionality of LCMV-reactive CD8+ T cells in the absence of DGKα or ζ
Another feature of CD8+ T cell-mediated antiviral immune responses is differentiation of naive CD8+ T cells into viral-specific, cytokine-producing CTLs. In response to LCMV infection, CD8+ T cells are known to produce large amounts of IFN-γ and TNF-α that are important for control of the infection (42). To determine whether LCMV-specific, DGK-deficient CD8+ T cells developed into cytokine producing effector cells postinfection, we harvested splenocytes from WT and DGK-deficient mice 7 d after immunization, and cultured them with MHC-I–restricted LCMV-specific peptides gp33–41, gp276–286, and NP396–404. IFN-γ and TNF-α production were determined by intracellular staining and FACS analysis. The percentages of cytokine-producing cells and gMFIs of produced IFN-γ, but not TNF-α, were increased in DGK-deficient CD8+ T cells (Fig. 2A, 2B). Because the total number of CD8+ T cells, as well as the proportion of overall number of cytokine-producing CD8+ T cells, was increased in the DGK-deficient mice, the overall number of cytokine-producing effector T cells was dramatically enhanced (Fig. 2C). Consistent with increased cytokine production upon LCMV infection, lower amount of LCMV transcript was detected in DGKα- or ζ-deficient mice as compared with WT controls, indicating that viral clearance was more effective in the absence of DGKα or ζ (Fig. 2D).
Intrinsic role for DGKζ-mediated DAG metabolism in antiviral CD8+ T cell expansion
The infection of aKO and zKO mice revealed a crucial role for DAG metabolism in regulating the CD8+ T cell response to viral infection. Because these mice are germline KO models, we cannot rule out the contribution of extrinsic factors (i.e., cytokines, growth factors, and costimulatory molecules) produced by cells other than T cells that may influence CD8+ T cells during LCMV infection. To investigate whether the enhanced antiviral response caused by DGKα or ζ deficiency involves a T cell-intrinsic mechanism, we bred aKO and zKO mice with P14 TCR transgenic mice to generate aKO-P14 and zKO-P14 mice. P14 mice have an MHC-I–restricted Vα2+Vβ8.1+ TCR specific to the gp33–41 epitope derived from LCMV (43). Sorted transgenic CD8+Vα2+CD44lo naive T cells from WT (C57B6/J)-P14, aKO-P14, and zKO-P14 mice (Thy1.2+) were adoptively transferred into congenic Thy1.1+ or Thy1.1+Thy1.2+ WT hosts. The recipient mice were infected with LCMV-Arm 1 d after adoptive transfer and analyzed 7 d postinfection (Fig. 3A). Comparing recipients injected with zKO-P14 or WT-P14 CD8+ T cells, there was no obvious difference of total number of splenocytes between these two groups (Fig. 3B). However, the percentage of donor-P14 cells within the CD8+Vα2+ population was increased in recipient mice receiving zKO-P14 T cells as compared with WT-P14 recipient mice (Fig. 3C). The absolute number of splenic zKO-P14 donor cells was about 7-fold higher than the WT control (Fig. 3D). Together, these observations suggest that zKO-P14 T cells possess an advantage in clonal expansion over WT-P14 T cells in competing with the endogenous LCMV-specific T cells. In contrast with zKO-P14 cells, aKO-P14 cells did not show enhanced clonal expansion in the recipients as compared with WT-P14 cells (Fig. 3E–G). When stimulated with peptide in vitro, both aKO- and zKO-P14 cells produced slightly more IFN-γ than WT-P14 cells (Fig. 3H). However, only aKO-P14 cells produced more TNF-α than WT-P14 cells. Collectively, data from the adoptive transfer experiments demonstrate that the enhanced antiviral responses of DGKζ-deficient, but not DGKα-deficient, T cells are cell intrinsic.
DGK-deficient mice contain fewer Ag-specific memory CD8+ T cells after LCMV infection
The experiments described earlier have revealed important roles for DGKα and ζ as negative regulators in CD8+ T cells during primary antiviral immune responses. We further investigated whether DGK-deficient CD8+ T cells can differentiate into long-lived Ag-specific memory cells. Four months postinfection, the total numbers of splenocytes in each group were similar (Fig. 4A). However, the percentages and total numbers of splenic CD8+TetG+ cells were decreased ∼50% in DGK-deficient mice (Fig. 4B, top panels, 4C). It has been reported that central memory T cells expand more effectively than effector memory T cells (44); however, the ratios of CD44hiCD62Lhi central memory and CD44hiCD62Llo effector memory within the CD8+TetG+ population were similar among these three groups (Fig. 4B, bottom panels). The proportion of KLRG1loIL7Rαhi long-lived memory precursor cells within the LCMV-specific CD8+ T cell pool was comparable among the WT, aKO, and zKO mice (Fig. 4D). Cumulatively, these data suggest that the generation and/or maintenance of an Ag-specific memory CD8+ T cell is impaired by DGK deficiency.
Impaired expansion of DGK-deficient memory CD8+ T cells after re-exposure to Ag
To assess the function of DGK activity in the memory CD8+ T cell response, we sorted CD8+CD44hi T cells from the mice (Thy1.2+) infected with LCMV-Arm 4 mo prior, normalized the numbers of LCMV-specific memory T cells using TetG, and adoptively transferred 5 × 103 TetG-reactive CD8+ memory cells into naive hosts (Thy1.1+). Recipient mice were infected with LCMV-Arm 1 d later and analyzed on 3, 5, and 7 d postinfection. The overall splenic cellularities in recipients of DGK-deficient memory CD8+ T cells were slightly reduced over a 3-d time course as compared with recipients of WT control cells (Fig. 5A). Furthermore, the recipient mice receiving DGK-deficient memory T cells showed reduction in the numbers of CD8+Thy1.2+ cells, as well as TetG-reactive CD8+Thy1.2+ cells (Fig. 5B, 5C). Importantly, the recipients of DGK-deficient memory cells were less efficient in viral clearance than those that received WT memory cells (Fig. 5D). Furthermore, the detection of higher virus transcript at days 3 and 5 in mice receiving DGK-deficient memory T cells indicates that the reduced number of reacting memory T cells at day 7 is due to impaired T cell expansion and not the result of enhanced viral clearance coupled with the quicker initiation of T cell contraction. Together, these data indicated that DGK activity promotes memory CD8+ T cell expansion during recall responses.
The mechanisms that control the formation and function of memory T cells have been intensively investigated. Recent reports demonstrated that mTOR inhibition promotes the generation of memory T cells (45–48), and our study revealed that DGKs function as negative regulators of mTOR activation through the inhibition of the DAG-RasGRP1-Ras-Erk1/2 pathway (49). Consistent with previous reports, we observed a higher level of phospho-S6 (a marker of mTOR activation) in DGK-deficient memory CD8+ T cells than in WT control (Fig. 5E), suggesting that DGK activity may promote memory CD8 T cell responses by inhibiting mTOR signaling.
DGK deficiency does not impair cytokine production by memory CD8+ T cells during recall response
The impaired proliferation of DGK-deficient memory CD8+ T cells after LCMV infection led us to investigate whether DGKs may also regulate the production of antiviral cytokines by reactivated memory T cells. In contrast with decreased expansion of memory CD8+ T cells, DGK-deficient memory cells produced more cytokines than the WT control cells after in vitro stimulation with LCMV-derived peptides (Fig. 6). Although we could not rule out that the increased cytokine production was influenced by delayed viral clearance in the recipient mice, our data suggest that DGKα and ζ exert differential effects on memory CD8+ T cell expansion and effector cytokine production.
CD8+ T cells play crucial roles in antiviral immunity. Understanding how CD8+ T cells are regulated during viral infection is instrumental for treatment of infectious diseases and for enhancing the efficacy of vaccination. Engagement of CD8+ T cells with APCs through the interaction between viral-specific TCRs and viral peptide–MHC-I complex is critical for the initiation of CD8+ T cell-mediated antiviral immunity. The signal strength from viral-specific TCRs can influence the effector and memory CD8+ T cell differentiation, maintenance, and activation (3–5, 7, 50). We demonstrate that both DGKα and ζ are important regulators for CD8+ T cell-mediated antiviral immune responses, and that DGK activity plays differential roles in primary and memory CD8+ T cell responses.
Our data have revealed that DGKα and ζ negatively control CD8+ T cell-mediated antiviral responses during primary infection. These observations are consistent with the role of DGKs as negative regulators of T cell activation in vitro by terminating DAG-mediated signaling. As a potent secondary messenger, DAG promotes signaling in both the NF-κB and Ras-Erk pathways. In the absence of either DGKα or ζ, TCR-induced activations of Ras-Erk1/2 and PKC-NF-κB pathways are enhanced (9, 51), which likely contribute to enhanced antiviral responses of DGK-deficient T cells during primary LCMV infection. Under the control of T cell-specific promoter, a dominant IκB severely impairs T cell proliferation (52). Furthermore, RelB−/− (a component of NF-κB signaling complex) mice show increased susceptibility to LCMV infection and impaired expansion of the CD8+ T cell pool. RelB has also been shown to associate with the TNF-α promoter (53). In addition, proper signaling through the Ras-Erk pathway has also been shown to be critical during an immune response. RasGRP1 mice show impaired Ag-specific CD8+ T cell expansion and fail to generate LCMV-specific, IFN-γ–producing CD8+ T cells in response to LCMV infection. This impaired response results in higher viral titers in the SP of infected mice (54). The increased expansion and cytokine production observed in DGK-deficient CD8+ T cells after LCMV infection could be directly related to the accumulation of DAG and enhanced activation of the Ras-Erk1/2 and NF-κB signaling in vivo. Although these data support that proper regulation of the signaling cascades downstream of DGK activity is crucial for the generation of an effective adaptive immune response, we cannot rule out that DGKs may regulate CD8+ T cell-mediated immune responses through the generation of PA. Similar to DAG, PA binds to and regulates multiple signal molecules. DGK-derived PA has been implicated in several signaling pathways to modulate cellular function and development of different cell lineages. Furthermore, although our study focuses on the role of DGKs in mature T cells postinfection, it is worth noting that DGKs and the Ras-Erk and NF-κB pathways are critical to immune cell development (55, 56). Although we could not rule out that DGKα or ζ deficiency may directly or indirectly affect the generation of LCMV-specific T cells in thymus, we found that they are involved in regulating mature CD8+ T cells during viral infection. The P14 adoptive transfer experiment suggests DGKζ functions in mature T cells to dampen antiviral immune responses and does not function in development to profoundly distort the TCR repertoire.
DGKα and ζ play a redundant role for T cell maturation in thymus (56). Although both DGKα and ζ exert similar roles in CD8+ T cells during anti-LCMV responses, some differences between these two DGK isoforms have been noted. DGKζ appears to play a stronger role than DGKα, because the enhanced anti-LCMV response seen in aKO mice appears to rely, at least partially, on T cell extrinsic factors. The different antiviral effectiveness between DGKα and DGKζ deficiency is not surprising because there are important differences in the structure and activation of these proteins. DGKα, which contains two EF hand motifs, requires the binding of Ca2+ to achieve full enzymatic activity, whereas DGKζ is not sensitive to such regulation. Furthermore, DGKζ contains a nuclear localization sequence within its myristoylated alanine-rich C-kinase substrate domain, which may give DGKζ regulatory functions not associated with DGKα. Finally, these enzymes may serve important nonenzymatic functions by acting as protein docking substrates, where differences in primary and tertiary protein structure determine the components of multiprotein signaling complexes (57).
Our data indicate that naive and memory CD8+ T cells can be differentially regulated by signaling molecules modulated by DGK activity during antiviral immune responses. In contrast with inhibiting primary antiviral immune responses, DGKα and ζ appear to promote expansion of viral-specific memory CD8+ T cells during secondary infection. The mechanisms underlying such differential roles of DGK activity in primary and memory antiviral immunity are currently unclear. It is known that the mTOR promotes primary but inhibits memory CD8+ T cell responses during LCMV infection (45). Recently, we have reported that DGKα and ζ synergistically inhibit TCR-induced mTOR activation by downregulating the DAG-RasGRP1-Ras-Erk1/2 pathway (49). Consistent with these observations, we have found increased phosphorylation of S6 in DGK-deficient memory CD8+ T cells. Thus, the opposing roles of DGKs and mTOR in primary and memory CD8+ T cells during LCMV infection suggests that in addition to regulating Ras-Erk1/2 and NF-κB signaling, DGKα and ζ may differentially control primary and memory CD8+ T cell responses during LCMV infection by inhibiting mTOR activation.
The authors have no financial conflicts of interest.
We thank Nancy Martin and Mike Cook in the Duke Cancer Center Flow Cytometry Core Facility for providing sorting services.
This work was supported by the National Institutes of Health (Grants R01AI076357, R01AI079088, and R21AI079873), the American Cancer Society (Grant RSG-08-186-01-LIB), and the American Heart Association (to X.-P.Z.).
Abbreviations used in this article:
- diacylglycerol kinase
- geometric mean fluorescence intensity
- lymphocytic choriomeningitis virus
- MHC class I
- mammalian target of rapamycin
- phosphatidic acid
- peripheral blood
- protein kinase C
- quantitative real-time PCR
- Received August 4, 2011.
- Accepted December 26, 2011.
- Copyright © 2012 by The American Association of Immunologists, Inc.