RasGRP1 Regulates Antigen-Induced Developmental Programming by Naive CD8 T Cells

The TCR plays a fundamental role in both the development of mature T cells in the thymus and their function in the periphery (1, 2). During their ontogeny, the recognition of self-peptides/self-MHC molecules (self-Ags) by the TCR of a developing thymocyte is translated into a directive for either cell suicide or survival and differentiation. After maturation and export to the periphery, the interaction of a naive CD8 T cell’s TCR with self-Ags promotes their survival and peripheral maintenance, whereas encounter with a specific foreign Ag (foreign peptide associated with self-MHC molecules) instructs a developmental program consisting of massive clonal expansion, effector cell formation, contraction through apoptosis, and differentiation into long-lived memory cells. The basis by which the TCR may instruct different outcomes onto a cell is through graded signaling strengths, the affinity/avidity for its ligands determining the magnitude of signaling. Tight regulation of T cell homeostasis is essential for productive immune responses necessary for both the clearance of pathogen and prevention of autoimmune-mediated self-destruction (3). The molecular mechanisms that regulate the recruitment of naive T cells into cognate Ag-induced developmental program and navigate their transition from naive to effector to memory cells are poorly understood.

TCR stimulation induces the rapid activation of small GTPase Ras whose signals are critical for orchestrating the positive selection of developing thymocytes (4, 5). Ras functions to link membrane receptor signaling to the internal MAPK cascades, cycling between an “off” (GDP-bound) and an “on” (GTP-bound) form. Ras guanine nucleotide exchange factors (Ras GEFs) greatly accelerate the release of GDP, facilitating Ras to bind GTP and the assumption of the active state. Conversely, Ras activity can be shut off by either its slow intrinsic GTPase activity or by pairing with GTPase-activating proteins. In thymocytes, the regulation of Ras activity by TCR signal transduction invokes at least two Ras GEFs, Sos and RasGRP1 (6–8). The respective functions of both Ras GEFs are dependent on relocating to membranes by two distinct mechanisms. RasGRP1 is recruited to membranes by binding the phospholipase Cγ1-product diacylglycerol (DAG) through its C1 domain, whereas Sos is recruited to the phosphorylated adaptor molecule linker for activated T cells by way of its constitutive association with the Src homology 2-domain–containing protein Grb2. RasGRP1 is likely the primary Ras GEF responsible for well-documented Ras activation observed following treatment of thymocytes with DAG analogs, such as PMA (6, 9, 10).

Identifying the roles of Ras activation in mature T cell function have been hindered because transgenic (Tg) mouse approaches of knocking down either Ras or its downstream MAPK cascade results in blocked thymocyte development and is accompanied with very few peripheral T cells (4). At least two different mechanisms have been postulated to regulate Ras activation upon TCR ligation in both human and mouse peripheral T cells (11, 12). Similar to the case of regulation of Ras in thymocytes, mature T cells are also thought to use the RasGRP1 and Sos for TCR-induced Ras/ERK activation (13, 14). However, the relative contribution of these two Ras GEFs to TCR-induced Ras signaling and their resulting impact on the activation and differentiation of mature CD8 T cells is unclear. Notably, defects in both Ras and ERK activation following TCR stimulation have been associated with anergized CD4 T cells and the inability to produce the T cell growth factor IL-2 (15, 16). A causal relationship between defective Ras signaling and T cell anergy has recently been demonstrated using novel techniques to transfect a constitutively active Ras into anergic T cells (17). Moreover, gain-of-function studies using the human Jurkat T cell line suggest that RasGRP1 promotes IL-2 transcription and thus may stave off an anergic state (13).

We have previously generated two lines of TCR Tg mice to study the role of RasGRP1 in CD8 T cell development under conditions of defined TCR signaling (10). Our results found that RasGRP1 is critical for thymocytes expressing the weakly selecting H-Y TCR, whereas those that express the strongly selecting 2C TCR are much less dependent on RasGRP1 for ERK activation and positive selection (10). The relatively efficient thymic development of 2C TCR CD8 T cells in mice lacking RasGRP1 along with their naive peripheral phenotype and near-normal cell numbers afforded us with the unique opportunity to study the role of RasGRP1 in Ag-specific T cell responses. In this study, we report that RasGRP1 plays a key role in the early activation events, such as CD25 upregulation and IL-2 secretion, following exposure to cognate Ag. RasGRP1^−/− 2C CD8 T cells exhibit greatly reduced sensitivity to TCR stimulation, failing to proliferate to either immobilized anti-TCR Abs or limiting doses of cognate Ag, and reduced burst size upon encounter with cognate Ag both in vitro or in vivo. These investigations highlight RasGRP1’s function in regulating the growth of mature T cells and suggest that RasGRP1 may prove valuable as a potential drug target, suppressing T cell responses in transplantation and autoimmune disease settings.

RasGRP1-deficient mice have been created and described previously by Dower et al. (6). C57BL/6J (B6), B6.PL-Thy1^a/Cy (Thy 1.1⁺), and B6.IL-2^−/− mice were acquired from The Jackson Laboratory (Bar Harbor, ME; http://www.jax.org/). RasGRP1-deficient 2C TCR Tg mice were created by crossing with gene-targeted RasGRP1^−/− mice that had been bred onto a B6 background at least 10 generations. To generate Thy1.1⁺/Thy1.2⁺ 2C TCR Tg animals, B6 mice bearing the 2C TCR transgene were mated with B6.PL-Thy1^a/Cy mice (H-2^b, Thy1.1⁺). All studies followed guidelines set by the Animal Care Committee at the University of British Columbia in conjunction with the Canadian Council on Animal Care.

Abs against CD3 (145-2C11), CD4 (GK1.5), CD8α (53-6.7), CD8β (53-5.8), TCR-β (H57-597), CD25 (PC61.5), CD44 (IM7), CD62L (MEL-14), CD69 ([¹H].2F3), CD122 (TM-b1), CD127 (AKR34), Thy1.1 (HIS51), Thy1.2 (53-2.1), HSA (M1/69), programmed cell death 1 (PD-1) (J43), PD-1 ligand 1 (MIH5), B220 (RAB-6B2), IL-2 (JES6-5H4), IL-4 (11B11), IL-10 (JES5-16E3), TNF-α (MP6-XT22), and IFN-γ (XMG1.2) were purchased from eBioscience (San Diego, CA). The clonotypic anti-2C TCR Ab was purified from the 1B2 hybridoma (from Dr. Herman Eisen, Massachusetts Institute of Technology, Cambridge, MA) and conjugated to either FITC or biotin using standard methods. For detection of intracellular cytokine, cells were stimulated in the presence of GolgiPlug (BD Biosciences, San Jose, CA; contains brefeldin A) to block cytokine secretion as previously described (18). Anti–Ki-67 (B56) and isotype control (MOPC-21) Abs were purchased from BD Biosciences, and the staining protocol employed was as published (18). Data were acquired with either a BD LSRII benchtop cytometer using FACSDiVa software (BD Biosciences) or FACSCalibur and CellQuest software (BD Biosciences). Data was analyzed with either CellQuest or FlowJo (Tree Star, Ashland, OR) software.

CD8 T cells were purified from spleens and lymph nodes using indirect magnetic bead-based separation according to manufacturer’s instructions (MACS, Miltenyi Biotec, Auburn, CA). Briefly, CD8 T cells were labeled with anti-CD8β (53-5.8) Ab coupled to biotin, followed by incubation with Streptavidin MicroBeads (Miltenyi Biotec) and column purification. For plate-bound anti-TCR Ab stimulations, tissue culture wells were coated with either 10 μg/ml of anti-CD3ε (145-2C11) Ab alone or 10 μg/ml of anti-CD3ε plus 5 μg/ml of anti-CD28 (37.51) mAbs. For cognate Ag stimulations, 2C CD8 T cells were incubated with irradiated B6 splenocytes that had been pulsed with 0.1 μg/ml of SIY peptide or the indicated dose of p2Ca (LSPFPFDL), a low affinity 2C TCR ligand when presented by H-2K^b MHC class I molecules (19, 20).

CFSE-labeled wild-type and RasGRP1^−/− 2C CD8 T cells (50,000; Thy1.2⁺) were cultured in vitro for 4 d with 1 million congenic CD8-depleted splenocytes (feeder; Thy1.1⁺) and limiting numbers of SIY-coated APCs in flat-bottom 96-well plates. For preparation of SIY-coated APCs, CD8-depleted splenocytes from congenic (H-2^b; Thy1.1⁺) mice were incubated with 0.1 μg/ml SIY peptide for 1 h at RT and subsequently washed 3 times with PBS. For assessment of IFN-γ production, 1 million SIY-loaded splenocytes (Thy1.1⁺) and the golgi-transport inhibitor GolgiPlug (BD Biosciences) were added to cultures 4 d poststimulation for a 5-h period. Samples were stained and electronically gated on 2C CD8 (Thy1.2⁺ CD8⁺) cells to assess proliferation and markers of differentiation. To assess the contribution of APC-derived IL-2 to 2C T cell proliferation, experiments were performed as described above except feeder cells and SIY-coated APCs were generated from splenocytes isolated from IL-2^−/− mice where indicated. All wild-type and RasGRP1^−/− Thy1.2⁺ CD8⁺ T cells expressed high levels of the 2C TCR as measured by anti-clonotypic 2C TCR mAb (1B2-4H6; data not shown).

Peripheral CD8 T cells were labeled with biotinylated anti-CD8β Ab (53-5.8) and purified using streptavidin MicroBeads (Miltenyi Biotec). For preparation of APCs, T cell-depleted B6 splenocytes, using anti-Thy1.2 Ab (53-2.1) and magnetic separation, were incubated with the indicated doses of SIY for 90 min at 37°C and subsequently washed three times with media to eliminate soluble peptide. Both CD8 T cells and APCs were prewarmed at 37°C for at least 15 min prior to being mixed together. To set up conjugation assays, 2.5 × 10⁵ purified 2C CD8 T cells from wild-type and mutant mice were incubated for 15 min at 37°C with 1 × 10⁶ APCs, subsequently fixed using 1% paraformaldehyde in PBS for 30 min at RT, and stained with anti-CD8α (53-6.7), anti-2C TCR, and anti-B220 Abs. For the assessment of conjugation formation, the frequency of CD8⁺ 2C TCR⁺ cells that labeled with anti-B220 Ab was determined by flow cytometry.

2C CD8 T cells were purified by labeling pooled splenic/lymph node single cell suspensions with biotinylated anti-CD8β (53-5.8) Ab, incubated with streptavidin MicroBeads (Miltenyi Biotec), and subjected to magnetic separation. For TCR stimulations, CD8 T cells (10⁷/ml) were prewarmed at 37°C for at least 15 min in serum-free media and subsequently treated with 10 μg/ml of anti-CD3ε Ab (145-2C11). Abs used to detect phosphorylated forms of ERK, linker for activation of T cells (LAT), and ZAP-70 and total amounts of ERK2 have been described previously (21). For IL-2 signaling, T cell blasts were generated by activating purified 2C CD8 T cells for 2 d with 0.1 μg/ml SIY peptide plus irradiated splenocytes for APCs and subsequently expanding them for 3 to 4 d with 20 U/ml IL-2. T cell blasts were serum-starved for 4 h prior to stimulation with either IL-2 (100 U/ml) or PMA (100 μM). At the indicated time, cells were rapidly precipitated by centrifugation, washed with ice-cold PBS, and lysed in 10 mM Tris (pH 7.5), 150 mM NaCl, 1% NP40, 0.1% SDS, protease inhibitors, and phosphatase inhibitors. Ras pulldown assays using Raf-GST fusion protein bound to glutathione-coupled beads were performed as described elsewhere (6, 13). A mixture of a pan-Ras (no. R02120, BD Transduction Laboratories, Lexington, KY) and anti–K-ras (no. 30, Santa Cruz Biotechnology, Santa Cruz, CA) Abs were used to measure Ras in both GST-Raf precipitates and total lysates. Blots were developed using ECL system (Amersham Biosciences, Piscataway, NJ). Densitometry was performed using ImageJ software (National Institutes of Health, Bethesda, MD).

A recombinant L. monocytogenes strain expressing a secreted form of the 2C TCR agonist peptide SIYRYYGL (rLM-SIY) was engineered to express the SIY peptide within a secreted dihydrofolate reductase fusion protein. The Ag cassette containing the SIY peptide with flanking OVA and α-KG amino acid residues has been previously shown to induce strong anti-SIY responses in B6 mice (22). After PCR amplification and cloning of the Ag cassette into the suicide vector pJJD180, the construct was introduced into the bacterial genome by homologous recombination as described previously (23).

To assess the capacity of rLM-SIY to induce 2C TCR CD8 T cell proliferation, 1 million purified wild-type (Thy1.2⁺/1.2⁺) 2C TCR CD8 T cells were labeled with CFSE and adoptively transferred into congenic (Thy1.1⁺/1.1⁺) B6 mice via i.v. injection. Twenty-four hours later, mice were left untreated, infected with 10,000 CFU wild-type LM (strain 10403S), or infected with 10,000 CFU rLM-SIY. One-week postinfection, donor 2C CD8 T cells were identified by staining with anti-Thy1.2 and anti-CD8 Abs and their proliferation measured by flow cytometry using a BD FACSCalibur or BD LSRII benchtop cytometer using CellQuest and FACSDiVa software, respectively (BD Biosciences). For coadoptive transfers experiments, a 50:50 mixture (10,000 of each genotype) of wild-type (Thy1.1⁺/1.2⁺) and RasGRP1^−/− (Thy1.2⁺/1.2⁺) 2C CD8 T cells were infused into syngenic (Thy1.1⁺/1.1⁺) B6 mice and infected the next day by i.v. injection using a dose of either 1,000 or 10,000 CFU of rLM-SIY in PBS. Alternatively, wild-type and mutant 2C CD8 T cells (20,000) were infused into separate recipient hosts and immunized with 1,000 CFU of rLM-SIY. Bacterial doses were verified by plating the injectant on brain-heart infusion agar (BD Biosciences). To assess proliferation at day 3 postinfection, a greater number of wild-type or RasGRP1^−/− 2C CD8 T cells (1 × 10⁶), to facilitate their detection, were adoptively transferred into Thy1.1 mice prior to infection with rLM-SIY.

To investigate the role of RasGRP1 in peripheral CD8 T cell function, we analyzed CD8 T cells from RasGRP1-deficient mice expressing the MHC class I-restricted 2C TCR transgene (24). For the study of cognate Ag-driven CD8 T cell responses, the 2C TCR recognizes the agonist peptide SIYRYYGL (SIY) bound to syngeneic H-2K^b MHC class I molecules with high affinity (25). Previously, we have observed that CD8 SP thymocytes were present at similar numbers in wild-type and RasGRP1^−/− 2C TCR thymi, demonstrating that RasGRP1 is not critical for the development of mature cells by this TCR (10). Although peripheral RasGRP1^−/− 2C CD8 T cells are modestly reduced (∼50%) relative to wild-type, they possess a mature and naive cell surface phenotype as determined by expression levels of 2C TCR, PD-1, PD-1 ligand 1, CD3, CD24 (HSA), CD8α, CD8β, CD25, CD44, CD62L, CD69, CD122, and CD127 (10 and data not shown]. To investigate for impairments in TCR signaling, we tested the ability of wild-type and RasGRP1^−/− 2C CD8 T cells to proliferate when cultured in wells coated with anti-TCR Ab. Proliferation was tracked 3 d poststimulation by either analyzing dilution of CFSE-endowed fluorescence by flow cytometry (Fig. 1A) or measuring incorporated [³H]thymidine during a 6 h pulse period 3 d poststimulation (Fig. 1B). By contrast to the SIY/APC response, RasGRP1^−/− 2C CD8 T cells failed to even enter cell cycle when cultured on immobilized anti-TCR Ab alone. Furthermore, complementing TCR signals with anti-CD28 Ab-mediated costimulation also failed to induce significant proliferation by RasGRP1-deficient 2C CD8 T cells (Fig. 1A, 1B). However, the proliferative responses by RasGRP1^−/− 2C CD8 T cells resembled wild-type T cells when either plate-bound anti-TCR Ab or anti-TCR plus anti-CD28 Abs stimulations were supplemented with exogenous IL-2. These data demonstrate that RasGRP1-deficient 2C CD8 T cells display an anergic phenotype that is reversible upon the addition of exogenous IL-2.

The findings made with RasGRP1^−/− 2C CD8 T cells using immobilized anti-TCR Ab as an agonist argue that RasGRP1 is important for early T cell activation and recruitment into cell cycle (Fig. 1A, 1B). However, our previous experiments, using soluble SIY peptide presented by syngeneic splenocytes to stimulate RasGRP1^−/− 2C CD8 T cells, had suggested early activation and proliferation were RasGRP1-independent events (10). It is possible that the apparent discrepancy between these observations may be a consequence of the fact that APCs deliver additional signals, such as coreceptor engagement, costimulatory molecule interactions, and cytokines, that anti-TCR plus anti-CD28 Abs alone do not provide. To investigate the role of RasGRP1 in mediating early T cell activation events and whether this role differs depending on the type of stimulation, wild-type and RasGRP1^−/− 2C CD8 T cells were stimulated with either SIY-pulsed splenocytes or anti-TCR Ab for 16 h and subsequently assessed for the induction of activation markers (Fig. 1C). A comparison of histograms and mean fluorescence intensities (MFIs) revealed that RasGRP1^−/− 2C CD8 T cells exhibited impaired upregulation of the activation markers as compared with wild-type (CD25: 201 versus 1280, 6.4× lower MFI; CD69: 295 versus 776, 2.6× lower MFI; and PD-1: 295 versus 630, 2.2× lower MFI). By contrast, RasGRP1^−/− 2C CD8 T cells exhibited a more normal activation marker expression profile when cells were treated with peptide/APC (CD25: 1182 versus 1748, 1.5× lower MFI; CD69: 642 versus 893, 1.4× lower MFI; PD-1: 370 versus 464, 1.3× lower MFI). These findings argue that Ag receptor-induced expression of activation markers is heavily dependent on RasGRP1 when costimulatory or supplementary signals are limiting, such as the case with plate-bound anti-TCR Ab stimulation alone. Next, we sought to determine whether RasGRP1^−/− 2C T cells show defects in the induction of IL-2 because its production by T cells is an early event upon encounter with cognate Ag and exogenously added cytokine rescues mutant T cell proliferation (Fig. 1D). Wild-type and RasGRP1^−/− 2C T cells were stimulated with SIY-pulsed splenocytes and monitored hourly for IL-2 expression by intracellular flow cytometry. At each time point poststimulation, RasGRP1^−/− 2C T cells displayed a significant reduction in the proportion of IL-2–positive cells. This finding implies that signaling by RasGRP1 regulates IL-2 transcription.

Previously, we had observed that RasGRP1-deficient 2C CD8 T cells undergo early proliferation following exposure to cognate Ag, soluble SIY peptide, and irradiated syngeneic splenocytes, but failed to sustain their growth (10). To investigate whether the impaired Ag-driven proliferation and developmental programming of RasGRP1^−/− 2C CD8 T cells may be related to their decreased capacity to produce IL-2, we compared the differentiation status of wild-type and mutant naive 2C CD8 T cells stimulated with cognate Ag in vitro either in the absence or presence of exogenous IL-2 (Fig. 2A, 2B). After 4 d of culture with SIY-pulsed irradiated syngeneic splenocytes alone, wild-type 2C CD8 T cells underwent considerable remodeling of their cell surface, elevating CD25, CD44, and the cytotoxic effector molecule granzyme B while simultaneously downregulating CD62L (Fig. 2A). Although RasGRP1-deficient 2C CD8 T cells lost CD62L expression, they exhibited markedly reduced expression of activation and differentiation markers relative to wild-type (CD25: 91 versus 301, 3.3× lower MFI; CD44: 450 versus 1646, 3.6× lower MFI; granzyme B: 138 versus 294, 2.1× lower MFI). Next, we tested whether culturing in the presence of both SIY-pulsed irradiated syngeneic splenocytes and exogenous IL-2 could rescue the differentiation of RasGRP1^−/− CD8 T cells (Fig. 2B). Wild-type and RasGRP1^−/− CD8 T cell effector profiles exhibited much less pronounced differences when exogenous IL-2 is added to cultures (CD25: 354 versus 777, 2.2× lower MFI; CD44: 1431 versus 1715, 1.2× lower MFI; granzyme B: 384 versus 392, similar MFI). Therefore, the differentiation of RasGRP1^−/− CD8 T cells, as measured by surface markers and granzyme B expression, can largely be restored by supplementation of medium with IL-2. Collectively, these experiments demonstrate that RasGRP1 plays a critical role in the efficient generation of CTL effectors and that its function can be compensated for by the addition of exogenous IL-2.

The stimulation of naive T cells with cognate Ag results in their commitment to a developmental program linked with the production of effector cytokines, such as IFN-γ (26). However, it has become clear that CD8 T cell effectors can produce a wider variety of cytokines and be subdivided into polarized IFN-γ–producing (Tc1) and IL-4–producing (Tc2) populations in an analogous fashion to Th1 and Th2 paradigm of CD4 T cells (27). To determine whether signaling by RasGRP1 contributes to the quantity or variety of cytokines elicited upon TCR engagement, CD8 T cell blasts, generated by culture with cognate Ag and exogenous IL-2 (Fig. 2B), were stimulated with SIY-pulsed splenocytes in vitro for 5 h. Regardless of genotype, the majority (∼99%) of CD8 T cell effectors produce IFN-γ after contact with cognate Ag (Fig. 2C). As shown for naive RasGRP1^−/− 2C T cells, RasGRP1^−/− T cell blasts also secrete less IL-2 than their wild-type counterparts. To investigate if RasGRP1 deficiency affects cytokine polarization of CD8 T cells, we assessed the production of the Tc2 cytokines IL-4 and IL-10 (Fig. 2C). These results demonstrate that the majority of RasGRP1^−/− CD8 T cell effectors produce IFN-γ and suggest that RasGRP1 function does not appear to influence Tc1 versus Tc2 polarization.

The poor responsiveness to immobilized anti-TCR Abs and decreased capacity to synthesize IL-2 suggest that RasGRP1^−/− 2C T cells may be less sensitive to cognate Ag stimulation. To test this hypothesis, a constant number of CFSE-labeled wild-type and RasGRP1^−/− 2C T cells were cultured with 1 million feeder splenocytes and limiting numbers of SIY-pulsed APCs (Fig. 3A). After 4 d of culture, 150,000 total events were collected and CFSE fluorescence of 2C T cells plotted on histograms with cell numbers’ scale indicated on the y-axis. As has been observed previously in an analogous experiment with P14 TCR CD8 T cells (28), increasing the number of cognate Ag-loaded APCs resulted in an increased frequency of TCR Tg CD8 T cells recruited into cell cycle and correlated with an increased burst size (larger scale on y-axis). In the absence of SIY-coated APCs, about the same number of wild-type and mutant 2C T cells were recovered (same scale on y-axis; Fig. 3A). By contrast, RasGRP1^−/− 2C T cells were almost 10-fold less sensitive to cognate Ag relative to wild-type because 2000 SIY-loaded APCs were required to recruit >40% of cells into cycle, whereas only 200 SIY-APCs were necessary for wild-type T cells to reach this level of recruitment. In addition, recruited RasGRP1^−/− 2C T cells retained more CFSE-endowed fluorescence suggesting these cells underwent fewer divisions than wild-type (Fig. 3A). Moreover, decreased proliferation or increased cell death could reduce the burst size of RasGRP1^−/− 2C CD8 T cells. The finding that RasGRP1^−/− 2C CD8 T cells do not show increased rates of cell death, as measured by Annexin V staining (data not shown), suggests that their rate of cellular division rather than cell death contributes to their reduced representation. These results imply that RasGRP1 strongly influences the Ag dose necessary to commit naive CD8 T cells to undergo proliferation.

We next examined the relationship between the T cell differentiation marker CD44 and cell division in wild-type and mutant 2C T cells stimulated with the indicated doses of SIY-coated APCs (Fig. 3B). Regardless of cell genotype, Ag-driven proliferation was accompanied with elevated levels of CD44. However, RasGRP1^−/− 2C T cells had fewer CD44hi cells particularly at lower doses of cognate Ag. To determine whether CD44 upregulation correlated with effector function in mutant T cells, cultures were prepared as above (Fig. 3A, 3B), stimulated with 1 million SIY-coated splenocytes for additional 5 h, and assessed for IFN-γ production (Fig. 3C). Again, a reduced fraction of RasGRP1^−/− 2C T cells were capable of secreting IFN-γ because a meager number of cells were mobilized to proliferate and differentiate when cognate Ag was limiting. Collectively, these findings argue that RasGRP1 regulates the recruitment of naive CD8 T cells to initiate Ag-induced developmental programming.

The phenotypes exhibited by RasGRP1^−/− 2C T cells may be a function of their decreased capacity to produce IL-2 or unrelated to the production of this autocrine growth factor. To explore whether decreased IL-2 production by RasGRP1-deficient T cells might affect the induction of activation markers, wild-type and RasGRP1^−/− 2C T cells were stimulated with anti-TCR Abs (± exogenous IL-2) or SIY-coated APCs (wild-type or IL-2^−/− splenocytes) for 16 h (Fig. 4A). The finding that exogenous IL-2 partially restores the expression of CD25, CD69, and PD-1 on anti-TCR Ab activated RasGRP1^−/− 2C T cells suggests that their decreased IL-2 secretion may be responsible for these deficits. By contrast, IL-2 sufficiency by SIY-coated APCs impacts the capacity of 2C T cells to upregulate CD69 levels but has relatively minor effects on CD25 and PD-1. Together, these observations suggest that RasGRP1 signaling influences IL-2 as well as other genes.

Next, we sought to determine whether the capacity of SIY-coated APCs to initiate RasGRP1^−/− 2C T cell proliferation was dependent on their capacity to secrete IL-2. To address this question, an experiment similar to Fig. 3 was performed (Fig. 4B). Wild-type and RasGRP1^−/− 2C T cells were incubated with 1 million feeder splenocytes (wild-type or IL-2^−/−) plus various numbers (0, 1,000, or 10,000) SIY-pulsed APCs (wild-type or IL-2^−/−) and 2C T cell proliferation analyzed by FACS after 4-d culture. Strikingly, SIY-coated splenocytes lacking IL-2 expression were much less efficient at initiating both wild-type and RasGRP1^−/− 2C T cell proliferation relative to wild-type APCs. These findings suggest that the provision of IL-2 by SIY-coated APCs is critical for recruiting RasGRP1^−/− 2C T cells into cell cycle.

T cell responsiveness to cognate foreign Ag is determined by T cell-APC conjugate formation and regulated by VAV1 and Src kinase-associated protein of 55 kDa (29). To investigate whether diminished Ag-responsiveness by RasGRP1-deficient 2C T cells is connected with faulty formation of T cell-APC conjugates, prewarmed wild-type and RasGRP1^−/− 2C T cells were mixed with APCs (T cell-depleted splenocytes) that had been pulsed with various dilutions of SIY peptide for 15 min, fixed, and the frequency of 2C T cells costaining with B220 determined (Fig. 4C). These studies found that RasGRP1^−/− 2C T cells did not exhibit an impaired ability to generate T cell-APC conjugates and suggest that RasGRP1 might not be involved in this process.

Next, we assessed whether RasGRP1^−/− 2C T cells might exhibit decreased TCR sensitivity in the presence of exogenous IL-2. Wild-type and RasGRP1^−/− 2C T cells were set up for proliferation assays using various doses of p2Ca peptide along with irradiated splenocytes and exogenous IL-2 for stimulation (Fig. 4D). Presentation of p2Ca by H-2K^b MHC class I molecules forms a low-affinity ligand recognized by the 2C TCR (19, 20). RasGRP1^−/− 2C T cells displayed blunted responses relative to wild-type 2C T cells despite the presence of exogenous IL-2. These observations suggest that RasGRP1 deficiency results in reduced TCR sensitivity that is not fully reversible upon addition of IL-2.

Prior studies have shown that thymocytes lacking a functional RasGRP1 protein exhibit a greatly diminished capacity to activate ERK upon treatment with either DAG analogs (6, 9, 10) or anti-TCR Abs (6, 9, 21). To examine the consequences of RasGRP1 deficiency on the ability of mature peripheral T cells to activate ERK, 2C CD8 T cells were purified using magnetic separation from wild-type and mutant mice and stimulated with anti-TCR Ab for the indicated times (Fig. 5A). By contrast to the hyporesponsiveness of RasGRP1^−/− immature thymocytes (6, 9, 21), RasGRP1^−/− 2C CD8 T cells strongly phosphorylate ERK at an early time (3 min) poststimulation at levels resembling wild-type (Fig. 5A). However, ERK activation in mutant 2C CD8 T cells is shorter in duration, exhibiting significant deficits at 30 and 60 min post TCR stimulation. Analysis of the upstream signaling molecules argues that LAT and ZAP-70 phosphorylation are also affected by RasGRP1 deficiency. These findings suggest that RasGRP1 may be critical for sustaining the kinetics of ERK activation following cognate Ag stimulation and that such prolonged signaling may be necessary for Ag-stimulated T cells to enter cell cycle.

Activation of the Ras/MAPK cascade in T cells not only accompanies Ag receptor ligation but is also invoked by IL-2 receptor signaling (30, 31). To examine whether RasGRP1 plays a role in IL-2-mediated Ras signaling, activated wild-type and RasGRP1^−/− 2C CD8 T cell blasts were serum-starved for 4 h prior to incubation with IL-2 for 10 min (Fig. 5B). Assessment of the activation status of Ras and ERK post-IL-2 stimulation revealed that the Ras/ERK cascade seemed largely intact in RasGRP1^−/− cells despite the observation that they exhibited a 2-fold reduction in levels of the IL-2 receptor α chain CD25 (Figs. 2B, 5B). By contrast, RasGRP1^−/− 2C T cell blasts displayed very weak Ras activation upon PMA treatment relative to wild-type (Fig. 5B). These experiments show that activation of Ras initiated by IL-2 receptor signaling does not directly depend on RasGRP1.

To study the consequences of RasGRP1 deficiency on Ag-driven proliferation, differentiation, and memory formation of 2C CD8 T cells in vivo, we engineered L. monocytogenes, a well-characterized pathogen model for eliciting T cell responses, to express the 2C agonist SIY peptide (Fig. 6A). The recombinant strain rLM-SIY, derived by integration of an Ag cassette into the bacterial chromosome between the lecithinase and lactate dehydrogenase operons using techniques described previously (23), expresses the SIY peptide as a secreted dihydrofolate reductase fusion under the control of a virulence promoter (P_hly). To test the capacity of rLM-SIY to induce 2C T cell proliferation, 1 million CFSE-labeled wild-type 2C T cells were infused to B6 mice (nonirradiated; Thy1.1+) and left untreated (naive) or infected with 10,000 CFU of either rLM-SIY or wild-type LM 10403S (Fig. 6B). One wk postinfection, the vast majority of splenic 2C CD8 T cells residing in either naive mice or those infected with wild-type LM failed to proliferate (95% and 90%, respectively). By contrast, analysis of donor 2C CD8 T cells in rLM-SIY–infected mice revealed that these cells had strongly proliferated, virtually completely losing their CFSE-endowed fluorescence. In addition, donor 2C T cell proliferation induced by infection with rLM-SIY was accompanied by a >100-fold increased recovery of donor 2C T cells relative to either naive B6 hosts or those infected with wild-type LM (data not shown). These findings demonstrate that rLM-SIY can effectively prime and induce the expansion of naive wild-type 2C T cells in vivo.

Naive T cells encountering cognate Ag and the appropriate costimulatory signals initiate a development program, undergoing massive clonal expansion (26, 28). To test the proliferative capacity and function of RasGRP1^−/− 2C CD8 T cells in vivo, equivalent numbers of CFSE-labeled wild-type (Thy1.1⁺/1.2⁺) and mutant (Thy1.2⁺/1.2⁺) 2C CD8 T cells were infused into the same mouse, ensuring cells of each genotype were exposed to a similar environment and an equivalent amount of cognate Ag and recipients infected the next day with 1000 CFU of rLM-SIY (Fig. 6C). Because recent experiments have shown that naive T cell precursor frequency profoundly affects the differentiation program taken by CD4 and CD8 T cells (32, 33), we injected greatly reduced numbers of 2C T cells (10,000/genotype) per mouse in comparison with the experiments presented in Fig. 6B to more accurately resemble physiological levels of naive T cells specific for a given Ag (34, 35). Using anti-Thy1.2 Ab to identify donor cells and anti-Thy1.1 Ab to discriminate genotype, RasGRP1^−/− 2C CD8 T cells were found to be represented at a 4-fold decreased proportion and total numbers as compared with wild-type counterparts 1 wk postinfection (Fig. 6D, 6E). In addition, similar findings were observed when wild-type and mutant 2C T cells were immunized in separate hosts and their cell numbers determined 1 wk later. Under these conditions, wild-type 2C T cells were found to be ∼4.7-fold greater than those lacking RasGRP1 (^+/+ = 1.21 ± 0.34 × 10⁶ versus^−/− = 0.258 ± 0.045 × 10⁶; p < 0.01 using a Student t test).

Analyses of cellular proliferation revealed that the majority of 2C T cells had undergone at least seven to eight rounds of cell division regardless of genotype and therefore lacked fluorescence bestowed by CFSE (Fig. 6F). To determine whether the reduced yield of RasGRP1^−/− 2C T cells at day 7 postinfection was a consequence of decreased early expansion, CFSE-labeled wild-type and mutant 2C T cells were adoptively transferred and their proliferation history assessed 3 d postinfection (Fig. 6G). At this time point, wild-type and RasGRP1^−/− 2C CD8 T cell populations exhibited a similar average number of cell divisions (5.3 versus 4.9) and fraction of cells not recruited into cell cycle (9.7% versus 13.1%). Next, we analyzed the expression of the proliferation-associated nuclear Ag Ki-67 in day 7 2C T cell effectors to investigate whether decreased cellular proliferation might be responsible for the reduced representation of mutant CD8 T cells (Fig. 6H) because tracking of cell division using CFSE fails to resolve additional information once the fluorescence of CFSE-labeled cells approaches the autofluorescence of unstained cells (36). A lower proportion of RasGRP1^−/− 2C CD8 T cells were found to express the Ki-67 marker as compared with wild-type (10.7 ± 2.4% versus 26.9 ± 5.9%). Together, these findings suggest that RasGRP1 sustains the proliferation of CD8 T cells responding toward cognate Ag in vivo.

The differentiation of naive T cells resulting from Ag-driven clonal expansion is tightly linked with phenotypic and functional changes. This transformation includes increased expression of effector/memory markers, like CD44, and the acquisition of effector functions, such as the capacity to produce the proinflammatory cytokines, like IFN-γ and TNF-α, and cytotoxicity (26, 28, 37). To investigate the cell surface phenotype of donor 2C T cells after cognate Ag exposure in vivo, splenocytes of host mice 7 d postinfection with rLM-SIY were stained with Abs against various markers in addition to Thy1.1, Thy1.2, and CD8 (Fig. 6I). Analysis of donor 2C T cells (filled, black histograms) revealed that cells of either genotype had also strongly upregulated CD44 expression as compared with naive 2C cells (gray histograms), albeit the mutant cells expressed modestly reduced CD44 levels relative to wild-type (2720 versus 3790 MFI). To test the function of donor 2C T cell effectors, splenocytes isolated from host mice 7 d postinfection were placed in culture, stimulated for 5 h with the SIY peptide, and subsequently assessed for the production of the proinflammatory cytokines IFN-γ and TNF-α (Fig. 6J). By contrast to wild-type, a significant fraction of RasGRP1^−/− 2C CD8 T cell effectors failed to produce cytokine (21%), and a diminished percentage coexpressed IFN-γ and TNF-α relative to wild-type (38% versus 83%). However, RasGRP1^−/− 2C CD8 T cell effectors expressed wild-type levels of granzyme B and were efficient killers of EL-4 targets pulsed with SIY peptide (data not shown).

Next, we wanted to test whether the expression of RasGRP1 influenced the capacity of 2C T cell effectors to differentiate into memory T cells. However, to our surprise, donor 2C T cells, either wild-type or mutant, completely disappeared within 2 wk post rLM-SIY infection (data not shown). In addition, we were unable to visualize long-lived memory 2C T cells regardless of number of 2C T cells (1 × 10³–2 × 10⁶) adoptively transferred, the infectious dose of rLM-SIY (2 × 10³–1 × 10⁵ CFU), the number of events acquired (≥2 × 10⁷), or whether mice were rechallenged with rLM-SIY. Because analogous experiments with OVA-specific, MHC class I-restricted TCR transgenic (OT-1) CD8 T cells and rLM-OVA infections yielded abundant numbers of long-lived memory OT-1 CD8 T cells (data not shown), we believe that the inability to detect memory 2C CD8 T cells is not a technical issue. Moreover, the fact that rLM-SIY infection gives rise to long-lived anti-SIY CD8 memory T cells in normal (non-TCR Tg) B6 mice (data not shown and 38) leads us to suspect that the inability to find long-lived memory 2C CD8 T cells following rLM-SIY infection is somehow linked to the 2C TCR.

Understanding the molecular controls of the immune system will yield insight into the bases of immunological diseases, improve the efficacy of vaccination, and lead to the identification of molecular targets suitable for therapeutic intervention. Ras signaling pathways are known to play key roles in both the development of thymocytes and the regulation of mature T cell function. At least two important Ras GEFs, RasGRP1 and Sos, are known to play critical roles in TCR signal transduction in developing thymocytes and mature T cells (5). The findings that thymocyte positive selection is relatively inefficient in the absence of RasGRP1 demonstrate that these Ras GEFs are not functionally redundant (6, 9, 10). Moreover, aberrant positive selection in RasGRP1-deficient mice is associated with an abundance of activated- and memory-phenotype peripheral CD4 and CD8 T cells (9, 18, 39), a likely result of chronic immunodeficiency (18) and/or a breakdown in T cell tolerance (9). Therefore, studies of mature T cells in non-TCR Tg RasGRP1^−/− mice are tenuous because their differentiation status, a factor that influences TCR signal transduction (40), does not match wild-type. In this study, we report the use of RasGRP1^−/− 2C TCR mice to investigate how this Ras GEF participates in peripheral T cell function and differentiation because these animals produce sufficient quantities of 2C CD8 T cells, exhibiting a mature, naive cell surface phenotype resembling those recovered from wild-type 2C TCR animals (10).

Sufficient exposure of naive T cells to cognate Ag commits them to an Ag-independent phase of proliferation that is intimately coupled to differentiation, composing of chromatin remodeling, gene expression changes, and the acquisition of effector functions (26, 41). We have found that RasGRP1-deficient 2C CD8 T cells require a much greater dose of cognate Ag to divide and when recruited into cell cycle undergo fewer rounds of proliferation (Fig. 3A). A net consequence of decreased sensitivity to TCR-stimulation is that few RasGRP1^−/− 2C CD8 T cells differentiate into effectors at limiting doses of cognate Ag. A possible mechanism by which RasGRP1 may lower activation threshold and promote the growth of naive T cells is through its regulation of IL-2R α chain (CD25) expression and IL-2 production (Fig. 1C, 1D) because it has been shown that blocking of IL-2 signaling in Ag-stimulated TCR Tg CD8 T cells can reduce the proportion of naive T cells recruited to proliferate and the extent to which they divide (28). In addition, we found that the addition of exogenous IL-2 can promote both the proliferation and effector cell formation of TCR-stimulated RasGRP1^−/− 2C CD8 T cells (Figs. 1, 2). These findings support a role of RasGRP1 in which it impacts T cell activation and growth, at least in part, through regulation of IL-2 production.

Anergy is a peripheral tolerance mechanism in which a T cell is rendered functionally inactive following encounter with cognate Ag. The hyporesponsive state of anergic T cells is acquired and maintained through intrinsic wholesale changes in TCR signaling. Moreover, a number of characteristics displayed by RasGRP1^−/− 2C CD8 T cells are reminiscent of T cell anergy including hyporesponsiveness to cognate Ag stimulation, decreased IL-2 production, and variable reversibility of this phenotype by the addition of IL-2 (42). In addition, analyses of TCR signaling pathways in anergic T cells have often revealed defects in Ras/ERK activation and AP-1-induced gene transcription (43). Gene expression profiling experiments pinpointed the overexpression of DAG kinases (DGKs), negative regulators of Ras signaling that convert DAG to phosphatidic acid, as a possible cause of aberrant TCR signaling in anergic T cells (17). Supporting this notion, DGKζ-deficient mice exhibit enhanced T cell responses and Ras/ERK activation (44), whereas gain-of-function experiments enforcing DGKζ expression in Jurkat cells have been found to inhibit TCR-induced activation of Ras, ERK, and AP-1 (45). A cause-and-effect relationship between the overexpression of DGKs and anergy was demonstrated by findings that the introduction of DGKα into mouse T cells inhibited MAPK signaling, IL-2 production, and the recruitment of RasGRP1 to membranes (17). Our data are consistent with the hypothesis that DGKs suppress T cell responsiveness by blocking RasGRP1’s ability to activate Ras.

Cellular proliferation may be a critical factor in the differentiation of T cells because it may provide the opportunity for chromatin remodeling and thus facilitate changes in gene expression (41, 46). To investigate whether the decreased proliferative potential of naive RasGRP1^−/− 2C CD8 T cells in vitro affected their ability to differentiate into effector and memory CD8 T cells in vivo, we engineered the model pathogen L. monocytogenes to express the 2C TCR agonist SIY peptide (Fig. 6A). Although RasGRP1^−/− 2C CD8 T cells were found to exhibit a 4-fold reduction in the generation of day 7 effectors relative to wild-type (Fig. 6D), it is possible that the experimental conditions may have partially masked their signaling defects because the presence of wild-type host CD4 T cells and their accompanying helper-associated function of IL-2 secretion may compensate for RasGRP1 loss. Moreover, decreased numbers of RasGRP1^−/− 2C T cell effectors was associated with reduced amounts of proliferation (Fig. 6H) and proinflammatory cytokine secretion (Fig. 6J), suggesting the possibility that signaling by RasGRP1 may be important for the formation of T cell memory. However, we were unable to test this hypothesis because we could not detect long-lived memory 2C T cells after rLM-SIY infection despite the observations that these cells had been induced to undergo numerous rounds of proliferation (Fig. 6F), elevate CD44 expression (Fig. 6I), and acquire the capacity to produce proinflammatory cytokines (Fig. 6J).

Critical parameters mediating the differentiation of naive CD8 T cells into effector and, eventually, memory T cells are thought to be the Ag dose, Ag duration, and the affinity of the TCR for a given cognate Ag (26). The findings that rLM-SIY infection in normal B6 mice results in the formation of long-lived memory anti-SIY CD8 T cells suggest that the 2C TCR may be somehow problematic for the development of CD8 T cell memory (38). Perhaps the dose or duration of Ag provided by rLM-SIY infection is suboptimal to promote 2C T cells to differentiate into memory cells. Moreover, it is likely that rLM-SIY results in a contracted Ag exposure because it exhibits greatly reduced virulence (ability to divide in vivo), a trait observed for multiple strains of recombinant LM (47), as compared with the wild-type strain 10403S. For example, i.v. tail injection of 10,000 CFUs of wild-type LM 10403S into B6 mice will yield millions of CFUs in both the spleen and liver 3 d postinfection, whereas the bacterial load following the same dose of rLM-SIY will border the limit of detection (data not shown). Notably, a recent study has demonstrated that 2C T cells can differentiate into long-lived memory T cells, persisting at least 225 d postinfection, using a recombinant influenza virus expressing the SIY peptide, named WSN-SIY (48). By contrast to rLM-SIY infection, Ag is detectable for at least 8 d postinfection after WSN-SIY infection (J. Chen, personal communication). An alternate explanation for the disappearance of 2C CD8 T cells following our infection experiments could be the rejection of donor T cells by the host. Although we have bred our 2C TCR Tg mice 10 generations onto the B6 background, it is plausible that some flanking DNA closely linked to the 2C TCR transgene could be facilitating graft rejection.

One intriguing question is why TCR signaling engages two Ras GEFs, each using intricate, independent mechanisms for their membrane recruitment and mobilization (5). Analyses on wild-type and mutant 2C CD8 T cells revealed that RasGRP1 is likely crucial for sustaining TCR-induced ERK activation (Fig. 5A) and that such a prolonged signal may be required for recruitment of naive T cells into cell cycle (Fig. 1A, 1B). Because Ras can also be activated through the IL-2 receptor (30, 31), we also sought to investigate whether IL-2 signaling induced Ras activation via RasGRP1. Consistent with the finding that cytokine receptors activate the Ras/MAPK pathway after their phosphorylation through binding of Shc adaptor protein and subsequently the recruitment of Grb2/Sos (49), we found that IL-2 stimulations of RasGRP1^−/− 2C T cell blasts induced Ras activation at levels resembling wild-type (Fig. 5B) and restored proliferation to immobilized anti-TCR Ab (Fig. 1A, 1B). Therefore, we suspect that RasGRP1 does not directly mediate Ras activation through the IL-2 receptor. However, our findings suggest that RasGRP1 may indirectly affect IL-2 receptor signaling by influencing CD25 expression and IL-2 secretion (Fig. 1C, 1D). Recent work in the human Jurkat T cell line suggests that an unusual type of cooperation exists between RasGRP1 and Sos whereby RasGRP1-derived Ras-GTP boosts Sos’s activity by binding an allosteric pocket on the molecule, creating a positive feedback loop and collaborating to induce robust Ras activation (50). However, our results demonstrate another mechanism of Ras activation, possibly Grb2/Sos, is functional in the complete absence of RasGRP1 because RasGRP1^−/− 2C CD8 T cells can activate Ras and ERK upon TCR ligation (Fig. 5A and data not shown).

We hypothesize that RasGRP1 is critical for optimal Ras/ERK signaling necessary for the exquisite sensitivity shown by T cells in responding to a very limited number of peptide/MHC complexes or low affinity ligands on the surface of an APC (Figs. 3A, 4D). It has been postulated that ERK positively regulates TCR signaling in mature T cells through a feedback loop mechanism by inactivating the Src homology region 2 domain-containing phosphatase 1 and, as a consequence, increasing LCK activity (51). The existence of such a feedback loop could explain our results with RasGRP1^−/− 2C T cells exhibiting reduced TCR-induced phosphorylation of LAT and ZAP-70, two molecules residing upstream of RasGRP1 (Fig. 5A). Thus, RasGRP1/ERK pathway may function as a rheostat, providing the fine-tuning of TCR signals and regulating TCR responsiveness.

The discovery of subcellular compartmentalization of Ras activation has garnered great attention and revealed another dimension of complexity to cellular signaling and possible cellular responses (5, 52). Although conventional wisdom had placed the site of Ras activity at the plasma membrane (PM), recent studies suggest that Ras activation also occurs at endomembranes, namely the Golgi, in a PLCγ-dependent fashion by way of RasGRP1 (53–55). Recently, subcellular compartmentalization of the Ras/RasGRP1/ERK pathway has been suggested to be pivotal in developing thymocytes undergoing selection (56). Using OT-1 TCR Tg preselection CD4⁺CD8⁺ thymocytes and specific peptides spanning the boundary of positive and negative selection, it was found that negatively selecting peptides recruited Ras, Raf-1, and RasGRP1 to the PM, whereas these molecules colocalized to internal cellular sites, perhaps the Golgi and/or other endomembranes, when thymocytes were stimulated with a peptide mediating positive selection (56). In addition, studies on peripheral OT-1 CD8 T cells correlated active ERK being localized to the PM when cells were induced to undergo apoptosis, whereas a condition promoting survival and proliferation was associated with ERK activation present at endosomal compartments (57). The rationale for alternate cellular sites of Ras activity may be to subject it to differential regulation and/or serve to pair it with a unique subset of effectors. Comparison of the PM versus endomembranes has found that Ras activation at the latter locale appears to have delayed or sustained kinetics relative to the former (53–55). A possible mechanism for prolonged Ras signaling at endomembranes may be through calcium signals, serving to recruit the Ras GTPase activating protein CAPRI to the PM to inactivate Ras at this site while at the same time mediating the translocation of RasGRP1 to internal cellular sites (53, 58). In conclusion, our results combined with the findings of others hint at the possibility that subcellular compartmentalization of Ras/RasGRP1/ERK pathway is necessary for sustained Ras/ERK signaling and regulates the commitment of naive CD8 T cells to Ag-induced developmental programming.

We thank Soo-Jeet Teh for technical assistance and the Wesbrook Animal Facility for animal husbandry. We are also grateful to Dr. D. H. Loh (Nippon Roche Research Center) for providing breeders of the 2C TCR Tg mouse and Dr. Herman Eisen (Massachusetts Institute of Technology) for supplying the 1B2 hybridoma.

Disclosures The authors have no financial conflicts of interest.