Engagement of NKG2D by Cognate Ligand or Antibody Alone Is Insufficient to Mediate Costimulation of Human and Mouse CD8+ T Cells
http://www.100md.com
免疫学杂志 2005年第4期
Abstract
CD8+ T cells require a signal through a costimulatory receptor in addition to TCR engagement to become activated. The role of CD28 in costimulating T cell activation is well established. NKG2D, a receptor found on NK cells, CD8+ -TCR+ T cells, and -TCR+ T cells, has also been implicated in T cell costimulation. In this study we have evaluated the role of NKG2D in costimulating mouse and human naive and effector CD8+ T cells. Unexpectedly, in contrast to CD28, NKG2D engagement by ligand or mAb is not sufficient to costimulate naive or effector CD8+ T cell responses in conventional T cell populations. While NKG2D did not costimulate CD8+ T cells on its own, it was able to modify CD28-mediated costimulation of human CD8+ T cells under certain contitions. It is, therefore, likely that NKG2D acts as a costimulatory molecule only under restricted conditions or requires additional cofactors.
Introduction
Innate and adaptive immune systems function coordinately to recognize and eliminate pathogens and to prevent pathology upon re-exposure to the same pathogen. The innate immune system recognizes molecular patterns on classes of pathogens to quickly alert the body to infection and contain the pathogen (1). In contrast, the adaptive immune system recognizes unique epitopes on specific pathogens or transformed cells to generate a tailored response to eliminate the pathogen or tumor and to provide for long term immunological memory (2). Given the different roles of innate and adaptive immunity, receptors used by cells of the two systems are frequently specialized for different functions. However, there are some receptors that are shared between cells of the innate and adaptive immune systems, suggesting that they may be important mediators of both types of immunity. One such receptor is NKG2D, which is expressed on mouse and human NK cells, -TCR+ T cells, human CD8+ -TCR+ T cells, and activated mouse CD8+ -TCR+ T cells (3, 4, 5, 6).
NKG2D forms a homodimer that pairs with the adapter protein DAP10 in resting NK cells and CD8+ T cells (4, 6). The only known signaling motif on the DAP10 molecule is a YXXM motif, which enables NKG2D/DAP10 to recruit the p85 subunit of PI3K and Grb2 upon ligand binding (4, 7). This motif is shared by other costimulatory receptors, such as CD28.
CD8+ T cells require two stimuli for activation. Signal 1 is provided through the TCR via recognition of cognate peptide:MHC, and signal 2 is provided through costimulatory molecules (8), the prototype of which is CD28. Because NKG2D shares the YXXM motif with CD28 and is capable of recruiting the p85 subunit of PI3K, it was proposed that it could play a costimulatory role in CD8+ T cells. Indeed, CD8+ T cell costimulation via NKG2D has been reported in both human and mouse models (5, 9). In CMV-specific human T cell clones, NKG2D was reported to costimulate proliferation, cytokine production, and cytolytic activity (9). Similarly, in a lymphocytic choriomeningitis virus (LCMV)5-specific mouse CD8+ T cell line, NKG2D cross-linking was reported to costimulate proliferation. We also characterized the role of NKG2D in costimulation of mouse and human CD8+ T cells. We report in this study that neither stimulation via mAb cross-linking nor NKG2D ligand engagement alone is sufficient to costimulate proliferation or effector functions in primary mouse or human CD8+ T cells, suggesting that NKG2D acts as a costimulatory molecule only under restricted conditions or requires additional cofactors.
Materials and Methods
Abs and flow cytometry
The following mAbs were used for immunofluorescent staining of mouse cells: anti-NKG2D (CX5 or 191004) (10), anti-CD8 (53-6.7), rat IgG2a (B39-4), rat IgG2b (A95-1), anti-pan-RAE-1 (186107), anti-B7.1 (16-10A1), and anti-CD11c (HL3). Biotinylated mAbs were detected with allophycocyanin-conjugated streptavidin (BD Biosciences), and pure mAbs were detected with FITC-conjugated goat anti-rat IgG (Caltag Laboratories). The following were used for staining human cells: anti-NKG2D (149810), anti-CD8 (25T8-5H7; gift from E. Reinherz, Dana Farber Cancer Institute, Boston, MA), anti-CD8 (Leu 2a), anti-CD28 (L293), and anti-CD80 (2D10; eBioscience). Unless otherwise indicated, mAbs were obtained from BD Biosciences, except for mAbs 149810, 191044, and 186107, which were produced in collaboration with Dr. J. P. Houchins (R&D Systems). Cell staining and flow cytometry were performed using standard procedures on a FACSCalibur instrument (BD Biosciences).
Mice
C57BL/6 mice were obtained from Charles River. OT.1 TCR-transgenic mice were provided by Dr. J. Cyster (University of California, San Francisco (UCSF), CA). Experiments were performed according to the guidelines of the UCSF committee on animal research. Mouse RAE-1 cDNA was cloned by RT-PCR using oligonucleotide primers (sense primer, atg gcc aag gca gca gtg acc aa; antisense primer, tca cat tgc aaa tgc aaa tgc aaa taat). The pFRTZ vector (11) (provided by Dr. S. Dymechi, Harvard University, Boston, MA) was modified to remove lacZ and insert the RAE-1 cDNA behind the human -actin promoter. The RAE-1 vector was injected into fertilized (C57BL/6xDBA/2) F2 oocytes by the UCSF Comprehensive Cancer Center Transgenic and Targeted Mutagenesis Shared Resource. RAE-1 transgenic mice were backcrossed for five generations onto the C57BL/6 background.
Cell lines
Mouse B16 melanoma cells (12) were transduced with pMX retroviruses (13) encoding mouse B7.1 and/or RAE-1. P815 cells were transfected by electroporation with human B7.1 in the pBJ1-neo vector and were transduced with MHC class I chain-related A (MICA)*0019 in the pMX-pie retrovirus using standard methods. MICB*002-transfected P815 cells were provided by Dr. K. Soderstrom (Stanford University, Stanford, CA).
Mouse CD8+ T cell purification
For naive mouse CD8+ T cells, lymphocytes were obtained from spleens and lymph nodes by mechanical disruption, and RBC were lysed in ACK buffer (Cambrex BioScience). Cells were resuspended in RPMI-T (RPMI 1640 with 10% heat-inactivated FCS, 2 mM L-glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin, and 50 μM 2-ME) and mixed with anti-CD4 (GK1.5) mAb for 30 min at 4°C. After washing, lymphocytes were incubated for 20 min at 4°C with goat anti-rat IgG and goat anti-mouse IgG magnetic beads (Qiagen), which bind mAb-coated CD4+ cells and B cells, respectively, followed by magnetic depletion of the bead-bound cells. The remaining cells were incubated with PE-conjugated anti-CD8 (53-6.7) for 30 min at 4°C. Anti-PE-conjugated MACS beads (Miltenyi Biotec) were used to positively select mouse CD8+ T cells according to manufacturer’s instructions. Purity was confirmed by flow cytometry.
Generation of mouse OT.1 CD8+ T cell effectors
Splenocytes (3 x 106) from OT.1 mice were incubated at 37°C with 1 nM SIINFEKL peptide (OVAp) in a 24-well plate with 1 ml of RPMI-T. Four days later, the cells were expanded 1:5 into RPMI-T containing 200 U/ml human rIL-2 (National Cancer Institute Biological Resources Branch Pre-Clinical Repository) and 4 ng/ml mouse rIL-7 (R&D Systems). Two days later, the cells were split 1:2 into the same medium. At 7–8 days, the cells remaining in culture (>98.5% CD8+; data not shown) were used as effectors.
Mouse CD8+ T cell proliferation assays
For mAb-induced proliferation assays, wells of 96-well plates were incubated with 1 mg/ml N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium chloride (Sigma-Aldrich) in 50% ethanol for 5 min at room temperature. The plate was washed in PBS, and mAb in carbonate buffer (8.4 g of NaHCO3 and 3.56 g of Na2CO3 in 1 liter of H2O) were incubated overnight at 4°C. Anti-mouse CD3 mAb (500A2; BD Biosciences) was used at the indicated concentrations. Anti-mouse CD28 mAb (37.51; BD Biosciences) and anti-mouse NKG2D mAb (CX5) were used at 5 μg/ml. The wells were washed and blocked in RPMI-T for 10 min at room temperature. T cells (1 x 105) were incubated in each well in RPMI-T medium for 48 h (or 4 days; Fig. 2B) at 37°C before pulsing each well with 1 μCi of [3H]thymidine (Amersham Biosciences). After 16 h, the plates were harvested, and the samples were analyzed by liquid scintillation in a Microcounter (Wallac). The counts from triplicate wells were averaged.
FIGURE 2. NKG2D does not costimulate naive mouse CD8+ T cell proliferation. A, Purified C57BL/6 CD8+ T cells were incubated on plates coated with varying amounts of anti-CD3 mAb, either alone (), with 5 μg/ml anti-CD28 mAb (), 5 μg/ml anti-NKG2D mAb (), or both mAbs (?). [3H]thymidine incorporation was assayed in a 72-h proliferation assay. Data are representative of four experiments. B, Proliferation of naive C57BL/6 CD8+ T cells was assessed as described in A after 5 days of culture. C, Splenocytes from C57BL/6 (dashed line) and RAE-1 (solid lines) transgenic mice were stained with anti-pan-RAE-1 mAb to assess the expression levels of NKG2D ligands on the transgenic splenocytes. The left panel shows RAE-1 expression on whole splenocytes, whereas the right panel shows expression on gated CD11c+ DC. Cells from C57BL/6 and RAE-1 transgenic mice stained with an isotype-matched control Ig were indistinguishable from C57BL/6 cells stained with anti-pan-RAE-1 (not shown). D, OT.1 T lymphocytes were cultured with irradiated C57BL/6 () or RAE-1-transgenic () splenocytes in the presence of the indicated concentration of OVAp. A 72-h proliferation assay was used to compare OT.1 T cell proliferation in the presence or the absence of NKG2D ligands. Data are representative of two experiments. For all panels, error bars represent 1 SD from the mean for triplicate wells.
For proliferation induced by the B16 transfectants, cells were irradiated at 20,000 rad using a cesium irradiator. Day 7 effector OT.1 CD8+ T cells (5 x 104) were incubated with 1 x 105 B16 cells with OVAp peptide at the indicated concentrations. Proliferation was measured as described above by [3H]thymidine incorporation 3 days after culture initiation.
For proliferation of naive CD8+ T cells, 1 x 105 OT.1 lymphocytes were incubated with 1 x 105 irradiated (2000 rad) splenocytes from C57BL/6 or RAE-1 transgenic mice in the presence of the indicated concentration of OVAp in RPMI-T. For proliferation of effector CD8+ T cells, naive OT.1 lymphocytes were incubated at a 1:1 ratio with either C57BL/6 or RAE-1 transgenic splenocytes in the presence of 100 nM OVAp for 3 days in RPMI-T at 37°C. The cells were then split 1:5 in RPMI-T with 200 U/ml human rIL-2 and 4 ng/ml mouse rIL-7. Seven days after primary stimulation, 1 x 105 OT.1 effectors from both primary stimulations were incubated with 1 x 105 irradiated (2000 rad) C57BL/6 or RAE-1 transgenic splenocytes in the presence of the indicated concentrations of OVAp. Proliferation was measured, as described above, by [3H]thymidine incorporation 3 days after culture initiation.
Cytotoxicity assays
Cytotoxicity assays using mouse NK cells were conducted as previously described (10). Redirected lysis assays against P815 targets using NKL cells as effectors were performed as previously described (6), except that the anti-human NKG2D mAb (149810) was used. For cytotoxicity assays using effector OT.1 T cells, parental RMA cells and those expressing RAE-1 were used as targets. Targets were adjusted to 1 x 107 cells/ml in RPMI-T and were labeled with Na2[51Cr]O4 for 2 h at 37°C in the presence or the absence of 1 x 10–6 M OVAp. The cells were washed, and 1 x 104 targets were added to each well of a 96-well plate. Day 7 OT.1 effector T cells were added to triplicate wells at the indicated E:T cell ratios, and the plate was incubated at 37°C for 5 h. Twenty-five microliters of cell-free supernatant was collected from each well, and radioactivity was measured in a Microcounter (Wallac). The percentage of specific 51Cr release was calculated according to the following formula: % specific lysis = 100 x (experimental cpm – spontaneous cpm)/(maximal cpm – spontaneous cpm).
Purification of small, resting PBLs
Buffy coats from healthy donors were purchased from the Stanford Blood Center. PBMCs were prepared according to standard procedures. To obtain small, resting PBLs, the PBMCs were washed in PBS and resuspended in 30% Percoll (Amersham Biosciences)/10% FBS in PBS. The cells were then overlaid on top of 40% Percoll/10% FBS in PBS, and the tubes were spun for 35 min at 2000 rpm (Sorvall RT6000B centrifuge) at 10°C with no brake. The pellet, which contained the small, resting PBLs (14), was harvested and washed, and the cells were resuspended in RPMI-T.
Human CD8+ T cell proliferation assays
To assess mAb-induced CD8+ T cell proliferation, in each well of a 96-well plate, 1 x 105 small, resting PBLs were incubated in RPMI-T with 1 x 104 irradiated (8500 rad) P815 cells in the presence of 100 ng/ml anti-human CD3 (Leu 4; BD Biosciences) and 5 μg/ml of a combination of two of the following mAbs as indicated: anti-human CD28 (L293; BD Biosciences), anti-human NKG2D (149810; R&D Systems), or anti-CD56 (DX32; used as a control mAb). The control mAb was used to ensure that the FcRs would be equally loaded with each mAb under the different conditions. The plates were incubated at 37°C for 48 h, then each well was pulsed with 1 μCi of [3H]thymidine (Amersham Biosciences) for 16 h. Plates were harvested, and samples were analyzed by liquid scintillation. Counts from triplicate wells were averaged.
For proliferation induced by the panel of P815 transfectants, 1 x 105 small, resting PBLs were incubated in RPMI-T with 1 x 104 irradiated P815 or transfectants/well of a 96-well plate in the presence of 150 ng/ml anti-CD3 (Leu 4). At 48 h, the plates were pulsed with [3H]thymidine for 16 h, harvested, and analyzed. For analysis of proliferation using CFSE (Molecular Probes), small, resting PBLs were isolated, washed, and resuspended in 4 μM CFSE in PBS at a concentration of 1 x 107 cells/ml for 8 min at room temperature. The CFSE was then quenched with an equal volume of FBS for 1 min. The cells were washed twice in RPMI-T, then 1.5 x 106 labeled cells were incubated with 2 x 105 irradiated P815 or transfectants in a 24-well plate in the presence of 75 ng/ml anti-CD3 (Leu 4). After 3 days at 37°C, the cells were harvested and stained for flow cytometry with the following mAbs for 30 min at 4°C: allophycocyanin-conjugated anti-CD8 (BD Biosciences) and PE-conjugated anti-CD28 (L293; BD Biosciences). After washing, the cells were resuspended in 1 μg/ml propidium iodide in PBS and 1% FBS, and the samples were analyzed by flow cytometry.
For proliferation of preactivated PBLs, small, resting PBLs were incubated at 37°C for 3 days in RPMI-T with human B7.1+ P815 transfectant and 50 ng/ml anti-CD3 (Leu 4). The cells were then expanded 1:5 into RPMI-T. Seven days after primary stimulation, CD4+ cells were depleted by incubating the cells with 5 μg of anti-CD4 (RPAT4) for 30 min at 4°C. Magnetic goat anti-mouse IgG beads (Qiagen) were used to deplete CD4+ cells as described for purification of mouse naive CD8+ T cells above. Cells (105) were then cultured with 1 x 104 irradiated P815 or transfected cells that had been preincubated with 50 ng/ml anti-CD3 (Leu 4) mAb along with 10 μg/ml isotype-matched control mAb, 5 μg/ml anti-B7.1 (L307)/5 μg/ml isotype-matched control mAb, 5 μg/ml anti-MICA (159209; R&D Systems)/5 μg/ml isotype-matched control mAb, or 5 μg/ml anti-MICA (159209)/5 μg/ml anti-B7.1 (L307) as indicated. Proliferation was measured as described for small, resting PBLs.
IFN- detection
Small, resting human PBLs were cultured with irradiated P815 or transfectants and anti-CD3 (Leu 4) as described above. Three and 7 days after culture initiation, cell-free supernatants were collected. A human IFN- ELISA kit (eBioscience) was used to quantify the concentration of cytokine.
Results
NKG2D expression on mouse CD8+ T cells after TCR stimulation
To determine when NKG2D could potentially contribute to primary CD8+ T cell stimulation, we examined the kinetics of NKG2D up-regulation after TCR engagement. CD8+ T cells were purified to >90% purity from spleens of 4- to 6-wk-old C57BL/6 mice. These cells were cultured on plate-bound anti-CD3, anti-CD3 plus anti-CD28, anti-CD3 plus anti-NKG2D, or all three mAbs, and NKG2D levels were monitored over time by flow cytometry. CD8+ T cells stimulated by anti-CD3 plus anti-CD28 began to up-regulate NKG2D as early as 2 days after stimulation, and the level of NKG2D increased over 5 days (Fig. 1A). The presence of anti-CD28 and/or anti-NKG2D mAbs did not affect the kinetics or extent of NKG2D up-regulation induced by anti-CD3 (data not shown). We also used OT.1 TCR-transgenic CD8+ T cells to examine the kinetics of NKG2D up-regulation after TCR stimulation via cognate peptide:MHC interactions. CD8+ T cells were purified to >90% purity from the spleen of a 4- to 6-wk-old OT-1 mouse. Mouse B16 melanoma cells were irradiated and pulsed with 1 nM OVAp (SIINFEKL), the cognate peptide for OT.1 TCR. These cells were cocultured with OT.1 CD8+ T cells, and the kinetics of NKG2D expression were monitored. Significant NKG2D expression was detected as early as 1 day after coculture, and NKG2D levels continued to rise over the first 3 days (Fig. 1B). These results demonstrated a detectable increase in cell surface NKG2D within 1–2 days after TCR stimulation.
FIGURE 1. Kinetics of NKG2D up-regulation on mouse CD8+ T cells after TCR signaling. A, Purified CD8+ T cells from C57BL/6 mice were cultured on anti-CD3 and anti-CD28 mAb-coated plates. At the indicated times, cells were stained with mAbs to CD8 and NKG2D. Histograms depict NKG2D levels relative to an isotype-matched Ig control on gated CD8+ cells. Data are representative of four experiments. B, Purified OT.1 CD8+ splenic T cells were cocultured with irradiated B16 cells pulsed with 1 nM OVAp. At the indicated times, cells were analyzed for the expression of CD8 and NKG2D. Histograms depict NKG2D levels relative to an isotype-matched Ig control on gated CD8+ cells. Data are representative of three experiments.
NKG2D engagement fails to costimulate proliferation of mouse naive CD8+ T cells
The potential of NKG2D to costimulate naive mouse CD8+ T cells has not been reported. Because NKG2D is expressed at the cell surface within 1–2 days after TCR engagement, it could potentially contribute to proliferation of T cells during a primary response. To assess this possibility, we performed a 3-day proliferation assay using purified splenic CD8+ T cells from C57BL/6 mice. The cells were stimulated with plate-bound anti-CD3, anti-CD3 plus anti-CD28, anti-CD3 plus anti-NKG2D, or all three mAbs. Anti-CD28 provided significant costimulation of the anti-CD3-induced proliferation. In contrast, anti-NKG2D failed to costimulate this primary CD8+ T cell response. Furthermore, anti-NKG2D did not synergize with anti-CD28 (Fig. 2A). These results are representative of four independent experiments.
Because NKG2D is not detectable at the cell surface until 48 h after TCR stimulation, the time at which the cells were pulsed with [3H]thymidine in the above experiments, it was possible that NKG2D would only contribute to CD8+ T cell proliferation measured at a later time. Therefore, we also measured mAb-induced proliferation 5 days after stimulation of purified C57BL/6 CD8+ T cells. As shown in Fig. 2B, the 5-day proliferation assay produced the same result as the 3-day assays. Anti-CD3 mAb alone induced minimal proliferation, whereas anti-CD28 mAb significantly costimulated the response. Anti-NKG2D mAb did not costimulate anti-CD3-induced proliferation, nor did it synergize with anti-CD28. These results are representative of two independent experiments.
The anti-NKG2D mAb used for these studies has been shown to have agonist function, in that plate-bound anti-NKG2D mAb induces cytokine production by mouse NK cells and can trigger NK cell-mediated cytolytic responses against FcR-bearing targets in an mAb-redirected cytotoxicity assay (15). However, it is possible that this anti-NKG2D mAb does not have agonist activity for T cells. Therefore, additional studies were conducted using APC expressing RAE-1, a natural high affinity ligand for mouse NKG2D. This was accomplished using splenic APC cells from mice bearing a RAE-1 transgene driven by a -actin promoter to give ubiquitous expression in all tissues. To confirm the presence of NKG2D ligands in the RAE-1 transgenic mice, splenocytes were analyzed for RAE-1 expression by flow cytometry. Transgenic splenocytes expressed high levels of RAE-1 in comparison with C57BL/6 splenocytes (Fig. 2C, left panel). Furthermore, CD11c+ dendritic cells, which are potent APCs, from the transgenic mice expressed high levels of RAE-1 in comparison with C57BL/6 dendritic cells that do not express RAE-1 (Fig. 2C, right panel).
Splenocytes from C57BL/6 or RAE-1 transgenic mice were irradiated and pulsed with increasing concentrations of OVAp. OT.1 T lymphocytes were cocultured with these splenocytes, and proliferation was measured 3 days later. Consistent with our previous results using anti-NKG2D mAb, OT.1 T cell proliferation was not enhanced by the presence of the RAE-1 on the APC (Fig. 2D). It should be noted that endogenous CD28 ligands were likely able to provide costimulation in this experiment; therefore, this result indicates that NKG2D engagement with ligand cannot enhance CD28-mediated costimulation. Taken together, the above results indicate that NKG2D does not costimulate a primary mouse CD8+ T cell response despite rapid up-regulation of cell surface NKG2D after TCR stimulation and availability of NKG2D ligands.
NKG2D engagement fails to costimulate proliferation of mouse effector CD8+ T cells
It has previously been reported that NKG2D costimulates preactivated effector mouse CD8+ T cells (5). We sought to examine this using OT.1 T cells 7 days after primary activation with OVAp. Day 7 OT.1 effectors were restimulated with plate-bound anti-CD3, anti-CD3 plus anti-CD28, anti-CD3 plus anti-NKG2D, or all three mAbs, and proliferation was measured 3 days later. Anti-CD28 costimulated the anti-CD3-induced proliferation, but surprisingly, anti-NKG2D did not (Fig. 3A). These results are representative of three independent experiments. We also observed the same results when we cocultured day 8 OT.1 effectors with irradiated C57BL/6 splenocytes pulsed with increasing concentrations of OVAp in the context of plate-bound anti-CD28, anti-NKG2D, or anti-CD28 plus anti-NKG2D. Anti-CD28 costimulated proliferation, whereas anti-NKG2D did not, and anti-CD28 plus anti-NKG2D did not synergize (data not shown). NKG2D was expressed at high levels on all of the day 7 or 8 preactivated CD8+ T cells (data not shown); therefore, failure to respond could not be explained by the absence of the receptor.
FIGURE 3. NKG2D does not costimulate effector mouse CD8+ T cell proliferation. A, Day 7 preactivated OT.1 effector CD8+ T cells were cultured on plates coated with varying amounts of anti-CD3 mAb, either alone (), with 5 μg/ml anti-CD28 mAb (), 5 μg/ml anti-NKG2D mAb (), or both mAbs (?) and were measured for [3H]thymidine incorporation 72 h later. Data are representative of two experiments. B, B16 transfectants were stained with mAbs to RAE-1 and B7.1 to confirm the expression of transfected molecules. C, Proliferation of day 7 OT.1 effector CD8+T cells was assayed after 72 h of coculture with irradiated B16 cells () or B16 transfectants expressing B7.1 (), RAE-1 (), or both (?) in the presence of various amounts of OVAp. Data are representative of four experiments. D, B16 () and RAE-1+ B16 () cells were used as targets in a cytotoxicity assay to verify that the RAE-1 transfectant was capable of stimulating the NKG2D receptor. IL-2-activated C57BL/6 NK cells were used as effectors at the indicated E:T cell ratios. E, Day 7 preactivated OT.1 CD8+ T cell effectors that were generated with OVAp in the presence of either C57BL/6 or RAE-1 transgenic splenocytes were stained for the expression of CD8 and NKG2D. Histograms depict NKG2D levels (solid lines) vs an isotype-matched Ig control (dashed lines) on gated CD8+ cells. F, Day 7 preactivated OT.1 CD8+ T cell effectors that were generated as described in E, were cultured for a secondary stimulation with C57BL/6 or RAE-1 transgenic splenocytes pulsed with the indicated concentration of OVAp. Primary stimulation on C57BL/6, followed by secondary on C57BL/6 splenocytes (?), primary on C57BL/6 and secondary on RAE-1 (), primary on RAE-1 and secondary on C57BL/6 (), and primary on RAE-1 and secondary on RAE-1 (?) are represented. [3H]thymidine incorporation was measured 72 h later. Data are representative of two experiments. For proliferation assays, error bars represent 1 SD from the mean for triplicate wells.
Similar studies were conducted with cells expressing natural ligands for CD28 and NKG2D rather than mAbs against these receptors. For these experiments, we used B16 melanoma cells transfected with B7.1 (CD80) and/or RAE-1. Expression of high levels of B7.1 and/or RAE-1 was confirmed by flow cytometry (Fig. 3B). Day 7 OT.1 effectors, generated by incubating OT.1 splenocytes in the presence of OVAp, were cocultured with irradiated B16 transfectants pulsed with various concentrations of OVAp. Expression of B7.1 yielded markedly increased proliferation compared with parental B16 cells (Fig. 3C). In contrast, expression of RAE-1 did not costimulate CD8+ T cell effectors. Furthermore, B7.1 and RAE-1 were not synergistic (Fig. 3C). These results are representative of three independent experiments.
To confirm that NKG2D ligands expressed on the B16 panel were capable of stimulating the NKG2D receptor, B16 cells were compared with RAE-1 transfectants as targets in an NK cytotoxicity assay. As shown in Fig. 3D, IL-2-activated mouse NK cells lysed RAE-1-expressing B16 cells at higher levels than parental B16 cells. Therefore, the RAE-1 B16 transfectant is capable of inducing NKG2D signaling, so the lack of NKG2D-mediated CD8+ T cell costimulation cannot be explained by an inability of the RAE-1 B16 transfectant to stimulate NKG2D.
Because the above results are inconsistent with prior studies (5), we sought to confirm them in an additional system. To activate OT.1 T lymphocytes, we cocultured them with splenocytes from either C57BL/6 or RAE-1 transgenic mice in the presence of OVAp. Day 7 T cell effectors generated from these stimuli expressed high levels of NKG2D, as assessed by flow cytometry (Fig. 3E). Therefore, the presence of an NKG2D ligand during primary stimulation did not result in significant down-modulation of NKG2D. Day 7 OT.1 effectors were then cocultured with either C57BL/6 or RAE-1 transgenic splenocytes in the presence of increasing amounts of OVAp, and proliferation was assayed 3 days after this secondary stimulation. Regardless of the presence of the NKG2D ligand on the APC during primary or secondary stimulation (Fig. 2C), OT.1 CD8+ T cells proliferated equivalently (Fig. 3F). These results are representative of two independent experiments. This assay indicates that NKG2D cannot enhance CD28-mediated costimulation. Taken together, the above data indicate that the expression of NKG2D on CD8+ effector T cells and NKG2D ligands on APCs is not sufficient to result in costimulation of primary, resting mouse CD8+ T cells, or activated, effector CD8+ T cells.
NKG2D engagement with cognate ligand fails to enhance cytotoxicity of mouse CTL
Although NKG2D did not costimulate the proliferation of effector CD8+ T cells, it might augment effector functions, such as cytotoxicity. To test this possibility, we examined the ability of OT.1 effector CD8+T cells to lyse peptide-pulsed parental RMA tumor cells vs RMA transfectants expressing RAE-1. The expression of high levels of RAE-1 on the transfectant was confirmed by flow cytometry at the time of the experiment (data not shown). OT.1 CTL specifically lysed both parental RMA cells and the RAE-1 transfectants equivalently in the presence, but not in the absence, of 10–6 M OVAp (Fig. 4A). Furthermore, equivalent lysis was observed over a range of E:T cell ratios in the presence of 10–7 and 10–8 M peptide (data not shown). This experiment is representative of two independent assays. To confirm that the RMA RAE-1 transfectant was capable of inducing NKG2D-induced cytotoxicity, IL-2-activated NK cells were used as effectors in a cytotoxicity assay with RMA or RMA-RAE-1 targets. NK cells efficiently lysed the targets expressing RAE-1, and this lysis was inhibited by a blocking mAb against NKG2D (Fig. 4B). These results indicate that the presence of an NKG2D ligand on a target cell is insufficient to augment cytotoxicity of NKG2D-expressing mouse CD8+ effector T cells.
FIGURE 4. NKG2D does not costimulate cytolytic activity of effector mouse CD8+ T cells. A, Day 7 preactivated OT.1 CD8+ T cell effectors were incubated with 51Cr labeled RMA cells (), RAE-1+ RMA (?), RMA pulsed with 10–6M OVAp (), or RAE-1+ RMA pulsed with OVAp () at several E:T cell ratios. A 51Cr release assay was used to measure the specific lysis of the two targets. Data are representative of two experiments. B, IL-2-activated C57BL/6 NK cells were used as effectors in a cytotoxicity assay to confirm that the RMA-RAE-1 transfectant was capable of inducing signaling through NKG2D. NK cells were incubated with 51Cr-labeled RMA (X) or RMA-RAE-1 cells in the presence of a blocking anti-NKG2D mAb () or an isotype-matched Ig control (). Error bars represent 1 SD from the mean for triplicate wells.
NKG2D engagement fails to costimulate proliferation of small, resting human PBLs
Human peripheral blood CD8+ T cells constitutively express NKG2D (6), and it has been reported that NKG2D engagement can costimulate proliferation and effector functions (9). We sought to confirm and extend these findings. First, we enriched for small, resting PBLs, and examined NKG2D expression levels by flow cytometry. As reported previously (6), nearly all human CD8+ T cells express high levels of NKG2D (Fig. 5A). We next assessed the ability of mAbs against CD28 and/or NKG2D to costimulate anti-CD3 mAb-induced proliferation. Using an assay we previously described to demonstrate anti-CD28 mAb costimulation (14), irradiated P815 cells, a mouse mastocytoma line expressing FcRs, were incubated with small, resting human PBLs in the presence of a low concentration of anti-CD3 alone or with anti-CD28 and/or anti-NKG2D mAbs. Proliferation was assessed 3 days later. P815 cells coated only with anti-CD3 mAb induced little proliferation, whereas those also coated with anti-CD28 mAb induced a high level of proliferation (Fig. 5B), consistent with previous findings (14). In contrast, anti-NKG2D mAb did not increase the low level of proliferation seen with anti-CD3 alone, and anti-NKG2D did not augment anti-CD28 mAb costimulation (Fig. 5B). These assays indicated that CD28, but not NKG2D, was a potent costimulatory molecule for small, resting PBLs. The anti-NKG2D mAb used in these studies has agonist activity; it induces human NK cell-mediated lysis of anti-NKG2D mAb-coated P815 target cells (Fig. 5C) and costimulates cytokine production by human NK cells activated through an activating killer cell Ig-like receptor (16).
FIGURE 5. NKG2D does not costimulate the proliferation of fresh small, resting human peripheral blood CD8+ PBLs. A, Small, resting human PBLs were isolated, and the expression of NKG2D on CD8+ T cells was detected by flow cytometry. Numbers in the quadrants display the percentage of cells. B, Small, resting PBLs were cocultured with P815 cells in the presence of the indicated mAbs to costimulate suboptimal (100 ng/ml) anti-CD3 mAb-induced proliferation through CD28 and/or NKG2D. In addition to the samples containing P815, PBL, and anti-CD3 (), the following controls are included: P815 and anti-CD3 (), P815 and PBL (), and PBL and anti-CD3 (). Proliferation was determined 72 h later by measuring [3H]thymidine incorporation. The control mAb against CD56 was used to maintain a constant concentration of mAb to ensure equal FcR loading in all samples. Data are representative of two experiments. C, The human NK line, NKL, was cocultured with 51Cr-labeled P815 cells in the presence of anti-NKG2D mAb () or an isotype-matched Ig control (). Redirected mAb-dependent cytotoxicity was determined in a 4-h cytotoxicity assay. D, Small, resting PBLs were cocultured with irradiated P815 transfectants that expressed ligands for CD28 and/or NKG2D and a suboptimal concentration of anti-CD3 mAb. Proliferation assay and controls were as described in B. Data are representative of six experiments with four blood donors. E, Top, The P815 transfectants were stained with anti-human B7.1, and the cells were examined by flow cytometry. MICA was expressed using an IRES-GFP encoding construct (bottom). Confirmation that GFP is representative of MICA expression was achieved by staining with a human NKG2D-Ig fusion protein. F, NKL cells were used as effectors in a 51Cr release assay to confirm that the P815-MICA*0019 transfectant was capable of inducing signaling through NKG2D. NKL cells were incubated with 51Cr-labeled P815 cells () or P815-MICA*0019 transfectants (), and the percent specific lysis was determined after 4 h. G, Small, resting peripheral blood T cells were labeled with CFSE and cocultured with the P815 panel in the presence of a low concentration (75 ng/ml) of anti-CD3 mAb, as described in D. After 3 d, the cells were labeled with allophycocyanin-conjugated anti-CD8 and PE-conjugated anti-CD28 mAb and analyzed by flow cytometry. The dot plots display CFSE vs CD28 for CD8+ gated cells. Proliferation was analyzed by CFSE dilution. Numbers in the quadrants indicate the percentage of cells. Data are representative of two experiments. In proliferation assays, error bars represent 1 SD from the mean for triplicate wells.
We next investigated whether cognate ligands for CD28 and/or NKG2D could costimulate small, resting PBLs. We cocultured small, resting PBLs with P815 cells or transfectants expressing human B7.1 and/or MICA*0019, natural ligands for CD28 and NKG2D, respectively, in the presence of a suboptimal amount of anti-CD3 mAb. Consistent with previous findings (14), the B7.1+ P815 transfectant induced significantly higher levels of proliferation than the parental P815 cells. However, the MICA*0019 transfectant did not costimulate anti-CD3 mAb-induced proliferation (Fig. 5D). These data are representative of six independent experiments with four donors. The P815 transfectants were examined by flow cytometry to confirm high expression of B7.1 and MICA (Fig. 5E). To confirm that the MICA*0019 transfectant was capable of inducing signaling through NKG2D, P815 cells and the P815 MICA*0019 transfectants were compared as targets in a cytotoxicity assay using human NK cells (Fig. 5F). The MICA*0019 transfectant was killed much more efficiently than parental P815 cells.
NKG2D engagement fails to costimulate proliferation of peripheral blood CD28–CD8+ T cells
Groh et al. (9) previously reported that NKG2D augmented proliferation and effector functions of human CD28– CD8+ peripheral blood T cells. Thus, it was possible that [3H]thymidine incorporation, which measures proliferation of all cultured T cells, was not sensitive enough to discern costimulation of CD28– CD8+ cells, which represent only 20% of the CD8+ PBLs in our cultures. Therefore, we repeated the proliferation assay shown in Fig. 5D, except we labeled the human small, resting peripheral blood T cells with CFSE before coincubation with anti-CD3 mAb and the irradiated P815 transfectants. Three days later, cells were stained for CD8 and CD28, and proliferation of CD8+CD28+ and CD28–CD8+ T cell populations was examined by flow cytometry. The B7.1+ P815 transfectant costimulated proliferation of CD8+CD28+ T cells, whereas the MICA*0019 transfectant did not. No synergy was observed between B7.1 and MICA*0019. In contrast, CD28–CD8+ T cells did not proliferate significantly to any of the P815 transfectants coated with anti-CD3 mAb, consistent with previous reports that these cells are hyporesponsive to anti-CD3 stimulation (17) (Fig. 5G). These findings are representative of two independent experiments. The CFSE results are consistent with the [3H]thymidine results, indicating that CD28 is able to costimulate the proliferation of human small, resting CD8+ PBLs, whereas NKG2D does not have an apparent costimulatory function.
NKG2D engagement fails to costimulate proliferation of preactivated human peripheral blood CD8+ T cells; however, it augments CD28 costimulation
Although it was reported previously (9) that anti-NKG2D mAb could costimulate the proliferation of freshly isolated CD8+ T cells from human peripheral blood, we considered the possibility that some of the T cells in the donor used for these studies might have been preactivated due to an unknown infection. To test the possibility that NKG2D costimulates preactivated human CD8+ T cells, we stimulated small, resting PBLs with anti-CD3 mAb-coated B7.1+ P815 transfectants, using the conditions described in Fig. 5D. Seven days after primary stimulation, CD4+ cells were depleted, and CD8+ T cells (>90% pure), all of which expressed high levels of NKG2D (not shown), were cocultured with P815 transfectants using suboptimal concentrations of anti-CD3. The B7.1 transfectant, but not the MICA*0019 transfectant, significantly costimulated activated human CD8+ T cell proliferation (Fig. 6). Interestingly, NKG2D synergized with CD28, such that the greatest proliferation was observed in response to the MICA*0019/B7.1 double transfectant. This synergy was reduced to proliferation levels seen with the B7.1 transfectant using an anti-MICA mAb. Furthermore, an anti-B7.1 mAb completely blocked costimulation induced by the B7.1 or MICA*0019/B7.1 transfectants, indicating that CD28 costimulation is required for basal costimulation (Fig. 6). These results are representative of assays using peripheral blood from three different donors in two independent experiments. The above results indicate that CD28, but not NKG2D, is able to costimulate the proliferation of small, resting PBLs and preactivated CD8+ PBLs. However, NKG2D signaling synergizes with CD28 to enhance the proliferation of preactivated human CD8+ T cells.
FIGURE 6. NKG2D does not costimulate preactivated human peripheral blood CD8+ T cells; however, NKG2D synergizes with CD28. Small, resting PBLs were cocultured with irradiated B7.1+ P815 cells in the presence of 50 ng/ml anti-CD3 mAb. Seven days later, the activated cells were tested in a secondary stimulation proliferation assay by coculture with irradiated P815 transfectants, as described in Fig. 5D, in the presence of 50 ng/ml anti-CD3 mAb. The cells were cultured in the presence of control mAbs (), anti-MICA mAb (), anti-B7.1 mAb (), or both anti-MICA and anti-B7.1 mAbs () to determine the specificity of costimulation. Proliferation was measured by [3H]-thymidine uptake at 72 h. These data are representative of assays using peripheral blood from three different donors in two experiments.
]NKG2D engagement fails to enhance IFN- production by primary human peripheral blood CD8+ T cells
It is possible that NKG2D costimulates effector functions of human CD8+ PBLs rather than proliferation. To address this possibility, IFN- production was measured by ELISA 3 and 7 days after coculture of small, resting PBLs with the P815 transfectants and anti-CD3 mAb. The suboptimal level of anti-CD3 used was insufficient to induce IFN- production in response to parental P815 cells (Fig. 7). IFN- was not produced after stimulation with MICA*0019 or MICB*002 transfectants. However, IFN- was produced after coculture with the B7.1 or B7.1/MICA*0019 transfectants on both days 3 and 7 (Fig. 7). These data indicate that CD28 is able to costimulate the production of IFN-, whereas NKG2D is not. As shown in Fig. 6, NKG2D synergized with CD28 to elicit more IFN- release after stimulation with the B7.1/MICA*0019 double transfectant in comparison with the B7.1 transfectant (Fig. 7). These data are representative of assays using peripheral blood from three different donors in two independent experiments. Therefore, although NKG2D can synergize with CD28 to costimulate IFN- production from small, resting human PBLs and proliferation of preactivated human CD8+T cells, we were unable to detect a role for NKG2D alone in costimulation of either proliferation or effector responses in primary human CD8+ T cells.
FIGURE 7. NKG2D does not costimulate IFN- secretion on small, resting PBLs, although it synergizes with CD28. Small, resting PBLs were cocultured with irradiated P815 transfectants and anti-CD3 as described in Fig. 5D. Three () and 7 () days after culture initiation, supernatants were analyzed for IFN- by ELISA. Error bars represent 1 SD for the mean of triplicate wells. These data are representative of assays using peripheral blood from three different donors in two experiments.
Discussion
The role of NKG2D as a costimulatory molecule on certain murine and human CD8+ T cells has been documented in several studies (5, 9, 18, 19). However, we report in this study that engagement of NKG2D alone is not sufficient to costimulate primary mouse or human CD8+ T cell activation. In agreement with previous reports (5), we found that NKG2D is rapidly induced on activated mouse CD8+ T cells and is constitutively expressed on all human peripheral blood CD8+ T cells (6). However, we were unable to demonstrate a role for NKG2D in costimulating proliferation, cytolytic activity or IFN- production of naive or effector splenic mouse CD8+ T cells or small, resting or preactivated human peripheral blood CD8+ T cells. In all these assays, CD28 costimulation was observed.
It has previously been reported that NKG2D can costimulate the proliferation of effector mouse CD8+ T cells equivalently to CD28 (5, 18). In these experiments, an LCMV-specific T cell line was generated and maintained via weekly restimulation with peptide in the presence of cytokines. To assay for costimulation, T cells were labeled with CFSE 7 days after restimulation and cultured on mAb-coated plates for 4 days before analysis by flow cytometry. In this context, NKG2D and CD28 were both shown to equivalently costimulate cells that had undergone more than one or more than three cell divisions. Using OT.1 CD8+ T cell effectors, we were unable to demonstrate NKG2D-mediated costimulation of proliferation, although we did observe potent CD28-mediated costimulation. Our effector cells were generated by a single in vitro exposure to OVAp, followed 4 days later by transfer to IL-2- and IL-7-containing media. Day 7 effectors were then used in proliferation assays. Using mAbs or cognate ligands, CD28 engagement provided significant costimulation of proliferation, whereas NKG2D did not. Because it was reported previously that the major effect of NKG2D costimulation was on cells that had undergone more than three cell divisions (5), we were concerned that this difference might be obscured by the measurement of all cells undergoing division in a [3H]thymidine assay. Therefore, we also performed proliferation assays with CFSE-labeled OT.1 T cell effectors. Nonetheless, we observed the same results; CD28 costimulated proliferation, whereas NKG2D did not (data not shown). Furthermore, proliferation assays were conducted comparing our CX5 anti-NKG2D mAb with the MI6 mAb, previously described by Jamieson et al. (5). We observed no evidence for NKG2D costimulation with either mAb (unpublished observations). Most significantly, cognate NKG2D ligands expressed by B16 transfectants or RAE-1 transgenic splenocytes were unable to costimulate effector CD8+ T cell proliferation. Although we cannot explain the discrepancy between our data and the previous report, it is possible that T cells in the LCMV-specific T cell line had been exposed to cytokines during infection that potentiated the costimulatory capability of NKG2D. Alternatively, weekly restimulation of these cells in the presence of cytokines may have selected out a small subpopulation on which NKG2D was costimulatory or may have potentiated NKG2D signaling.
Consistent with the hypothesis that a small subpopulation of mouse effector CD8+ T cells might be responsible for the previously reported costimulatory properties of NKG2D, it has recently been reported that extrathymically derived IL-2- or IL-15-activated CD8+CD44high cells can be stimulated via NKG2D engagement (20). These cells represent a small subpopulation that can be activated by IL-2 in the absence of TCR stimulation. After IL-2 stimulation, they up-regulate NK cell markers in addition to NKG2D, and they express DAP12. Furthermore, in male H-Y TCR-transgenic mice, in which the majority of thymocytes are negatively selected due to expression of the male Ag, lymphocytes of similar phenotype and functionality can be found in the periphery (21). These cells also express DAP12, and their cytolytic function can be activated directly by NKG2D. Interestingly, the cytolytic activity of CD8+ T cell effectors generated from the periphery of female H-Y transgenic mice were not costimulated by NKG2D engagement (21), consistent with our results. These data, in conjunction with our findings, suggest that although activated conventional primary peripheral mouse CD8+ T cells express NKG2D, they are not costimulated by NKG2D engagement. Instead, a small subpopulation of mouse CD8+ T cells is apparently able to proliferate in response to inflammatory cytokines, such as IL-2 and IL-15, leading to the acquisition of NKG2D activity. This activity might be due to the expression of DAP12 in these cells (20, 21), because it has previously been demonstrated that activated T cells can be directly stimulated through NKG2D, if DAP12 expression is enforced as a transgene in mouse T cells (18).
NKG2D-mediated costimulation of human CD8+CD28–T cells was also previously reported (9). These experiments, which were conducted predominantly with long term human CMV-specific T cell clones, demonstrated that CD8+CD28– cells lysed human CMV-infected targets in an MHC-restricted manner, and this lysis could be inhibited by blocking NKG2D. Furthermore, NKG2D engagement via ligand or mAb cross-linking costimulated proliferation, cytolysis, and cytokine secretion in a peptide:MHC-dependent manner. In addition, the authors demonstrated that anti-NKG2D mAb could costimulate anti-CD3 mAb-mediated proliferation of freshly isolated CD8+CD28– PBLs (9). Our experiments with a long term polyclonal human CD8+ T cell line indeed demonstrated NKG2D-mediated costimulation of anti-CD3-induced cytokine production (unpublished observations). However, we were unable to show NKG2D costimulation using small, resting PBLs isolated from the blood of several healthy donors. With all donors tested, neither mAb cross-linking nor ligand engagement of NKG2D was able to costimulate proliferation, although CD28 engagement was consistently costimulatory.
We were concerned that [3H]thymidine incorporation assays might not detect NKG2D costimulation if only CD28–CD8+ T cells responded to NKG2D signals. Therefore, we used CFSE labeling of small, resting PBLs to distinguish between the proliferation of CD8+CD28– cells and CD8+CD28+ T cells. We found that only CD8+CD28+ cells proliferated, and these cells were costimulated only by CD28, not NKG2D. The lack of proliferation of CD8+CD28–T cells stimulated with anti-CD3 mAb has been previously reported (17). Collectively, we cannot explain the discrepancy between our data and the previous results reported with freshly isolated, human peripheral blood CD8+CD28– T cells. However, the T cell clones used in the previous reports may have represented a small subpopulation in which NKG2D is costimulatory. Alternatively, long term culture of CD8+T cells in IL-2 may render them permissive for NKG2D costimulation, consistent with our unpublished observations.
Recently, it has been reported that culturing human CD8+ T cells in IL-15 or high concentrations of IL-2 can induce TCR-independent lymphokine-activated killer (LAK) activity that is mediated by NKG2D stimulation (19, 22, 23). However, only cytolytic activity was directly induced by NKG2D signaling in these LAK cells, whereas proliferation and cytokine secretion required TCR stimulation (19). Interestingly, Jabri and colleagues have demonstrated NKG2D-mediated costimulation of proliferation and IFN- production of freshly isolated intraepithelial T lymphocytes, and this costimulation was enhanced by the inclusion of IL-15 (19). Furthermore, intraepithelial CD8+ T cells from celiac patients, where endogenous IL-15 concentrations are high, demonstrated NKG2D-mediated LAK activity (22). Perhaps human T cells isolated from inflamed tissues, but not those isolated from blood, are able to signal through NKG2D. Consistent with the idea that T cells isolated from abnormal tissues may preferentially use NKG2D as a stimulatory molecule, several reports have indicated that tumor-infiltrating lymphocytes from tumors of diverse origin contain T cells that are responsive to NKG2D stimuli (24, 25, 26). However, all of these findings suggest that at least some of the tumor-infiltrating T cells are responsive to NKG2D engagement in an MHC-independent manner. Thus, human T cell NKG2D reactivity against tumors may predominantly be a LAK phenomenon, rather than TCR-dependent Ag-specific T cell recognition.
Although we were unable to demonstrate NKG2D-mediated costimulation of human CD8+ T cells, we did observe synergy between CD28 and NKG2D in two assays. NKG2D synergized with CD28 to induce IFN- secretion from small, resting human PBLs, and NKG2D synergized with CD28 to costimulate proliferation of preactivated human CD8+ T cells. Anti-B7.1 completely abrogated proliferation induced by the B7.1/MICA double transfectant, indicating that CD28 signaling is required for minimal costimulation. These results indicate that NKG2D can contribute to both preactivated human CD8+ T cell proliferation and resting CD8+ cytokine secretion, but only in conjunction with CD28 signaling.
The possibility that NKG2D may costimulate CD8+ T cells has generated much excitement in the fields of tumor biology, autoimmunity, and infectious disease (27, 28). NKG2D may be a potential candidate for therapeutic targeting to block autoimmune T cells or enhance T cell responses to tumors or pathogen-infected tissues. Indeed, we have recently reported that anti-NKG2D treatment can block diabetes in NOD mice, possibly due to blockade of CD8+ T cell activation (29). However, in this report we demonstrate that mAb- or ligand-mediated engagement of NKG2D alone is not sufficient to costimulate proliferation or effector functions of primary, resting, or activated human or mouse CD8+ T cells. Thus, it will be important to identify the subpopulations and cytokine conditions that render CD8+ T cells permissive for NKG2D-mediated signaling.
Acknowledgments
We thank Amanda Jamieson and David Raulet for providing MI-6 mAb for comparison with CX5. We thank Thomas Pertel for excellent technical assistance, Jason Dietrich for generation of the RAE-1 transgenic mice, and Nigel Killeen for helpful discussions.
Footnotes
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.
1 This work was supported by National Institutes of Health Grants CA89189 and CA95137. L.R.E. is supported by the Damon Runyon Cancer Research Foundation (DRG-1759-03), K.O. is supported by a Human Frontier Science Program Long-Term Fellowship, J.A.H. is an Irvington Foundation Fellow, and L.L.L. is an American Cancer Society Research Professor.
2 Current address: Department of Pathology, Stanford University, B257 Beckman Center, Stanford, CA 94305-5323.
3 Current address: Department of Experimental Medicine and Pathology, University of Rome La Sapienza, Viale Regina Elena 324 00161 Rome, Italy.
4 Address correspondence and reprint requests to Dr. Lewis L. Lanier, Department of Microbiology and Immunology and The Cancer Research Institute, University of California, 513 Parnassus Avenue, HSE 1001, Box 0414, San Francisco, CA 94143. E-mail address: lanier@itsa.ucsf.edu
5 Abbreviations used in this paper: LCMV, lymphocytic choriomeningitis virus; LAK, lymphokine-activated killer; OVAp, SIINFEKL peptide; MICA, MHC class I chain-related A.
Received for publication September 15, 2004. Accepted for publication November 7, 2004.
References
Janeway, C. A., Jr, R. Medzhitov. 2002. Innate immune recognition. Annu. Rev. Immunol. 20:197.
Ahmed, R., D. Gray. 1996. Immunological memory and protective immunity: understanding their relation. Science 272:54.
Diefenbach, A., A. M. Jamieson, S. D. Liu, N. Shastri, D. H. Raulet. 2000. Ligands for the murine NKG2D receptor: expression by tumor cells and activation of NK cells and macrophages. Nat. Immunol. 1:119.
Wu, J., Y. Song, A. B. Bakker, S. Bauer, T. Spies, L. L. Lanier, J. H. Phillips. 1999. An activating immunoreceptor complex formed by NKG2D and DAP10. Science 285:730.
Jamieson, A. M., A. Diefenbach, C. W. McMahon, N. Xiong, J. R. Carlyle, D. H. Raulet. 2002. The role of the NKG2D immunoreceptor in immune cell activation and natural killing. Immunity 17:19.
Bauer, S., V. Groh, J. Wu, A. Steinle, J. H. Phillips, L. L. Lanier, T. Spies. 1999. Activation of NK cells and T cells by NKG2D, a receptor for stress-inducible MICA. Science 285:727.
Chang, C., J. Dietrich, A. G. Harpur, J. A. Lindquist, A. Haude, Y. W. Loke, A. King, M. Colonna, J. Trowsdale, M. J. Wilson. 1999. Cutting edge: KAP10, a novel transmembrane adapter protein genetically linked to DAP12 but with unique signaling properties. J. Immunol. 163:4651.
Dymecki, S. M.. 1996. Flp recombinase promotes site-specific DNA recombination in embryonic stem cells and transgenic mice. Proc. Natl. Acad. Sci. USA 93:6191.
Townsend, S. E., J. P. Allison. 1993. Tumor rejection after direct costimulation of CD8+ T cells by B7-transfected melanoma cells. Science 259:368.
Kinsella, T. M., G. P. Nolan. 1996. Episomal vectors rapidly and stably produce high-titer recombinant retrovirus. Hum. Gene Ther. 7:1405.
Azuma, M., M. Cayabyab, J. H. Phillips, L. L. Lanier. 1993. Requirements for CD28-dependent T cell-mediated cytotoxicity. J. Immunol. 150:2091.
Zompi, S., J. A. Hamerman, K. Ogasawara, E. Schweighoffer, V. L. Tybulewicz, J. P. Di Santo, L. L. Lanier, F. Colucci. 2003. NKG2D triggers cytotoxicity in mouse NK cells lacking DAP12 or Syk family kinases. Nat. Immunol. 4:565
Zingoni, A., T. Sornasse, B. G. Cocks, Y. Tanaka, A. Santoni, L. L. Lanier. 2004. Cross-talk between activated human NK cells and CD4+ T cells via OX40-OX40L interactions. J. Immunol. :173.
Azuma, M., J. H. Phillips, L. L. Lanier. 1993. CD28– T lymphocytes: antigenic and functional properties. J. Immunol. 150:1147.
Diefenbach, A., E. Tomasello, M. Lucas, A. M. Jamieson, J. K. Hsia, E. Vivier, D. H. Raulet. 2002. Selective associations with signaling proteins determine stimulatory versus costimulatory activity of NKG2D. Nat. Immunol. 3:1142.
Roberts, A. I., L. Lee, E. Schwarz, V. Groh, T. Spies, E. C. Ebert, B. Jabri. 2001. NKG2D receptors induced by IL-15 costimulate CD28-negative effector CTL in the tissue microenvironment. J. Immunol. 167:5527.
Dhanji, S., H. S. Teh. 2003. IL-2-activated CD8+CD44high cells express both adaptive and innate immune system receptors and demonstrate specificity for syngeneic tumor cells. J. Immunol. 171:3442.
Dhanji, S., S. J. Teh, D. Oble, J. J. Priatel, H. S. Teh. 2004. Self-reactive memory-phenotype CD8 T cells exhibit both MHC-restricted and non-MHC restricted cytotoxicity: a role for the t cell receptor and natural killer cell receptors. Blood 104:2116.
Meresse, B., Z. Chen, C. Ciszewski, M. Tretiakova, G. Bhagat, T. N. Krausz, D. H. Raulet, L. L. Lanier, V. Groh, T. Spies, et al 2004. Coordinated induction by IL15 of a TCR-independent NKG2D signaling pathway converts CTL into lymphokine-activated killer cells in celiac disease. Immunity 21:357.
Verneris, M. R., M. Karami, J. Baker, A. Jayaswal, R. S. Negrin. 2004. Role of NKG2D signaling in the cytotoxicity of activated and expanded CD8+ T cells. Blood 103:3065.
Conejo-Garcia, J. R., F. Benencia, M. C. Courreges, P. A. Gimotty, E. Khang, R. J. Buckanovich, K. A. Frauwirth, L. Zhang, D. Katsaros, C. B. Thompson, et al 2004. Ovarian carcinoma expresses the NKG2D ligand Letal and promotes the survival and expansion of CD28– antitumor T cells. Cancer Res. 64:2175.
Girardi, M., D. E. Oppenheim, C. R. Steele, J. M. Lewis, E. Glusac, R. Filler, P. Hobby, B. Sutton, R. E. Tigelaar, A. C. Hayday. 2001. Regulation of cutaneous malignancy by T cells. Science 294:605.
Maccalli, C., D. Pende, C. Castelli, M. C. Mingari, P. F. Robbins, G. Parmiani. 2003. NKG2D engagement of colorectal cancer-specific T cells strengthens TCR-mediated antigen stimulation and elicits TCR independent anti-tumor activity. Eur. J. Immunol. 33:2033.
Raulet, D. H.. 2003. Roles of the NKG2D immunoreceptor and its ligands. Nat. Rev. Immunol. 3:781.
Cerwenka, A., L. L. Lanier. 2001. Natural killer cells, viruses and cancer. Nat. Rev. Immunol. 1:41.
Ogasawara, K., J. A. Hamerman, L. R. Ehrlich, H. Bour-Jordan, P. Santamaria, J. A. Bluestone, L. L. Lanier. 2004. NKG2D blockade prevents autoimmune diabetes in NOD mice. Immunity 20:757.(Lauren I. Richie Ehrlich,)
CD8+ T cells require a signal through a costimulatory receptor in addition to TCR engagement to become activated. The role of CD28 in costimulating T cell activation is well established. NKG2D, a receptor found on NK cells, CD8+ -TCR+ T cells, and -TCR+ T cells, has also been implicated in T cell costimulation. In this study we have evaluated the role of NKG2D in costimulating mouse and human naive and effector CD8+ T cells. Unexpectedly, in contrast to CD28, NKG2D engagement by ligand or mAb is not sufficient to costimulate naive or effector CD8+ T cell responses in conventional T cell populations. While NKG2D did not costimulate CD8+ T cells on its own, it was able to modify CD28-mediated costimulation of human CD8+ T cells under certain contitions. It is, therefore, likely that NKG2D acts as a costimulatory molecule only under restricted conditions or requires additional cofactors.
Introduction
Innate and adaptive immune systems function coordinately to recognize and eliminate pathogens and to prevent pathology upon re-exposure to the same pathogen. The innate immune system recognizes molecular patterns on classes of pathogens to quickly alert the body to infection and contain the pathogen (1). In contrast, the adaptive immune system recognizes unique epitopes on specific pathogens or transformed cells to generate a tailored response to eliminate the pathogen or tumor and to provide for long term immunological memory (2). Given the different roles of innate and adaptive immunity, receptors used by cells of the two systems are frequently specialized for different functions. However, there are some receptors that are shared between cells of the innate and adaptive immune systems, suggesting that they may be important mediators of both types of immunity. One such receptor is NKG2D, which is expressed on mouse and human NK cells, -TCR+ T cells, human CD8+ -TCR+ T cells, and activated mouse CD8+ -TCR+ T cells (3, 4, 5, 6).
NKG2D forms a homodimer that pairs with the adapter protein DAP10 in resting NK cells and CD8+ T cells (4, 6). The only known signaling motif on the DAP10 molecule is a YXXM motif, which enables NKG2D/DAP10 to recruit the p85 subunit of PI3K and Grb2 upon ligand binding (4, 7). This motif is shared by other costimulatory receptors, such as CD28.
CD8+ T cells require two stimuli for activation. Signal 1 is provided through the TCR via recognition of cognate peptide:MHC, and signal 2 is provided through costimulatory molecules (8), the prototype of which is CD28. Because NKG2D shares the YXXM motif with CD28 and is capable of recruiting the p85 subunit of PI3K, it was proposed that it could play a costimulatory role in CD8+ T cells. Indeed, CD8+ T cell costimulation via NKG2D has been reported in both human and mouse models (5, 9). In CMV-specific human T cell clones, NKG2D was reported to costimulate proliferation, cytokine production, and cytolytic activity (9). Similarly, in a lymphocytic choriomeningitis virus (LCMV)5-specific mouse CD8+ T cell line, NKG2D cross-linking was reported to costimulate proliferation. We also characterized the role of NKG2D in costimulation of mouse and human CD8+ T cells. We report in this study that neither stimulation via mAb cross-linking nor NKG2D ligand engagement alone is sufficient to costimulate proliferation or effector functions in primary mouse or human CD8+ T cells, suggesting that NKG2D acts as a costimulatory molecule only under restricted conditions or requires additional cofactors.
Materials and Methods
Abs and flow cytometry
The following mAbs were used for immunofluorescent staining of mouse cells: anti-NKG2D (CX5 or 191004) (10), anti-CD8 (53-6.7), rat IgG2a (B39-4), rat IgG2b (A95-1), anti-pan-RAE-1 (186107), anti-B7.1 (16-10A1), and anti-CD11c (HL3). Biotinylated mAbs were detected with allophycocyanin-conjugated streptavidin (BD Biosciences), and pure mAbs were detected with FITC-conjugated goat anti-rat IgG (Caltag Laboratories). The following were used for staining human cells: anti-NKG2D (149810), anti-CD8 (25T8-5H7; gift from E. Reinherz, Dana Farber Cancer Institute, Boston, MA), anti-CD8 (Leu 2a), anti-CD28 (L293), and anti-CD80 (2D10; eBioscience). Unless otherwise indicated, mAbs were obtained from BD Biosciences, except for mAbs 149810, 191044, and 186107, which were produced in collaboration with Dr. J. P. Houchins (R&D Systems). Cell staining and flow cytometry were performed using standard procedures on a FACSCalibur instrument (BD Biosciences).
Mice
C57BL/6 mice were obtained from Charles River. OT.1 TCR-transgenic mice were provided by Dr. J. Cyster (University of California, San Francisco (UCSF), CA). Experiments were performed according to the guidelines of the UCSF committee on animal research. Mouse RAE-1 cDNA was cloned by RT-PCR using oligonucleotide primers (sense primer, atg gcc aag gca gca gtg acc aa; antisense primer, tca cat tgc aaa tgc aaa tgc aaa taat). The pFRTZ vector (11) (provided by Dr. S. Dymechi, Harvard University, Boston, MA) was modified to remove lacZ and insert the RAE-1 cDNA behind the human -actin promoter. The RAE-1 vector was injected into fertilized (C57BL/6xDBA/2) F2 oocytes by the UCSF Comprehensive Cancer Center Transgenic and Targeted Mutagenesis Shared Resource. RAE-1 transgenic mice were backcrossed for five generations onto the C57BL/6 background.
Cell lines
Mouse B16 melanoma cells (12) were transduced with pMX retroviruses (13) encoding mouse B7.1 and/or RAE-1. P815 cells were transfected by electroporation with human B7.1 in the pBJ1-neo vector and were transduced with MHC class I chain-related A (MICA)*0019 in the pMX-pie retrovirus using standard methods. MICB*002-transfected P815 cells were provided by Dr. K. Soderstrom (Stanford University, Stanford, CA).
Mouse CD8+ T cell purification
For naive mouse CD8+ T cells, lymphocytes were obtained from spleens and lymph nodes by mechanical disruption, and RBC were lysed in ACK buffer (Cambrex BioScience). Cells were resuspended in RPMI-T (RPMI 1640 with 10% heat-inactivated FCS, 2 mM L-glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin, and 50 μM 2-ME) and mixed with anti-CD4 (GK1.5) mAb for 30 min at 4°C. After washing, lymphocytes were incubated for 20 min at 4°C with goat anti-rat IgG and goat anti-mouse IgG magnetic beads (Qiagen), which bind mAb-coated CD4+ cells and B cells, respectively, followed by magnetic depletion of the bead-bound cells. The remaining cells were incubated with PE-conjugated anti-CD8 (53-6.7) for 30 min at 4°C. Anti-PE-conjugated MACS beads (Miltenyi Biotec) were used to positively select mouse CD8+ T cells according to manufacturer’s instructions. Purity was confirmed by flow cytometry.
Generation of mouse OT.1 CD8+ T cell effectors
Splenocytes (3 x 106) from OT.1 mice were incubated at 37°C with 1 nM SIINFEKL peptide (OVAp) in a 24-well plate with 1 ml of RPMI-T. Four days later, the cells were expanded 1:5 into RPMI-T containing 200 U/ml human rIL-2 (National Cancer Institute Biological Resources Branch Pre-Clinical Repository) and 4 ng/ml mouse rIL-7 (R&D Systems). Two days later, the cells were split 1:2 into the same medium. At 7–8 days, the cells remaining in culture (>98.5% CD8+; data not shown) were used as effectors.
Mouse CD8+ T cell proliferation assays
For mAb-induced proliferation assays, wells of 96-well plates were incubated with 1 mg/ml N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium chloride (Sigma-Aldrich) in 50% ethanol for 5 min at room temperature. The plate was washed in PBS, and mAb in carbonate buffer (8.4 g of NaHCO3 and 3.56 g of Na2CO3 in 1 liter of H2O) were incubated overnight at 4°C. Anti-mouse CD3 mAb (500A2; BD Biosciences) was used at the indicated concentrations. Anti-mouse CD28 mAb (37.51; BD Biosciences) and anti-mouse NKG2D mAb (CX5) were used at 5 μg/ml. The wells were washed and blocked in RPMI-T for 10 min at room temperature. T cells (1 x 105) were incubated in each well in RPMI-T medium for 48 h (or 4 days; Fig. 2B) at 37°C before pulsing each well with 1 μCi of [3H]thymidine (Amersham Biosciences). After 16 h, the plates were harvested, and the samples were analyzed by liquid scintillation in a Microcounter (Wallac). The counts from triplicate wells were averaged.
FIGURE 2. NKG2D does not costimulate naive mouse CD8+ T cell proliferation. A, Purified C57BL/6 CD8+ T cells were incubated on plates coated with varying amounts of anti-CD3 mAb, either alone (), with 5 μg/ml anti-CD28 mAb (), 5 μg/ml anti-NKG2D mAb (), or both mAbs (?). [3H]thymidine incorporation was assayed in a 72-h proliferation assay. Data are representative of four experiments. B, Proliferation of naive C57BL/6 CD8+ T cells was assessed as described in A after 5 days of culture. C, Splenocytes from C57BL/6 (dashed line) and RAE-1 (solid lines) transgenic mice were stained with anti-pan-RAE-1 mAb to assess the expression levels of NKG2D ligands on the transgenic splenocytes. The left panel shows RAE-1 expression on whole splenocytes, whereas the right panel shows expression on gated CD11c+ DC. Cells from C57BL/6 and RAE-1 transgenic mice stained with an isotype-matched control Ig were indistinguishable from C57BL/6 cells stained with anti-pan-RAE-1 (not shown). D, OT.1 T lymphocytes were cultured with irradiated C57BL/6 () or RAE-1-transgenic () splenocytes in the presence of the indicated concentration of OVAp. A 72-h proliferation assay was used to compare OT.1 T cell proliferation in the presence or the absence of NKG2D ligands. Data are representative of two experiments. For all panels, error bars represent 1 SD from the mean for triplicate wells.
For proliferation induced by the B16 transfectants, cells were irradiated at 20,000 rad using a cesium irradiator. Day 7 effector OT.1 CD8+ T cells (5 x 104) were incubated with 1 x 105 B16 cells with OVAp peptide at the indicated concentrations. Proliferation was measured as described above by [3H]thymidine incorporation 3 days after culture initiation.
For proliferation of naive CD8+ T cells, 1 x 105 OT.1 lymphocytes were incubated with 1 x 105 irradiated (2000 rad) splenocytes from C57BL/6 or RAE-1 transgenic mice in the presence of the indicated concentration of OVAp in RPMI-T. For proliferation of effector CD8+ T cells, naive OT.1 lymphocytes were incubated at a 1:1 ratio with either C57BL/6 or RAE-1 transgenic splenocytes in the presence of 100 nM OVAp for 3 days in RPMI-T at 37°C. The cells were then split 1:5 in RPMI-T with 200 U/ml human rIL-2 and 4 ng/ml mouse rIL-7. Seven days after primary stimulation, 1 x 105 OT.1 effectors from both primary stimulations were incubated with 1 x 105 irradiated (2000 rad) C57BL/6 or RAE-1 transgenic splenocytes in the presence of the indicated concentrations of OVAp. Proliferation was measured, as described above, by [3H]thymidine incorporation 3 days after culture initiation.
Cytotoxicity assays
Cytotoxicity assays using mouse NK cells were conducted as previously described (10). Redirected lysis assays against P815 targets using NKL cells as effectors were performed as previously described (6), except that the anti-human NKG2D mAb (149810) was used. For cytotoxicity assays using effector OT.1 T cells, parental RMA cells and those expressing RAE-1 were used as targets. Targets were adjusted to 1 x 107 cells/ml in RPMI-T and were labeled with Na2[51Cr]O4 for 2 h at 37°C in the presence or the absence of 1 x 10–6 M OVAp. The cells were washed, and 1 x 104 targets were added to each well of a 96-well plate. Day 7 OT.1 effector T cells were added to triplicate wells at the indicated E:T cell ratios, and the plate was incubated at 37°C for 5 h. Twenty-five microliters of cell-free supernatant was collected from each well, and radioactivity was measured in a Microcounter (Wallac). The percentage of specific 51Cr release was calculated according to the following formula: % specific lysis = 100 x (experimental cpm – spontaneous cpm)/(maximal cpm – spontaneous cpm).
Purification of small, resting PBLs
Buffy coats from healthy donors were purchased from the Stanford Blood Center. PBMCs were prepared according to standard procedures. To obtain small, resting PBLs, the PBMCs were washed in PBS and resuspended in 30% Percoll (Amersham Biosciences)/10% FBS in PBS. The cells were then overlaid on top of 40% Percoll/10% FBS in PBS, and the tubes were spun for 35 min at 2000 rpm (Sorvall RT6000B centrifuge) at 10°C with no brake. The pellet, which contained the small, resting PBLs (14), was harvested and washed, and the cells were resuspended in RPMI-T.
Human CD8+ T cell proliferation assays
To assess mAb-induced CD8+ T cell proliferation, in each well of a 96-well plate, 1 x 105 small, resting PBLs were incubated in RPMI-T with 1 x 104 irradiated (8500 rad) P815 cells in the presence of 100 ng/ml anti-human CD3 (Leu 4; BD Biosciences) and 5 μg/ml of a combination of two of the following mAbs as indicated: anti-human CD28 (L293; BD Biosciences), anti-human NKG2D (149810; R&D Systems), or anti-CD56 (DX32; used as a control mAb). The control mAb was used to ensure that the FcRs would be equally loaded with each mAb under the different conditions. The plates were incubated at 37°C for 48 h, then each well was pulsed with 1 μCi of [3H]thymidine (Amersham Biosciences) for 16 h. Plates were harvested, and samples were analyzed by liquid scintillation. Counts from triplicate wells were averaged.
For proliferation induced by the panel of P815 transfectants, 1 x 105 small, resting PBLs were incubated in RPMI-T with 1 x 104 irradiated P815 or transfectants/well of a 96-well plate in the presence of 150 ng/ml anti-CD3 (Leu 4). At 48 h, the plates were pulsed with [3H]thymidine for 16 h, harvested, and analyzed. For analysis of proliferation using CFSE (Molecular Probes), small, resting PBLs were isolated, washed, and resuspended in 4 μM CFSE in PBS at a concentration of 1 x 107 cells/ml for 8 min at room temperature. The CFSE was then quenched with an equal volume of FBS for 1 min. The cells were washed twice in RPMI-T, then 1.5 x 106 labeled cells were incubated with 2 x 105 irradiated P815 or transfectants in a 24-well plate in the presence of 75 ng/ml anti-CD3 (Leu 4). After 3 days at 37°C, the cells were harvested and stained for flow cytometry with the following mAbs for 30 min at 4°C: allophycocyanin-conjugated anti-CD8 (BD Biosciences) and PE-conjugated anti-CD28 (L293; BD Biosciences). After washing, the cells were resuspended in 1 μg/ml propidium iodide in PBS and 1% FBS, and the samples were analyzed by flow cytometry.
For proliferation of preactivated PBLs, small, resting PBLs were incubated at 37°C for 3 days in RPMI-T with human B7.1+ P815 transfectant and 50 ng/ml anti-CD3 (Leu 4). The cells were then expanded 1:5 into RPMI-T. Seven days after primary stimulation, CD4+ cells were depleted by incubating the cells with 5 μg of anti-CD4 (RPAT4) for 30 min at 4°C. Magnetic goat anti-mouse IgG beads (Qiagen) were used to deplete CD4+ cells as described for purification of mouse naive CD8+ T cells above. Cells (105) were then cultured with 1 x 104 irradiated P815 or transfected cells that had been preincubated with 50 ng/ml anti-CD3 (Leu 4) mAb along with 10 μg/ml isotype-matched control mAb, 5 μg/ml anti-B7.1 (L307)/5 μg/ml isotype-matched control mAb, 5 μg/ml anti-MICA (159209; R&D Systems)/5 μg/ml isotype-matched control mAb, or 5 μg/ml anti-MICA (159209)/5 μg/ml anti-B7.1 (L307) as indicated. Proliferation was measured as described for small, resting PBLs.
IFN- detection
Small, resting human PBLs were cultured with irradiated P815 or transfectants and anti-CD3 (Leu 4) as described above. Three and 7 days after culture initiation, cell-free supernatants were collected. A human IFN- ELISA kit (eBioscience) was used to quantify the concentration of cytokine.
Results
NKG2D expression on mouse CD8+ T cells after TCR stimulation
To determine when NKG2D could potentially contribute to primary CD8+ T cell stimulation, we examined the kinetics of NKG2D up-regulation after TCR engagement. CD8+ T cells were purified to >90% purity from spleens of 4- to 6-wk-old C57BL/6 mice. These cells were cultured on plate-bound anti-CD3, anti-CD3 plus anti-CD28, anti-CD3 plus anti-NKG2D, or all three mAbs, and NKG2D levels were monitored over time by flow cytometry. CD8+ T cells stimulated by anti-CD3 plus anti-CD28 began to up-regulate NKG2D as early as 2 days after stimulation, and the level of NKG2D increased over 5 days (Fig. 1A). The presence of anti-CD28 and/or anti-NKG2D mAbs did not affect the kinetics or extent of NKG2D up-regulation induced by anti-CD3 (data not shown). We also used OT.1 TCR-transgenic CD8+ T cells to examine the kinetics of NKG2D up-regulation after TCR stimulation via cognate peptide:MHC interactions. CD8+ T cells were purified to >90% purity from the spleen of a 4- to 6-wk-old OT-1 mouse. Mouse B16 melanoma cells were irradiated and pulsed with 1 nM OVAp (SIINFEKL), the cognate peptide for OT.1 TCR. These cells were cocultured with OT.1 CD8+ T cells, and the kinetics of NKG2D expression were monitored. Significant NKG2D expression was detected as early as 1 day after coculture, and NKG2D levels continued to rise over the first 3 days (Fig. 1B). These results demonstrated a detectable increase in cell surface NKG2D within 1–2 days after TCR stimulation.
FIGURE 1. Kinetics of NKG2D up-regulation on mouse CD8+ T cells after TCR signaling. A, Purified CD8+ T cells from C57BL/6 mice were cultured on anti-CD3 and anti-CD28 mAb-coated plates. At the indicated times, cells were stained with mAbs to CD8 and NKG2D. Histograms depict NKG2D levels relative to an isotype-matched Ig control on gated CD8+ cells. Data are representative of four experiments. B, Purified OT.1 CD8+ splenic T cells were cocultured with irradiated B16 cells pulsed with 1 nM OVAp. At the indicated times, cells were analyzed for the expression of CD8 and NKG2D. Histograms depict NKG2D levels relative to an isotype-matched Ig control on gated CD8+ cells. Data are representative of three experiments.
NKG2D engagement fails to costimulate proliferation of mouse naive CD8+ T cells
The potential of NKG2D to costimulate naive mouse CD8+ T cells has not been reported. Because NKG2D is expressed at the cell surface within 1–2 days after TCR engagement, it could potentially contribute to proliferation of T cells during a primary response. To assess this possibility, we performed a 3-day proliferation assay using purified splenic CD8+ T cells from C57BL/6 mice. The cells were stimulated with plate-bound anti-CD3, anti-CD3 plus anti-CD28, anti-CD3 plus anti-NKG2D, or all three mAbs. Anti-CD28 provided significant costimulation of the anti-CD3-induced proliferation. In contrast, anti-NKG2D failed to costimulate this primary CD8+ T cell response. Furthermore, anti-NKG2D did not synergize with anti-CD28 (Fig. 2A). These results are representative of four independent experiments.
Because NKG2D is not detectable at the cell surface until 48 h after TCR stimulation, the time at which the cells were pulsed with [3H]thymidine in the above experiments, it was possible that NKG2D would only contribute to CD8+ T cell proliferation measured at a later time. Therefore, we also measured mAb-induced proliferation 5 days after stimulation of purified C57BL/6 CD8+ T cells. As shown in Fig. 2B, the 5-day proliferation assay produced the same result as the 3-day assays. Anti-CD3 mAb alone induced minimal proliferation, whereas anti-CD28 mAb significantly costimulated the response. Anti-NKG2D mAb did not costimulate anti-CD3-induced proliferation, nor did it synergize with anti-CD28. These results are representative of two independent experiments.
The anti-NKG2D mAb used for these studies has been shown to have agonist function, in that plate-bound anti-NKG2D mAb induces cytokine production by mouse NK cells and can trigger NK cell-mediated cytolytic responses against FcR-bearing targets in an mAb-redirected cytotoxicity assay (15). However, it is possible that this anti-NKG2D mAb does not have agonist activity for T cells. Therefore, additional studies were conducted using APC expressing RAE-1, a natural high affinity ligand for mouse NKG2D. This was accomplished using splenic APC cells from mice bearing a RAE-1 transgene driven by a -actin promoter to give ubiquitous expression in all tissues. To confirm the presence of NKG2D ligands in the RAE-1 transgenic mice, splenocytes were analyzed for RAE-1 expression by flow cytometry. Transgenic splenocytes expressed high levels of RAE-1 in comparison with C57BL/6 splenocytes (Fig. 2C, left panel). Furthermore, CD11c+ dendritic cells, which are potent APCs, from the transgenic mice expressed high levels of RAE-1 in comparison with C57BL/6 dendritic cells that do not express RAE-1 (Fig. 2C, right panel).
Splenocytes from C57BL/6 or RAE-1 transgenic mice were irradiated and pulsed with increasing concentrations of OVAp. OT.1 T lymphocytes were cocultured with these splenocytes, and proliferation was measured 3 days later. Consistent with our previous results using anti-NKG2D mAb, OT.1 T cell proliferation was not enhanced by the presence of the RAE-1 on the APC (Fig. 2D). It should be noted that endogenous CD28 ligands were likely able to provide costimulation in this experiment; therefore, this result indicates that NKG2D engagement with ligand cannot enhance CD28-mediated costimulation. Taken together, the above results indicate that NKG2D does not costimulate a primary mouse CD8+ T cell response despite rapid up-regulation of cell surface NKG2D after TCR stimulation and availability of NKG2D ligands.
NKG2D engagement fails to costimulate proliferation of mouse effector CD8+ T cells
It has previously been reported that NKG2D costimulates preactivated effector mouse CD8+ T cells (5). We sought to examine this using OT.1 T cells 7 days after primary activation with OVAp. Day 7 OT.1 effectors were restimulated with plate-bound anti-CD3, anti-CD3 plus anti-CD28, anti-CD3 plus anti-NKG2D, or all three mAbs, and proliferation was measured 3 days later. Anti-CD28 costimulated the anti-CD3-induced proliferation, but surprisingly, anti-NKG2D did not (Fig. 3A). These results are representative of three independent experiments. We also observed the same results when we cocultured day 8 OT.1 effectors with irradiated C57BL/6 splenocytes pulsed with increasing concentrations of OVAp in the context of plate-bound anti-CD28, anti-NKG2D, or anti-CD28 plus anti-NKG2D. Anti-CD28 costimulated proliferation, whereas anti-NKG2D did not, and anti-CD28 plus anti-NKG2D did not synergize (data not shown). NKG2D was expressed at high levels on all of the day 7 or 8 preactivated CD8+ T cells (data not shown); therefore, failure to respond could not be explained by the absence of the receptor.
FIGURE 3. NKG2D does not costimulate effector mouse CD8+ T cell proliferation. A, Day 7 preactivated OT.1 effector CD8+ T cells were cultured on plates coated with varying amounts of anti-CD3 mAb, either alone (), with 5 μg/ml anti-CD28 mAb (), 5 μg/ml anti-NKG2D mAb (), or both mAbs (?) and were measured for [3H]thymidine incorporation 72 h later. Data are representative of two experiments. B, B16 transfectants were stained with mAbs to RAE-1 and B7.1 to confirm the expression of transfected molecules. C, Proliferation of day 7 OT.1 effector CD8+T cells was assayed after 72 h of coculture with irradiated B16 cells () or B16 transfectants expressing B7.1 (), RAE-1 (), or both (?) in the presence of various amounts of OVAp. Data are representative of four experiments. D, B16 () and RAE-1+ B16 () cells were used as targets in a cytotoxicity assay to verify that the RAE-1 transfectant was capable of stimulating the NKG2D receptor. IL-2-activated C57BL/6 NK cells were used as effectors at the indicated E:T cell ratios. E, Day 7 preactivated OT.1 CD8+ T cell effectors that were generated with OVAp in the presence of either C57BL/6 or RAE-1 transgenic splenocytes were stained for the expression of CD8 and NKG2D. Histograms depict NKG2D levels (solid lines) vs an isotype-matched Ig control (dashed lines) on gated CD8+ cells. F, Day 7 preactivated OT.1 CD8+ T cell effectors that were generated as described in E, were cultured for a secondary stimulation with C57BL/6 or RAE-1 transgenic splenocytes pulsed with the indicated concentration of OVAp. Primary stimulation on C57BL/6, followed by secondary on C57BL/6 splenocytes (?), primary on C57BL/6 and secondary on RAE-1 (), primary on RAE-1 and secondary on C57BL/6 (), and primary on RAE-1 and secondary on RAE-1 (?) are represented. [3H]thymidine incorporation was measured 72 h later. Data are representative of two experiments. For proliferation assays, error bars represent 1 SD from the mean for triplicate wells.
Similar studies were conducted with cells expressing natural ligands for CD28 and NKG2D rather than mAbs against these receptors. For these experiments, we used B16 melanoma cells transfected with B7.1 (CD80) and/or RAE-1. Expression of high levels of B7.1 and/or RAE-1 was confirmed by flow cytometry (Fig. 3B). Day 7 OT.1 effectors, generated by incubating OT.1 splenocytes in the presence of OVAp, were cocultured with irradiated B16 transfectants pulsed with various concentrations of OVAp. Expression of B7.1 yielded markedly increased proliferation compared with parental B16 cells (Fig. 3C). In contrast, expression of RAE-1 did not costimulate CD8+ T cell effectors. Furthermore, B7.1 and RAE-1 were not synergistic (Fig. 3C). These results are representative of three independent experiments.
To confirm that NKG2D ligands expressed on the B16 panel were capable of stimulating the NKG2D receptor, B16 cells were compared with RAE-1 transfectants as targets in an NK cytotoxicity assay. As shown in Fig. 3D, IL-2-activated mouse NK cells lysed RAE-1-expressing B16 cells at higher levels than parental B16 cells. Therefore, the RAE-1 B16 transfectant is capable of inducing NKG2D signaling, so the lack of NKG2D-mediated CD8+ T cell costimulation cannot be explained by an inability of the RAE-1 B16 transfectant to stimulate NKG2D.
Because the above results are inconsistent with prior studies (5), we sought to confirm them in an additional system. To activate OT.1 T lymphocytes, we cocultured them with splenocytes from either C57BL/6 or RAE-1 transgenic mice in the presence of OVAp. Day 7 T cell effectors generated from these stimuli expressed high levels of NKG2D, as assessed by flow cytometry (Fig. 3E). Therefore, the presence of an NKG2D ligand during primary stimulation did not result in significant down-modulation of NKG2D. Day 7 OT.1 effectors were then cocultured with either C57BL/6 or RAE-1 transgenic splenocytes in the presence of increasing amounts of OVAp, and proliferation was assayed 3 days after this secondary stimulation. Regardless of the presence of the NKG2D ligand on the APC during primary or secondary stimulation (Fig. 2C), OT.1 CD8+ T cells proliferated equivalently (Fig. 3F). These results are representative of two independent experiments. This assay indicates that NKG2D cannot enhance CD28-mediated costimulation. Taken together, the above data indicate that the expression of NKG2D on CD8+ effector T cells and NKG2D ligands on APCs is not sufficient to result in costimulation of primary, resting mouse CD8+ T cells, or activated, effector CD8+ T cells.
NKG2D engagement with cognate ligand fails to enhance cytotoxicity of mouse CTL
Although NKG2D did not costimulate the proliferation of effector CD8+ T cells, it might augment effector functions, such as cytotoxicity. To test this possibility, we examined the ability of OT.1 effector CD8+T cells to lyse peptide-pulsed parental RMA tumor cells vs RMA transfectants expressing RAE-1. The expression of high levels of RAE-1 on the transfectant was confirmed by flow cytometry at the time of the experiment (data not shown). OT.1 CTL specifically lysed both parental RMA cells and the RAE-1 transfectants equivalently in the presence, but not in the absence, of 10–6 M OVAp (Fig. 4A). Furthermore, equivalent lysis was observed over a range of E:T cell ratios in the presence of 10–7 and 10–8 M peptide (data not shown). This experiment is representative of two independent assays. To confirm that the RMA RAE-1 transfectant was capable of inducing NKG2D-induced cytotoxicity, IL-2-activated NK cells were used as effectors in a cytotoxicity assay with RMA or RMA-RAE-1 targets. NK cells efficiently lysed the targets expressing RAE-1, and this lysis was inhibited by a blocking mAb against NKG2D (Fig. 4B). These results indicate that the presence of an NKG2D ligand on a target cell is insufficient to augment cytotoxicity of NKG2D-expressing mouse CD8+ effector T cells.
FIGURE 4. NKG2D does not costimulate cytolytic activity of effector mouse CD8+ T cells. A, Day 7 preactivated OT.1 CD8+ T cell effectors were incubated with 51Cr labeled RMA cells (), RAE-1+ RMA (?), RMA pulsed with 10–6M OVAp (), or RAE-1+ RMA pulsed with OVAp () at several E:T cell ratios. A 51Cr release assay was used to measure the specific lysis of the two targets. Data are representative of two experiments. B, IL-2-activated C57BL/6 NK cells were used as effectors in a cytotoxicity assay to confirm that the RMA-RAE-1 transfectant was capable of inducing signaling through NKG2D. NK cells were incubated with 51Cr-labeled RMA (X) or RMA-RAE-1 cells in the presence of a blocking anti-NKG2D mAb () or an isotype-matched Ig control (). Error bars represent 1 SD from the mean for triplicate wells.
NKG2D engagement fails to costimulate proliferation of small, resting human PBLs
Human peripheral blood CD8+ T cells constitutively express NKG2D (6), and it has been reported that NKG2D engagement can costimulate proliferation and effector functions (9). We sought to confirm and extend these findings. First, we enriched for small, resting PBLs, and examined NKG2D expression levels by flow cytometry. As reported previously (6), nearly all human CD8+ T cells express high levels of NKG2D (Fig. 5A). We next assessed the ability of mAbs against CD28 and/or NKG2D to costimulate anti-CD3 mAb-induced proliferation. Using an assay we previously described to demonstrate anti-CD28 mAb costimulation (14), irradiated P815 cells, a mouse mastocytoma line expressing FcRs, were incubated with small, resting human PBLs in the presence of a low concentration of anti-CD3 alone or with anti-CD28 and/or anti-NKG2D mAbs. Proliferation was assessed 3 days later. P815 cells coated only with anti-CD3 mAb induced little proliferation, whereas those also coated with anti-CD28 mAb induced a high level of proliferation (Fig. 5B), consistent with previous findings (14). In contrast, anti-NKG2D mAb did not increase the low level of proliferation seen with anti-CD3 alone, and anti-NKG2D did not augment anti-CD28 mAb costimulation (Fig. 5B). These assays indicated that CD28, but not NKG2D, was a potent costimulatory molecule for small, resting PBLs. The anti-NKG2D mAb used in these studies has agonist activity; it induces human NK cell-mediated lysis of anti-NKG2D mAb-coated P815 target cells (Fig. 5C) and costimulates cytokine production by human NK cells activated through an activating killer cell Ig-like receptor (16).
FIGURE 5. NKG2D does not costimulate the proliferation of fresh small, resting human peripheral blood CD8+ PBLs. A, Small, resting human PBLs were isolated, and the expression of NKG2D on CD8+ T cells was detected by flow cytometry. Numbers in the quadrants display the percentage of cells. B, Small, resting PBLs were cocultured with P815 cells in the presence of the indicated mAbs to costimulate suboptimal (100 ng/ml) anti-CD3 mAb-induced proliferation through CD28 and/or NKG2D. In addition to the samples containing P815, PBL, and anti-CD3 (), the following controls are included: P815 and anti-CD3 (), P815 and PBL (), and PBL and anti-CD3 (). Proliferation was determined 72 h later by measuring [3H]thymidine incorporation. The control mAb against CD56 was used to maintain a constant concentration of mAb to ensure equal FcR loading in all samples. Data are representative of two experiments. C, The human NK line, NKL, was cocultured with 51Cr-labeled P815 cells in the presence of anti-NKG2D mAb () or an isotype-matched Ig control (). Redirected mAb-dependent cytotoxicity was determined in a 4-h cytotoxicity assay. D, Small, resting PBLs were cocultured with irradiated P815 transfectants that expressed ligands for CD28 and/or NKG2D and a suboptimal concentration of anti-CD3 mAb. Proliferation assay and controls were as described in B. Data are representative of six experiments with four blood donors. E, Top, The P815 transfectants were stained with anti-human B7.1, and the cells were examined by flow cytometry. MICA was expressed using an IRES-GFP encoding construct (bottom). Confirmation that GFP is representative of MICA expression was achieved by staining with a human NKG2D-Ig fusion protein. F, NKL cells were used as effectors in a 51Cr release assay to confirm that the P815-MICA*0019 transfectant was capable of inducing signaling through NKG2D. NKL cells were incubated with 51Cr-labeled P815 cells () or P815-MICA*0019 transfectants (), and the percent specific lysis was determined after 4 h. G, Small, resting peripheral blood T cells were labeled with CFSE and cocultured with the P815 panel in the presence of a low concentration (75 ng/ml) of anti-CD3 mAb, as described in D. After 3 d, the cells were labeled with allophycocyanin-conjugated anti-CD8 and PE-conjugated anti-CD28 mAb and analyzed by flow cytometry. The dot plots display CFSE vs CD28 for CD8+ gated cells. Proliferation was analyzed by CFSE dilution. Numbers in the quadrants indicate the percentage of cells. Data are representative of two experiments. In proliferation assays, error bars represent 1 SD from the mean for triplicate wells.
We next investigated whether cognate ligands for CD28 and/or NKG2D could costimulate small, resting PBLs. We cocultured small, resting PBLs with P815 cells or transfectants expressing human B7.1 and/or MICA*0019, natural ligands for CD28 and NKG2D, respectively, in the presence of a suboptimal amount of anti-CD3 mAb. Consistent with previous findings (14), the B7.1+ P815 transfectant induced significantly higher levels of proliferation than the parental P815 cells. However, the MICA*0019 transfectant did not costimulate anti-CD3 mAb-induced proliferation (Fig. 5D). These data are representative of six independent experiments with four donors. The P815 transfectants were examined by flow cytometry to confirm high expression of B7.1 and MICA (Fig. 5E). To confirm that the MICA*0019 transfectant was capable of inducing signaling through NKG2D, P815 cells and the P815 MICA*0019 transfectants were compared as targets in a cytotoxicity assay using human NK cells (Fig. 5F). The MICA*0019 transfectant was killed much more efficiently than parental P815 cells.
NKG2D engagement fails to costimulate proliferation of peripheral blood CD28–CD8+ T cells
Groh et al. (9) previously reported that NKG2D augmented proliferation and effector functions of human CD28– CD8+ peripheral blood T cells. Thus, it was possible that [3H]thymidine incorporation, which measures proliferation of all cultured T cells, was not sensitive enough to discern costimulation of CD28– CD8+ cells, which represent only 20% of the CD8+ PBLs in our cultures. Therefore, we repeated the proliferation assay shown in Fig. 5D, except we labeled the human small, resting peripheral blood T cells with CFSE before coincubation with anti-CD3 mAb and the irradiated P815 transfectants. Three days later, cells were stained for CD8 and CD28, and proliferation of CD8+CD28+ and CD28–CD8+ T cell populations was examined by flow cytometry. The B7.1+ P815 transfectant costimulated proliferation of CD8+CD28+ T cells, whereas the MICA*0019 transfectant did not. No synergy was observed between B7.1 and MICA*0019. In contrast, CD28–CD8+ T cells did not proliferate significantly to any of the P815 transfectants coated with anti-CD3 mAb, consistent with previous reports that these cells are hyporesponsive to anti-CD3 stimulation (17) (Fig. 5G). These findings are representative of two independent experiments. The CFSE results are consistent with the [3H]thymidine results, indicating that CD28 is able to costimulate the proliferation of human small, resting CD8+ PBLs, whereas NKG2D does not have an apparent costimulatory function.
NKG2D engagement fails to costimulate proliferation of preactivated human peripheral blood CD8+ T cells; however, it augments CD28 costimulation
Although it was reported previously (9) that anti-NKG2D mAb could costimulate the proliferation of freshly isolated CD8+ T cells from human peripheral blood, we considered the possibility that some of the T cells in the donor used for these studies might have been preactivated due to an unknown infection. To test the possibility that NKG2D costimulates preactivated human CD8+ T cells, we stimulated small, resting PBLs with anti-CD3 mAb-coated B7.1+ P815 transfectants, using the conditions described in Fig. 5D. Seven days after primary stimulation, CD4+ cells were depleted, and CD8+ T cells (>90% pure), all of which expressed high levels of NKG2D (not shown), were cocultured with P815 transfectants using suboptimal concentrations of anti-CD3. The B7.1 transfectant, but not the MICA*0019 transfectant, significantly costimulated activated human CD8+ T cell proliferation (Fig. 6). Interestingly, NKG2D synergized with CD28, such that the greatest proliferation was observed in response to the MICA*0019/B7.1 double transfectant. This synergy was reduced to proliferation levels seen with the B7.1 transfectant using an anti-MICA mAb. Furthermore, an anti-B7.1 mAb completely blocked costimulation induced by the B7.1 or MICA*0019/B7.1 transfectants, indicating that CD28 costimulation is required for basal costimulation (Fig. 6). These results are representative of assays using peripheral blood from three different donors in two independent experiments. The above results indicate that CD28, but not NKG2D, is able to costimulate the proliferation of small, resting PBLs and preactivated CD8+ PBLs. However, NKG2D signaling synergizes with CD28 to enhance the proliferation of preactivated human CD8+ T cells.
FIGURE 6. NKG2D does not costimulate preactivated human peripheral blood CD8+ T cells; however, NKG2D synergizes with CD28. Small, resting PBLs were cocultured with irradiated B7.1+ P815 cells in the presence of 50 ng/ml anti-CD3 mAb. Seven days later, the activated cells were tested in a secondary stimulation proliferation assay by coculture with irradiated P815 transfectants, as described in Fig. 5D, in the presence of 50 ng/ml anti-CD3 mAb. The cells were cultured in the presence of control mAbs (), anti-MICA mAb (), anti-B7.1 mAb (), or both anti-MICA and anti-B7.1 mAbs () to determine the specificity of costimulation. Proliferation was measured by [3H]-thymidine uptake at 72 h. These data are representative of assays using peripheral blood from three different donors in two experiments.
]NKG2D engagement fails to enhance IFN- production by primary human peripheral blood CD8+ T cells
It is possible that NKG2D costimulates effector functions of human CD8+ PBLs rather than proliferation. To address this possibility, IFN- production was measured by ELISA 3 and 7 days after coculture of small, resting PBLs with the P815 transfectants and anti-CD3 mAb. The suboptimal level of anti-CD3 used was insufficient to induce IFN- production in response to parental P815 cells (Fig. 7). IFN- was not produced after stimulation with MICA*0019 or MICB*002 transfectants. However, IFN- was produced after coculture with the B7.1 or B7.1/MICA*0019 transfectants on both days 3 and 7 (Fig. 7). These data indicate that CD28 is able to costimulate the production of IFN-, whereas NKG2D is not. As shown in Fig. 6, NKG2D synergized with CD28 to elicit more IFN- release after stimulation with the B7.1/MICA*0019 double transfectant in comparison with the B7.1 transfectant (Fig. 7). These data are representative of assays using peripheral blood from three different donors in two independent experiments. Therefore, although NKG2D can synergize with CD28 to costimulate IFN- production from small, resting human PBLs and proliferation of preactivated human CD8+T cells, we were unable to detect a role for NKG2D alone in costimulation of either proliferation or effector responses in primary human CD8+ T cells.
FIGURE 7. NKG2D does not costimulate IFN- secretion on small, resting PBLs, although it synergizes with CD28. Small, resting PBLs were cocultured with irradiated P815 transfectants and anti-CD3 as described in Fig. 5D. Three () and 7 () days after culture initiation, supernatants were analyzed for IFN- by ELISA. Error bars represent 1 SD for the mean of triplicate wells. These data are representative of assays using peripheral blood from three different donors in two experiments.
Discussion
The role of NKG2D as a costimulatory molecule on certain murine and human CD8+ T cells has been documented in several studies (5, 9, 18, 19). However, we report in this study that engagement of NKG2D alone is not sufficient to costimulate primary mouse or human CD8+ T cell activation. In agreement with previous reports (5), we found that NKG2D is rapidly induced on activated mouse CD8+ T cells and is constitutively expressed on all human peripheral blood CD8+ T cells (6). However, we were unable to demonstrate a role for NKG2D in costimulating proliferation, cytolytic activity or IFN- production of naive or effector splenic mouse CD8+ T cells or small, resting or preactivated human peripheral blood CD8+ T cells. In all these assays, CD28 costimulation was observed.
It has previously been reported that NKG2D can costimulate the proliferation of effector mouse CD8+ T cells equivalently to CD28 (5, 18). In these experiments, an LCMV-specific T cell line was generated and maintained via weekly restimulation with peptide in the presence of cytokines. To assay for costimulation, T cells were labeled with CFSE 7 days after restimulation and cultured on mAb-coated plates for 4 days before analysis by flow cytometry. In this context, NKG2D and CD28 were both shown to equivalently costimulate cells that had undergone more than one or more than three cell divisions. Using OT.1 CD8+ T cell effectors, we were unable to demonstrate NKG2D-mediated costimulation of proliferation, although we did observe potent CD28-mediated costimulation. Our effector cells were generated by a single in vitro exposure to OVAp, followed 4 days later by transfer to IL-2- and IL-7-containing media. Day 7 effectors were then used in proliferation assays. Using mAbs or cognate ligands, CD28 engagement provided significant costimulation of proliferation, whereas NKG2D did not. Because it was reported previously that the major effect of NKG2D costimulation was on cells that had undergone more than three cell divisions (5), we were concerned that this difference might be obscured by the measurement of all cells undergoing division in a [3H]thymidine assay. Therefore, we also performed proliferation assays with CFSE-labeled OT.1 T cell effectors. Nonetheless, we observed the same results; CD28 costimulated proliferation, whereas NKG2D did not (data not shown). Furthermore, proliferation assays were conducted comparing our CX5 anti-NKG2D mAb with the MI6 mAb, previously described by Jamieson et al. (5). We observed no evidence for NKG2D costimulation with either mAb (unpublished observations). Most significantly, cognate NKG2D ligands expressed by B16 transfectants or RAE-1 transgenic splenocytes were unable to costimulate effector CD8+ T cell proliferation. Although we cannot explain the discrepancy between our data and the previous report, it is possible that T cells in the LCMV-specific T cell line had been exposed to cytokines during infection that potentiated the costimulatory capability of NKG2D. Alternatively, weekly restimulation of these cells in the presence of cytokines may have selected out a small subpopulation on which NKG2D was costimulatory or may have potentiated NKG2D signaling.
Consistent with the hypothesis that a small subpopulation of mouse effector CD8+ T cells might be responsible for the previously reported costimulatory properties of NKG2D, it has recently been reported that extrathymically derived IL-2- or IL-15-activated CD8+CD44high cells can be stimulated via NKG2D engagement (20). These cells represent a small subpopulation that can be activated by IL-2 in the absence of TCR stimulation. After IL-2 stimulation, they up-regulate NK cell markers in addition to NKG2D, and they express DAP12. Furthermore, in male H-Y TCR-transgenic mice, in which the majority of thymocytes are negatively selected due to expression of the male Ag, lymphocytes of similar phenotype and functionality can be found in the periphery (21). These cells also express DAP12, and their cytolytic function can be activated directly by NKG2D. Interestingly, the cytolytic activity of CD8+ T cell effectors generated from the periphery of female H-Y transgenic mice were not costimulated by NKG2D engagement (21), consistent with our results. These data, in conjunction with our findings, suggest that although activated conventional primary peripheral mouse CD8+ T cells express NKG2D, they are not costimulated by NKG2D engagement. Instead, a small subpopulation of mouse CD8+ T cells is apparently able to proliferate in response to inflammatory cytokines, such as IL-2 and IL-15, leading to the acquisition of NKG2D activity. This activity might be due to the expression of DAP12 in these cells (20, 21), because it has previously been demonstrated that activated T cells can be directly stimulated through NKG2D, if DAP12 expression is enforced as a transgene in mouse T cells (18).
NKG2D-mediated costimulation of human CD8+CD28–T cells was also previously reported (9). These experiments, which were conducted predominantly with long term human CMV-specific T cell clones, demonstrated that CD8+CD28– cells lysed human CMV-infected targets in an MHC-restricted manner, and this lysis could be inhibited by blocking NKG2D. Furthermore, NKG2D engagement via ligand or mAb cross-linking costimulated proliferation, cytolysis, and cytokine secretion in a peptide:MHC-dependent manner. In addition, the authors demonstrated that anti-NKG2D mAb could costimulate anti-CD3 mAb-mediated proliferation of freshly isolated CD8+CD28– PBLs (9). Our experiments with a long term polyclonal human CD8+ T cell line indeed demonstrated NKG2D-mediated costimulation of anti-CD3-induced cytokine production (unpublished observations). However, we were unable to show NKG2D costimulation using small, resting PBLs isolated from the blood of several healthy donors. With all donors tested, neither mAb cross-linking nor ligand engagement of NKG2D was able to costimulate proliferation, although CD28 engagement was consistently costimulatory.
We were concerned that [3H]thymidine incorporation assays might not detect NKG2D costimulation if only CD28–CD8+ T cells responded to NKG2D signals. Therefore, we used CFSE labeling of small, resting PBLs to distinguish between the proliferation of CD8+CD28– cells and CD8+CD28+ T cells. We found that only CD8+CD28+ cells proliferated, and these cells were costimulated only by CD28, not NKG2D. The lack of proliferation of CD8+CD28–T cells stimulated with anti-CD3 mAb has been previously reported (17). Collectively, we cannot explain the discrepancy between our data and the previous results reported with freshly isolated, human peripheral blood CD8+CD28– T cells. However, the T cell clones used in the previous reports may have represented a small subpopulation in which NKG2D is costimulatory. Alternatively, long term culture of CD8+T cells in IL-2 may render them permissive for NKG2D costimulation, consistent with our unpublished observations.
Recently, it has been reported that culturing human CD8+ T cells in IL-15 or high concentrations of IL-2 can induce TCR-independent lymphokine-activated killer (LAK) activity that is mediated by NKG2D stimulation (19, 22, 23). However, only cytolytic activity was directly induced by NKG2D signaling in these LAK cells, whereas proliferation and cytokine secretion required TCR stimulation (19). Interestingly, Jabri and colleagues have demonstrated NKG2D-mediated costimulation of proliferation and IFN- production of freshly isolated intraepithelial T lymphocytes, and this costimulation was enhanced by the inclusion of IL-15 (19). Furthermore, intraepithelial CD8+ T cells from celiac patients, where endogenous IL-15 concentrations are high, demonstrated NKG2D-mediated LAK activity (22). Perhaps human T cells isolated from inflamed tissues, but not those isolated from blood, are able to signal through NKG2D. Consistent with the idea that T cells isolated from abnormal tissues may preferentially use NKG2D as a stimulatory molecule, several reports have indicated that tumor-infiltrating lymphocytes from tumors of diverse origin contain T cells that are responsive to NKG2D stimuli (24, 25, 26). However, all of these findings suggest that at least some of the tumor-infiltrating T cells are responsive to NKG2D engagement in an MHC-independent manner. Thus, human T cell NKG2D reactivity against tumors may predominantly be a LAK phenomenon, rather than TCR-dependent Ag-specific T cell recognition.
Although we were unable to demonstrate NKG2D-mediated costimulation of human CD8+ T cells, we did observe synergy between CD28 and NKG2D in two assays. NKG2D synergized with CD28 to induce IFN- secretion from small, resting human PBLs, and NKG2D synergized with CD28 to costimulate proliferation of preactivated human CD8+ T cells. Anti-B7.1 completely abrogated proliferation induced by the B7.1/MICA double transfectant, indicating that CD28 signaling is required for minimal costimulation. These results indicate that NKG2D can contribute to both preactivated human CD8+ T cell proliferation and resting CD8+ cytokine secretion, but only in conjunction with CD28 signaling.
The possibility that NKG2D may costimulate CD8+ T cells has generated much excitement in the fields of tumor biology, autoimmunity, and infectious disease (27, 28). NKG2D may be a potential candidate for therapeutic targeting to block autoimmune T cells or enhance T cell responses to tumors or pathogen-infected tissues. Indeed, we have recently reported that anti-NKG2D treatment can block diabetes in NOD mice, possibly due to blockade of CD8+ T cell activation (29). However, in this report we demonstrate that mAb- or ligand-mediated engagement of NKG2D alone is not sufficient to costimulate proliferation or effector functions of primary, resting, or activated human or mouse CD8+ T cells. Thus, it will be important to identify the subpopulations and cytokine conditions that render CD8+ T cells permissive for NKG2D-mediated signaling.
Acknowledgments
We thank Amanda Jamieson and David Raulet for providing MI-6 mAb for comparison with CX5. We thank Thomas Pertel for excellent technical assistance, Jason Dietrich for generation of the RAE-1 transgenic mice, and Nigel Killeen for helpful discussions.
Footnotes
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.
1 This work was supported by National Institutes of Health Grants CA89189 and CA95137. L.R.E. is supported by the Damon Runyon Cancer Research Foundation (DRG-1759-03), K.O. is supported by a Human Frontier Science Program Long-Term Fellowship, J.A.H. is an Irvington Foundation Fellow, and L.L.L. is an American Cancer Society Research Professor.
2 Current address: Department of Pathology, Stanford University, B257 Beckman Center, Stanford, CA 94305-5323.
3 Current address: Department of Experimental Medicine and Pathology, University of Rome La Sapienza, Viale Regina Elena 324 00161 Rome, Italy.
4 Address correspondence and reprint requests to Dr. Lewis L. Lanier, Department of Microbiology and Immunology and The Cancer Research Institute, University of California, 513 Parnassus Avenue, HSE 1001, Box 0414, San Francisco, CA 94143. E-mail address: lanier@itsa.ucsf.edu
5 Abbreviations used in this paper: LCMV, lymphocytic choriomeningitis virus; LAK, lymphokine-activated killer; OVAp, SIINFEKL peptide; MICA, MHC class I chain-related A.
Received for publication September 15, 2004. Accepted for publication November 7, 2004.
References
Janeway, C. A., Jr, R. Medzhitov. 2002. Innate immune recognition. Annu. Rev. Immunol. 20:197.
Ahmed, R., D. Gray. 1996. Immunological memory and protective immunity: understanding their relation. Science 272:54.
Diefenbach, A., A. M. Jamieson, S. D. Liu, N. Shastri, D. H. Raulet. 2000. Ligands for the murine NKG2D receptor: expression by tumor cells and activation of NK cells and macrophages. Nat. Immunol. 1:119.
Wu, J., Y. Song, A. B. Bakker, S. Bauer, T. Spies, L. L. Lanier, J. H. Phillips. 1999. An activating immunoreceptor complex formed by NKG2D and DAP10. Science 285:730.
Jamieson, A. M., A. Diefenbach, C. W. McMahon, N. Xiong, J. R. Carlyle, D. H. Raulet. 2002. The role of the NKG2D immunoreceptor in immune cell activation and natural killing. Immunity 17:19.
Bauer, S., V. Groh, J. Wu, A. Steinle, J. H. Phillips, L. L. Lanier, T. Spies. 1999. Activation of NK cells and T cells by NKG2D, a receptor for stress-inducible MICA. Science 285:727.
Chang, C., J. Dietrich, A. G. Harpur, J. A. Lindquist, A. Haude, Y. W. Loke, A. King, M. Colonna, J. Trowsdale, M. J. Wilson. 1999. Cutting edge: KAP10, a novel transmembrane adapter protein genetically linked to DAP12 but with unique signaling properties. J. Immunol. 163:4651.
Dymecki, S. M.. 1996. Flp recombinase promotes site-specific DNA recombination in embryonic stem cells and transgenic mice. Proc. Natl. Acad. Sci. USA 93:6191.
Townsend, S. E., J. P. Allison. 1993. Tumor rejection after direct costimulation of CD8+ T cells by B7-transfected melanoma cells. Science 259:368.
Kinsella, T. M., G. P. Nolan. 1996. Episomal vectors rapidly and stably produce high-titer recombinant retrovirus. Hum. Gene Ther. 7:1405.
Azuma, M., M. Cayabyab, J. H. Phillips, L. L. Lanier. 1993. Requirements for CD28-dependent T cell-mediated cytotoxicity. J. Immunol. 150:2091.
Zompi, S., J. A. Hamerman, K. Ogasawara, E. Schweighoffer, V. L. Tybulewicz, J. P. Di Santo, L. L. Lanier, F. Colucci. 2003. NKG2D triggers cytotoxicity in mouse NK cells lacking DAP12 or Syk family kinases. Nat. Immunol. 4:565
Zingoni, A., T. Sornasse, B. G. Cocks, Y. Tanaka, A. Santoni, L. L. Lanier. 2004. Cross-talk between activated human NK cells and CD4+ T cells via OX40-OX40L interactions. J. Immunol. :173.
Azuma, M., J. H. Phillips, L. L. Lanier. 1993. CD28– T lymphocytes: antigenic and functional properties. J. Immunol. 150:1147.
Diefenbach, A., E. Tomasello, M. Lucas, A. M. Jamieson, J. K. Hsia, E. Vivier, D. H. Raulet. 2002. Selective associations with signaling proteins determine stimulatory versus costimulatory activity of NKG2D. Nat. Immunol. 3:1142.
Roberts, A. I., L. Lee, E. Schwarz, V. Groh, T. Spies, E. C. Ebert, B. Jabri. 2001. NKG2D receptors induced by IL-15 costimulate CD28-negative effector CTL in the tissue microenvironment. J. Immunol. 167:5527.
Dhanji, S., H. S. Teh. 2003. IL-2-activated CD8+CD44high cells express both adaptive and innate immune system receptors and demonstrate specificity for syngeneic tumor cells. J. Immunol. 171:3442.
Dhanji, S., S. J. Teh, D. Oble, J. J. Priatel, H. S. Teh. 2004. Self-reactive memory-phenotype CD8 T cells exhibit both MHC-restricted and non-MHC restricted cytotoxicity: a role for the t cell receptor and natural killer cell receptors. Blood 104:2116.
Meresse, B., Z. Chen, C. Ciszewski, M. Tretiakova, G. Bhagat, T. N. Krausz, D. H. Raulet, L. L. Lanier, V. Groh, T. Spies, et al 2004. Coordinated induction by IL15 of a TCR-independent NKG2D signaling pathway converts CTL into lymphokine-activated killer cells in celiac disease. Immunity 21:357.
Verneris, M. R., M. Karami, J. Baker, A. Jayaswal, R. S. Negrin. 2004. Role of NKG2D signaling in the cytotoxicity of activated and expanded CD8+ T cells. Blood 103:3065.
Conejo-Garcia, J. R., F. Benencia, M. C. Courreges, P. A. Gimotty, E. Khang, R. J. Buckanovich, K. A. Frauwirth, L. Zhang, D. Katsaros, C. B. Thompson, et al 2004. Ovarian carcinoma expresses the NKG2D ligand Letal and promotes the survival and expansion of CD28– antitumor T cells. Cancer Res. 64:2175.
Girardi, M., D. E. Oppenheim, C. R. Steele, J. M. Lewis, E. Glusac, R. Filler, P. Hobby, B. Sutton, R. E. Tigelaar, A. C. Hayday. 2001. Regulation of cutaneous malignancy by T cells. Science 294:605.
Maccalli, C., D. Pende, C. Castelli, M. C. Mingari, P. F. Robbins, G. Parmiani. 2003. NKG2D engagement of colorectal cancer-specific T cells strengthens TCR-mediated antigen stimulation and elicits TCR independent anti-tumor activity. Eur. J. Immunol. 33:2033.
Raulet, D. H.. 2003. Roles of the NKG2D immunoreceptor and its ligands. Nat. Rev. Immunol. 3:781.
Cerwenka, A., L. L. Lanier. 2001. Natural killer cells, viruses and cancer. Nat. Rev. Immunol. 1:41.
Ogasawara, K., J. A. Hamerman, L. R. Ehrlich, H. Bour-Jordan, P. Santamaria, J. A. Bluestone, L. L. Lanier. 2004. NKG2D blockade prevents autoimmune diabetes in NOD mice. Immunity 20:757.(Lauren I. Richie Ehrlich,)