Protein kinase C-θ (PKCθ) is critical for TCR-initiated signaling in mature T cells, but initial reports found no requirement for PKCθ in thymocyte development. Thymocytes and peripheral T cells utilize many of the same signaling components and, given the significant role of PKCθ in peripheral T cells, it was surprising that it was not involved at all in TCR signaling in thymocytes. We decided to re-evaluate the role of PKCθ in thymocyte development using the well-characterized class II-restricted n3.L2 TCR-transgenic TCR model. Analysis of n3.L2 PKCθ−/− mice revealed a defect in thymocyte-positive selection, resulting in a 50% reduction in the generation of n3.L2 CD4 single-positive thymocytes and n3.L2 CD4 mature T cells. Competition between n3.L2 WT and n3.L2 PKCθ−/− thymocytes in bone marrow chimeras revealed a more dramatic defect, with a >80% reduction in generation of n3.L2 CD4 single-positive thymocytes derived from PKCθ−/− mice. Inefficient positive selection of n3.L2 PKCθ−/− CD4 single-positive cells resulted from “weaker” signaling through the TCR and correlated with diminished ERK activation. The defect in positive selection was not complete in the PKCθ−/− mice, most likely accounted for by compensation by other PKC isoforms not evident in peripheral cells. Similar decreased positive selection of both CD4 and CD8 single-positive thymocytes was also seen in nontransgenic PKCθ−/− mice. These findings now place PKCθ as a key signaling molecule in the positive selection of thymocytes as well as in the activation of mature T cells.
Protein kinase Cθ (PKCθ)3 is a serine/threonine kinase important for TCR signaling and mature T cell function (1, 2, 3, 4). PKCθ is localized to the immunological synapse in mature T cells (5, 6) and has been linked to multiple pathways downstream of TCR signaling, including but not limited to calcium activation and activation of ERK and I-κB kinase (7, 8, 9, 10, 11, 12, 13). The downstream effects of these multiple signaling roles for PKCθ are revealed by the multiple activation defects of PKCθ−/− T cells. Mature T cells from PKCθ−/− mice do not produce IL-2 and do not proliferate in response to Ag stimulation in vitro and are deficient in switching to the Th2 phenotype in vivo (7, 9, 14, 15). Furthermore, development of NK1.1 and CD25+FoxP3+ regulatory T (Treg) cells was diminished in PKCθ−/− mice (16).
Thymic development of α/β TCR-expressing T cells in PKCθ−/− mice has not yet been tested using a class II-restricted transgenic TCR. Given the central role for PKCθ in mature T cell signaling, and given that PKCθ is up-regulated during transition from the double-negative (DN) to the double-positive (DP) stage (17), it was not unreasonable to expect that PKCθ would participate in thymocyte development. However, initial examination of nontransgenic PKCθ−/− mice found no obvious differences in α/β T cell development (7, 9), and positive and negative selection appeared unaffected using the class I-restricted transgenic TCR H-Y in female and male PKCθ−/− mice, respectively (7). However, in nontransgenic mice, TCR repertoire shifts can obscure moderate differences in TCR signaling strength. We hypothesized that by using a class II-restricted transgenic TCR to fix the TCR repertoire, a role for PKCθ in thymocyte TCR signaling would be revealed.
The CD4-restricted n3.L2-transgenic TCR has been previously described and is specific for Hb64-76/I-Ek (18, 19, 20). n3.L2+ thymocytes are positively selected in H-2k mice (18, 19). Useful reagents for this TCR have been established, including a clonotypic Ab (Cab) and purified refolded I-Ek with antigenic peptide (pI-Ek), both of which can be used for in vitro stimulation. In addition, a panel of altered peptide ligands (APLs) has been established for the n3.L2 TCR, which enable an analysis of TCR signaling strength. Thus, these reagents make the n3.L2 system ideal for examining thymic development of CD4+ T cells. To address whether CD4+ T cell development was altered in PKCθ−/− mice, n3.L2 TCR-transgenic, PKCθ−/−, and H-2k mice were generated.
In this study, we report that generation of CD4 single-positive (CD4SP) n3.L2 thymocytes was significantly reduced in n3.L2 PKCθ−/− mice as compared with n3.L2 WT mice. This developmental defect of CD4SP thymocytes in n3.L2 PKCθ−/− mice resulted from diminished positive selection and correlated with reduced TCR-stimulated phosphorylation of ERK kinase. Once we had established the defect in positive selection in the n3.L2-transgenic system, we evaluated nontransgenic PKCθ−/− mice and confirmed reduced positive selection of both CD4SP and CD8SP thymocytes. We hypothesized that the reduction in generation of CD4SP thymocytes might explain the previously reported deficit of CD25+FoxP3+ Treg cells in PKCθ−/− mice (16). However, the failure of CD25+FoxP3+ cell development was apparent even when controlling for the reduced development of CD4SP thymocytes. Thus, here we provide evidence for a previously unrecognized role for PKCθ in TCR signaling in thymocytes.
Materials and Methods
PKCθ−/− mice backcrossed onto a B6 background for at least 15 generations were a gift from D. Littman (Skirball Institute of Biomolecular Medicine, New York University School of Medicine, New York, NY) (14). n3.L2-transgenic mice have been previously described and are on the background of B6.AKR-H-2k (B6.K) (19, 20). Breeding the PKCθ−/− mice to the n3.L2 B6.K mice generated n3.L2 PKCθ−/− mice on a B6 background that is H-2k and would thus positively select the n3.L2 TCR. Non-TCR-transgenic mice were generated by breeding the PKCθ−/− mice to nontransgenic B6.K mice. Generation of a pure population of n3.L2+ CD4/CD8 DP thymocytes was accomplished by crossing the n3.L2-transgenic mouse to the nonselecting H-2b, RAG-1-deficient strain (C57BL/6J-RAG1tm1 Mom; The Jackson Laboratory) and has been previously described (19). For biochemistry experiments, n3.L2 PKCθ−/−RAG1−/− H-2b mice were generated by breeding n3.L2 PKCθ−/− B6.K mice to n3.L2 RAG1−/− H-2b mice. The phenotypes of mice were determined either by PCR for PKCθ, RAG1, and n3.L2 or by flow cytometry for H-2k or H-2b. All mice were housed in specific pathogen-free environments at the Washington University and maintained according to approved practices. All experiments were performed according to protocols approved by the Washington University Animal Studies Committee. Mice were used between the ages of 6 and 12 wk.
Single-cell suspensions of thymocytes, splenocytes, and lymph node (axillary and inguinal) cells were incubated on ice with the indicated Abs in FACS buffer (0.5% BSA and 0.1% sodium azide in PBS). The clonotypic Ab Cab recognizes the n3.L2 TCR and has been previously described (18). Fluorescently conjugated Abs purchased from commercial sources were FITC-anti-CD4, allophycocyanin-anti-CD4, allophycocyanin/Cy7-anti-CD4, PE-anti-CD8, PE-anti-CD69, PE-anti-CD24, FITC-anti-CD25, PE-anti-CD25, PE-anti-CD44, PE-anti-γ/δ TCR, PE-anti-H-2b (Biolegend); PE-anti-CD4, PerCP-anti-CD8, PE-anti-Bcl-2, PE-anti-NK1.1, FITC-anti-H-2k, PE/Cy5-strepavidin (BD Biosciences), FITC-anti-CD8, allophycocyanin-anti-CD8, PE-anti-FoxP3, PE-anti-IL-7Rα, allophycocyanin/Alexa Fluor750-anti-CD8, and PE/Cy7-anti-CD45.1 (eBioscience). Intracellular staining for FoxP3 and Bcl-2 was performed according to the manufacturer’s instructions. Stained cells were analyzed on a FACScan, FACSCalibur, or FACSCanto and data were analyzed using either CellQuest (BD Biosciences) or FlowJo (Tree Star) software. Statistical significance was determined using an unpaired, two-tailed Student’s t test with p < 0.05 being significant (GraphPad Prism version 4 software). Because data were collected in separate experiments (see Fig. 6), each using a single pair of age- and sex-matched mice, paired, two-tailed Student’s t test was used for analysis.
In vitro stimulation assays
Plates (Corning Costar) for stimulation were prepared by coating with avidin (10 μg/ml) in PBS, followed by incubation with biotinylated refolded I-Ek with antigenic peptide (pI-Ek; 14 μg/ml) or biotinylated Cab. Biotinylated pI-Ek was refolded as previously described (21). Unconjugated anti-CD3 (145-2C11; BD Biosciences) and Cab (10 μg/ml) were used in some experiments; no difference between using biotinylated Cab and unconjugated Cab was seen (data not shown).
Plates for stimulation with peptide-pulsed APCs were prepared by plating 5 × 105 DCEK Hi7 cells/well of a 24-well plate and incubating overnight with the indicated concentrations of stimulatory peptides. The fibroblast cell line DCEK Hi7 has been previously described (22). DCEK Hi7 cells were maintained in selection medium containing RPMI 1640, 10% bovine growth serum, 2 mM glutamine (GlutaMAX; Invitrogen Life Technologies), 50 μg/ml gentamicin, 50 μM 2-ME, 0.5 mg/ml hygromycin, and 0.5 mg/ml G418. Peptides were synthesized, purified, and analyzed as previously described (18). The peptide sequence of the antigenic peptide Hb64–76 is as follows: GKKVITAFNEGLK. A weak agonist peptide was generated by substituting a threonine for the asparagine at position 72, such that its sequence is GKKVITAFTEGLK. An antagonist peptide was generated by substituting an isoleucine for the asparagine at position 72, such that its sequence is GKKVITAFIEGLK.
Plates were washed three times before addition of thymocytes. Thymocytes were incubated in I10 (IMDM, 10% FCS, 50 μM 2-ME, 2 mM glutamine, 50 μg/ml gentamicin, 100 μM non-essential amino acids, 1 mM sodium pyruvate, and 10 mM HEPES) overnight in plates. Thymocytes were washed and stained as indicated. Thymocytes were gated for n3.L2+ DP thymocytes and the percentage of cells positive for CD69 was determined using unstimulated cells as a negative control.
Cell death assay.
Thymocytes were incubated overnight in I10 on plates coated with the indicated concentrations of anti-CD3 (145-2C11) and anti-CD28 (37.51; BD Biosciences Pharmingen). Cells were harvested and stained with PerCP-anti-CD4, PE-anti-CD8, annexin V–AlexaFluor 488 (Invitrogen/Molecular Probes). Binding buffer for annexin V was 10 mM HEPES, 140 mM sodium chloride, and 2.5 mM calcium chloride, as recommended by the manufacturer. Data were acquired by flow cytometry and analyzed.
Proliferation and IL-2 production.
Cells isolated from spleens and lymph nodes were purified using CD4+ MACS beads (Miltenyi Biotec) according to the manufacturer’s instructions, except that the column buffer used was 10% bovine growth serum in PBS. Cell purity was consistently >90% as assessed by cytometry (data not shown). Cells were stained for CD4 and for expression of the n3.L2 TCR and populations were normalized such that an equivalent number of CD4+n3.L2+ T cells were added to each well. APCs used for proliferation and IL-2 production were irradiated splenocytes (2000 rad) isolated from B6.K mice.
For proliferation assays, 3–5 × 104 CD4+n3.L2+ T cells were added to each well of a 96-well plate along with 5 × 105 APCs pulsed with the indicated concentrations of antigenic Hb64–76 peptide. Cells were incubated in I10 for a total of 72 h, to which 0.4 μCi/well [3H]thymidine was added for the last 24 h.
For IL-2 production, 5 × 105 CD4+ T cells were added to each well of a 24-well plate along with 2.5 × 106 APCs pulsed with the indicated concentration of antigenic Hb64–76 peptide. Cells were incubated in I10 for ∼36 h, after which time cells were removed from plates, restimulated with PMA and ionomycin in the presence of brefeldin A, and stained for CD4 and n3.L2 on ice and fixed with 2% paraformaldehyde. Cells were then permeabilized with 0.5% saponin and stained for IL-2 using PE-anti-IL-2 (Biolegend). Data were acquired by flow cytometry and analyzed (23). Cells were gated on CD4+n3.L2+ T cells and the percentage of cells positive for IL-2 was determined using stained, unstimulated cells as a negative control.
Generation of bone marrow chimeras
B6.K or B6.K/CD45.1 mice were lethally irradiated (1000 rad) and retro-orbitally injected with 100 μl of HBSS containing unfractionated bone marrow cells from n3.L2 PKCθ−/−RAG1−/− H-2b and n3.L2 RAG1−/− H-2k mice mixed in a 1:1 ratio. The ratio of H-2Kb:H-2Kk bone marrow cells was confirmed by flow cytometric analysis. Thymus and lymph nodes were harvested from recipient mice 5–6 wk after transfer and analyzed by flow cytometry for expression of CD4, CD8, n3.L2 TCR, H-2Kk, and H-2Kb. Statistical significance was determined using a paired, two-tailed, nonparametric test (Wilcoxon-signed rank test), as data were not normally distributed.
Western blot assays were performed essentially as previously described (24). Thymocytes were stimulated on plates coated with 10 μg/ml anti-CD3 (145-2C11) and 10 μg/ml anti-CD28 (37.51) for the indicated periods of time. Stimulation was stopped by lysing cells on plates with lysis buffer (1% Nonidet P-40, 150 mM NaCl, 25 mM HEPES, 1 mM EDTA (pH 7.4), 10 μg/ml leupeptin, 10 μg/ml aprotinin, 1 mM PMSF, and 1 mM sodium orthovanadate). Samples were incubated on ice for 20 min, and particulate matter was removed by centrifugation at 21,000 × g for 10 min at 4°C. Supernatants were mixed with 2× SDS Laemmli sample buffer (Sigma-Aldrich). All samples were then boiled and separated by SDS-PAGE (12% acrylamide, Protogel; National Diagnostics). Proteins were transferred to nitrocellulose membranes (Bio-Rad), blocked with 1:1 PBS:blocking buffer (LI-COR), and then probed with the indicated primary Abs. Abs used for immunoblots included monoclonal anti-ERK (BD Biosciences Pharmingen) and rabbit polyclonal anti-phosphorylated ERK (Cell Signaling Technology). The anti-ERK and anti-phosphoERK Abs recognize both the p42 and p44 MAPKs. Secondary Abs for immunoblots included IRDye 800- conjugated goat anti-mouse IgG (Rockland Immunochemicals) and Alexa Fluor 680-conjugated goat anti-rabbit IgG (Invitrogen/Molecular Probes). Bound Abs were detected with the Odyssey infrared imaging system (LI-COR) according to the manufacturer’s instructions. Integrated intensity of each band was determined using Odyssey software (LI-COR).
Calcium-imaging assays were performed as previously described (23). In brief, thymocytes were loaded with 1 μM fura 2-AM (Invitrogen/Molecular Probes) in Ringer’s imaging solution (150 mM NaCl, 10 mM glucose, 5 mM HEPES, 5 mM KCl, 1 mM MgCl2, and 2 mM CaCl2) and washed immediately before use. The fura 2-loaded thymocytes (2–3 × 106/sample) were added to 8-chambered coverglass slides (Lab-Tek, Nalge Nunc International) that had been coated with peptide-pulsed DCEK Hi7 cells. Calcium imaging was performed on a Zeiss Axiovert 200M microscope equipped with a xenon arc lamp. Fura 2-loaded thymocytes were excited using 340- and 380-nm excitation filters (71000a set; Chroma Technology) and a polychroic mirror (73100bs; Chroma Technology). Fluorescence was passed through a 510 ± 40-nM wide-band emission filter (Chroma Technology) and captured with a Cascade 512B camera (Roper Scientific). Ratio measurements of fluorescence emission (340:380) were recorded at 3-s intervals. Data are presented as the 340:380 ratio, which is proportional to the intracellular calcium concentration. Peak calcium level, mean calcium level, and degree of oscillatory behavior were calculated as previously described (23). Degree of oscillatory behavior was determined by calculating the SD upon linear regression analysis of calcium levels assessed after the initial calcium spike. SDs were calculated by determining the SD of the vertical distances of the data points from the regression line (GraphPad Prism) (23).
EMSA for NF-κB activation.
EMSA to assess activation of NF-κB was performed as previously described (25). Thymocytes were stimulated in serum-free IMDM by incubating 30–40 × 106 cells with soluble, biotinylated anti-CD3 (145-2C11) and anti-CD4 (RM4-5; 10 μg/ml each) and cross-linking with avidin (10 μg/ml). Samples were incubated at 37°C for the indicated time periods. Stimulation with TNF (a gift from R. Schreiber, Washington University in St. Louis, St. Louis, MO) at 20 ng/ml was used as a positive control. Stimulation was stopped by the addition of an excess volume of ice-cold IMDM. Twenty micrograms of nuclear extract was incubated with IRDye700-labeled DNA probe for NF-κB (Ig κ promoter; LI-COR) and IRDye800-labeled Y-box (Eα promoter; LI-COR) DNA probe for the constitutively active NF-Y complex. Protein-DNA complexes were resolved on a 4–15% gradient gel (Bio-Rad) and imaged with the Odyssey infrared imaging system (LI-COR). Integrated intensity of each band was determined using Odyssey software. The intensity of the NF-κB band was normalized to the intensity of the NF-Y band to control for loading of the gel.
Reduction of mature CD4SP thymocytes in n3.L2 PKCθ−/− mice
To determine whether PKCθ was required for maturation of n3.L2+ T cells, thymocyte subpopulations were analyzed by flow cytometry (Fig. 1). The percentage and absolute number of CD4SP thymocytes was significantly decreased in n3.L2 PKCθ−/− mice, although the total numbers of thymocytes isolated were equivalent (Fig. 1, A and B). The decrease in CD4SP thymocytes appeared to be offset by a slight, although not statistically significant, increase in CD4/CD8 DP thymocytes (Fig. 1, A and B). The percentage of TCRhigh cells was reduced in n3.L2 PKCθ−/− mice (Fig. 1,C), and the percentage of TCRhigh cells that were CD4SP was diminished in n3.L2 PKCθ−/− mice (Fig. 1,D). Finally, the absolute number of n3.L2+ CD4SP cells was reduced by ∼50% (Fig. 1,E). The decrease in TCRhigh cells appeared to be offset by an accumulation of TCRint cells in n3.L2 PKCθ−/− thymocytes, consistent with the slight increase in DP thymocytes (Fig. 1, B–D).
The bulk of the CD4SP thymocyte population contains a few cells other than naive α/β TCR CD4SP thymocytes, such as NKT cells, recirculating memory cells, Treg cells, and γ/δ TCR T cells. It was possible that the decrease in n3.L2 CD4SP thymocytes in n3.L2 PKCθ−/− mice reflected a decrease in another of these lineages. Excluding these non-naive CD4+ cells from analysis (26) revealed that 92% of n3.L2 CD4SP thymocytes were naive α/β T cells in n3.L2 WT mice and 94% of n3.L2 CD4SP thymocytes were naive α/β T cells in n3.L2 PKCθ−/− mice (data not shown). Although there was a slight decrease in the number of NKT, γ/δ, Treg, and memory T cells in the n3.L2 CD4SP population in n3.L2 PKCθ−/− mice, this difference was not large enough to explain the 50% decrease in the number of n3.L2 CD4SP cells recovered.
The reduction in numbers of CD4SP thymocytes was reflected in a decreased percentage and number of peripheral n3.L2+ mature T cells (Fig. 2). The total number of T cells isolated from lymph nodes was reduced in n3.L2 PKCθ−/− mice compared with n3.L2 WT, although total splenocyte numbers were equivalent (Fig. 2, B and D). The number of n3.L2+ CD4+ mature T cells present in spleen and lymph nodes of n3.L2 PKCθ−/− mice was reduced ∼50% as compared with n3.L2 WT mice, similar to the 50% reduction of n3.L2+ CD4SP thymocytes (Fig. 2, B and D). Exclusion of NKT, γ/δ, Treg, and memory T cells from the n3.L2 CD4+ T cell population when determining cell numbers did not alter the observed decrease in n3.L2 CD4+ T cells in n3.L2 PKCθ−/− mice (data not shown). Thus, production of n3.L2 CD4SP thymocytes and n3.L2 CD4+ mature T cells was reduced by half in the absence of PKCθ, demonstrating a significant role for PKCθ in the development of CD4+ T cells.
Mature n3.L2 PKCθ−/− T cells deficient in proliferative response and IL-2 production
The developmental defect of thymocytes observed in n3.L2 PKCθ−/− mice varied from previous reports (7, 9). To determine whether mature T cells from n3.L2 PKCθ−/− mice exhibited activation defects similar to those previously demonstrated (7, 9), we assessed IL-2 production and proliferation following TCR stimulation. n3.L2 PKCθ−/− mature T cells did not produce IL-2 (Fig. 3,A) or proliferate (Fig. 3 B) in response to Ag stimulation as robustly as n3.L2 WT T cells. Thus, n3.L2 PKCθ−/− mature T cells that did develop exhibited the same activation defects that have been previously reported (7, 9).
Decrease in n3.L2 PKCθ−/− CD4SP thymocytes due to diminished positive selection
Given the slight increase in DP thymocytes and a significant decrease in CD4SP thymocytes in n3.L2 PKCθ−/− mice (Fig. 1), we hypothesized that DP thymocytes were not efficiently transitioning from the DP to the CD4SP stage due to inefficient positive selection. Positive selection is associated with changes in the expression of surface molecules, including an increase of CD69 and a decrease of CD24 (27, 28). CD69 expression was lower and CD24 expression was higher on both n3.L2+ PKCθ−/− DP and n3.L2+ PKCθ−/− CD4SP thymocytes, compared with n3.L2+ WT DP and n3.L2+ WT CD4SP thymocytes (Fig. 4). These differences in CD69 and CD24 expression indicate that positive selection is reduced in n3.L2 PKCθ−/− mice.
Anti-CD3-mediated up-regulation of CD69 reduced in n3.L2 PKCθ−/− thymocytes
If inefficient positive selection of n3.L2 PKCθ−/− CD4SP thymocytes was due to diminished TCR signaling, then in vitro TCR-mediated stimulation of CD69 up-regulation should have been decreased in n3.L2 PKCθ−/− thymocytes. To test this prediction, thymocytes from n3.L2 WT and n3.L2 PKCθ−/− mice were incubated overnight on plate-bound anti-CD3, anti-TCR, or recombinant I-Ek with peptide (pI-Ek; Fig. 5,A). Regardless of stimulus, up-regulation of CD69 was diminished on n3.L2 PKCθ−/− DP thymocytes, as assessed by both the percentage and the mean fluorescence intensity (MFI) of cells positive for CD69 (Fig. 5 A and data not shown). Up-regulation of CD69 after stimulation with PMA and ionomycin, used as a positive control, was equivalent in PKCθ−/− and WT DP thymocytes (82 and 80%, respectively). The reduction of CD69 up-regulation following anti-CD3 or anti-TCR stimulation is consistent with the hypothesis that TCR signaling is diminished in n3.L2 PKCθ−/− thymocytes.
To further test the hypothesis that the strength of TCR signaling was altered in n3.L2 PKCθ−/− thymocytes, we used APLs previously identified for the n3.L2 TCR (29). If TCR signaling in n3.L2 PKCθ−/− thymocytes was weaker, then up-regulation of CD69 upon stimulation with APLs should be reduced in n3.L2 PKCθ−/− thymocytes, compared with n3.L2 WT thymocytes. However, if TCR signaling was stronger in n3.L2 PKCθ−/− thymocytes, then APL-stimulated up-regulation of CD69 of n3.L2 PKCθ−/− should be increased, compared with n3.L2 WT thymocytes. To differentiate between these possibilities, thymocytes were stimulated with an agonist, a weak agonist, or an antagonist peptide. Stimulation with all APLs revealed diminished up-regulation of CD69 in n3.L2 PKCθ−/− DP thymocytes, compared with n3.L2 WT DP thymocytes, as assessed by both percentage and MFI of cells positive for CD69 (Fig. 5, B and C, and data not shown). The dose response of the agonist peptide was more apparent when assessing the percentage of cells positive for CD69 at low concentrations of agonist peptide (Fig. 5 B). These findings support the hypothesis of weaker TCR signaling in n3.L2 PKCθ−/− thymocytes.
Diminished positive selection of CD4SP and CD8SP thymocytes in nontransgenic PKCθ−/− mice
To examine whether development of CD8SP thymocytes was also dependent on PKCθ, thymocyte populations from nontransgenic PKCθ−/− and B6.K mice were compared (Fig. 6). A decrease in the percentage of cells that were CD4SP or CD8SP and a slight increase in the DP population were again seen (Fig. 6,A). The numbers of CD4SP and CD8SP cells were reduced in PKCθ−/− mice as compared with WT mice by ∼50% (Fig. 6,B). Exclusion of NK1.1+, γ/δ TCR+, CD25+, and CD44+ cells from the analysis revealed that 85% of CD4SP thymocytes in WT mice were naive T cells, whereas 92% of CD4SP thymocytes in PKCθ−/− mice were naive T cells (data not shown). This slight reduction of non-naive T cell numbers was insufficient to explain the diminished numbers of CD4SP thymocytes in PKCθ−/− mice. These data revealed that the diminished positive selection of CD4SP thymocytes in n3.L2 PKCθ−/− mice was not limited to the n3.L2-transgenic system and that the disturbance of thymocyte maturation in PKCθ−/− mice was not restricted to the CD4 lineage. Furthermore, CD69 expression was decreased and CD24 expression was increased on PKCθ−/− CD4SP and CD8SP cells as compared with WT CD4SP and CD8SP thymocytes (Fig. 6, C–F), as was seen in n3.L2 PKCθ−/− CD4SP thymocytes. Diminished thymic production of CD4SP and CD8SP cells was reflected in a decrease in the total number of CD4+ and CD8+ cells isolated from the lymph nodes of PKCθ−/− mice, although no difference in splenocyte numbers was apparent (data not shown). Thus, a dependence for thymocyte development on PKCθ for both CD4SP and CD8SP cells was seen in the nontransgenic PKCθ−/− mice.
Reduction of PKCθ−/− CD4SP thymocytes not due to decreased survival or to change in DN thymocyte population
The decrease in CD4SP thymocytes in PKCθ−/− mice could also be due to decreased survival of PKCθ−/− thymocytes. PKCθ−/− thymocytes could theoretically be more likely to undergo death by neglect or be more sensitive to activation-induced cell death (AICD). To address these possibilities, PKCθ−/− and WT thymocytes were incubated overnight with or without plate-bound anti-CD3 and anti-CD28 stimulation. Cell death was assessed by surface staining with annexin V. The percentage of annexin V-positive PKCθ−/− thymocytes was not increased compared with WT thymocytes, indicating that PKCθ−/− thymocytes were not more likely to undergo apoptosis either from neglect or from AICD (Fig. 7 A). There was also no difference in survival between WT and PKCθ−/− peripheral cells when cells were incubated without stimulation and assessed for cell death using annexin V labeling every 24 h for 96 h (data not shown). Thus, the reduction of CD4+ cell numbers in PKCθ−/− was not due to differential survival.
Additionally, positive selection has been shown to induce the up-regulation of molecules required for survival of SP thymocytes, including IL-7Rα and Bcl-2 (27, 30). To determine whether altered expression of IL-7Rα or Bcl-2 contributed to decreased numbers of CD4SP PKCθ−/− thymocytes, the expression of IL-7Rα and Bcl-2 on thymocytes from PKCθ−/− and WT mice was examined (Fig. 7, B and C). There was a slight decrease in the percentage of PKCθ−/− CD4SP cells that expressed IL-7Rα, but Bcl-2 expression was equivalent in CD4SP thymocytes from WT and PKCθ−/− mice. Similar results were obtained when thymocytes from n3.L2 WT and n3.L2 PKCθ−/− mice were assessed for IL-7Rα and Bcl-2 expression (data not shown). Thus, the 50% reduction of CD4SP thymocytes in PKCθ−/− mice was associated with a failure to progress from the DP to the SP stage, not with decreased survival or increased sensitivity to AICD.
It has been proposed that PKCθ may play a role in Notch signaling, which is required for progression through the DN to the DP stages of thymocyte development (31). To determine whether thymocyte development before positive selection was affected in PKCθ−/− mice, we assessed the expression of CD44 and CD25 on DN thymocytes isolated from n3.L2 PKCθ−/− and n3.L2 WT mice. There was no difference in the distribution of CD25 and CD44 DN populations between thymocytes isolated from n3.L2 WT and n3.L2 PKCθ−/− mice, nor between CD25 and CD44 DN populations in nontransgenic PKCθ−/− and B6.K mice (data not shown). Therefore, thymic maturation appears to progress normally in n3.L2 PKCθ−/− mice until the DP stage.
PKCθ−/− thymocytes are at a significant competitive disadvantage in mixed bone marrow chimeras
Positive selection of n3.L2 PKCθ−/− thymocytes was diminished compared with WT thymocytes but was not absent. Other PKC isoforms may have partially compensated for the loss of PKCθ and partially rescued positive selection of PKCθ−/− thymocytes. However, if PKCθ is the dominant PKC isoform required during positive selection, a more dramatic defect might be revealed by direct competition between PKCθ−/− and WT thymocytes. We therefore isolated bone marrow from n3.L2 PKCθ−/−RAG1−/− H-2b and n3.L2 RAG1−/− H-2k mice, mixed the bone marrow cells in a 1:1 ratio, and transferred the cells into lethally irradiated nontransgenic H-2k recipients. Both n3.L2 PKCθ−/−RAG1−/− H-2b and n3.L2 RAG1−/− H-2k DP thymocytes have equivalent chances to be positively selected in the recipient H-2k background; because the donor bone marrow cells were n3.L2 RAG1−/−, there was no risk for graft-versus-host disease. Thymi and lymph nodes were harvested 5–6 wk after transfer and analyzed for the percentage of n3.L2 DP and n3.L2 CD4SP cells that were positive for either H-2Kb, and thus derived from PKCθ−/− mice, or H-2Kk, and derived from PKCθ-sufficient mice.
There was an 80% decrease in the percentage of n3.L2 CD4SP thymocytes derived from PKCθ−/− mice compared with the percentage of PKCθ−/− n3.L2 DP thymocytes (Fig. 8), indicating that n3.L2 RAG1−/− PKCθ-sufficient thymocytes significantly outcompeted n3.L2 RAG1−/−PKCθ−/− thymocytes during positive selection. The observation that ∼20% of n3.L2 DP thymocytes were derived from PKCθ−/− mice when 50% of transferred bone marrow cells were PKCθ−/− suggests that PKCθ may additionally function in bone marrow engraftment, thymocyte seeding, or thymocyte development before the DP stage. Since there was no difference in the distribution of CD44 and CD25 populations in PKCθ−/− DN thymocytes (data not shown), it seems likely that the additional function of PKCθ occurs before the DN1 stage of thymocyte development. MHC class I molecules are not up-regulated until the later stages of positive selection (17) and we therefore noted n3.L2 CD4SP thymocytes that expressed MHC class I at low levels for which it was not possible to accurately determine H-2Kk or H-2Kb expression (Fig. 8 A). However, the proportion of H-2Kk and H-2Kb cells in the class Ilow population would be predicted to be reflected in the proportion of H-2Kk and H-2Kb cells in the class Ihigh population of CD4SP thymocytes, as well as in the proportion of H-2Kk and H-2Kb cells in the mature CD4+ cells from the lymph nodes. Because the proportion of H-2Kk and H-2Kb cells in n3.L2 CD4 mature T cells was not significantly different from the proportion of H-2Kk and H-2Kb cells in the n3.L2 CD4SP thymocytes, we do not think that the exclusion of MHC class Ilow cells significantly altered our analysis. Comparison of the proportion of PKCθ−/−-derived thymocytes in the CD4SP and DP populations allowed an analysis of the role of PKCθ during positive selection, independent of other possible functions of PKCθ. The large decrease in the percentage of PKCθ−/−-derived cells in the n3.L2 CD4SP population, compared with the percentage of PKCθ−/−-derived cells in the n3.L2 DP population, suggests that PKCθ is the dominant PKC isoform required for TCR signaling during positive selection.
Decreased production of Treg cells in n3.L2 PKCθ−/− mice
Defective production of CD25+FoxP3+ Treg cells in PKCθ−/− mice has been previously reported (16). However, the authors had assumed an equivalent number of CD4SP thymocytes in PKCθ−/− and WT mice and therefore did not account for a possible overall reduction in the number of CD4SP thymocytes when they assessed that PKCθ−/− mice generated fewer Treg cells. To determine whether reduced thymic production of Treg cells in PKCθ−/− mice could be explained by the observed decrease in production of CD4SP thymocytes, we determined the percentage of CD4SP thymocytes that were CD25+FoxP3+ from WT and PKCθ−/− mice (Fig. 9). Even when controlling for the reduction of CD4SP cells in PKCθ−/− mice, there was a dramatic (>90%) decrease in the percentage and number of CD4SP thymocytes that were Treg cells (Fig. 9). There was also an approximate 50% decrease in the percentage and number of Treg cells isolated from the spleens and lymph nodes of PKCθ−/− mice (Fig. 9). Despite the decrease in Treg cell production, PKCθ−/− mice do not suffer from any spontaneous autoimmune disease. Possible explanations for this include the activation defect in mature PKCθ−/− T cells and that the numbers of Treg cells present in the periphery are sufficient to prevent disease. The decrease in thymic production of Treg cells cannot be explained by the observed decrease in generation of CD4SP thymocytes, as the reduction in Treg cell generation is profound even when controlling for the number of CD4SP thymocytes.
Failure to sustain ERK activation in n3.L2 PKCθ−/− thymocytes
To further investigate TCR signaling events that underlie the diminution of positive selection in n3.L2 PKCθ−/− mice, we examined n3.L2 RAG1−/− H-2b thymocytes. n3.L2 thymocytes are not positively selected on the H-2b background and accumulate in the DP stage.
Numerous reports have demonstrated the requirement of sustained ERK activation for positive selection of thymocytes (32, 33, 34, 35, 36). To determine whether the reduction of positive selection in PKCθ−/− thymocytes was associated with reduced ERK activation, nonselected DP thymocytes from n3.L2 PKCθ−/−RAG1−/− H-2b or n3.L2 WT RAG1−/− H-2b mice were stimulated on plate-bound anti-CD3 and anti-CD28 Ab (Fig. 10 A). Activation of ERK was assessed by immunoblot for phosphorylated ERK. Phosphorylation of ERK was not sustained for as long in n3.L2 PKCθ−/−RAG1−/− H-2b thymocytes as in n3.L2 WT RAG1−/− H-2b thymocytes. Failure to sustain ERK activity thus correlated with diminished positive selection in PKCθ−/− thymocytes.
Calcium signaling is also central to TCR signaling in thymocytes during positive selection (36, 37). A previous report has shown that calcium signaling can be regulated by PKCθ (13). Therefore, calcium signaling was assessed using single-cell imaging (23). Nonselected DP thymocytes from n3.L2 PKCθ−/−RAG1−/− H-2b and n3.L2 RAG1−/− H-2b mice were loaded with fura 2 and incubated on APCs pulsed with antigenic peptide at two concentrations (near-maximal stimulation and suboptimal). At the lower concentration of agonist peptide, the calcium mobilization in PKCθ−/− thymocytes peaked at a higher level and was more oscillatory than that of the WT thymocytes (Fig. 10 B). However, at an increased concentration of Ag, the difference in peak calcium levels and oscillatory behavior was not statistically significant. Thus, TCR proximal signaling events leading to calcium signaling were not inhibited in PKCθ−/− thymocytes and may have been enhanced at suboptimal concentrations of agonist peptide. These results are consistent with findings from Manicassamy et al. (13), in which PMA activation of PKCθ could inhibit calcium signaling in mature T cells.
In mature T cells, PKCθ is required for the activation of NF-κB (7, 9). In some systems, NF-κB has been linked to survival of DP thymocytes (39), although NF-κB activity may be dispensable for positive selection (40). To determine whether NF-κB activation was dependent upon PKCθ in thymocytes in our system and to address the conflicting reports of the requirement for NF-κB in positive selection, we assessed the activation of NF-κB in PKCθ−/− DP thymocytes by EMSA. Nonselected DP thymocytes from n3.L2 PKCθ−/−RAG1−/− H-2b and n3.L2 WT RAG1−/− H-2b mice were stimulated with cross-linked anti-CD3 and anti-CD4, or TNF, which stimulates NF-κB activation independently of PKCθ (41). Activation of NF-κB occurred with the same kinetics in PKCθ−/− thymocytes as in WT mice, indicating that in our system TCR stimulation of NF-κB activation is not dependent upon PKCθ (Fig. 10 C). These results are consistent with previous publications (7, 9). Furthermore, there was no correlation between diminished positive selection and activation of NF-κB, consistent with the hypothesis that NF-κB activation is not required for signaling to positive selection (40).
PKCθ plays multiple roles in TCR signaling and T cell function (for reviews, see Refs. 3, 4, 42, 43). Given the central role of PKCθ in TCR signaling in mature T cells, it was surprising that no effect on thymocyte development was appreciated initially. However, after reexamining thymocyte development in PKCθ−/− mice using the class II- restricted transgenic TCR n3.L2, we now report that positive selection is less efficient in the absence of PKCθ. Both n3.L2- transgenic and nontransgenic PKCθ−/− mice produced ∼50% fewer CD4SP thymocytes than did comparable WT mice. We also noted a statistically significant decrease in the number of CD8SP thymocytes generated in the nontransgenic PKCθ−/− mice, indicating that a dependence of thymocyte development on PKCθ was not restricted to the CD4 lineage or the n3.L2-transgenic system. Additionally, analysis of changes in the expression of markers associated with positive selection, such as up-regulation of CD69 and down-regulation of CD24, revealed that positive selection was less efficient in PKCθ−/− mice. Our results may differ from those previously published not only because of the use of a different transgenic TCR, but also because the strain examined here had been extensively backcrossed on the B6 background, whereas previous reports examined PKCθ−/− mice on mixed backgrounds (7, 9).
The decrease in the numbers of CD4SP thymocytes in PKCθ−/− mice was not due to decreased cell survival or to reduced expression of molecules required for survival, such as IL-7Rα or Bcl-2. IL-7Rα is not required for positive selection, but is up-regulated as a consequence of selection (44). Although IL-7Rα expression was slightly reduced on CD4SP thymocytes from n3.L2 PKCθ−/− and PKCθ−/− mice, Bcl-2 expression was equivalent. Because up-regulation of Bcl-2 is dependent upon IL-7R signaling (45, 46), it seems unlikely that the slight diminution of IL-7Rα expression was physiologically relevant. Consistent with normal expression of Bcl-2, thymocytes from PKCθ−/− mice did not exhibit increased susceptibility to apoptosis and in fact may have been slightly less likely than thymocytes from WT mice to undergo AICD (Fig. 7). Reduced AICD would be predicted to increase the number of CD4SP thymocytes in PKCθ−/− mice and may partially obscure the results of diminished positive selection in PKCθ−/− mice.
Partial compensation of PKCθ function by other PKC isoforms may also prevent a complete block of thymocyte development in PKCθ−/− mice (3). Functional redundancy of PKC isoforms appears to be different in thymocytes and mature T cells. For instance, PKCθ is specifically required for activation of NF-κB in mature T cells (7, 9), but not in thymocytes (Fig. 10 and Ref. 7). This difference suggests that either another PKC isoform can substitute for PKCθ in the activation of NF-κB in thymocytes but not in mature T cells or that NF-κB activation is PKC independent in thymocytes. Interestingly, PKCθ−/− thymocytes are at a significant competitive disadvantage with PKC-sufficient cells (Fig. 8), suggesting that even though other PKC isoforms may partially compensate for PKCθ, PKCθ is still the dominant isoform required for TCR signaling in thymocytes. Direct competition between PKCθ−/− and PKCθ-sufficient bone marrow stem cells in bone marrow chimeras also revealed a potential role for PKCθ in either bone marrow engraftment, early hematopoietic stem cell development, or possibly thymocyte seeding that could be further explored.
TCR signal strength determines cell fate during thymocyte development (47, 48). DP thymocytes that express TCRs incapable of recognizing self-MHC molecules fail to receive any signal through peptide-MHC (pMHC)-TCR engagement and are eliminated through death by neglect. DP thymocytes expressing α/β TCRs capable of interacting with self-MHC molecules do receive a signal through pMHC-TCR engagement and are positively selected. Thymocytes bearing autoreactive TCRs receive a strong signal through pMHC-TCR engagement and are deleted by negative selection. To undergo positive but not negative selection, a T cell must therefore bear a TCR that receives enough of a signal from TCR-pMHC engagement to survive but not one strong enough to induce cell death, otherwise termed a weak signal. Complex intracellular signaling events must therefore be orchestrated to encode the strength of pMHC-TCR engagement in developing thymocytes.
Using defined APLs for the n3.L2 TCR, we tested whether the absence of PKCθ resulted in a stronger or weaker signal through the TCR. Because CD69 up-regulation was reduced with all APLs tested, we concluded that TCR signaling was diminished in the absence of PKCθ. Furthermore, we demonstrated a decrease in TCR-stimulated ERK phosphorylation in PKCθ−/− thymocytes, as compared with WT thymocytes. Phosphorylation and activation of ERK appears to be one mechanism by which DP thymocytes encode TCR signal strength. Positively selecting weaker ligands induces lower levels of ERK phosphorylation than do negatively selecting stronger ligands (34, 49, 50). Multiple previous studies have demonstrated that sustained ERK activity is required for positive selection (32, 34, 51). A change in ERK phosphorylation would therefore be predicted to change the outcome of thymic development. PKCθ has been previously demonstrated to positively regulate ERK phosphorylation (12, 52, 53, 54, 55). The mechanism of PKCθ regulation of ERK may be through activation of Ras-GRP (12). We demonstrate here that TCR-mediated phosphorylation of ERK is diminished in PKCθ−/− thymocytes. Reduced activation of ERK in PKCθ−/− thymocytes likely explains the observed decrease in positive selection.
However, not all TCR-mediated signaling events were diminished in PKCθ−/− DP thymocytes. We also examined calcium mobilization and NF-κB activation in PKCθ−/− thymocytes because both signaling pathways have been shown to be dependent upon PKCθ in mature T cells and both may be involved in signaling to positive selection (7, 9, 13, 36, 56, 57, 58). Since a previous study has examined calcium activation in PKCθ−/− thymocytes and found no inhibition in a bulk population of thymocytes (13), we examined calcium signaling at a single- cell level to allow for a more detailed analysis. We also found no inhibition of calcium entry in n3.L2 PKCθ−/− DP thymocytes. Intriguingly, at a low concentration of antigenic peptide, calcium levels seemed to be slightly enhanced in PKCθ−/− thymocytes as compared with WT thymocytes. However, there was no statistically significant difference in calcium responses in PKCθ−/− and WT thymocytes at a higher concentration of peptide. Since calcium entry was not inhibited in PKCθ−/− thymocytes, it seems unlikely that a slight change in calcium signaling explains the observed decrease in positive selection. Also consistent with previously published results, TCR-stimulated NF-κB activation was intact in PKCθ−/− DP thymocytes (7). The decreased TCR signaling in PKCθ−/− DP thymocytes therefore appears to be a specific defect in activation of ERK. The observation that not all TCR signaling pathways were PKCθ dependent most likely explains the observation that up-regulation of some markers associated with positive selection, such as IL-7Rα, were unchanged in PKCθ−/− thymocytes.
Development of CD25+FoxP3+ Treg cells is severely inhibited in PKCθ−/− mice. When the developmental defect of Treg cells in PKCθ−/− mice was initially described, the authors assumed that the generation of CD4SP thymocytes was unaffected. We therefore tested the possibility that the decrease in production of CD4SP thymocytes might explain the decrease in production of Treg cells in PKCθ−/− mice. However, even when controlling for the reduction of CD4SP thymocytes, there was still a profound decrease in the number of Treg cells produced in PKCθ−/− mice. The molecular mechanism for the differential requirement for PKCθ in the development of α/β T cells and FoxP3+ Treg cells is currently unclear. It has been proposed that defective activation of NF-κB in PKCθ−/− thymocytes results in the defective production of Treg cells (16), but we find no inhibition of NF-κB in PKCθ−/− thymocytes in our system. Activation of the Raf-MAPK pathway also appears to be required for full development of the Treg cell compartment (59). Thus, the defective activation of ERK in PKCθ−/− thymocytes may be sufficient to explain the diminished production of Treg cells as well as the inefficient positive selection of FoxP3-negative α/β thymocytes.
By using a class II-restricted transgenic TCR, a previously unrecognized dependence of thymocyte TCR signaling on PKCθ has been revealed. Diminished ERK activation following in vitro TCR ligation in PKCθ−/− thymocytes correlated with inefficient positive selection of PKCθ−/− thymocytes in vivo. Inefficient positive selection of PKCθ−/− thymocytes was not limited to either the n3.L2- transgenic system or to the CD4 lineage, as we found similar results in nontransgenic PKCθ−/− mice on a B6.K background for both CD4SP and CD8SP thymocytes. Therefore, PKCθ is important for both thymocyte and mature T cell TCR signal transduction.
PKCθ−/− mice were generously provided by Dan Littman. We thank Dan Littman and Wojciech Swat for critical reviews of this manuscript. We also thank Darren Kreamalmeyer for technical assistance with management of the mouse colony, Stephen Horvath for technical assistance with genotyping the mice, Jennifer Racz for technical assistance with the generation of bone marrow chimeras, and Andrea Bredemeyer and Beth Helmink for assistance with the EMSAs.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the Pediatric Infectious Diseases Society-St. Jude Fellowship Award for Basic Research (to S.C.M.). P.M.A. is supported by grants from the National Institutes of Health.
Abbreviations used in this paper: PKCθ, protein kinase Cθ; Cab, clonotypic Ab; APL, altered peptide ligand; pI-Ek, I-Ek with the antigenic peptide; DP, double positive; SP, single positive; MFI, mean fluorescence intensity; AICD, activation-induced cell death; Treg, regulatory T; WT, wild type; DN, double negative; pMHC, peptide-MHC.