bone marrow-derived dendritic cell
human serum albumin
lysosome-associated membrane protein-1
We have recently demonstrated that the p38 and ERK1/2 MAP kinases play reciprocal roles in the control of CD1d-mediated antigen presentation. Although the use of specific inhibitors for these pathways clearly had an effect, the effects were not complete, leading to speculations that additional pathways were involved. Here, we show that inhibiting protein kinase C δ (PKCδ) substantially impairs antigen presentation by murine CD1d1 to NKT cells. This effect was accompanied by marked changes in the intracellular localization of CD1d. Expression of a dominant-negative mutant of PKCδ in CD1d+ cells resulted in nearly undetectable endogenous antigen presentation, substantially impaired CD1d recycling, a decrease in MAPK activation, and a decrease in the ability to present low (but not high) concentrations of α-galactosylceramide at the cell surface. These data strongly suggest that PKCδ is a critical regulator of CD1d-mediated antigen presentation and is involved in multiple steps of the process.
Understanding the regulation of antigen presentation is the key to being able to design appropriate strategies aimed at generating an effective immune response against a pathogen, tumor, or other environmental insults. Although the discovery of antigenic peptides presented by MHC class I or class II molecules has gone a long way in helping us to achieve this goal, the lipid antigen presentation system of the CD1 family members is not as well understood. There are two groups of CD1 molecules: Group 1 consists of the CD1a, b, and c glycoproteins, whereas CD1d is the sole member of Group 2 1. It is the latter group that selects a unique T cell subpopulation called “natural killer T” (NKT) cells 1. These cells produce both Th1 and Th2 cytokines promptly after activation and, as such, are important immunoregulatory cells.
Cells are able to sense changes in the extracellular environment (as well as within) by various cell signaling pathways. These pathways are activated by specific stimuli (e.g. stress), resulting in a cellular response. The role(s) that signal transduction pathways play in antigen presentation has only recently been addressed. We have shown that the MAP kinases p38 and extracellular signal-regulated kinase 1 and 2 (ERK1/2) reciprocally regulate CD1d-mediated antigen presentation 2. Although pharmacologic inhibitors specific for these individual MAPK pathways had clear effects on the ability of CD1d+ antigen-presenting cells to stimulate NKT cells, such effects were not complete. Thus, it was likely that other pathways were involved in the control of antigen presentation by CD1d.
Protein kinase C (PKC) molecules comprise a family of serine/threonine kinases involved in cell growth regulation and apoptosis. These are divided into conventional (PKCα, βI, βII, γ), atypical (e.g. PKCζ), and novel (e.g. PKCδ) PKC, with different requirements for Ca++ and/or diacylglycerol for their activity 3, 4. PKCδ plays a role in a number of different cellular functions, and “links up” with other signal transduction pathways, including MAPK 3, 4, by serving as an upstream regulator of these kinases via either direct or indirect mechanisms. The purpose of this study was to determine if PKCδ contributed to the regulation of CD1d-mediated antigen presentation. Our results show that this signal transduction pathway is critical to such control.
Results and discussion
Pharmacologic inhibition of PKCδ impairs CD1d-mediated antigen presentation
PKCδ is specifically inhibited by the drug rottlerin 5. In order to determine if PKCδ played a role in antigen presentation by CD1d, mouse LMTK-CD1d1 cells were pretreated for 2 h with various concentrations of rottlerin, fixed in paraformaldehyde and co-cultured with a panel of murine NKT cell hybridomas. In a dose-dependent manner, rottlerin was able to substantially inhibit antigen presentation by CD1d (Fig. 1A). At a dose of 8 µM, the inhibition was almost complete, with no effect on the surface level of CD1d (Fig. 1C and data not shown). Similarly, rottlerin-treated LMTK-CD1d1 cells poorly stimulated fresh NKT cells (Fig. 1B). This suggested that there may be alterations in CD1d at the intracellular level. Thus, LMTK-CD1d1 cells were pretreated with vehicle (DMSO) or rottlerin and then analyzed by confocal microscopy. CD1d molecules normally traffic through the endocytic pathway 1 and, as expected, co-localized with lysosome-associated membrane protein-1 (LAMP-1) in vehicle-treated cells (Fig. 1D). In contrast, following rottlerin treatment, there was a substantial change in the intracellular distribution of CD1d, with minimal LAMP-1 co-localization, consistent with the ability of this PKCδ-specific inhibitor to impair antigen presentation to NKT cells. Murine DC express high levels of CD1d and are able to effectively stimulate NKT cells 1. Thus, DC derived from the bone marrow of normal C57BL/6 mice were treated in the presence or absence of various concentrations of rottlerin as above and co-cultured with NKT cell hybridomas. Rottlerin substantially inhibited antigen presentation by the murine DC, without a change in CD1d cell surface expression (Fig. 1E and data not shown). Therefore, these results support those obtained with the CD1d+ fibroblasts, and further suggest that PKCδ regulates antigen presentation by CD1d.
Expression of a dominant-negative PKCδ impairs CD1d-mediated antigen presentation
As another approach to analyze the role of PKCδ in regulating CD1d-mediated antigen presentation, LMTK-CD1d1 cells transduced with a control plasmid or with a dominant-negative (DN) form of PKCδ (K376A; DN PKCδ; 6) were co-cultured with NKT cell hybridomas. There was a substantial impairment in the antigen presenting ability of LMTK-CD1d1 cells expressing the DN PKCδ as compared to the control (Fig. 2A). It was clear that the DN PKCδ was inhibiting the activity of the endogenous PKCδ, as there was impaired phosphorylation of its target, MARCKS, when the cells were treated with PMA, which activates the kinase activity of PKCδ (Fig. 2B). Although the level of CD1d on the cell surface of the DN mutant was actually reduced relative to the control cells, there was still a significant amount remaining (Fig. 2C), as another LMTK-CD1d1 control line with comparable CD1d surface expression as the DN mutant was able to stimulate NKT cells at a high level (data not shown). Furthermore, when exogenous α-galactosylceramide (α-GalCer) was used to stimulate NKT cells, it was found that the DN PKCδ-expressing LMTK-CD1d1 cells could stimulate NKT cells almost as well as the control cells (Fig. 2D). Thus, it is highly unlikely that the reason we observed such a decrease in NKT cell activation by the LMTK-CD1d1 cells expressing the DN PKCδ mutant was due solely to less cell surface CD1d. In contrast to these observations, and consistent with an important role for PKCδ in antigen presentation by CD1d, when LMTK-CD1d1 cells were treated with PMA or bryostatin (both drugs capable of stimulating PKCδ), CD1d-mediated antigen presentation was enhanced (data not shown).
The intracellular localization of CD1d in LMTK-CD1d1 cells expressing the DN PKCδ mutant was also altered as compared to control cells, with a significant reduction in the amount of CD1d co-localization with LAMP-1 (Fig. 3A; p = 0.0153) but enhanced co-localization with the Golgi-specific marker, golgi phosphoprotein 4 (Fig. 3C), although this was not found to be statistically significant using a correlation coefficient analysis. No difference in the localization of CD1d in early endosomes (as detected by the EEA-1 marker) between these cells was observed (data not shown). However, the recycling rate of CD1d1 in the DN PKCδ-expressing cells was substantially impaired relative to the control (Fig. 3B). Therefore, these data strongly suggest that PKCδ regulates the recycling rate of CD1d. The defects in recycling, coupled with its altered intracellular localization upon PKCδ inhibition, likely result in a reduced ability of CD1d to be loaded with the proper endogenous NKT cell-stimulating ligand in the correct compartment 1.
As mentioned above, CD1d molecules are structurally similar to both MHC class I and MHC class II molecules. Therefore, it was important to determine whether PKCδ could also affect antigen presentation by these two pathways. To answer this question, LMTK-CD1d1 cells were infected with vaccinia virus (VV) for 6 h, with the last 2 h in the presence of various concentrations of rottlerin. The cells were then fixed in paraformaldehyde, washed and co-cultured with splenocytes from C3H mice infected 6 days previously with VV as a source of VV-specific CTL. The production of IFN-γ was used as a measure of virus-specific T cell recognition. For the analysis of antigen presentation by MHC class II molecules, LMTK-CD1d1-DR4 cells were pulsed with human serum albumin (HSA) in the presence or absence of various concentrations of rottlerin. As shown, rottlerin did not substantially alter antigen presentation by MHC class I molecules (Fig. 1F). In contrast, presentation of HSA by MHC class II molecules to the HLA-DR4-restricted 17.9 T cell hybridoma was reduced in the presence of rottlerin in a dose-dependent manner, albeit not to the same degree as with CD1d (Fig. 1G and data not shown). Thus, these results suggest that PKCδ regulates antigen presentation by molecules that acquire their antigens in the endocytic pathway, but not by those that present endogenous peptide ligands to CTL.
It has been reported that PKCδ plays a role in the control of MAPK-mediated signaling, mainly upstream of p38 and ERK1/2 7–9, and we have recently demonstrated that both p38 and ERK1/2 participate in the regulation of CD1d-mediated antigen presentation 2. Interestingly, the phosphorylation of both p38 and ERK1/2 was decreased in the DN PKCδ-expressing cells as compared to the control (Fig. 3D). Considering the greater apparent sensitivity of ERK1/2 pathway inhibition in the control of antigen presentation by (and intracellular localization of) CD1d 2, it is possible that PKCδ regulation of ERK1/2 activity contributed (at least in part) to our results. Similarly, the inhibition of ERK1/2 pathway control of FcϵRI signaling in mast cells when the activity of PKCδ is substantially reduced 4 is consistent with our data. Therefore, these results suggest that when PKCδ activity is inhibited, MAPK activation is also compromised.
In sum, we have shown that PKCδ is a critical regulator of antigen presentation by CD1d, playing important roles at multiple steps in the process, including intracellular trafficking, recycling, and effects on other signal transduction pathways. As we have also found that antigen presentation by MHC class II molecules is also sensitive to PKCδ inhibition, further studies aimed at investigating the interactions between the players involved in these different pathways will lead to a better understanding of the overall regulation of antigen presentation by molecules utilizing the endocytic pathway.
Very little is known about how signal transduction pathways regulate antigen presentation. Following our previous report demonstrating that the p38 and ERK1/2 MAPK reciprocally control CD1d-mediated antigen presentation 2, we can now include PKCδ as a critical regulator of this system. With PKCδ linked to other signal transduction pathways that include MAPK, dissecting the role for this serine/threonine kinase in the presentation by CD1d to NKT cells (and/or processing of lipid antigens) will likely have important applications to infectious diseases, cancer, and autoimmunity. The additional finding that antigen presentation by MHC class II (but not MHC class I) molecules is also sensitive to PKCδ inhibition broadens the overall biological significance of this signal transduction pathway in the generation of immune responses.
Materials and methods
Female C57BL/6 and C3H/HeJ mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and used at approximately 6 wk of age. All procedures were approved by the Indiana University School of Medicine Institutional Animal Care and Use Committee.
Cell lines, antibodies, and other reagents
The murine fibroblast lines LMTK-VC and LMTK-CD1d1 10, and the murine NKT cell hybridomas DN32.D3 11, 12, N37–1A12, N38–2C12, and N38–3C3 13, have all been described previously. The PKCδ-specific inhibitor rottlerin was purchased from EMD Biosciences, Inc. (San Diego, CA). Antibodies specific for murine CD1d (1H6 mAb 14, and 1B1 from BD Pharmingen, San Diego, CA), FcRγII [2.4G2; American Type Culture Collection (ATCC), Richmond, VA], and LAMP-1 (1D4B; ATCC) were used in flow cytometry or confocal microscopy. Streptavidin-allophycocyanin (SA-APC; BD Pharmingen) was used to detect biotinylated 1B1. For indirect immunofluorescence, PE-conjugated rabbit anti-mouse Ig (Dako, Carpinteria, CA), FITC-conjugated donkey anti-mouse IgG or Texas Red-labeled donkey anti-rat IgG (Jackson ImmunoResearch, West Grove, PA) were used. For analyzing co-localization of CD1d with early endosomes or Golgi, rabbit anti-EEA-1 (Affinity BioReagents, Golden, CO) or anti-GOLPH4 (Abcam, Cambridge, MA) antibodies were used, respectively, followed by the Texas Red conjugate used above. Measuring IL-2 production from the NKT cell hybridoma panel by ELISA was done as described 14. α-GalCer was prepared as described 15. For fresh NKT cells, liver mononuclear cells from C57BL/6 mice were harvested and electronically sorted using α-GalCer-loaded CD1d1 dimers as reported previously 16. Co-cultures with fresh NKT cells were incubated for 24 h and IFN-γ production was measured by ELISA.
LMTK-CD1d1 cells were transduced with a retrovirus expressing the DN form of murine PKCδ (K376A; 6) subcloned into the pMSCV-puro retrovirus vector (Clontech, Mountain View, CA) and prepared in E-86 ecotropic packaging cells (ATCC). Transduced cells were selected in the presence of 2 µg/mL puromycin.
Cytotoxic T lymphocyte and MHC class II-mediated antigen presentation assays
LMTK-CD1d1 cells were mock-infected or infected with VV (MOI = 5) for 6 h. During the last 2 h, rottlerin was added at various concentrations, and the cells were then washed, fixed and co-cultured with splenocytes from C3H/HeJ mice on day 6 post VV infection (1 × 106 PFU/mouse, i.p.) for 24 h. IFN-γ production was measured by ELISA. For analyses of antigen presentation by MHC class II molecules, LMTK cells expressing CD1d1 and/or HLA-DR4 10 were pulsed overnight in the presence or absence of 10 µg HSA (Sigma). The cells were then treated with various concentrations of rottlerin for 2 h, fixed and co-cultured with the 17.9 murine T cell hybridoma specific for HSA presented by HLA-DR4 (kindly provided by Dr. J. Blum, Indiana University). Following an overnight incubation at 37°C, IL-2 secretion was measured by ELISA.
Generation of bone marrow-derived DC
Bone marrow cells from C57BL/6 mice were cultured in RPMI 1640 supplemented with 2 mM L-glutamine, 50 µM 2-mercaptoethanol, 10% FBS and antibiotics, as well as 10 ng/mL each of murine GM-CSF and IL-4. On day 6, 1 µg/mL LPS was added. On day 7, the plates were gently flushed (3–4 times) to remove the loosely adherent cells, which were subsequently used in analyses as bone marrow-derived DC (BMDC).
Flow cytometry and confocal microscopy
Cells used for flow cytometry were treated with the indicated antibodies as described 14. BMDC were first exposed to 2.4G2 supernatant for 30 min on ice, before being stained with PE-conjugated mAb for direct immunofluorescence. Confocal microscopy was performed as described 2. Analysis of the relative level of CD1d co-localization with LAMP-1 or GOLPH4 was performed using MetaMorph software (version 5; Molecular Devices, Sunnyvale, CA).
CD1d recycling assay
The CD1d recycling assay was done essentially as described 17. Briefly, CD1d molecules on control and DN PKCδ-expressing LMTK-CD1d1 cells pretreated for 30 min with 25 µg/mL cycloheximide were blocked using saturating amounts of purified rat anti-mouse CD1d mAb (1B1) on ice for 30 min, washed and incubated at 37°C in the presence of cycloheximide. At the indicated time points, an aliquot of the cells was harvested and kept on ice. After the last time point, the cells were washed, fixed in 1% paraformaldehyde, stained with PE-conjugated 1B1 and analyzed by FACS for the determination of “% CD1d recycled”.
Western blot analysis
Western blots for the analysis of p38 and ERK1/2 activation were done as previously described 2, using antibodies specific for the phosphorylated and “total” forms of these MAPK (Cell Signaling Technology, Beverly, MA). For the detection of phospho-MARCKS, serum-starved control or DN PKCδ-expressing LMTK-CD1d1 cells were treated in the presence or absence of PMA (100 nM) for 30 min. The phosphorylation of MARCKS was detected by using specific antibodies (Cell Signaling Technology). GAPDH expression was measured to ensure equal loading.
For statistical analyses, Student's t-test was performed using GraphPad PRISM software (version 4.00 for Windows; GraphPad, San Diego, CA). A p value below 0.05 was considered significant. The error bars in the bar graphs show the standard deviation from the mean.
This work was supported by NIH grants RO1 AI46455, RO1 CA89026, and PO1 AI056097 to R.R.B., and NSF CHE-0194682 from the National Science Foundation to J.G.-H. M.R.P. and G.J.R. were supported by NIH training grant T32 DK07519. R.R.B. is a Scholar of the Leukemia and Lymphoma Society.