|Home | About | Journals | Submit | Contact Us | Français|
The chemokine receptor CXCR3 is expressed on the surface of both resting and activated T- lymphocytes. We describe here a study of the endocytosis of CXCR3 using T-lymphocytes and CXCR3 transfectants. Chemokine-induced CXCR3 downregulation occurred in a rapid, dose-dependent manner, with CXCL11 the most potent and efficacious ligand. Endocytosis was mediated in part by arrestins, but appeared to occur independently of clathrin and caveolae. In contrast to other chemokine receptors, which are largely recycled to the cell surface within an hour, cell surface replenishment of CXCR3 occurred over several hours and was dependent upon mRNA transcription, de novo protein synthesis and transport through the ER and Golgi. Confocal microscopy and Western blotting confirmed the fate of endocytosed CXCR3 to be degradation, mediated in part by lysosomes and proteosomes.
Site-directed mutagenesis of the CXCR3 C-terminus revealed that internalization and degradation were independent of phophorylation, ubiquitination or a conserved LL motif. CXCR3 was found to be efficiently internalized in the absence of ligand, a process involving a YXXL motif at the extreme of the C-terminus. Although freshly isolated T-lymphocytes expressed moderate cell surface levels of CXCR3, they were only responsive to CXCL11 with CXCL9 and CXCL10 only having significant activity on activated T-lymphocytes. Thus, the activities of CXCR3 are tightly controlled following mRNA translation. Since CXCR3+ cells are themselves a source of IFN-γ, which potently induces the expression of CXCR3 ligands, such tight regulation of CXCR3 may serve as a control to avoid the unnecessary amplification of activated T-lymphocyte recruitment.
The chemokine receptor CXCR3 is expressed on a wide variety of cells including activated T-lymphocytes, natural killer cells, malignant B-lymphocytes, endothelial cells and thymocytes (1-6). Three major CXCR3 ligands, CXCL9, CXCL10 and CXCL11, have been identified, all of which are induced by IFN-γ and are therefore thought to promote Th1 immune responses (7-9). Recent studies have shown that the CXCR3 ligands exhibit unique temporal and spatial expression patterns, suggesting that they have non-redundant functions in vivo. Moreover, the CXCR3 ligands share low sequence homology (around 40% amino acid identity) and exhibit differences in their potencies and efficacies at CXCR3 with CXCL11 being the dominant ligand in several assays (8, 10). CXCR3 and its ligands have been implicated as playing an important role in the induction and perpetuation of several human inflammatory disorders including atherosclerosis (11), autoimmune diseases (12), transplant rejection (13, 14) and viral infections (15). Consequently, the mechanisms underlying the regulation of CXCR3 expression at the cell surface are of considerable interest.
The number of receptors on a cell surface results from a balance between the rate of internalization and the rate of replacement (recycling and synthesis of nascent receptor). Following ligand binding, there are two major routes whereby G protein coupled receptors (GPCRs), as typified by chemokine receptors, are internalized into cells. The first and most well-defined route involves the binding of arrestin to the phosphorylated receptor, which in turn initiates the internalization process by binding to clathrin. The receptor-arrestin complex is then sequestered in clathrin-coated pits. This pathway is often considered a default system for degradation and recycling of receptors (16, 17). The second pathway involves invaginations of the cell membrane known as caveolae and functions independently of clathrin coated pits (18). Although the rate of internalization of a receptor is an important factor in determining its level at the cell surface, the rate of recycling and the rate of synthesis of new receptors are also important. Until recently, the mechanisms of the recycling process were poorly understood and internalized receptors were thought to have several potential fates. The concept of two different classes of receptor (as distinguished by their recycling) has been introduced recently, in which class A receptors traffic to recycling endosomes and are rapidly returned to the cell surface (16). In contrast, class B receptors are dephosphorylated in endosomes followed by slow recycling back to the plasma membrane. Sequentially, the receptors pass through late endosomes and the Golgi and finally are transported back to the cell surface. Another potential fate is that of degradation which may be perceived to down regulate receptor expression. To-date, protein synthesis has not been shown to play a role in GPCR replenishment (19-21).
Here we show that CXCR3 is internalized both constitutively and following incubation with CXCL11, resulting in degradation of the receptor. We also show that in the absence of detectable recycling, cell surface replenishment of CXCR3 is dependent upon de novo protein synthesis.
All chemicals unless otherwise stated were purchase from Sigma-Aldrich (Poole, UK). Chemokines were purchased from PeproTech EC (London, UK). Filipin, sucrose, nystatin, monensin and nocodazole were purchased from Sigma-Aldrich or Calbiochem (Nottingham, UK). Brefeldin A, actinomycin D and bafilomycin A1 were obtained from Tocris (Avonmouth, UK) Cycloheximide was from ICN Biomedicals Inc, (Aurora, Ohio). The mouse anti-human CXCR3 mAb (clone 49801.111) and the mouse isotype-matched control IgG1 (MOPC 21 clone) were obtained from Sigma-Aldrich. The anti-HA.11 antibody was from Covance (Berkeley, California), and the anti-tubulin α antibody was from Abcam (Cambridge, UK). Secondary antibodies were obtained from Dako Cytomation, (Ely, UK). Plasmids encoding dominant negative mutants of β-arrestin 1 and β-arrestin 2 were kind gifts of Dr Marc Caron, Duke University Medical Center Durham, NC. Plasmids encoding the fusion proteins GFP-DIII and GFP-D3Δ2 were kind gifts of Dr Alexandre Benmarah, Institute Cochin, Paris, France.
The murine pre-B cell line L1.2 was maintained as previously described in RPMI supplemented medium (22). L1.2 cells stably transfected with pCDNA3 containing the CXCR3A cDNA HA-tagged at the N-terminus (10) were cultured in the same medium with the addition of 1 mg/ml Geneticin (G418) to maintain selection. Mutant CXCR3 constructs were generated by site-directed mutagenesis using the Quikchange mutagenesis kit (Stratagene, La Jolla, CA) with the pCDNA3 HA-CXCR3A plasmid as template. Transient transfection of L1.2 cells with plasmids was carried out by electroporation as previously described (23). Cells were cultured for 24h in medium supplemented with 10 mM sodium butyrate prior to use, to enhance cell surface receptor expression. Mouse embryonic fibroblasts (MEFs) derived from both WT and mice deficient in β-arrestins 1 and 2 were a kind gift of Dr Robert Lefkowitz, Duke University Medical Center, North Carolina and were maintained as previously described (24). Transfection of MEFs was by electroporation as previously described (25). Cells were cultured over night in medium supplemented with 10 mM sodium butyrate prior to use, to enhance cell surface receptor expression.
For the generation of activated T-lymphocytes, peripheral blood mononuclear cells (PBMCs) were isolated from blood sampled from healthy donors according to the Royal Brompton Hospital Ethics Committee approved protocol as previously described (26). Lymphocytes were separated from monocytes by allowing the latter to adhere to a tissue culture flask for 2h at 37°C and were activated by culture in presence of 100 IU/ml IL-2 and 2mg/ml concanavalin A for 7-10 days. Purified T-lymphocytes were isolated from whole blood using the Rosette-Sep Human T cell Enrichment Cocktail kit (Stem Cell Technologies, Grenoble, France) which typically gave a population more than 95% pure. Nucleofection of purified T-lymphocytes with plasmids encoding GFP-DIII and GFP-D3Δ2 was achieved by using an Amaxa nucleofector, according to the manufacturer's instructions, using program U-014 which typically gave around 60% cell viability as deduced by staining with the dye To-Pro3 (Invitrogen, Paisley, UK).
Internalization assays were essentially carried out as previously described by Sauty et al (27). Activated T-lymphocytes or L1.2-CXCR3 cells were incubated with serum-free medium for 1 hour at 37°C and then resuspended in medium without serum at 5×106 cells/ml. Cells were then incubated with chemokines (50nM) for various times at 37°C, and washed in ice cold PBS containing 1% FCS and 1% NaN3 prior to flow cytometry analysis. Cell surface expressed CXCR3 was detected using an anti-CXCR3-A antibody and FITC conjugated anti-mouse IgG. Samples were quantified on a FACSCalibur and data processed with CellQuest software version 3.1 (Becton Dickinson, San Jose, CA) with dead cells excluded from analysis. The relative CXCR3 surface expression was calculated as 100 × [mean channel of fluorescence (stimulated) - mean channel of fluorescence (negative control)/mean channel of fluorescence (medium) - mean channel of fluorescence (negative control)] (%). Pilot experiments staining CXCR3 tranfectants on ice in either the presence or absence of 50nM CXCL11 confirmed that binding of ligand by CXCR3 did not significantly reduce detection by the primary antibody (data not shown). Where inhibitors were employed, cells were incubated for 1 hour at 37°C with, filipin (5μg/ml), nystatin (50μg/ml), monensin (50μM), sucrose (0.4M) or cycloheximide (10μg/ml) before assays of receptor downregulation were performed.
Receptor downregulation was initiated as described above. After 30 minutes incubation with chemokines, the cells were washed three times in medium without FCS and resuspended in medium without FCS and incubated at 37°C. To remove CXCL11 from endogenous GAGs, activated T-lymphocytes were washed once in pre-warmed 0.5M NaCl/RPMI as previously described (27) then twice in RPMI and incubated at 37°C. Samples were taken at different time points and cells were washed in PBS buffer containing 1% FCS and 1% NaN3. Cells were stained with antibodies as described. Where inhibitors were employed, brefeldin A (5μM), actinomycin D (5μM), bafilomycin A1 (100nM) or cycloheximide (10μg/ml) were added to the cells during the recovery phase, following induction of CXCR3 down regulation by ligand.
The H9 human T cell lymphoma line was washed in RPMI and resuspended at a concentration of 5×106 cells/ml in serum free RPMI and were incubated at 37°C with 50nM CXCL11. Samples were removed either before or at the indicated times following the addition of CXCL11. Internalization buffer (1% FCS, 1% NaN3 in PBS) was added to the samples removed and tubes were incubated on ice until all time points were collected. Each timepoint was divided into either isotype control or antibody staining tubes. Cells were washed twice in cold PBS prior to fixation in 4% PFA for 20 minutes on ice. Cells were then washed in PBS and permeabilised in 0.5% saponin buffer containing anti-human LAMP-1 antibody (BD Pharmingen, 1:20 dilution) or an equal concentration of mouse IgG1 isotype control. After incubation with the primary antibody, cells were washed in saponin buffer prior to incubation with goat anti-mouse IgG alexa fluor 568 (Invitrogen, 1:100 dilution in 0.5% saponin buffer). Cells were washed and then pre-blocked with mouse IgG before being incubated with mouse anti-human CXCR3 FITC (R&D Systems, 1:10 dilution in 0.5% saponin) or isotype control. Cells were then resuspended in 4% PFA and were spun onto poly-L-lysine coated glass coverslips in 24 well tissue culture plates at 1200rpm for 5 minutes. The supernatant was removed and the coverslips washed twice in PBS and once in deionised water before being removed from the wells allowed to air dry and mounted onto slides in Vectorshield hardset fluorescence mounting medium (Vector Laboratories, Peterborough, UK). Analysis was carried out by confocal microscopy using a Leica TCS NT confocal microscope with a 40x oil objective. Image analysis was carried out using Leica LCS Lite software Version 2.61 and the images manipulated for presentation using Adobe Photoshop version 6.0.
L1.2 transfectants expressing WT CXCR3 and CXCR3-AAA, −K324R, −Δ4, and −Δ34 constructs were washed in RPMI and resuspended at 5×106 cells/ml in serum free RPMI containing 10μg/ml cycloheximide. Where indicated, 40μM MG132 or 200μM chloroquine was also added. The cells were preincubated for 30 min at 37°C before the addition of 50nM CXCL11. Samples were taken at the indicated time points, the cells washed in ice cold PBS and resuspended in lysis buffer containing 1% N-dodecyl-β maltoside, 10% glycerol, 1/1000 protease inhibitor cocktail in PBS (28). Equal quantities of cell lysates were separated on 4-12% SDS-PAGE gels and were electrophoretically transferred onto a nitrocellulose membrane, which was subsequently blocked with 5% milk in 0.01M PBS with 0.05% Tween 20. The blots were independently probed with either anti-HA (Covance, 1:1000 dilution) or anti-α tubulin (Abcam, 1:10000 dilution) as a loading control. Following washing and probing with a secondary HRP-conjugated polyclonal goat anti-mouse immunoglobulins (1:1000 dilution), blots were developed by enhanced chemiluminescence (GE Healthcare, UK).
Chemotaxis assays using either CXCR3 transfectants or purified T-lymphocytes were performed essentially as previously described (10, 29) using ChemoTX plates with a 5μm pore size, purchased from Neuroprobe (Gaithersburg, MD). For T-lymphocyte migration, enumeration was carried out using a hemocytometer and cell migration to buffer alone was subtracted from the resulting data, with individual results expressed as a percentage of the total cells applied to the filter. For L1.2 transfectant chemotaxis, the same apparatus was used, although at the end of the assay, cells were transferred from the lower chamber to a white 96-well microtiter plate using a funnel plate (Neuroprobe), and cells were detected with CellTiter Glo (Promega, Southampton UK). Luminescence was measured using a TopCount microplate scintillation & luminescence counter (Perkin-Elmer, Monza, Italy). Data are described are expressed as a chemotactic index, relative to migration observed to medium alone.
125I-CXCL11 and 125I-CXCL10 were purchased from PerkinElmer Life Sciences (Boston, MA, USA. Ligand binding was performed as previously described using centrifugation through oil to separate bound chemokine from free chemokine (26). Data are presented following the subtraction of non-specific binding, taken as the counts obtained when the labelled chemokine was displaced by a 1000 fold excess of homologous cold chemokine.
Data were analysed using Prism 4.0 (GraphPad Software, San Diego, CA) using ANOVA with Bonferroni's Multiple comparisons test.
We initially employed activated PBMCs to investigate the process of CXCR3 downregulation following incubation of the cells with the three natural ligands described to date for CXCR3, namely CXCL9/Mig, CXCL10/IP-10 and CXCL11/I-TAC. Activated PBMCs, cultured for 7-10 days with Concanavalin A and IL-2, readily expressed CXCR3 on their cell surface as detected by flow cytometry using a specific mAb. Incubation of PBMCs with all three CXCR3 ligands induced a dose-dependent loss of CXCR3 from the cell surface (Figure 1A) as deduced by staining with the same CXCR3-specific mAb.
Notably, CXCL11 was the most efficacious ligand, with a 50nM concentration of CXCL11 reducing cell-surface staining to less than 20% of their starting levels. We subsequently examined the kinetics of this response, using 50nM concentrations of each ligand (Figure 1B). Loss of cell surface CXCR3 occurred rapidly, with optimal downregulation observed by 30 minutes in agreement with a previous study (27). Similar data was also obtained using a previously described L1.2 cell line stably expressing CXCR3 (data not shown) with CXCL11 again the most efficacious of the ligands, although the maximum level of receptor down regulation observed was reduced to approximately 50% of starting levels as we have previously observed with L1.2 cells expressing the related receptor CCR3 (30). This likely reflects less efficacious coupling of the human receptor to murine intracellular machinery in the transfectants. In all subsequent experiments, we therefore incubated cells with 50nM CXCL11 to achieve optimal CXCR3 downregulation.
Two major pathways are known by which chemokine receptors are internalized; either via clathrin coated vesicles following the clathrin mediated endocytic pathway or via caveolae. Hypertonic sucrose medium has been shown to block the assembly of clathrin coated pits (31) whereas internalization via caveolae can be inhibited with filipin or nystatin (32). Monensin is an inhibitor of vesicle acidification, a process essential for the sorting events occurring during endocytosis of GPCRs such as the β2-adrenergic receptor (33). We assessed the activity of these inhibitors on CXCR3 downregulation in PBMCs and L1.2 CXCR3 transfectants (Figures 1C and 1D). Neither filipin nor nystatin had any inhibitory effect on CXCR3 downregulation in either cell type, suggesting that caveolae are not involved in the endocytosis of CXCR3. Whilst sucrose had little effect on ligand-induced CXCR3 down regulation in PBMCs (Figure 1C) it was observed to significantly reduce the levels of CXCR3 internalization in L1.2 CXCR3 cells following treatment with CXCL11 (Figure 1D). In L1.2 transfectants, monensin treatment significantly reduced CXCL11-induced internalization of the receptor, suggesting that vesicular acidification is necessary for the sorting events occurring following CXCR3 endocytosis. In PBMCs, monensin had a modest inhibitory effect on CXCR3 endocytosis which did not reach statistical significance. Since it has been previously demonstrated that cholesterol and lipid rafts are requited for the maintenance of chemokine receptor conformation (34, 35) we also sought to examine the effects of filipin and nystatin on ligand binding. Whilst nystatin treatment altered the density of CXCR3 tranfectants making them unable be centrifuged through oil in our ligand binding assay (data not shown) treatment of cells with filipin was observed to have little effect on CXCL11 binding (Figure 1E).
Since sucrose and monensin were without effect on CXCR3 internalization in PBMCs we sought to confirm our findings by using an alternative strategy to inhibit clathrin. T-lymphocytes were purified from blood and underwent nucleofection either in buffer alone (mock nucleofection) or in buffer containing plasmid encoding a GFP-tagged construct, DIII (DIII transfection). This construct inhibits clathrin-coated pit assembly and therefore clathrin-dependent internalization (36, 37). Twenty-four hours after nucleofection, cells were harvested and incubated at 37°C either in the presence or absence of CXCL11 prior to staining for CXCR3 expression. A significant percentage of mock-nucleofected T-lymphocytes were shown to express CXCR3 (Figure 2A, upper left quadrant) which was seen to be reduced following CXCR3 treatment (Figure 2C, upper left quadrant). Nucleofection of T-lymphocytes led to the identification of two populations of CXCR3 positive cells, a major population not expressing the DIII-GFP fusion protein (Figure 2B, upper left quadrant) and a minor population expressing the DIII-GFP fusion protein (Figure 2B, upper right quadrant). CXCR3 treatment was seen to significantly reduce numbers of cells within both populations (Figure 2D, upper right and left quadrants). A similar lack of effect upon CXCL11-induced CXCR3 endocytosis was also seen following nucleofection of T-lymphocytes with the control protein GFP-D3Δ2 which corresponds to the GFP-DIII construct lacking all AP-2 binding sites (data not shown). Collectively this suggests that ligand-driven endocytosis of CXCR3 in T-lymphocytes occurs independently of clathrin.
The clathrin dependent pathway for endocytosis of GPCRs typically involves the binding of arrestins to the intracellular face of the phosphorylated receptor. To examine whether CXCR3 internalization is dependent upon arrestins, we transiently transfected mouse embryonic fibroblasts (MEFs) obtained from both wildtype (WT) and β-arrestin 1 and 2 deficient mice. Internalization was then induced by incubation with CXCL11 and CXCR3 downregulation assessed as before by flow cytometry. CXCR3 downregulation in WT MEFs was similar to that seen in L1.2 transfectants, with CXCL11 reducing cell surface levels to around 50% of their starting levels (Figure 2E). In MEFs from β-arrestin 1 and 2 deficient mice, CXCR3 downregulation in response to ligand was observed, but at a reduced level, with only a 20% reduction of CXCR3 cell surface levels in response to CXCL11, suggestive of an incomplete requirement for β arrestin in the down modulation process. Similarly, transfection of L1.2 CXCR3 transfectants with either empty plasmid or plasmids encoding the V53D and V54D dominant forms of β-arrestin 1 and β-arrestin 2 were without effect upon CXCR3 downregulation induce by CXCL11 (Figure 2F). Collectively, the data suggest the existence of a β-arrestin-independent pathway for the endocytosis of CXCR3.
After the induction of downregulation by ligand, the recovery of cell surface CXCR3 levels was relatively slow in both PBMCs and L1.2 transfectants (Figure 3A & B), with only about 70-80% recovery of the original CXCR3 cell surface levels observed a full 3 hours after incubation with CXCL11. This was in contrast to another Th1-expressed chemokine receptor, CXCR6, which showed 100% recovery of cell surface levels within an hour of ligand-induced downregulation (Figure 3C) and is typical of receptor recycling to the cell surface as described for other chemokine receptors (19, 38, 39). The slow recovery of cell surface CXCR3 levels suggested to us that upon ligand-induced internalization, CXCR3 is either slowly recycled, as is the case for Class B GPCRs such as the vasopressin type 2 receptor (40) or alternatively, is degraded. In the case of degradation, cell surface replenishment would therefore require de novo synthesis of receptor.
To examine this latter hypothesis, we pre-incubated the cells for 1 hour with cycloheximide, induced CXCR3 internalization with CXCL11 and let the cells recover in the presence of cycloheximide. Treatment with cycloheximide ablated the recovery of CXCR3 in both PBMCs and L1.2 transfectants whilst recovery of cell surface CXCR6 levels to 80% of the starting levels was observed at the 2 hour point. This latter value was approximately half of the levels of staining seen at the same time point with CXCR6 transfectants that had not been treated with cycloheximide, suggesting that both recycling and de novo synthesis cooperate in maintaining CXCR6 cell surface levels in the transfectant system used (Figure 3C).
Thus, in contrast to CXCR6 and other chemokine receptors described in the literature (19, 20, 41), cell surface replenishment of CXCR3 is dependent upon de novo protein synthesis. If this postulate is true, then CXCR3 cell surface replenishment should also be dependent upon mRNA transcription and efficient transport through the ER and Golgi. We therefore pre-incubated PBMCs or L1.2 transfectants for 1 hour in the presence or absence of actinomycin D (an inhibitor of transcription), or brefeldin A and bafilomycin A1, which have been shown to inhibit function of the ER and Golgi-apparatus respectively and therefore inhibit transport of receptors through these compartments (41, 42). Internalization was induced with 50nM CXCL11 and the expression of CXCR3 was monitored at 3 hours post internalization. In PBMCs cell surface replenishment of CXCR3 was also significantly inhibited, although not reduced to basal levels (Figure 4A). In L1.2 transfectants, cell surface CXCR3 levels remained at baseline following incubation with CXCL11 in the presence of either actinomycin D, brefeldin A or bafilomycin A1 (Figure 4B). Collectively, the data suggest that the observed recovery of CXCR3 at the cell surface is dependent upon newly-synthesised receptor trafficking though functional Golgi-apparatus in the cell, in contrast to chemokine receptors such as CCR4, CCR5 and CXCR6 which appear to be replenished by a recycling mechanism (19, 20).
Since the C-terminus of several GPCRs has been implicated in the internalization process, we sought to examine the role of this motif in the internalization of CXCR3. Site-directed mutagenesis of the CXCR3 cDNA was performed to generate 4 mutant constructs (Figure 5A). The first of these mutated a triple LLL motif to AAA, thereby losing two potential LL motifs previously reported to be involved in CXCR2 internalization (43). The second mutation targeted the sole intracellular lysine residue, K324. Ubiquitination of internalized GPCRs has been shown to target them for degradation, a process whereby the 74 amino acid ubiquitin is covalently attached to intracellular lysine residues. The two remaining mutations introduced premature stop codons within the cDNA, truncating the receptor by either 4 amino acids (Δ4 construct) or 34 amino acids (Δ34 construct). These removed a conserved YXXL motif at the extreme C-terminus and the entire repertoire of C-terminal serine and threonine residues respectively, the latter construct allowing us to examine the requirement for phosphorylation of CXCR3 in the internalization process. All four mutants were transiently expressed in L1.2 cells, and cell surface expression monitored by flow cytometry. All four mutants trafficked to the cell membrane although the Δ34 mutant was expressed at levels significantly below those of WT CXCR3 (Figure 5B). Conversely, the Δ4 mutant was consistently expressed at greater levels greater than WT CXCR3, although this did not reach significance. All four constructs were able to mediate chemotaxis of cells in response to CXCL11, with the typical bell-shaped responses having optima around the 3nM concentration (Figure 5C). Likewise, internalization of CXCR3 in response to CXCL11 was unimpaired by mutation, with the 50 and 100nM concentrations of ligand inducing significant internalization compared to untreated cells (Figure 5D). Since the Δ4 construct appeared to be expressed at higher levels than WT CXCR3 (Figure 5B), we postulated that CXCR3 might be internalized constitutively, i.e. in the absence of ligand, and the loss of the 4 most C-terminal residues might inhibit this process. We subsequently examined the expression of both WT CXCR3 and the Δ4 construct over a 6 hour period, following pre-treatment with cycloheximide to inhibit de novo synthesis. WT CXCR3 was seen to be quite rapidly lost from the cell surface in the absence of ligand, with approximately half of the original cell-surface levels of CXCR3 observed after 4 hours incubation. In comparison, internalization of the Δ4 construct was less efficacious, with the remaining cell surface levels of mutant receptor at the 6 hour time point, significantly greater than those of WT CXCR3.
We subsequently turned our attention to the fate of CXCR3 following its internalization, using confocal microscopy to examine intracellular localisation of the receptor in permeablized T-lymphocytes. A predominantly granular intracellular staining pattern for CXCR3 was evident in untreated cells, (Figure 6A) identical to that previously described by Gasser and colleagues (44). Likewise, a similar pattern was seen for staining with the late endosome marker lamp-1 (Figure 6B) with little co-localisation of signals seen (Figure 6C). Treatment with CXCL11 for 15 minutes resulted in clustering of lamp-1+ vesicles with apparent co-localisation of CXCR3 with lamp-1 in some but not all cells (Figure 6D-F). This may reflect either a rapid loss of CXCR3 immunoreactivity following trafficking to lysosomes or the fact that this pathway is not the sole route of CXCR3 degradation. Little if any colocalization of CXCR3 with lamp-1 staining was observed in cells 60 minutes following treatment with CXCL11, suggesting that degradation of CXCR3 may be complete by this point (Figure 6G-I).
To further examine CXCR3 degradation, we utilized Western blotting methodologies. HA-tagged CXCR3 was expressed transiently in L1.2 cells and following pre-treatment with cycloheximide to inhibit de novo synthesis, the cells were incubated for varying time periods in the presence or absence of CXCL11. Cell lysates were then examined by SDS-PAGE, followed by Western blotting. As can be seen in Figure 7A, CXCR3 appears as a band of approximately 50 kDa. Following 3 hrs incubation at 37°C, either in buffer alone or supplemented with CXCL11, the band representing CXCR3 was seen to reduce considerably in intensity, suggestive of a degradative fate. Additional pre-treatment of CXCR3 transfectants with either the proteosome inhibitor MG132 or the lysosomal inhibitor chloroquine, prior to treatment with CXCL11, was observed to inhibit the degradative process (Figure 7B). We subsequently examined the panel of 4 C-terminal CXCR3 mutants to examine the effects of mutation upon degradation. Compared to untreated cells, obvious degradation of each construct was observed, suggesting that none of the C-terminal motifs we examined are critical for targeting CXCR3 for degradation (Figure 7C). Thus it appears that CXCR3 is readily degraded in the presence or absence of ligand, by pathways involving both the proteosome and lysosomes, and that ubiquitination of CXCR3 is not a fundamental part of this process.
Previous reports have detailed the strict control of CXCR3 mRNA expression in freshly isolated PBMCs. Although a significant proportion of freshly isolated PBMCs express CXCR3 at the cell surface, mRNA transcripts remain undetectable and the cells are unresponsive to CXCL9 and CXCL10 in assays of chemotaxis and calcium flux (1, 2). This unresponsive phenotype is reversed following culture of PBMCs for several days in a medium containing IL-2 and a mitogen such as PHA, and correlates with mRNA induction and increased levels of CXCR3 at the cell surface. We revisited these data using freshly isolated T-lymphocytes used either immediately after isolation or following culture for 10 days in medium supplemented with IL-2 and ConA. As previously described for PBMCs, freshly isolated T-lymphocytes expressed modest levels of cell surface CXCR3 which were significantly upregulated following 10 days of culture (Figure 8A). Modest numbers of freshly isolated cells were seen to migrate in response to increasing concentrations of CXCL9 and CXCL10, responses which were significantly enhanced following 10 days of culture (Figure 8B and C). In contrast, chemotactic responses of both freshly isolated and cultured T-lymphocytes to CXCL11 were robust, notably at the optimal concentration of 10nM (Figure 8D). The greater efficacy of CXCL11 on freshly isolated cells was also evident when internalization assays were performed, with CXCL11 but neither CXCL9 nor CXCL10 inducing significant internalization of CXCR3 (Figure 8E). Radioligand binding assays competing 125I-CXCL10 and 125I-CXCL11 with homologous unlabelled ligand (Figure 8F) suggested a greater number of binding sites for CXCL11 on freshly isolated T-lymphocytes than were evident for CXCL10, despite both ligands reported to have similar nanomolar affinities at CXCR3 (8, 10, 45). Culture for 10 days in the presence of IL-2 and ConA resulted in a trend for an increase in the numbers of binding sites for both ligands, although this was not statistically significant. Thus, despite freshly isolated cells expressing CXCR3 at the cell surface, the responses to CXCR3 ligands appear to be muted, with only CXCL11 inducing significant biological function.
Although there is growing information regarding the mechanisms of GPCR internalization and recycling, no data concerning the fate of internalized CXCR3 have been published to-date. Once internalized, a GPCR can experience one of two fates, namely dissociation of ligand and recycling of functional receptor back to the plasma membrane or degradation. These fates are not mutually exclusive, as CXCR4 has been shown to undergo both processes following engagement with ligand (46, 47). In this manuscript, we provide several lines of evidence to suggest that the fate of endocytosed CXCR3 is predominantly one of degradation, with de novo synthesis of CXCR3 required for the recovery of CXCR3 cell surface levels. Indeed, little if any CXCR3 was seen to reappear at the cell surface when transcription, translation or the trafficking of nascent proteins though the golgi-ER was perturbed with appropriate pharmacological inhibitors.
Whilst the related receptor CXCR4 has been shown to undergo lysosomal degradation in a ubiquitin-dependent manner (47, 48), both the lysosome and the proteosome appear to facilitate the degradation of CXCR3, as deduced by sensitivity to both MG132, a 26 S proteasome inhibitor (49) and chloroquine, an inhibitor of intra-lysosomal catabolism (50). Taken in the context of our confocal microscopy data, it suggests that internalized CXCR3 traffics to late endosomes/lysosomes which may communicate in part with the proteosome for CXCR3 degradation. Cooperation between both the lysosomal and proteosomal pathways has been described for the degradation of other receptors including the growth hormone receptor (GHR) (51) and the IL-2 receptor/ligand complex (52). In the case of the GHR, endocytosis occurs in the absence of receptor ubiquitination but still requires intact proteasomal activity, suggesting that an adaptor protein targets the receptor to the proteosome (53). In the case of CXCR3, ubiquitination of the receptor appears not to be required for either internalization or degradation, as mutation of the sole intracellular lysine residue had little effect upon either process. This may suggest the existence of an additional motif within the CXCR3 intracellular regions which targets it for degradation by this route. Alternatively, CXCR3 may be envisaged to interact with an adapter protein, which itself undergoes ubiquitination, targeting both proteins to the proteosome. Such a process has been described for the lectin Siglec-7, which is targeted to the proteosome as a complex with SOCS (suppressor of cytokine signalling) -3 (54).
Two main routes have been described for the internalization of GPCRs following their activation. The best characterised pathway employs clathrin coated pits. In this pathway the phosphorylated receptor is bound by arrestins and located to clathrin coated pits, where the complex is internalized in vesicles which are subsequently released from the plasma membrane by dynamin and transported to endosomes, where dephosphorylation of the receptor occurs and re-sensitized receptor is recycled to the plasma membrane (55). The clathrin mediated pathway has been demonstrated for the internalization of other chemokine receptors of the CXC class, notably CXCR1 (56) CXCR2 (57) and CXCR4 (38). A second pathway of internalization depends on caveolae (58), cholesterol rich, highly organised membrane structures which have been shown to be involved in the internalization of other GPCRs including the chemokine receptors CCR4 and CCR5 (19, 20). Although caveolae have been described in macrophages (59), there is still some debate as to whether lymphocytes contain caveolae (60, 61). Evidence for the utilization of either pathway of receptor internalization is often provided through the over-expression of dominant negative constructs (e.g. arrestin, dynamin and clathrin mutants) or pharmacological inhibitors to invoke or preclude the use of a particular pathway (62). In both human PBMCs and an established transfectant system expressing human orthologue of CXCR3 (10), cell surface levels of CXCR3 were rapidly reduced in a concentration and time dependant manner following exposure to ligand. In CXCR3 transfectants, use of inhibitors suggested that the pathway mediating ligand-induced endocytosis did not appear to involve caveolae but involved clathrin. In contrast, treatment of activated PBMCs with hypertonic sucrose did not inhibit the internalization of CXCR3, and the use of an inhibitor of clathrin coated pit assembly had no effect upon the down regulation of CXCR3 in purified T-lymphocytes. Likewise, we found no absolute requirement for arrestin in the internalization process. In MEFs from mice deficient in β-arrestins 1 and 2 (24), internalization of CXCR3 was significantly reduced, but not completely abolished, whilst the use of dominant negative arrestin mutants was without effect upon CXCR3 internalization in our transfectant system. Collectively, this suggests an alternative pathway for the endocytosis of CXCR3, one that is independent of clathrin and/or arrestin. This finding is in agreement with a previous study where CXCL11-induced internalization of CXCR3 in a transfectant system was found to occur in a dynamin and β-arrestin-2 independent manner (28).
The cellular motifs controlling ligand-driven internalization and targeting it for subsequent degradation remain elusive. Removal of potential phosphorylation sites in the C-terminus by truncation had no effect on CXCL11-induced internalization, as described previously for CXCR3 transfectants in both HEK-293 and 300-19 cell lines (28, 63). Likewise, mutation of the LLL motif was without effect on CXCL11-induced internalization again in agreement with a study using 300-19 transfectants (28) but in disagreement with a study using HEK-293 transfectants where some inhibition of CXCL11 induced internalization was observed (63). This likely reflects differences in the intracellular machinery to which CXCR3 is coupled in either cell system.
Of interest was the finding that CXCR3 is constitutively degraded in the absence of ligand, a robust process that was mediated to a significant extent by a conserved canonical YXXϕ motif at the extreme of the C-terminus. Since such motifs have been implicated in the sorting of transmembrane proteins to endosomes and lysosomes (64) we hypothezise that the YSGL motif interacts with currently unknown intracellular protein(s) and controls the constitutive internalization of CXCR3. Supportive of our hypothesis, a distal YKKL motif within the C-terminus of the GPCR PAR1 directs constitutive receptor internalization which is clathrin and dynamin dependent but independent of arrestins (65, 66).
Cell surface levels of CXCR3 are tightly regulated by both constitutive and ligand driven degradation and the replenishment of cell surface CXCR3 appears not to be dependent upon recycling as has been shown for other chemokine receptors, but upon de novo synthesis of CXCR3 protein and its subsequent transportation through the Golgi apparatus. To our knowledge, this is the first example of a GPCR, where protein synthesis is essential for the replenishment of the receptor on the cell surface following stimulation with ligand. This strict control is in addition to other mechanisms of post-translational regulation of CXCR3 function. Although expressing significant amounts of CXCR3 on the cell surface, freshly isolated T-lymphocytes were poorly responsive to the CXCR3 ligands CXCL9 and CXCL10 as previously described (2), with the exception of the ligand CXCL11. This phenotype was corrected upon activation of the T-lymphocytes by prolonged incubation with IL-2 and a mitogen such as concanavalin-A, a process which corresponds with CXCR3 mRNA induction (2), increased cell surface expression of the receptor and the acquisition of robust functional responses to all three ligands. CXCL10 and CXCL11 have previously been described as allotopic ligands of CXCR3, with activated T-cells expressing a significant population of CXCR3 molecules that can bind 125I-CXCL11 but not 125I-CXCL10 (45). The binding of CXCL10 is thought to be controlled at the level of G protein coupling, since treatment of cell membranes with GTPγS or pertussis toxin resulted in a total loss of CXCL10 binding. In contrast, CXCL11 can bind to both coupled and uncoupled CXCR3 (45). Supportive of this, both resting and activated T-lymphocytes were observed to bind significantly more 125I-CXCL11 than125I-CXCL10. Recently published data from studies using mice deficient in the G protein alpha subunits Gαi2 and Gαi3 found that whilst T-lymphocytes from mice lacking Gαi2 subunits exhibited no chemotaxis to CXCR3 ligands, T- lymphocytes from mice lacking Gαi3 displayed significant increases in both migration and GTPγS binding and migration as compared with wild type T-lymphocytes (67). This suggests that in the mouse, Gαi2 subunits are crucial for CXCR3 signalling, and that Gαi3 subunits can act as intracellular inhibitors of CXCR3 function, thereby modulating CXCR3 responsiveness. Examining our findings in the light of this data, we can hypothesize that upregulation of CXCR3 itself does not necessarily result in responsiveness to CXCL10 and that CXCR3 function in the human is likely be modulated at the intracellular level by interaction with G proteins.
CXCR3 has previously been reported to be expressed in an intracellular compartment within T- lymphocytes, which can rapidly be mobilized to the cell surface by treatment with arachidonic acid (44). This rapid, transient mobilization of receptor has been postulated to enable the cells to respond timely to changes in the microenvironment in vivo. Such a capacity for increased cell surface expression is likely to be counter-balanced by the degradative fate of CXCR3 we describe here. It is noteworthy that the CXCR3 ligands are all readily induced by IFN-γ (7-9) and that the Th1-polarised lymphocytes specifically attracted by these chemokines are themselves a source of IFN-γ (68). It can be postulated that such fine tuning of CXCR3 activity by degradation of internalized receptor serves to avoid the unnecessary amplification of T-lymphocyte recruitment in vivo, which would have undesirable consequences for the host. Generation of an artificial CXCR3 ligand which promotes the cellular degradation of CXCR3 in the absence of intracellular signalling may represent an alternative strategy for the therapeutic modulation of CXCR3 with potential benefit in a wide variety of disease processes.
We are grateful to Dr Robert Lefkowitz for kindly providing the WT and β-arrestin deficient MEFs, to Dr Marc Caron for the provision of plasmids encoding arrestin mutants and to Dr Alexandre Benmerah for providing plasmids encoding the DIII and D3Δ2 constructs. We thank Professor Mark Marsh, University College London, and Dr Richard Colvin, Massachusetts General Hospital, for helpful discussions.
We are grateful to the British Heart Foundation (PG/2000055 and FS/05/021) the Arthritis Research Campaign (174240) and the Wellcome Trust (Project Grant 076036/Z/04/Z) for their support of this work.
Publisher's Disclaimer: This is an author-produced version of a manuscript accepted for publication in The Journal of Immunology (The JI). The American Association of Immunologists, Inc. (AAI), publisher of The JI, holds the copyright to this manuscript. This version of the manuscript has not yet been copyedited or subjected to editorial proofreading by The JI; hence, it may differ from the final version published in The JI (online and in print). AAI (The JI) is not liable for errors or omissions in this author-produced version of the manuscript or in any version derived from it by the U.S. National Institutes of Health or any other third party. The final, citable version of record can be found at www.jimmunol.org