The regulated expression, intracellular trafficking and membrane recycling of the P2Y-like receptor GPR17 in Oli-neu oligodendroglial cells

GPR17 is a G-protein-coupled receptor that is activated by two classes of molecules: uracil-nucleotides and cysteinyl-leukotrienes. GPR17 is required for initiating the differentiation of oligodendrocyte precursors but has to be down-regulated to allow cells to undergo terminal maturation. Although a great deal has been learned about GPR17 expression and signaling, no information is currently available about the trafficking of native receptors after the exposure of differentiating oligodendrocytes to endogenous agonists. Here, we demonstrate that neuron-conditioned medium induces the transcriptionally mediated, time-regulated expression of GPR17 in Oli-neu, an oligodendrocyte precursor cell line, making these cells suitable for studying the endocytic traffic of the native receptor. Agonist-induced internalization, intracellular trafficking, and membrane recycling of GPR17 were analyzed by biochemical and immunofluorescence assays using an ad hoc-developed antibody against the extracellular N-terminal of GPR17. Both UDP-glucose and LTD(4) increased GPR17 internalization, although with different efficiency. At early time points, internalized GPR17 co-localized with transferrin receptor, whereas at later times it partially co-localized with the lysosomal marker Lamp1, suggesting that a portion of GPR17 is targeted to lysosomes upon ligand binding. An analysis of receptor recycling and degradation demonstrated that a significant aliquot of GPR17 is recycled to the cell surface. Furthermore, internalized GPR17 displayed a co-localization with the marker of the "short loop" recycling endosomes, Rab4, while showing very minor co-localization with the "long loop" recycling marker, Rab11. Our results provide the first data on the agonist-induced trafficking of native GPR17 in oligodendroglial cells and may have implications for both physiological and pathological myelination.


SUMMARY
GPR17 is a G-protein-coupled receptor that is activated by two classes of molecules: uracilnucleotides and cysteinyl-leukotrienes. GPR17 is required for initiating the differentiation of oligodendrocyte precursors but has to be downregulated to allow cells to undergo terminal maturation. Although a great deal has been learned about GPR17 expression and signaling, no information is currently available about the trafficking of native receptors following the exposure of differentiating oligodendrocytes to endogenous agonists. Here, we demonstrate that neuron-conditioned medium induces the transcriptionally-mediated, time-regulated expression of GPR17 in Oli-neu, an oligodendrocyte precursor cell line, making these cells suitable for studying the endocytic traffic of the native receptor. Agonist-induced internalization, intracellular trafficking and membrane recycling of GPR17 were analyzed by biochemical and immunofluorescence assays using an ad hoc-developed antibody against the extracellular N-terminal of GPR17. Both UDPglucose and LTD 4 increased GPR17 internalization, although with different efficiency. At early time points, internalized GPR17 colocalized with transferrin receptor, whereas at later times it partially co-localized with the lysosomal marker Lamp1, suggesting that a portion of GPR17 is targeted to lysosomes upon ligand binding. An analysis of receptor recycling and degradation demonstrated that a significant aliquot of GPR17 is recycled to the cell surface. Furthermore, internalized GPR17 displayed a colocalization with the marker of the "short loop" recycling endosomes, Rab4, while showing very minor co-localization with the "long loop" recycling marker, Rab11. Our results provide the first data on the agonist-induced trafficking of native GPR17 in oligodendroglial cells and may have implications for both physiological and pathological myelination.
The physiological role of GPR17 has been deeply investigated in both in vivo and in vitro systems, and a number of studies have revealed its crucial role in oligodendrocyte precursor cell (OPC) differentiation (2, [7][8][9][10][11]. Receptor expression, almost absent in early OPCs, gradually increases in more mature precursors, reaches a plateau in immature/preoligodendrocytes and then gradually decreases during terminal differentiation. In line with these findings, GPR17 is co-expressed with the early oligodendrocyte marker NG2 and markers of pre/immature oligodendrocyte phenotype (such as O4 and DM- 20) but is downregulated in cells expressing myelin proteins such as myelin basic protein (MBP), which is highly synthesized in fully mature cells (7,10,11). Consistent with the role of GPR17 in oligodendrocyte ontogenesis, its activation by natural agonists promotes OPC differentiation under physiological conditions (2, 10). In contrast, the inhibition of GPR17 expression causes an impairment in oligodendrocyte differentiation and myelination in both in vivo (7) and in vitro systems (10). Altogether, these studies indicate that GPR17 is an integral signaling component controlling oligodendrocyte ontogenesis and suggest that the appropriate activation and deactivation of GPR17 are crucial steps in OPC maturation.
As it has been reported for many GPCRs, after ligand binding, GPR17 may undergo endocytosis and subsequent sorting into lysosomes for degradation and/or into recycling endosomes for re-incorporation into the plasma membrane. The balance of this dynamic intracellular trafficking is physiologically relevant because it modulates receptor levels at the cell surface. This process has important implications for the activation or silencing of GPR17-signaling pathway(s), and in turn, for OPC differentiation (12)(13)(14)(15)(16). It may even be hypothesized that GPR17 endocytosis may represent a key event necessary to allow OPCs to proceed to myelination. A similar process has been associated with the specification of other cell lineages, where the downregulation of membrane receptors has been proposed to be necessary to allow cells to proceed towards terminal differentiation (17). Interestingly, the abnormal upregulation of GPR17 has been associated with defective myelination during development and with multiple sclerosis (7). Thus, the characterization of the mechanisms involved in the expression of GPR17 in the plasma membrane may help us to better understand the molecular mechanisms of the contribution of GPR17 to oligodendrogenesis and may set the background for interpreting the consequences of GPR17 dysfunction in disease.
At present, however, there are very few studies available on the trafficking of GPR17, both under basal conditions and upon activation. In 1321N1 cells heterologously expressing hGPR17, it has been demonstrated that the GPR17 agonists UDPglucose and LTD 4 determine receptor desensitization/re-sensitization (6). On the other hand, a previous study has failed to demonstrate the direct activation of GPR17 by agonists, proposing that the receptor may function exclusively as a negative regulator for the cysLT 1 R response to LTD 4 treatment (18). Furthermore, Benned-Jensen and Rosenkilde (19) reported that mouse or human GPR17 is activated by uracil nucleotides but apparently not by LTD 4 or LTC 4 , and showed that LTD 4 did not significantly increase the internalization of FLAG-tagged hGPR17 in transiently transfected HEK293 cells. Moreover, despite the crucial role of GPR17 in OPCs (see above) and the evidence that the receptor is clearly downregulated in cells achieving terminal maturation, no studies are yet available on the agonist-induced regulation of native GPR17 in cells of the oligodendroglial lineage.
In this study, therefore, we decided to analyze the endocytic trafficking of native GPR17 following activation with uracil nucleotides or cysLTs using a physiological expression system. Although OPC primary cultures would represent an ideal system, the necessity to isolate them from tissue for each experiment and the relatively low number of cells obtained from each preparation markedly limited their use in the detailed biochemical analysis planned for the present work. To avoid this problem, we selected Oli-neu cells, an OPC cell line immortalized from E16 mouse brains, because these cells can be induced to recapitulate several features of differentiating OPCs in vitro (20,21). Therefore, we first established the conditions for stimulating the endogenous expression of GPR17 in Oli-neu and then used these cells for the analysis of the endocytic trafficking of native GPR17. We found that Oli-neu cells express significant amounts of GPR17 following incubation with medium conditioned from neuronal primary cultures. The expression of the receptor was transient and depended on the maturational stages of Oli-neu towards a more differentiated phenotype: GPR17 synthesis was upregulated during the early stages of differentiation and was downregulated when the cells reached later stages of maturation, as shown in primary OPCs (2, 10,11). Furthermore, the GPR17 gene promoter appeared to be highly activated by factors released in the medium by neurons and/or astrocytes. Finally, we demonstrate that UDP-glucose and LTD 4 , although with low efficiency, stimulated clathrin-mediated endocytosis of GPR17. After internalization, the receptor is delivered into early endosomes and then sorted, either to lysosomes for degradation or recycled to the cell surface via the small G-protein Rab4dependent pathway.

Antibodies against the N-terminal region of GPR17-
An antibody was raised in rabbit using a synthetic peptide corresponding to the NH 2 -terminal sequence of mouse GPR17 (MNGLEAALPSLTDNSSLAYSEQC) coupled to keyhole-limpet hemocyanin (22). The antibodies (from here on referred to as anti-Nt-GPR17) were affinity-purified and tested for their specificity for in vivo immunolabeling.
Cloning of the GPR17 promoter region and the generation of luciferase constructs-The human GPR17 promoter was analyzed with the Genomatix suite. A small fragment of 909 bp was hypothesized to be a regulatory region, and it was cloned into a pGL4.17 vector (Promega), upstream to the luc2 reporter gene, using the following primers: Fw 5'-CGCTCGAGTTCCCATGTTGCTGGATGTA-3', and: Rv 5'-CAAGCTTCGCTGAGTGTTTCCTCTGCT-3', containing a cassette for the restriction enzymes XhoI and HindIII, respectively. The generated vector was called pGL4-h909.
Cell culture and differentiation-The Oli-neu murine OPC cell line was kindly provided by Prof. J. Trotter (University of Mainz, Germany) and cultured in Sato medium containing 1% horse serum (HS), as previously described (20). The neuron-conditioned medium (CM) was obtained from primary cultured neurons prepared from the cerebral cortex of 18-dayold rat embryos and maintained in neurobasal medium supplemented with B27 as previously described (23). The CM was collected after 7-10 days and passed through a 22-µm filter. The Oli-neu cells were cultured in normal medium or normal medium plus CM in a 1:1 ratio. COS-7 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% Fetal bovine serum. 4 above). The gels were exposed to phosphor screens and analyzed with a Storm phosphoimager (Molecular Dynamics). Band intensity was quantified with imageQuanta software.

RNA interference and plasmid transfections-For
RNA interference, the Oli-neu cells were plated the day before and incubated in CM for 6 h prior to transfection. The cells were transfected using the transfection reagent INTERFERin TM without siRNA (mock transfection control), 2.5 nM of scrambled siRNA (negative control) or a siRNA designed to silence mouse GPR17 (Qiagen; target sequences), in accordance with the manufacturer's instructions. Then, 24 or 48 h after transfection, the cells were fixed for immunofluorescence analysis or the cellular proteins extracted for immunoblotting. For forkhead box protein (Fox)O1 knock-down, Oli-neu cells were transfected with 50nM siRNA designed to silence mouse FoxO1 (Cell Signaling Technology, Millipore). Thirty six hr after transfection cells were incubated for 24 h with CM and then fixed, as described above. For cDNA transfection, the day prior to transfection, Oli-neu cells were seeded onto 24x24-mm glass coverslips (75,000 cells/coverslip) and incubated in CM. After 24 h, the cells were transfected with plasmids (0.5 µg/coverslip) encoding the NH 2terminally EGFP-tagged Rab5a (24), Rab4a, or Rab11a (25), a generous gift of Dr. C. Bucci (Dept. of Science, University of Salento, Italy) and B. Chini (Institute of Neuroscience, Milan, Italy). cDNAs were mixed with jetPEI reagent (PolyplusTransfection) in a 1:2 ratio, incubated for 20 min at room temperature and gently added to the cells. After 24 h of transfection, the medium was removed, and fresh CM was added. Analysis of GPR17 internalization was performed approximately 48 h after transfection. For FoxO1 overexpression, Oli-neu were transfected with cDNAs encoding GFP-FoxO1 or FLAG-FoxO1alanine-aspartic acid-alanine (ADA) obtained from Dr. D. Accili (Columbia University, New York, New York) and provided by Addgene. After 48 h incubation in Sato medium, cells were fixed and processed for immunofluorescence. For the analysis of promoter induction, Oli-neu cells were transfected with the normalizing vector pGL4-TK-Renilla (Promega) along with either the reporter construct pGL4-h909 or the corresponding empty vector (as a negative control) in a 1:50 ratio. A total of 250 ng of DNA was transfected with the JetPEI reagent. The day after transfection, the cells were incubated for 24 or 48 h with the selected treatments. The Dual Reporter Luminometer Assay (Promega) was performed according to the manufacturer's instructions.

Pharmacological treatments and cell solubilization-
Agonists were added to differentiating Oli-neu cells at the following concentrations: 100 µM UDP-glucose and 50 nM LTD 4 . Cells were then incubated at 37°C for the indicated times. When required, cells were preincubated for 10 min at 4°C with antagonists (10 µM cangrelor; 1 µM montelukast). Cells were lysed in buffer A (150 mM NaCl, 2 mM EGTA, 50 mM Tris-HCl, pH 7.5, and a Sigma-Aldrich protease inhibitor cocktail diluted 1:1,000) containing 1% Triton X-100, and then centrifuged 20,000 x g for 20 min at 4°C.

Immunoprecipitation, Endoglycosidase digestions and Western blotting-
For the immunoprecipitation of GPR17, aliquots of the cell extracts (100-200 µg of protein) were incubated for 1 h with protein A or G Sepharose beads. The beads were removed by centrifugation, and the 'pre-cleared' supernatants were added to protein A or G beads that had been preincubated with the primary antibodies or non-immune IgG for 2 h at 4°C. After overnight incubation, the beads were extensively washed with 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.3% (w/v) Triton X-100. For endoglycosidase digestions, proteins isolated by immunoprecipitation were eluted from protein A beads by heating at 65°C in a solution containing 0.5% SDS and 1% β-mercaptoethanol. Aliquots of each sample were then diluted with Nonidet P-40 (1% final concentration) and 1/10 concentrated endoglycosidase F (EndoF, also called PNGaseF) or endoglycosidase H (EndoH) reaction buffers (G7 or G5, New England Biolabs) in accordance with the manufacturer's instructions and digested with 250 units of EndoF or EndoH for 1 h at 37°C or 0°C. The reactions were stopped by the addition of Laemmli sample buffer and analysed by Western blotting, as previously described (22), using anti-rabbit IgG light chains or anti-mouse IgG conjugated to peroxidase (diluted 1:50,000) as secondary antibodies. The peroxidase was revealed using a chemiluminescent substrate (Pierce). For quantitative analysis, unsaturated autoradiograms were acquired using an ARCUS II scanner (Agfa-Gevaert, Mortsel, Germany), and the density of each band was quantified using NIH Image J software (National Technical Information Service, Springfield, VA, USA).

RNA extraction,
real-time and reversetranscription PCR-Total RNA was purified from tissue, platelets or cell cultures using an RNeasy Plus Mini Kit (Qiagen) in accordance with the manufacturer's instructions. The target sequences for the quantitative, real-time PCR were amplified from 1 µg cDNA; for the analysis of gene expression, we used the ABI Prism® 7000 Sequence Detection System, SDS software version 1.2.3 (Applied Biosystems, CA, USA). The target sequences were amplified using a pre-programmed thermal profile of enzyme activation at 50°C, followed by 40 identical cycles of denaturation at 95°C for 15 s and annealing and amplification at 60°C for 1 min; these conditions were previously determined to be optimal by Applied Biosystems. The TaqMan® primer and probe assays used were GPR17 (Mm02619401_s1) and the endogenous control glyceraldehyde-3phosphate dehydrogenase (GADPH: Mm99999915_g1). Control reactions that were performed in the absence of Quantiscript reverse transcriptase (Qiagen) excluded the possibility of genomic traces. The results were calculated using the 2 −ΔΔCt method, allowing for the normalization to GAPDH with the calibrator set to a value of 1 (26). For reverse-transcription PCR, 1 µg of RNA was transcribed to cDNA using the QuantiTect Reverse Transcription Kit (Qiagen), and aliquots (1 µl) of the total cDNA were amplified in 40 cycles of PCR prepared using TopTaq Master Mix Kit (Qiagen). The primers are indicated in Table 1. The DNA fragments were then analyzed by means of agarose gel electrophoresis.
Biotinylation assay-Cells were incubated in neuronconditioned medium for 48 or 72 h and then biotinylated using 0.3 mg/ml of sulfo-NHS-SS-Biotin (Thermo Scientific) dissolved in PBS with 0.1 mM CaCl 2 and 1 mM MgCl 2 for 30 min at 4°C. The labeled cells were washed three times for 10 min with 50 mM glycine in TBS (25 mM Tris, 85 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 ) to quench free biotin. When required, the cells were subsequently incubated in Oli-neu medium in the absence or presence of the drugs for the times indicated in the results. After a 4°C wash with ice-cold PBS containing 0.1 mM CaCl 2 and 1 mM MgCl 2 , the remaining surface biotin labeling was removed by incubating the cells twice with 50 mM DTT or MesNA, a membrane-impermeable reducing agent, at 4°C. Both reducing agents were then neutralized with iodoacetamide (10 mM) in PBS with 0.1 mM CaCl 2 and 1 mM MgCl 2 , and the cells were lysed in buffer A containing 1% Triton X-100 and a protease inhibitor cocktail. After centrifugation (20,000 x g for 20 min at 4°C), supernatants containing equal amounts of protein were incubated with streptavidin beads to isolate the biotinylated proteins. Following extensive washes in extraction buffer, the proteins were eluted from the streptavidin beads and analyzed by SDS-PAGE followed by Western blotting. Unsaturated autoradiograms were acquired using an ARCUS II scanner (Agfa-Gevaert, Mortsel, Germany), and the density of each band was quantified using NIH Image J software (National Technical Information Service, Springfield, VA, USA). Data were collected from at least three experiments.
Antibody labeling of cell surface-exposed GPR17 in living cells-Cells that were cultured on coverslips as described above were chilled at 4°C and washed once in ice-cold PBS containing 0.1 mM CaCl 2 , 1 mM MgCl and 1% HS. Cells were then incubated at 4°C for 45 min with rabbit anti-Nt-GPR17 antibody (10 µg IgG/24x24 mm coverslips) diluted in ice-cold PBS with 0.1 mM CaCl 2 , 1 mM MgCl and 1% HS. Following the incubation period, the cells were washed 5 times with PBS containing 0.1 mM CaCl 2 , 1 mM MgCl 2 and 1% HS and either fixed with formaldehyde or incubated in Sato medium (controls) in medium supplemented with agonists in the absence or presence of antagonists, 0.45 M sucrose or 80 µM dynasore as reported in the Results section. After an incubation at 37°C for different times, the cells were chilled at 4°C to block endocytosis, washed with PBS containing 0.1 mM CaCl 2 , 1 mM MgCl 2 , and when required, with 50 mM glycine in HBSS (glycine buffer). Finally, the cells were fixed and processed for immunofluorescence.
Immunofluorescence-The cells were fixed for 8-10 min with 4% paraformaldehyde in phosphate buffer, pH 7.3, containing 4% sucrose at 37°C and permeabilized for 5 min at room temperature in PBS containing 0.3% Triton X-100. After immunostaining as previously described (27), images were recorded using either a Zeiss LSM510 Meta or an MRC-1024 laser-scanning microscope (Bio-Rad) equipped with a 60x objective, or an Axiovert 200M (Zeiss) confocal system equipped with a spinning disc (PerkinElmer) with a 63x or 40x objectives. To compare the doublestained patterns, images from the fluorescein or rhodamine channels were acquired separately and superimposed. The images were processed using Photoshop (Adobe Systems, Mountain View, CA, USA).

Image analyses and quantitation-
We quantified the internalization of GPR17 in Oli-neu cells by measuring the immunofluorescence intensity of anti-Nt-GPR17 in cells before (time 0) or following incubations with or without agonists at 37°C followed by washes with glycine-buffer to strip the antibody bound to the remaining cell surface exposed receptors. At least three independent experiments were performed for each condition, and approximately 180 cells were examined in each experiment. All images were recorded with an MRC-1024 laser-scanning microscope (Bio-Rad) using identical parameters. Serial sections were acquired for each field, and the images were exported as TIFF files for analysis with Image J software. Values are expressed as the fold increase from the intensity detected in the cells at time 0 (background).

Statistical Analysis.
We analyzed the data using GraphPad Prism software and expressed the results as mean values ± s.e. The significance of the differences was assessed using either a two-tailed, non-paired Student's t-test, or a two-way ANOVA, and the customary threshold of p<0.05 was used to declare statistical significance (confidence intervals 95%).

Endogenous expression and biochemical characterization of GPR17 in Oli-neu
To investigate whether Oli-neu cells express GPR17, cell cultures maintained in normal conditions (Sato medium) were analyzed by confocal immunofluorescence microscopy following immunolabeling with antibodies directed against the C-terminal of GPR17 (anti-Ct-GPR17; 1) or protein markers typical of different stages of oligodendrocyte differentiation. In line with previous data, Oli-neu cells showed very few cell processes and expressed classical markers of an immature oligodendrocyte phenotype (Olig2 and NG2, data not shown) when grown in Sato medium. At this stage, very few (approximately 4%) GPR17-positive cells were detected, while the majority of cells did not show immunoreactivity for GPR17 or MAG (Fig. 1A). As a number of studies have reported that Oli-neu cells can be induced to acquire a more mature phenotype if they are placed under differentiating conditions (20,28), we decided to incubate Oli-neu cells in selected conditions to promote their differentiation and possibly modulate their expression of GPR17. However, many of the tested treatments (e.g., ATP, LIF and IGF administration, data not shown) were ineffective, or induced low levels of GPR17 expression after 72 h (e.g., di-butyryl-cAMP, data not shown). Conversely, when the cells were cultured in the presence of medium collected from primary cultures of cortical neurons (from here on referred to as CM), a sustained increase in the levels of GPR17 expression was observed. Notably, after a 24-hour incubation in CM, immunoreactivity for GPR17 was detected in a large number of Oli-neu cells. Twentyfour hours later (+CM 48 h), approximately 80% of cells appeared intensively immunolabeled for GPR17. These GPR17-positive cells showed an increased number of processes, and some started to co-express the myelin protein, MAG (Fig. 1A). To confirm the specificity of the immunostaining for GPR17, 48 hdifferentiated cells were transfected with a small interfering (si) RNA specific to silence mouse GPR17. As shown in Figure 1B, GPR17 labeling was largely reduced upon treatment with the specific siRNA but remained unchanged when the cells were transfected with a non-specific (scrambled) siRNA.
The expression of GPR17 in CM-stimulated Oli-neu cells was also analyzed in cell extracts by Western blotting. As shown in Figure 1C, a broad band of approximately 46-48 kDa and a sharper band of 38 kDa were both detected with the anti-Ct-GPR17 antibody in differentiated but not in undifferentiated Oli-neu cells. The expression of these two polypeptides was largely reduced upon receptor knockdown with the specific siRNA (Fig. 1D). In contrast, a polypeptide of approximately 57 kDa, which was also recognized by the antibody in undifferentiated and differentiated Oli-neu cells, was unaffected by GPR17 knockdown, suggesting that this band is not related to GPR17, but rather, corresponds to a non-specific signal. Altogether, these data further demonstrate the expression of GPR17 in differentiating Oli-neu cells and indicate that the fully mature, post-translationally modified receptor displays a molecular mass of 46-48 kDa, whereas the detected band of 38 kDa likely represents a precursor form. These conclusions are also supported by two additional observations (Fig. 2). First, pulse-chase experiments with 35 S-Met/Cys followed by immunoprecipitation of 35 S-labeled GPR17 from cell extracts indicated the presence of two major 35 S-labeled polypeptides of 46-48 kDa and 38 kDa, respectively, after labeling (pulse), while the 38-kDa polypeptide disappeared after the chase, suggesting that it indeed corresponds to a precursor form of the receptor that is converted to the mature receptor during the chase experiment ( Fig. 2A). Second, endoglycosidase F digestion (to remove all N-linked carbohydrates) converted both polypeptides of 46-48 kDa and 38 kDa into a single band of 33 kDa (Fig.  2B). In contrast, the 46-48-kDa band remained unchanged following endoglycosidase H digestion, which only removes high-mannose N-linked carbohydrates, while the 38 kDa band was largely converted to the 33 kDa band. This result indicates that the 38-kDa polypeptide is a precursor form of GPR17 carrying high mannose oligosaccharide chains (Fig. 2B).
Next, we evaluated the expression of GPR17 in Oli-neu cells at different time points after the administration of CM. Previous data have demonstrated that in OPCs, GPR17 expression increases during differentiation from the precursor stage to pre-oligodendrocytes and subsequently decreases during the transition from preoligodendrocytes to mature oligodendrocytes (2, 7, 10, 11). Importantly, in Oli-neu cells incubated in CM, the protein levels of GPR17 showed a similarly transient expression. As shown in Figure 3, after 48-72 h incubation in CM, a strong immunoreactivity for GPR17 was detected in approximately 80-90% of cells. This signal declined at later times, and few GPR17-highly positive cells were detected after 96-120 h. In contrast, the number of cells expressing detectable levels of myelin proteins increased strongly after 72 h in CM, and almost 80% of the cells were highly positive for myelin proteins after 96 h (Fig. 3). These differences in GPR17 and MAG expression were also confirmed by Western blotting followed by densitometric analysis. As shown in Figure 3B and C, the peak of GPR17 protein expression preceded that of MAG and then rapidly declined, whereas the levels of the myelin protein continued to increase for up to 96 h. Changes in GPR17 protein expression after incubation with CM were consistent with the changes in the level of GPR17 mRNA during Oli-neu cell differentiation. As revealed by real-time PCR (Fig.3D), the mRNA coding for GPR17 was upregulated during differentiation and increased 10fold following a 48-hour incubation in CM. In accordance with the protein expression levels, the levels of GPR17 mRNA declined rapidly and returned to a lower level after 96 h. In contrast, the expression of most of the other members of the purinergic or cysLT receptor families detected in Oli-neu cells were almost unchanged, with the only exception being P2Y 2 , whose mRNA levels were increased upon incubation with CM (Fig. 4). Interestingly, P2Y 12 and cysLT 1 R mRNAs were undetected at any time.
To evaluate whether the increase in GPR17 expression in Oli-neu cells during differentiation occurred via a mechanism involving gene activation, one of the putative promoter regions identified on the human GPR17 gene (h909) and corresponding to highly conserved orthologous sequences on the mouse and rat genes (Fig. 5A) was cloned into a luciferase reporter vector and transfected in Oli-neu cells to measure promoter activity upon the exposure of cells to various stimuli. As reported in Figure 5B, a statistically significant stimulation of the GPR17 promoter and luciferase activity was observed after a 48 h exposure of Oli-neu cells to CM. A slight, but not statistically significant, stimulation of GPR17 promoter activity was detected after the exposure of Oli-neu cells to neurobasal medium supplemented with B27 (which is normally used to culture cortical neurons, indicated as +NB in Fig. 5B). In accordance with these results, western blot analysis of cell extracts from parallel cultures revealed a slight increase in the GPR17 protein in Oli-neu cells after incubation in neurobasal medium supplemented with B27 (+NB in Fig. 5C) at later times (72-96 h), while receptor protein levels were strongly increased after a 48 h incubation in CM suggesting for the first time that the GPR17 promoter is highly activated by factors released by neurons and/or astrocytes. To shed light on the mechanisms underlying the induction of GPR17 in Oli-neu, we turned our attention to recent data demonstrating that the transcription factor FoxO1 binds to the GPR17 promoter and increases receptor expression in a subset of hypothalamic neurons critical for initiating food intake (29). To investigate whether FoxO1 is also implicated in the expression of GPR17 in differentiating Oli-neu, we first examined the levels of FoxO1 in Oli-neu incubated in Sato medium or after administration of CM. Western blotting analysis demonstrated the presence of a faint band of approximately 78 kDa in extracts of cells incubated for 24 h, suggesting that very low levels of FoxO1 are normally expressed in Oli-neu (Fig. 5D). We then investigated the possible role of FoxO1 on GPR17 expression by mRNA silencing as well as by protein overexpression. After transfection with either a siRNA targeting mouse FoxO1 or scrambled siRNA, Oli-neu were incubated for 36 h in CM to induce GPR17 expression. Confocal images revealed no differences between specific or negative siRNA transfected cells (data not shown). Next, we transfected Oli-neu with either wild type GFP-FoxO1 or a constitutively active FoxO1 (FLAG-FoxO1ADA) in which phosphorylation sites have been mutated in order to prevent FoxO1 inactivation (29 and references therein). After 48 h incubation of cells in Sato medium, high levels of FoxO1 were detected in the cytoplasm or in the the nucleus of approximately 10 % of the cells but no increases in GPR17 expression were observed in transfected compared to untransfected cells, thus suggesting that FoxO1 does not regulate GPR17, at least in Oli-neu oligodendroglial cells and under the tested experimental conditions (Fig. 5E).
Altogether. these data demonstrate that GPR17 expression in Oli-neu cells can be modulated during differentiation, as observed in OPC primary cultures or in vivo (2, 7, 10, 11), and support the conclusion that these cells are an appropriate model for investigating GPR17 trafficking.

Agonist-induced internalization of native GPR17
In light of these results, we investigated the internalization of GPR17 in differentiated Oli-neu cells after the administration of either agonists or antagonists. Cells were incubated with sulfo-NHS-SS-Biotin under conditions that prevent plasma membrane internalization (4°C). Preliminary experiments revealed that GPR17 and transferrin receptors (TfR) exposed to the cell surface were efficiently biotinylated, while, as expected, cytosolic proteins such as actin were not (Supplemental data Fig. 6). Furthermore, biotin labeling of GPR17 and TfR was abolished by stripping with disulfide reducing reagents (such as DTT or MesNA), further indicating that no membrane internalization occurred at 4°C. Conversely, when biotinylated Oli-neu cells were incubated for 15 min at 37°C, a portion of biotinylated receptors (10.82% ± 6.63% of total labeling, n = 3 independent experiments: Fig. 6C) was protected from stripping with reducing agents, suggesting that GPR17, similar to many other GPCRs, undergoes constitutive internalization. To investigate agonist-induced endocytosis, biotinylated Oli-neu cells were incubated with micromolar concentrations of uracil nucleotides (100 µM UDP-glucose) and nanomolar concentrations of cysLTs (50 nM LTD 4 ); these agonist concentrations were shown previously to activate the receptors in OPC primary cultures (10). Cells were exposed to the agonists for 5, 15 and 30 min at 37°C, and at the end of each time point, incubated with a reducing agent as described above. The results of western blotting indicated that the agonists activated the endocytosis of native GPR17 in a time-dependent manner (Fig. 6D), as it has been previously reported for the receptor in a heterologous expression system (6). Quantitative analysis of the internalization of biotinylated receptors following a 15-min incubation with the agonists demonstrated that UDP-glucose was more efficient at stimulating the endocytosis of GPR17 compared to LTD 4 . The amount of internalized receptor was indeed increased to 392.9±97.2% (n = 3; P = 0.0241) or to 157.9± 12.21% (P<0.0091) compared to the constitutively internalized receptor following an incubation with UDP-glucose or LTD 4 , respectively ( Fig. 6F and G). Moreover, the effects of the two agonists were counteracted by the addition of cangrelor or montelukast, two antagonists previously shown to inhibit receptor activation mediated by UDP-glucose or LTD 4 , respectively (1, 2, 10).

The distribution of GPR17 in the endocytic compartments following agonist administration
To analyze the trafficking of endocytosed GPR17 after the administration of UDP-glucose or LTD 4 , we employed protocols for tracking cell surface receptors by using antibodies directed against extracellular epitopes (30). To this end, we raised antibodies directed against the extracellular Nterminal (Nt) domain of GPR17 and tested their ability to recognize the receptor on the plasma membranes of living cells. An anti-Nt-GPR17 antibody was found to primarily label the plasma membrane of differentiated Oli-neu cells after incubation at 4°C (Fig. 7A). The specificity of the antibody was demonstrated by i) the absence of labeling on the plasma membrane of undifferentiated cells (data not shown) and ii) the abolishment of immunostaining after antibody pre-adsorption with the antigen (Fig. 7B).
This antibody was then used to follow the endocytic pathways taken by GPR17. Receptors exposed to the cell surface were labeled with the anti-Nt-GPR17 antibody at 4°C and then incubated at 37°C with or without agonists for different times. To allow for a better localization of the internalized receptor, following an incubation at 37°C, the cells were washed with glycine buffer (pH 2.8) to strip the antibodies bound to the receptors remaining at the cell surface. Cells were then fixed, permeabilized and labeled with anti-rabbit IgG conjugated to Cy3, or double-labeled using monoclonal antibodies directed against TfR or Lamp1 followed by species-specific secondary antibodies. As a first step, we examined whether labeling of cells with anti-Nt-GPR17 antibody could "per se" increase the basal internalization of the receptor. Quantitative analysis of the intensity of immunofluorescence detected on Olineu after labeling at 4°C or after labeling followed by a 15 min incubation at 37 °C without agonists indicated that the antibody increased the internalization of GPR17 only slightly compared to biotin (18.30% ± 1.29% of total labeling, n = 158 cells from two independent experiments), thus suggesting that it has little effect on GPR17.
After 2 min of incubation at 37°C with agonists, immunostaining for the anti-Nt-GPR17 was still largely detected at the cell surface although the labeling appeared more clustered, suggesting that receptor stimulation results in a more efficient recruitment of the receptor to endocytic patches and vesicles. At later times (8 and 12 min at 37°C), an increasing number of dot-like structures that were immunolabeled for GPR17 appeared concentrated in the cytoplasm of cells following the administration of agonists. We traced the endocytic route of GPR17 by double-labeling with antibodies against protein markers of early/recycling endosomes (TfR) or lysosomes (Lamp1). As shown in Figure 8, after 8 min of endocytosis, GPR17 partially co-localized with TfR in small vesicles that were dispersed throughout the cytoplasm, while at later times (15 and 30 min), both GPR17 and TfR were found to co-localize in larger dot-like structures located in the perinuclear region, presumably recycling endosomes (Fig. 9). In addition, after 30 min of LTD 4 or UDP-glucose administration, GPR17 immunoreactivity was accumulated in a number of Lamp1-positive vesicles (Fig. 9). These data suggest that GPR17 follows the same intracellular route following the administration of LTD 4 or UDP-glucose and that it may be degraded in lysosomal compartments or recycled to the cell surface.

GPR17 endocytosis is mediated by clathrin
Before analyzing the degradation and/or recycling of GPR17 receptors, we examined whether GPR17 undergoes clathrin dependent-or independentendocytosis (31 and refs therein). To this aim, after labeling with the anti Nt-GPR17-antibody, Oli-neu cells were incubated for 12 min at 37°C with the agonists in the presence of hypertonic sucrose to prevent the assembly of clathrin lattices. As shown in Figure 10, the intracellular accumulation of cell surface-labeled GPR17 was largely inhibited by treatment with 0.45 M sucrose, which is also demonstrated by the quantitative analysis of the intensity of receptor immunoreactivity recovered in cells after stimulation in the presence or absence of hypertonic sucrose (Fig. 10). This analysis demonstrates that the internalization of GPR17 was increased 2.4-or 3.5-fold with respect to background following incubation with LTD 4 or UDP-glucose, respectively; incubation in hypertonic sucrose largely reduced this agonist-activated receptor endocytosis. These data are in line with the biochemical analysis of biotinylated receptors (Fig. 6), confirming that the intracellular accumulation of GPR17 is greater in UDP-glucose than in LTD 4 -stimulated cells and demonstrating that the endocytosis of GPR17 is clathrin-mediated. In line with these data, we found that dynasore, an inhibitor of dynamin polymerization, largely reduced the internalization of GPR17, further supporting the conclusion that GPR17-endocytosis is a clathrin and dynamin dependent process (Fig. 11).

A portion of GPR17 is re-routed to the plasma membrane via Rab4/positive recycling endosomes
We next asked whether receptors underwent recycling and/or degradation following agonistactivated internalization. To test this, agonistactivated receptor internalization was measured following the biotinylation of the cell surface, stimulation at 37°C for 15 min and biotin stripping as described above. To reveal the re-routing of receptors to the cell surface, parallel samples of Oli-neu cells were re-incubated at 37°C in normal medium for 15 and 30 min at 37°C to allow for the sorting of internalized receptors to the cell surface or degradative compartments. At selected time points, cells were lysed without or following biotin cleavage. Biotinylated receptors were then isolated from cell extracts by streptavidin-beads and analyzed by western blotting as described above. Recycling was measured as the amount of biotinylated receptors that were not protected by MesNA cleavage and thus reappeared at the cell surface. Degradation was calculated by subtracting the amount of receptors recycled to the cell surface and the amount of receptors that were still present intracellularly from the total amount of internalized receptors following agonist administration (time 0). The results clearly indicate that a significant amount of GPR17 is recycled to the plasma membrane (Fig. 12). Densitometric analysis of western blots from three independent experiments demonstrated that, although approximately 50%-45% of endocytosed GPR17 underwent degradation after agonist administration, a consistent portion of the receptor (approximately 25% and approximately 30% after LTD 4 or UDP-glucose, respectively) was also recycled to the cell surface (Fig. 12). The amounts of recycled or degraded receptors were not significantly different after stimulation with either agonist.
To further characterize the recycling pathways taken by GPR17, we compared its distribution to that of the small G-proteins Rab4, Rab5 and Rab11 at different times after the administration of agonists. A number of important studies have demonstrated the relevance of these Rab proteins as regulators of the recycling/endosomal compartments. It has also been demonstrated that early/sorting endosomes in the cell periphery contain primarily Rab5 and Rab4, while recycling endosomes in the perinuclear area contain primarily Rab4 or Rab11 (32)(33)(34)(35). To investigate the distribution of GPR17 and Rab proteins, we transfected Oli-neu cells with cDNAs (0.5-0.2 µg) encoding for Rab4a, Rab5a, or Rab11a bearing a green fluorescent protein EGFP tag at the NH2terminal. Following transfection, the cells were differentiated with CM for 48 h and then processed as described above to investigate the internalization and intracellular distribution of GPR17. Cells expressing moderate levels of GFP-Rabs (approximately 5% of total cells) were selected for analysis. Under this condition, Rab5, Rab4 and Rab11 show an intracellular distribution similar to that previously reported (32). Based on the confocal images (Fig. 13), GPR17 largely co-localized with GFP-Rab5 and Rab4 in small vesicles that were primarily scattered in the cytoplasm or located at the periphery of the cells at early times of internalization. At later times, GPR17 partially co-localized with Rab4 primarily in large, dot-like structures that were distributed in the perinuclear region (Fig. 14). A very minor colocalization of GPR17 with Rab11a could also be detected, as reported in Figure 15. Altogether, these data support the conclusion that a portion of GPR17 is also recycled to the plasma membrane following internalization upon agonist stimulation through a fast recycling loop as revealed by co-localization with Rab4.

DISCUSSION
To investigate the intracellular trafficking of GPR17 in a native expression model, we have first characterized its expression in Oli-neu cells, an immortalized OPC cell line, and showed that medium derived from cortical neuron primary cultures stimulates the transient expression of GPR17, as previously observed in maturing cultured OPCs (2, 7, 10, 11). Using Oli-neu cells, we were then able to investigate the endocytic trafficking of the receptor following the administration of agonists and to demonstrate that (i) GPR17 undergoes clathrinmediated endocytosis and (ii) the internalized receptor is partially recycled to the cell surface and partially sorted to lysosomes for degradation. Furthermore, although UDP-glucose and LTD 4 bind to distinct sites on GPR17 (36), we found that both agonists mediate the trafficking of GPR17 to similar endocytic compartments and no obvious differences were observed between LTD 4 or UDP-glucose stimulation; the only exception is that the latter agonist causes a more efficient internalization of the receptor.

GPR17 expression in Oli-neu cells is highly regulated by neuron-conditioned medium
The Oli-neu cell line is derived from brain cultures enriched in OPCs after immortalization with the neu oncogene, and when maintained in normal Sato medium, the majority of cells express markers of immature oligodendrocyte phenotypes such as the O4 antigen and NG2 proteoglycan (20, 21 and our unpublished results); at this stage, only 4% of the cells express detectable amounts of GPR17. Because our previous study demonstrated that a sub-population of NG2-positive precursor cells upregulates GPR17 when induced to differentiate to post-mitotic premyelinating oligodendrocytes (11), we tested the conditions necessary to promote Oli-neu cell differentiation and GPR17 expression. Our immunofluorescence and biochemical data show that medium from neuronal primary cultures, while fostering Oli-neu cells towards a more differentiated phenotype, upregulates the expression of GPR17. According to previous studies in OPCs and also in Oli-neu cells, GPR17 is upregulated at early stages of differentiation and is then turned-down in cells with a more mature phenotype that synthesize high levels of myelin proteins (10,11). Furthermore, our data demonstrate that this is a transcriptionally mediated effect that is induced by factors released by neurons and/or astrocytes (which are also present, although in a minor amount, in cortical neuronal primary cultures) because neuron-conditioned (but not neurobasal unconditioned medium) efficiently activated the GPR17 gene promoter. A number of extrinsic signals (in cooperation with intrinsic mechanisms) are indeed expected to regulate oligodendroglial cell differentiation (and possibly modulate the expression of GPR17), including axonal surface ligands and secreted molecules, as well as axonal activity and glutamate released from axons (37 and refs therein). In line with evidence indicating GPR17 downregulation after the O4 stage (10 and the present data) and according to recent findings that have clearly coupled glutamate release with axonal myelination (38), an event known to occur at quite late stages of oligodendrocytes differentiation, we did not observe any activation of the GPR17 gene promoter or any increase in GPR17 synthesis in Olineu cells probed with glutamate or glutamate-receptor antagonists (data not shown). Alternatively, other factors may be involved. In this respect, we have also tested the action of ATP, platelet derived growth factor, leukemia inhibitory factor and insulin-like growth factor. These molecules are known to play important roles in OPC survival, differentiation and myelination, but they were unable to upregulate the expression of GPR17 in Oli-neu cells. Recently, Acilli and co-authors (29) demonstrated that in a subset of hypothalamic neurons secreting Agouti-related peptide, GPR17 expression is regulated by FoxO1. Microarray analysis indicated that GPR17 is highly decreased in FoxO1 deficient AgRP neurons, while FoxO1 overexpression increased the expression of GPR17 mRNA in neuro2A neuroblastoma cells. Our present results, however, indicate that the mechanism(s) underlying GPR17 expression in Olineu is different from that reported for AgRP neurons and FoxO1 does not appear to be involved. This conclusion comes from three observations: i) the amount of FoxO1 in Oli-neu is very low even after incubation with CM, ii) its knockdown does not reduce GPR17-expression and iii) its overexpression (even of the constitutively active forms) does not induce expression of GPR17. Although FoxO1 does not regulate GPR17 in Oli-neu, FoxO proteins may have important function in OPC proliferation and oligodendrogenesis. Recent studies have indeed demonstrated that, after hypoxia, these transcription factors, mainly FoxO1, can affect the expression of p27Kip1, a key regulator of oligodendrogenesis (39).
Concerning Oli-neu, we cannot exclude the possibility that factors present in the CM could counteract the signaling of the neu-oncogene, which has been used to immortalize the precursor cells isolated from E16 mouse brains and foster their proliferation (20). Exit from the mitotic cycle may trigger the expression of GPR17 and promote differentiation as has been reported for OPCs (10,11). This hypothesis is supported by the observation that an erbB inhibitor (PD174,265) induced the expression of GPR17 and the rapid (24 h) differentiation of Olineu cells (data not shown). Further work is required to identify additional molecules that may promote GPR17 gene activation and receptor protein synthesis.

Agonist-dependent GPR17 trafficking
Our previous work has shown that GPR17 is a "dualistic" receptor that is activated by uracil nucleotides and cysLTs (1, 2, 10). Moreover, studies in heterologously transfected cells demonstrated that, similar to other GPCRs, recombinant GPR17 can undergo a loss of function upon prolonged exposure to its endogenous agonists, as shown by desensitization, i.e., the inability to further respond to agonist stimulation, followed by re-sensitization upon agonist removal (6). However, the relevance of these findings to the regulation of native GPR17 in physiological systems has not been investigated previously. Moreover, in our previous study, the role of endocytic trafficking in agonist-induced GPR17 desensitization/re-sensitization was only marginally analyzed. The dynamics of intracellular trafficking of the receptor is, nevertheless, physiologically relevant, as it modulates receptor levels at the cell surface (12,(14)(15)(16)40). The internalization and subsequent sorting of the receptor into degradative or recycling compartments may have important implications for the activation/silencing of signaling pathway(s). In the case of GPR17, it may represent a key event necessary to achieve GPR17 downregulation, and in turn, to allow immature oligodendroglial cells to proceed to a more mature phenotype. A similar process has been associated with the differentiation of erythrocytes, where the downregulation of erythropoietin receptors via lysosomal degradation has been proposed to be necessary to allow cells to proceed towards terminal maturation (17). Therefore, it was important to investigate in detail the trafficking of GPR17 in a more physiological model such as differentiating Olineu cells. Furthermore, given the dualistic nature of GPR17 and the likely presence of two distinct binding sites, one for uracil nucleotides and one for cysLTs, in its structure (36), it was important to verify whether the two distinct classes of natural agonists modulate the endocytic pathways taken by the receptor in different ways.
In this study, we demonstrate that native GPR17 in Oli-neu cells is internalized in a time-and dose-dependent manner following the administration of agonists, although the efficacy of LTD 4 appeared to be less than that of UDP-glucose. This conclusion is supported by quantitative analysis performed by using biochemical (biotinylation of cell surface receptors) and immunocytochemical (immunolabeling of cell surface receptors with a specific antibody) approaches aimed at revealing the internalization of GPR17. Moreover, we demonstrate that, at least in our cell system, LTD 4 acts by directly activating GPR17 and not via the CysLT1R; this conclusion is based on the finding that this receptor is not expressed in Oli-neu cells. These data are in line with our previous data obtained in a heterologous expression system in which CysLTR1 is also not expressed (6), but we cannot exclude the possibility that GPR17 may be differentially regulated and therefore act as a negative regulator of CysLTR1 when both receptors are coexpressed in the same system as previously reported (18). However, this possibility will never occur in OPCs and pre-oligodendrocytes where CyLTR1 has not been detected (10), thus further confirming that in oligodendroglial cells, cysLTs act by directly promoting the activation of GPR17. Finally, the mechanisms and functional consequences of this difference in the internalization of GPR17 following exposure to either uracil nucleotides or cysLTs are not yet known. Differences in the efficiency of receptor internalization may be due to differences in binding to arrestins and/or the activation of different intracellular signals.
Although we did observe some quantitative differences in the efficiency of internalization, we did not observe any significant differences in the pathway taken by internalized receptors after UDP or LTD 4 stimulation. Using a protocol that has allowed us to visualize the fate of internalized receptors in endocytic compartments by means of confocal microscopy analysis, we demonstrated that GPR17 is internalized via clathrin-mediated endocytosis after both LTD 4 and UDP-glucose administration. At early times after internalization, the receptor reaches early endosomes independent of the agonist used for stimulation. At later times, it accumulated in dot-like structures and partially co-localized with a marker of lysosomes, thus suggesting that both agonists stimulated receptor degradation. This conclusion is also supported by the quantification of the amounts of biotinylated receptor recovered in cells after stimulation and agonist retrieval. In addition, these experiments indicate that while an aliquot of the receptor is degraded, a consistent portion of it is still recycled to the cell surface. Furthermore, using an immunocytochemical approach, we investigated whether internalized GPR17 is recycled via a rapid or slow recycling loop by comparing the distribution of GPR17 with that of the Rab proteins involved in endocytic traffic. After transfection in Oli-neu cells, Rab proteins exhibited a localization similar to that reported previously (32). Rab5 was mainly concentrated in vesicles that were dispersed in the cytoplasm and Rab4 was located in small vesicles and in dotted structures at the perinuclear region, whereas, given its distribution and cycling in the trans-Golgi-network, Rab11 was mainly detected in the perinuclear region. Our results demonstrate that GPR17 is transported into the early/sorting endosomes and can be recycled to the membrane via the fast recycling loop. This conclusion is supported by the observation that soon after internalization, GPR17 co-localized with Rab5 and Rab4 in a compartment at the periphery of the cell. At later times, GPR17 also accumulated in a Rab4positive compartment known to be involved in the rapid recycling loop. Only a very small amount of colocalization was observed with Rab11, a marker of the slow recycling loop.
In conclusion, we demonstrate that following agonist activation, GPR17 undergoes degradation and recycling. The balance of receptor sorting into recycling and/or degradative compartments may have important implications in the downregulation of GPR17 signaling pathway(s) in differentiating OPCs.      A. Schematic representation of the human (h), mouse (m) and rat (r) GPR17 gene. The region upstream to the coding sequence has been analyzed with the Genomatix suite. Small blocks (from 600bp to 1kb) have been identified as putative promoter regions. Five regions in human, two in mouse, and two in rat were found (in grey). Only the region 1 is highly conserved among the three species. The human sequence 1 (909 bp) has been cloned in a pGL4 vector for the reporter assay. B. Results of dual reporter luciferase assays showing the GPR17 promoter activity in Oli-neu cells after transfection with the reporter construct pGL4-h909 or the corresponding empty vector. This was followed by incubation in Sato medium (control), neurobasal medium plus B27 (+NB) or CM for 24-48 h. Data are from three independent experiments and represent the mean ± s.e. Statistical analysis was performed using a two-way ANOVA (*** P<0.0001). C. Cell extracts (30 µg protein) from Oli-neu cells incubated with either medium alone (-) or CM or neurobasal medium plus B27 (+NB) for the selected time points were analyzed by western blotting using antibodies against GPR17. D, Cell extracts from COS-7 (30 µg) or Oli-neu cells incubated with either Sato medium alone (-) or supplemented with CM for the selected time points were analyzed by Western blotting using antibodies against FoxO1 or GPR17. E, Oli-neu cells transiently transfected with cDNAs encoding for GFP-FoxO1 (FoxO1) or FLAG-FoxO1ADA were cultured in Sato medium for 48, fixed and then labeled with anti-Ct-GPR17 antibodies and anti-FLAG antibody followed by secondary antibodies. Cells were examined using an Axiovert 200M confocal system equipped with a spinning disc. Arrows indicate FoxO1transfected cells. Note that wild type protein is accumulated mainly in cytoplasm whereas the constitutively active form (FoxO1ADA) is accumulated in the nucleus. The images are representatives from three independent experiments. Bar = 20 µm.

Figure 6. Agonist-induced internalization of GPR17. Internalization of GPR17 was measured in differentiated
Oli-neu cells after cell surface labeling at 4°C with the thiol-cleavable sulfo-NHS-SS-biotin. A, after labeling cell were washed without (-) or with (+) DTT to remove the biotin from surface and cell extracts (250 µg) were incubated with streptavidin. The proteins bound to streptavidin beads and 5% of the cell extracts used for streptavidin-bead incubation (C.Ex) and also 5% of the supernatants from the streptavidin incubation (unboundproteins, Ub) were analyzed by western blotting with antibodies against TfR, GPR17 and actin. B, after labeling Oli-neu cells were incubated for 15 min at 37°C to allow for receptor internalization and then incubated without (-) or with (+) DTT. The cell extracts were analyzed as described above. In C, quantitative analysis of western blots revealed that 10.82±6.63% of the total biotinylated GPR17 underwent internalization under basal conditions. Data are from three independent experiments (the mean ± s.e.). ***, P<0.0001. In D and E, after labeling cells were extracted or incubated with either UDP-glucose (100 µM, UDPg) or LTD 4 (50 nM). At selected time points, the cells were cooled at 4°C, incubated with the membrane-impermeable reducing agent MesNA to remove biotin from the remaining cell surface receptors and solubilized. Aliquots (250 µg) of cell extracts were incubated with streptavidin-beads and proteins bound to streptavidin (E) and also 10 µg of total cell extracts (D) were analyzed by western blotting with antibodies against GPR17. F, after biotin labeling, Olineu cells were incubated at 37°C in the absence (-) or presence of UDP-glucose; UDP-glucose and cangrelor (10 µM, Cng); LTD 4 ; or LTD 4 and montelukast (1 µM, MTL). After 15 min, cells were cooled at 4°C, treated with reducing agents and solubilized. Cell extracts (200 µg) were incubated with streptavidin-beads and bound proteins were analyzed by western blotting with anti-GPR17 antibodies. The image shows representative blots from three independent experiments. G, graphs represent the quantitative analysis of blots and the data are expressed as the percentage of biotinylated GPR17 internalized in the absence of agonists (-) and represent the mean ± s.e. (* P<0.05; **, P<0.01, two-tailed, non-paired Student's t-test).

Figure 7.
Labeling of cell surface-exposed GPR17 with an antibody raised against the N-terminus of GPR17. Cells were cultured with CM for 48 h and then incubated for 45 min at 4°C with the anti-Nt-GPR17 antibody without or with pre-incubation with the antigen. After washing, the cells were fixed and labeled with a monoclonal antibody against MAG followed by an incubation with anti-mouse IgG conjugated to Cy3 (MAG, red) and an anti-rabbit IgG conjugated to fluorescein to reveal the distribution of the anti-GPR17 antibody (GPR17, green). Note that anti-Nt-GPR17 labels the surface of cells that do not express or express small amounts of MAG. Immunostaining for GPR17 was completely abolished by pre-incubating the antiserum with the peptide used for rabbit immunization. The merged images are shown: Bars = 10µm. Figure 8. Endocytosis of surface labeled GPR17. Differentiated Oli-neu cells that were surface-labeled with the affinity purified anti-Nt-GPR17 antibody were incubated without (control, CTR) or with agonists (50 nM LTD 4 or 100 µM UDP-glucose, UDPglc). After an 8-min incubation at 37°C, cells were washed with glycine buffer to remove the antibodies bound to the remaining cell surface exposed receptors, fixed and immunostained with a monoclonal antibody against TfR followed by incubation with anti-mouse IgG conjugated to fluorescein (TfR) and anti-rabbit IgG conjugated to Cy3 (GPR17). Confocal microscopy images show immunostaining for GPR17 was accumulated in small structures dispersed within the cytoplasm. Arrows indicate vesicles labeled for GPR17 and TfR. Images are representative of four independent experiments. The merged images are shown in A, B, C. Bars = 10 µm. . After glycine washing, cells were fixed and labeled with monoclonal antibodies against TfR or Lamp1 followed by incubation with anti-rabbit IgG conjugated to Cy3 (GPR, red) and anti-mouse or anti-rat IgG conjugated to fluorescein (TfR or LAMP, green). Merged images represent individual confocal sections (0.5 µΜ) and are representative of four independent experiments. Note a subset of cytoplasmic structures co-labeled (yellow) for GPR17 and TfR or Lamp1. Bar = 10 µm. Figure 10. GPR17 is internalized via clathrin-dependent endocytosis. In A, differentiated Oli-neu cells were surface-labeled with anti-Nt-GPR17 antibodies and then incubated for 12 min at 37°C without (control, CTR) or with agonists (LTD 4 or UDP-glucose, UDPglc) in Sato medium minus (-) or supplemented (+) with 450 mM sucrose. After glycine washing, cells were fixed and labeled with monoclonal antibodies against Lamp1 followed by incubation with anti-rat IgG conjugated to fluorescein (LAMP) and anti-rabbit IgG conjugated to Cy3 (GPR17). Confocal microscopy images show no accumulation of GPR17-immunostaining in the cells incubated in hypertonic sucrose. Images are representative of three independent experiments. The merged images are shown. Bar = 10 µm. In B, to quantify the effects of clathrin lattice inhibition on the endocytosis of GPR17, confocal images were collected from Oli-neu cells incubated with or without hypertonic sucrose (+), and the pixel intensity was determined by ImageJ software. The graph represents the increase of GPR17 immunoreactivity in the cells over background (background = the immunofluorescence detected in the cells after labeling at 4°C followed by acid stripping). The values are the mean (± s.e.) of three independent experiments.*, P>0.05; **, P<0.01; ***, P<0.0001 (two-tailed, non-paired Student's t-test).

Figure 11. Dynamin inhibition affects GPR17 endocytosis.
Oli-neu cells were labeled with anti-Nt-GPR17 antibodies, pre-treated at 4°C with or without the dynamin inhibitor dynasore and then incubated at 37°C for 12 min in Sato medium (Ctr) or medium supplemented with dynasore (+/-dyn) and UDP-glucose (UDP-g) or LTD 4 . After fixation, cells were double-immunolabeled with monoclonal antibodies against Lamp1 followed by incubation with anti-rat IgG conjugated to fluorescein (LAMP, green) and anti-rabbit IgG conjugated to Cy3 (GPR17, red). Note the large reduction of GPR17 internalization in Oli-neu incubated with the inhibitor of dynamin (Bar = 20 µm).

Figure 12. Recycling and degradation of GPR17 after agonist-induced endocytosis.
Oli-neu cells after cell surface labeling with sulfo-NHS-SS-biotin at 4°C were incubated at 37°C with UDP-glucose (100 µM, UDPg) or LTD 4 (50 nM). After 15 min, cells were cooled at 4°C, incubated with MesNA to remove biotin from the remaining cell surface receptors and solubilized (time 0) or further incubated in normal medium. At selected times, cells were either solubilized (-) or re-incubated with MesNA to remove biotinylated receptors re-exposed to the cell surface. In A, aliquots (200 µg) of cell extracts were incubated with streptavidin-beads, and the proteins bound to streptavidin were analyzed by western blotting with antibodies against GPR17. In B, graphs represent the quantitative analysis of GPR17 recycling and degradation, which is reported as a percentage of the amount of receptor internalized after 15 min at 37°C with LTD 4 or UDP-glucose; the values are the mean of four independent experiments ± s.e. Figure 13. At early time points, endocytosed GPR17 accumulates in Rab5-or Rab4-positive vesicles. Olineu cells transiently transfected with cDNA encoding for GFP-Rab5a (Rab5) or GFP-Rab4a (Rab4) were labeled with anti-Nt-GPR17 antibodies and then stimulated with LTD 4 or UDP-glucose (UDPglc) for 15 min. After glycine washing, the cells were fixed, and the endocytosed GPR17 was visualized by labeling with Cy3conjugated antibodies (red). The cells were then examined using an Axiovert 200M confocal system equipped with a spinning disc. Note the co-localization of Rab proteins with GPR17 (arrows) in small vesicles scattered throughout the cytoplasm. The images are representatives from three independent experiments. Bar = 5 µm.  Olineu cells transiently transfected with cDNA encoding for GFP-Rab11a (Rab11) were labeled with anti-Nt-GPR17 antibodies and then stimulated with LTD 4 or UDP-glucose (UDPglc) for 30 min. After glycine washing, the cells were fixed, and the endocytosed GPR17 was visualized by labeling with Cy3-conjugated antibodies (red, GPR). The cells were examined using an Axiovert 200M confocal system equipped with a spinning disc. The merged and single channel images shown are representatives from three independent experiments. Bar = 5 µm.