Uridine 5′-Diphosphate Induces Chemokine Expression in Microglia and Astrocytes through Activation of the P2Y₆ Receptor Free

Microglia and astrocytes play important roles in the brain inflammation that accompanies brain injury. In response to such injury, microglia and astrocytes release inflammatory mediators such as cytokines and chemokines stimulating subsequent brain inflammation (1–3). Chemokines are implicated in the pathogenesis of a number of neurologic diseases, including Alzheimer’s disease, cerebral ischemia, and multiple sclerosis (4). Chemokines are major effector molecules recruiting blood inflammatory cells to the injury site (5), and recently, several studies have reported that blood inflammatory cells also enter the injured brain and participate in brain inflammation (6–8). In particular, CCL2 (MCP-1) and CCL3 (MIP-1α) are important chemokines associated with monocyte infiltration into the injured brain (9–12).

Nucleotides have been explored as inflammatory inducers of brain inflammation (13). Although the roles of adenine nucleotides such as ATP in inflammatory processes were earlier intensively studied, research focus shifted to uridine nucleotides (UTP and UDP) after discovery of specific membrane receptors for these factors (14, 15). In the injured brain, uridine nucleotides are rapidly secreted or leaked to the extracellular space. For example, kainic acid increased the concentration of extracellular UTP (16). In astrocytes, increases in intracellular calcium induced UTP release to extracellular levels >20 times normal concentrations (17, 18). Such increases in extracellular UTP also resulted in increases in extracellular UDP, because secreted UTP is converted to UDP by ectonucleoside diphosphokinase (19).

Extracellular nucleotides including uridine nucleotides exert their effects through the activation of P2 purinergic receptors. To date, seven ionotropic (P2X_1–7) and eight metabotropic receptors (P2Y_{1, 2, 4, 6, 11–14}) have been discovered. It has been reported that both P2Y purinergic receptors and P2X₇ receptors play roles in inflammatory processes such as phagocytosis and cytokine/chemokine production (16, 20–24). However, involvement details of specific P2Y receptors and the exact roles of these receptors in brain inflammation are largely unknown. Microglia and astrocytes express P2Y_{1, 2, 6, 12–14} and P2Y_{1, 2, 4, 6, 13, 14}, respectively (14, 16, 18, 25–28). Each P2Y receptor has different affinities for specific nucleotides; ATP has a relatively high affinity for P2Y_{1, 2, 4, 12,} and ₁₃; ADP for P2Y_{1, 12,} and ₁₃; UTP for P2Y_{2, 4,} and ₆; and UDP for P2Y₆ (14).

The results of the current study showed that UDP significantly induces chemokine synthesis, particularly that of CCL2 and CCL3, in primary cultured microglia, astrocytes, and brain slices by activation of P2Y₆ receptors. Furthermore, UDP-treated astrocytes were major recruiters of blood monocytes.

FBS was purchased from Hyclone (Logan, UT). RNAzol B was from iNtRON Biotechnology (Sungnam, Korea) and reverse transcriptase from Promega (Madison, WI). Real-time master mix was from Kapa Biosystems (Cape Town, South Africa). Oligonucleotide primers were purchased from IDT (Coralville, IA). Small interfering RNAs (siRNAs) including control siRNA and siGLO were purchased from Dharmacon (Lafayette, CO). Opti-MEM and RNAiMAX were from Invitrogen (Basel, Switzerland). Transwell plates (pore size 5.0 μm) were from Corning (Lowell, MA). BAPTA-AM was purchased from Biomol International (Plymouth Meeting, PA). Cyclosporin A (CsA) and inhibitor of NFAT-calcineurin association-6 (INCA-6) were from Calbiochem (San Diego, CA). ATP, ADP, UDP, UMP, uridine, LPS, apyrase, suramin, MRS2578, U73122, MEM, and other reagents were from Sigma-Aldrich (St. Louis, MO).

Primary microglia were cultured from the cerebral cortices of 1–3-d-old Sprague Dawley (SD) rats (Samtako, Osan, Korea) as described previously (29, 30). All animal procedures were approved by the Ajou University Institutional Animal Experimentation Committee. Briefly, cortices were triturated into single cells using Pasteur pipettes in MEM containing 10 mM HEPES (pH 7.4), 2 mM glutamine, and 10% (v/v) FBS and plated into 75 cm² T-flasks (half a hemisphere in each flask) using the same culture medium. After 2 wk of culture, microglia were detached from flasks by gentle shaking and filtered through nylon mesh to remove astrocytes and cell clumps. Microglia were counted and plated (5 × 10⁴ cells/cm²). Isolated microglia were allowed to attach to plates for 3 to 4 h, and unattached cells were removed by changing the medium. Microglial purity was confirmed by morphology analysis and/or by staining using a microglial marker isolectin B₄ (from Griffonia simplicifolia) (31). More than 95% of cells were isolectin B₄ positive (data not shown) and subsequently used in experiments.

Primary astrocytes remaining in the flask were harvested with 0.1% (w/v) trypsin, plated (3 × 10⁴ cells/cm²), and cultured with 5% (v/v) FBS-containing MEM for 1 to 2 d and serum-free MEM for 2 to 3 d. Then, astrocytes were detached with trypsin once again and plated (2 to 3 × 10⁴ cells/cm²) with 5% (v/v) FBS-containing MEM. Astrocytes were incubated at least overnight to minimize any effects induced by changing medium before they were used in experiments. More than 95% of purified astrocytes were GFAP positive (data not shown).

Adult astrocytes were cultured as described previously (32) with some modifications. Briefly, cortices obtained from 6–8-wk-old SD rats were chopped with a razor blade, incubated with 0.1% (w/v) trypsin (Hyclone) for 10 min, and washed with MEM containing 10% (v/v) FBS. Single cells were obtained by titration, layered over a 0.4 M sucrose solution, and centrifuged at 400 × g for 10 min. The pellet was washed and resuspended in MEM containing 10 mM HEPES (pH 7.4), 2 mM glutamine, and 10% (v/v) FBS and cultured in 75-cm² culture flasks (one half hemisphere per two flasks). Three days later, floating cells were removed by changing the medium. When cells reached confluence 3 to 4 wk later, they were subcultured and incubated at least overnight before experiments.

For experiments employing inhibitors of P2Y₆, calcium, phospholipase C (PLC), and/or mNFAT, cultures were pretreated for 5 min with the appropriate inhibitors before UDP treatment.

Cortical slices were prepared using a modified Stoppini method (33). Briefly, 1-d-old SD rats were anesthetized and then decapitated. Brains were rapidly removed and cortices separated by thin forceps in culture medium. Coronal cortical slices (400-μm thick) were prepared using a McIlwain tissue chopper (Mickle Laboratory Engineering, Goose Green, U.K.). Slices were placed into 24-well plates in which each well was filled with 500 μl culture medium (MEM containing 25% [v/v] HBSS, 25% [v/v] heat-inactivated horse serum [Hyclone], 6.5 mg/ml glucose, 1 mM l-glutamine, 10 U/ml penicillin-G, and 10 mg/ml streptomycin). Using ethidium homodimer-1 (EthD-1; Molecular Probes, Eugene, OR), cell viability in slices was assayed.

mRNA expression was detected with quantitative real-time PCR (qPCR). Total RNA was isolated using RNAzol B (iNtRON Biotechnology) and cDNA prepared using reverse transcriptase from avian myeloblastosis virus, according to the manufacturer’s (Promega) instructions. For qPCR, cDNA and forward/reverse primers (200 nM) were added to 2× KAPA SYBR Fast Master Mix (Kapa Biosystems). qPCR were performed on RG-6000 real-time amplification instrument (Corbett Research, Sydney, Australia). qPCR reaction protocol was denaturation at 95°C for 30 s, followed by 40 cycles of 95°C for 3 s, 55°C for 20 s, and 72°C for 3 s. Threshold cycle number of each gene was calculated, and GAPDH was used as a reference gene. The Δ-δ threshold cycle values of CCL2 and CCL3 were presented as relative fold induction. The following primers were used for amplification of rat CCL2, CCL3, P2Y₆, P2X₇, Iba-1, GFAP, and GAPDH: CCL2, 5′-ATG CAG GTC TCT GTC ACG CT-3′ (sense) and 5′-CTA GTT CTC TGT CAT ACT GG-3′ (antisense); CCL3, 5′-ATG AAG GTC TCC ACC ACT-3′ (sense) and 5′-TCA GGC ATT CAG TTC CAG-3′ (antisense); P2Y₆, 5′-GTG GTA TGT GGA GTC GTT TG-3′ (sense) and 5′-CTG TAG GAG ATC GTG TG GTT-3′; P2X₇, 5′-GTG CCA TTC TGA CCA GGG TTG TAT AAA-3′ (sense) and 5′-GCC ACC TCT GTA AAG TTC TCT CCG ATT-3′ (antisense); Iba-1, 5′-TTG ATC TGA ATG GCA ATG GA-3′ (sense) and 5′-CCT CCA ATT AGG GCA ACT CA-3′ (antisense); GFAP, 5′-TTC CTG TAC AGA CTT TCT CC-3′ (sense) and 5′-CCC TTC AGG ACT GCC TTA GT-3′ (antisense); and GAPDH, 5′-TCC CTC AAG ATT GTC AGC AA-3′ (sense) and 5′-AGA TCC ACA ACG GAT ACA TT-3′ (antisense).

Levels of chemokines CCL2 and CCL3 secreted to the media were measured using ELISA kits according to the manufacturer’s (Antigenix America) instructions. We also analyzed CXCL2/3 (MIP-2) secretion. But because UDP increased CXCL2/3 secretion in microglia but not in astrocytes (data not shown), we focused on CCL2 and CCL3.

For knockdown of P2Y₆ receptor expression, we used ON-TARGET plus siRNAs targeting the rat P2Y₆ gene (Dharmacon, Lafayette, CO). Primary astrocytes (2 × 10⁴ cells/cm²) and microglia (5 × 10⁴ cells/cm²) were seeded into 12-well plates and used at 50–70% or 90% confluence, respectively. One day before transfection, medium was replaced by Opti-MEM (Invitrogen). Transfection was performed using 20–40 nM siRNA and the RNAiMAX transfection reagent according to the manufacturer’s instructions (Invitrogen). Transfection efficiency was measured by fluorescence microscopy using siGLO Red (Dharmacon). At least 90% of cells were transfected (data not shown). Four to 5 d post-transfection, cells were used in knockdown experiments.

NFAT binding activities were measured using an EMSA, as described previously (34). Primary microglia (1.5 × 10⁶ cells) were harvested and suspended in 900 μl hypotonic solution (10 mM HEPES, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, and 0.5 mM PMSF [pH 7.9]) for 15 min, followed by incubation in hypotonic solution containing 0.5% (v/v) Nonidet P-40 for 5 min. Cells were centrifuged at 500 × g for 10 min at 4°C and the pellet containing the nuclear fraction resuspended in a solution containing 20 mM HEPES, 20% (v/v) glycerol, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, and 1 mM PMSF (pH 7.9). Samples were centrifuged at 10,000 × g for 10 min, and crude nuclear fractions (supernatants) were collected and stored at −70°C until use.

The NFATc consensus oligonucleotide (5′-CGC CCA AAG AGG AAA ATT TGT TTC ATA-3′) and mutant oligonucleotide (5′-CGC CCA AAG CTT AAA ATT TGT TTC ATA-3′) (Santa Cruz Biotechnology, Santa Cruz, CA) were end-labeled using T4 polynucleotide kinase (Promega) and [γ-³²P] ATP (PerkinElmer, Waltham, MA). The labeled DNA probe (∼0.5 ng) was incubated for 30 min with 2 μg nuclear proteins in a reaction mixture containing 21.4 mM EDTA, 21.4 mM EGTA, 20% (v/v) glycerol, 0.29 mM ZnSO₄, 10 ng/ml poly(deoxyinosinic-deoxycytidylic) acid, 1 mM DTT, 0.4 mg/ml BSA, and 8 mM MgCl₂. The reaction mixture was resolved on a 6% (w/v) polyacrylamide gel. For supershift assays, nuclear extracts were preincubated with a 1-μg combination of anti-NFATc1, anti-NFATc2, and control IgG Abs (Affinity Bioreagents, Golden, CO) for 10 min at 4°C before probe addition. The reaction mixture was subjected to electrophoresis on a 6% (w/v) polyacrylamide gel.

Primary microglia were seeded onto coverslips (Fisher Scientific, Pittsburgh, PA) and incubated overnight. After incubation with UDP (100 μM), cells were fixed in 4% (w/v) paraformaldehyde for 15 min, permeabilized in 0.1% (v/v) Triton X-100, and blocked with 1% (v/v) bovine serum. Next, microglia were incubated with Abs to NFATc1 (1:2000; Affinity Bioreagents), NFATc2 (1:200; Affinity Bioreagents), or anti-ionized calcium-binding adaptor molecule (Iba-1; 1:1000, Wako Pure Chemicals) overnight at 4°C. Alexa 555-conjugated anti-mouse donkey IgG and Alexa 488-conjugated anti-rabbit donkey IgG (Invitrogen) were used to detect NFATcs and Iba-1, respectively. Immunostained cells were mounted using VECTASHIELD containing DAPI (Vector Laboratories, Burlingame, CA). Confocal microscopy images were obtained (LSM510; Carl Zeiss).

Cortical slices were placed in a 96-well plate (one slice per well) filled with 80 μl HEPES-buffered solution (130 mM NaCl, 5 mM KCl, 1.5 mM CaCl₂, 1 mM MgCl₂, 10 mM glucose, and 10 mM HEPES [pH 7.4]) immediately after slicing and incubated with or without ectonucleotidase inhibitor (10 μM dipyridamole) for 10 min to 1 h. The levels of UDP secreted into the solution were measured using a fluorescence probe based on a perylene-dpa-Zn platform (Probe-1) provided by Dr. Juyoung Yoon (Department of Chemistry and Nano Science, Ewha Woman’s University, Suwon, Korea) (35). HEPES-buffered solution (80 μl) was mixed with Probe-1 (10 μM), and fluorescence levels were measured at 485/535 nm (excitation/emission) with a Victor3 fluorescence reader (PerkinElmer). Probe-1 has greater specificity for UDP than UTP and has little specificity for ATP, ADP, AMP, UMP, or GTP (35).

Rat monocytes were isolated and labeled as described (6), with some modifications. Briefly, blood was collected from 5-wk-old SD rats by cardiac puncture. Blood was mixed with PBS containing 2.5% (w/v) dextran (blood/PBS = 1:4) to sediment erythrocytes. The plasma layer was spun down, and the pellet was suspended in PBS, incubated with 10 μM carboxyfluorescein diacetate (CFDA; Molecular Probes), loaded onto a Ficoll gradient (Amersham Pharmacia Biotech, San Francisco, CA), and centrifuged (400 × g) for 30 min. The monocyte/lymphocyte layer at the interface was carefully collected and washed three times with sterile PBS. Cells were suspended in HBSS containing calcium (140 mg/l) and seeded in a Petri dish for 30 min. Unattached lymphocytes were removed and adherent monocytes collected and suspended in MEM containing 1% (v/v) FBS.

Primary astrocytes were cultured in 24-well plates (3 × 10⁴ cells/cm²), and medium was replaced with 600 μl/well MEM with 1% (v/v) FBS 1 d before the transmigration assay was conducted. Primary astrocytes were either untreated or preactivated with UDP (100 μM) 1 h prior to addition of monocytes. Freshly purified CFDA-labeled monocytes (4 × 10⁵ cells/100 μl) were added to the top of each transwell insert of 5-μm pore size (Corning) and allowed to migrate 8–10 h. The number of CFDA-positive monocytes transmigrating into the bottom chamber were counted in three randomly selected fields (×20 objective) under a fluorescent microscope (Zeiss Axiovert 200M; Carl Zeiss).

Data were analyzed by Student t test, one-way ANOVA, or two-way ANOVA followed by post hoc comparisons (Student-Newman-Keuls approach) using the Statistical Package for Social Sciences version 8.0 (SPSS, Chicago, IL).

In the injured brain, nucleotides such as ATP, ADP, UTP, and UDP are released at high levels from neurons and astrocytes (16, 36, 37). As microglial activation accompanies brain injury, we examined whether extracellular nucleotides could alter microglial function. Thus, we treated microglia with 100 μM ATP, ADP, UTP, or UDP and examined the expression levels of chemokines CCL2 and CCL3 by ELISA (Fig. 1A). Microglia released CCL2 and CCL3 in the range of 1–20 pg/μg protein. UDP was the most powerful inducer of CCL2, and ADP, UTP, and UDP induced CCL3 to similar extents, whereas ATP barely induced both chemokines (Fig. 1A). Based on these results, we selected UDP for use in further experiments.

We next examined the effect of UDP metabolites on chemokine expression, because UDP can be metabolized into UMP and uridine in culture media. In qPCR, UDP significantly induced microglial CCL2 and CCL3 mRNA expression. However, both UMP and uridine had little effect on chemokine expression (Fig. 1B). Accordingly, apyrase, an enzyme degrading UDP into UMP or uridine, also inhibited UDP-induced CCL2 and CCL3 mRNA expression (Fig. 1C). These results suggest that UDP per se is the active component inducing chemokine expression in microglia.

In the following experiments, UDP (100 μM) induced expression of mRNA encoding CCL2 and CCL3 within 30 min, and such expression reached maximal levels at 1 h, returning to basal levels at 3–12 h (Fig. 2A). Secretion of CCL2 and CCL3 increased within 1 h, peaked at 3–6 h, and was maintained for up to 24 h (Fig. 2B). UDP induced a concentration-dependent increase in the expression of chemokines at both the mRNA and protein level over a concentration range of 10 μM to 1 mM (Fig. 2C, 2D).

UDP is known to be a specific agonist of the P2Y₆ purinergic receptor (14), and P2Y₆ is known to be expressed in microglia (25, 26). Thus, we investigated whether P2Y₆ mediated UDP-induced production of chemokines in cultured microglia. MRS2578, a specific antagonist of the P2Y₆ receptor (38), significantly inhibited UDP-induced mRNA and protein expression of both chemokines at 2–5 μM (Fig. 3A, 3B). No cytotoxic effects of MRS2578 were evident at 5 μM (data not shown).

As the P2Y₆ receptor is a Gq/G₁₁-coupled receptor activating PLC and subsequently increasing intracellular calcium (39), we examined the effects of a PLC inhibitor (U73122) and an intracellular calcium chelator (BAPTA-AM) on UDP-induced chemokine expression in microglia. As expected, U73122 (1–5 μM) and BAPTA-AM (10–50 μM) significantly reduced UDP-induced chemokine mRNA expression (Fig. 3C, 3D).

P2Y₆ siRNA (siP2Y₆) also significantly inhibited UDP-induced chemokine secretion in primary microglia, whereas nontargeting control siRNA (siCont) had virtually no effect (Fig. 3E, 3F). Taken together, these results indicate that UDP induces chemokine expression in microglia by activation of P2Y₆ receptors.

Next, we investigated the involvement of transcription factors mediating UDP-induced chemokine expression. NF-κB is widely known as a transcription factor involved in chemokine expression by various cell types including microglia (40–43). However, UDP barely activated NF-κB in primary microglia compared with LPS (data not shown). As P2Y₆ is a Gq-coupled receptor, and as intracellular calcium increase is required for UDP-induced chemokine expression, as shown above (Fig. 3), we hypothesized that NFAT would be involved in this process. In EMSA, microglial NFAT activation was detected at 15–30 min post-treatment with either UDP or ionomycin (positive control) (Fig. 4A). The NFAT–probe complex was not formed with NFAT mutant (mut) probe changing AGG to CTT in the binding motif (Fig. 4B, lane 3). Moreover, the UDP-induced NFAT–probe complex was effectively competed by excessive cold wild-type probe (Fig. 4B, lane 4) but not by the mutant cold probe (Fig. 4B, lane 5). The NFAT–probe complex was disappeared or supershifted in the presence of Abs against NFATc1 and c2 but not in the presence of control IgG Abs (Fig. 4C). We also found that both NFATc1 and c2 were translocated from the cytosol to the nucleus within 15 min after UDP treatment, and MRS2578 inhibited translocation of NFATs (Fig. 4D). The specificity of NFATc1 and NFATc2 immunoreactivity was confirmed by staining with control IgG Abs in the absence of NFAT Abs (Fig. 4D).

The involvement of NFAT in UDP-induced chemokine expression was further confirmed using inhibitors of NFAT (INCA-6) (44) and calcineurin (CsA). Calcineurin is a calcium-dependent phosphatase activating NFAT (45). qPCR showed that 10–20 μM INCA-6 significantly reduced UDP-induced CCL2 and CCL3 mRNA expression (Fig. 5A). As expected, CsA in the range of 10–50 μM also attenuated UDP-induced mRNA expression of CCL2 and CCL3 in qPCR (Fig. 5B). These results indicated that NFAT mediates UDP-induced chemokine production in microglia, functioning as a downstream transcription factor of the P2Y₆ receptor.

Several lines of evidence indicate that astrocytes are the main source of chemokines in many neurologic diseases, although microglia play central roles in initiation of inflammatory processes (1, 46). Thus, we examined whether UDP induced chemokine expression in astrocytes by activation of P2Y₆, as in microglia. Primary astrocytes express P2Y₆ (25) and, in response to UDP (100 μM), astrocytes increased expression of mRNAs encoding CCL2 and CCL3 within 30 min (Fig. 6A). The elevated mRNA levels were maintained for up to 3 h and then rapidly decreased to the basal level by 12 h (Fig. 6A). CCL2 secretion increased continuously from 3 h (∼10 pg/μg protein) to 24 h (∼120 pg/μg protein) post-treatment. However, CCL3 was barely produced (<0.5 pg/μg protein), even 24 h after UDP treatment (Fig. 6B). We further tested whether adult astrocytes also express chemokines in response to UDP. In adult cultures, most cells (∼90%) were GFAP positive, but Iba-1–positive cells were also detected (data not shown). As was the case in neonatal astrocyte cultures, UDP (100 μM) also enhanced CCL2 levels in adult astrocytes, increasing production from 90 to 260 pg/μg protein between 3 and 24 h. CCL3 production was also increased by UDP (from 0.4 to 2.6 pg/μg protein between 3 and 24 h) (Fig. 6C).

Involvement of P2Y₆ in UDP-induced chemokine expression was assayed in two ways. First, MRS2578 (2–10 μM) inhibited chemokine expression (Fig. 7A). Furthermore, use of siP2Y₆ supported the involvement of P2Y₆ in UDP-induced chemokine expression in astrocytes as in microglia. In astrocytes transfected with siP2Y₆, P2Y₆ expression was dramatically reduced to levels <10% of those in siCont-treated cells, without any effect on expression of P2X₇, another purinergic receptor (Fig. 7B). P2Y₆ siRNA significantly decreased UDP-induced mRNA expression encoding CCL2 and CCL3, whereas control siRNA had little effect on mRNA levels encoding either P2Y₆ or the chemokines (Fig. 7C). Accordingly, siP2Y₆ reduced CCL2 secretion in response to extracellular UDP (Fig. 7D). In additional experiments, we found that UDP decreased P2Y₆ mRNA levels within 1 h, and this was sustained for up to 3 h (data not shown). Taken together, these results indicate that P2Y₆ mediates UDP-induced chemokine expression in astrocytes.

Next, we investigated whether P2Y₆ regulated chemokine expression in slice cultures of rat brain cortex. It is well known that the process of slice preparation caused traumatic injury (47). First, we analyzed the levels of UDP secreted from cortical slices using Probe-1. This probe specifically detects UDP and UTP (with higher sensitivity for UDP) and not other nucleotides, including ATP, ADP, and UMP/uridine (35). Within 10 min after slicing, ∼6.4 ± 3.5 μM UDP was detected, and the levels increased to 21.7 ± 4.7 μM after 1 h (Table I). In the presence of dipyridamole, an inhibitor of ectonucleotidase that inhibits degradation of UTP/UDP into UMP/uridine, UDP levels further increased to 228.5 ± 22.7 μM at 10 min (Table I).

In slice cultures, the mRNAs encoding CCL2 and CCL3 increased within 30 min of slicing and reached the first peak at 6 h (Fig. 8A). CCL2 mRNA expression further increased up to 72 h, whereas CCL3 mRNA levels rapidly decreased at 12 h (Fig. 8A). Analysis of chemokine secretion showed that CCL2 levels significantly increased from 3 to 72 h (0.4 to 30 pg/μg protein) compared with those of CCL3 (0.5 to 0.9 pg/μg protein) (Fig. 8B). We further found that exogenously added UDP (100 μM) did not enhance chemokine secretion from the slices (Fig. 8C). These results suggest that the amount of UDP released from slices was sufficient to induce chemokine expression.

We next investigated involvement of P2Y₆ in CCL2 expression from slice cultures using the inhibitor MRS2578. When slices were cultured with MRS2578 (10 μM), both mRNA expression encoding CCL2 (Fig. 8D) and chemokine secretion (Fig. 8E) were significantly reduced. We confirmed that MRS2578 had little effect on astrocyte and microglial viability in slices, as shown by both EthD-1 staining and analysis of mRNA levels encoding GFAP and Iba-1, markers of astrocytes and microglia, respectively (Fig. 8F). These results suggest that P2Y₆ may mediate chemokine expression in the injured brain.

As CCL2 is well known to function in monocyte recruitment, we examined whether UDP-treated astrocytes were able to recruit monocytes. To this end, monocytes labeled with CFDA-succinimidyl ester and astrocytes were cultured in the upper and lower chambers, respectively, of transwells, and the numbers of monocytes migrating to lower chambers were counted (Fig. 9). Monocyte levels detected in lower chambers increased ∼1.6-fold when astrocytes were treated with UDP (Fig. 9A, 9B). However, in the presence of brefeldin A (BFA), an inhibitor of protein trafficking from the endoplasmic reticulum to the Golgi, migration of monocytes toward UDP-treated astrocytes significantly decreased (Fig. 8A, 8B). Astrocytes treated with UDP for only 5 min still induced monocyte migration (Fig. 9B). The extent of monocyte migration correlated with the extent of CCL2 secretion because astrocytes treated with UDP either overnight or for 5 min produced similar levels of CCL2, and astrocytes treated with UDP in the presence of BFA did not produce CCL2 (Fig. 9C). We also examined the effect of UDP-treated microglia on transmigration of monocytes. In accordance with differences in secretion levels of CCL2 from astrocytes and microglia (>100 pg/μg protein from astrocytes versus <20 pg/μg protein from microglia), UDP-activated microglia barely induced monocyte migration (data not shown). Taken together, the results suggest that UDP released from injured brain cells may activate astrocytes and microglia to secrete chemokines. Furthermore, astrocytes could play major roles in recruiting blood inflammatory cells into the brain.

In this study, we have demonstrated a novel pathway for induction of chemokine expression in microglia and astrocytes. UDP released from damaged cells could induce chemokine synthesis in microglia and astrocytes through activation of P2Y₆. In fact, significant amounts of UDP were rapidly released from cortical slices (Table I), and exogenously added UDP had no additional effect on chemokine expression.

In response to UDP, microglia rapidly (within 3 h) produced small amounts of chemokines (<20 pg/μg protein), whereas astrocytes cultured from both neonatal and adult rats continuously (over 9–24 h) produced large amounts of chemokines, particularly CCL2. Furthermore, UDP-treated microglia barely enhanced monocyte migration (data not shown) whereas UDP-treated astrocytes strongly recruited monocytes. These findings led us to speculate that the functions of chemokines produced by microglia and astrocytes may differ. Chemokines not only recruit inflammatory cells to injury sites but also exert several other functions. Recently, neuroprotection effects of chemokines such as CCL2, CCL5 (RANTES), CXCL8 (IL-8), and CXCL2/3 (MIP-2) have been reported (48–50). Chemokines also modulate ion channel activities: CCL2 activates nonselective cation channels and inhibits delayed-rectifier potassium channels (51), and CCL3 modulates the activities of transient receptor potential channels (51). Chemokines may sensitize immune cells to inflammatory stimuli. Thus, when LPS was injected into the brain of CCL2^−/− mice, less IL-1β and TNF-α were produced compared with levels seen in normal mice (52). Therefore, rapidly released small amounts of chemokines from UDP-activated microglia may function to protect neurons or to modulate channel activity in the injured brain rather than to recruit blood inflammatory cells. In contrast, large amounts of chemokines produced by astrocytes may play major roles in recruitment of blood inflammatory cells. In animal models of the ischemic and traumatic brain, the duration (9–24 h) of chemokine production from astrocytes was correlated with the time (12–24 h) required for infiltration of neutrophils and monocytes (53). These findings support the hypothesis that astrocyte-derived chemokines play important roles in recruitment of blood inflammatory cells.

Recently, several lines of evidence have indicated that purinergic receptors activated by extracellular nucleotides mediate various inflammatory responses, including production of reactive oxygen species and expression of cytokines and chemokines (54, 55). In the current study, we showed that UDP-activated P2Y₆ mediated chemokine production in both microglia and astrocytes, based on several experimental approaches. First, among ATP, ADP, UTP, and UDP (all at 100 μM), UDP, with the highest affinity for P2Y₆ (14), was the most potent chemokine inducer (Fig. 1). Second, siRNA directed against P2Y₆ abrogated UDP-induced chemokine expression (Figs. 3, 7). Third, in slice cultures, a P2Y₆ antagonist reduced chemokine expression (Fig. 8). Recently, growing evidence has suggested that P2Y₆ may mediate inflammatory responses in several types of cells. In monocytes and intestinal and lung epithelial cells, P2Y₆ mediated secretion of CXCL8 and CCL20 (MIP-3α) (22, 23, 56, 57). In promonocytic U937 cells transfected with human P2Y₆, UDP induced expression of CCL2, CCL10 (IP-10), and CXCL8 (58). In microglia, the P2Y₆ receptor is known to mediate phagocytosis of damaged cells (16).

P2X₇ is also a well-known purinergic receptor mediating CCL2 expression in astrocytes (59). We found that, in astrocytes, UDP induced CCL2 to a level 30–40% of that seen after 3-O-(4-benzoyl)-ATP (an agonist of the P2X₇ receptor) addition (data not shown). However, in microglia, P2X₇ might not play a major role in CCL2 expression because ATP (even in the millimolar range) and 3-O-(4-benzoyl)-ATP only weakly induced CCL2 expression compared with UDP (Fig. 1A and data not shown). P2Y₂ also regulates CCL2 secretion in rat alveolar and peritoneal macrophages (24). However, P2Y₂ may not mediate CCL2 synthesis in microglia because ATP, a strong agonist of P2Y₂, did not increase CCL2 production (Fig. 1A). Therefore, we speculate that P2Y₆ may mediate chemokine expression in many cell types in the presence of even micromolar UDP levels.

Next, we investigated the molecular mechanism of P2Y₆-mediated chemokine expression. P2Y₆ is a Gq/G₁₁-coupled receptor that activates PLC, subsequently activating protein kinase C and increasing intracellular Ca²⁺ (38). We found that, in microglia, a PLC inhibitor and an intracellular calcium chelator reduced UDP-induced chemokine expression. Although NF-κB is a major transcription factor regulating chemokine expression (43), UDP barely activated NF-κB in primary microglia compared with the activation seen after LPS stimulation (data not shown). It has been reported that NFAT regulated CCL2 and CCL4 expression in human mast cells (60, 61). Furthermore, in vascular smooth muscle cells and astrocytes, P2Y₆ activation was associated with NFAT upregulation (62, 63). We found that, in microglia, UDP activated NFATc1 and NFATc2, and inhibitors of NFAT and calcineurin blocked UDP-induced chemokine expression (Figs. 4, 5). Therefore, calcineurin-dependent NFAT activation appears to play an important role in chemokine expression in microglia, acting downstream of UDP-P2Y₆ activation.

To the best of our knowledge, this is the first report to show that UDP induces chemokine production in both microglia and astrocytes by activation of P2Y₆. Furthermore, we suggest different roles for chemokines released from microglia and astrocytes because both the kinetics and extent of chemokine production from microglia and astrocytes differ. Chemokines from microglia may induce surrounding tissues to react to danger signals, whereas chemokines from astrocytes may recruit blood inflammatory cells to protect the brain from infection and repair injury. It is conceivable that this novel mechanism of chemokine induction described in this study, involving P2Y₆ activation, is an important pathway regulating specific inflammatory responses for treatment of various neurologic injuries associated with brain inflammatory processes.

The fluorescent Probe-1 was kindly provided by Dr. Juyoung Yoon (Ewha Womans University, Seoul, South Korea).

The authors have no financial conflicts of interest.