Adenosine A2A Receptor mRNA Regulation by Nerve Growth Factor Is TrkA-, Src-, and Ras-dependent via Extracellular Regulated Kinase and Stress-activated Protein Kinase/c-Jun NH2-terminal Kinase*

We have shown previously that nerve growth factor (NGF) down-regulates adenosine A2A receptor (A2AAR) mRNA in PC12 cells. To define cellular mechanisms that modulate A2AAR expression, A2AAR mRNA and protein levels were examined in three PC12 sublines: i) PC12nnr5 cells, which lack the high affinity NGF receptor TrkA, ii) srcDN2 cells, which overexpress kinase-defective Src, and iii) 17.26 cells, which overexpress a dominant-inhibitory Ras. In the absence of functional TrkA, Src, or Ras, NGF-induced down-regulation of A2AAR mRNA and protein was significantly impaired. However, regulation of A2AAR expression was reconstituted in PC12nnr5 cells stably transfected with TrkA. Whereas NGF stimulated the mitogen-activated protein kinases p38, extracellular regulated kinase 1 and 2 (ERK1/ERK2), and stress-activated protein kinase/c-Jun NH2-terminal kinase (SAPK/JNK) in PC12 cells, these kinases were activated only partially or not at all in srcDN2 and 17.26 cells. Inhibiting ERK1/ERK2 with PD98059 or inhibiting SAPK/JNK by transfecting cells with a dominant-negative SAPKβ/JNK3 mutant partially blocked NGF-induced down-regulation of A2AAR expression in PC12 cells. In contrast, inhibiting p38 with SB203580 had no effect on the regulation of A2AAR mRNA and protein levels. Treating SAPKβ/JNK3 mutant-transfected PC12 cells with PD98059 completely abolished the NGF-induced decrease in A2AAR mRNA and protein levels. These results reveal a role for ERK1/ERK2 and SAPK/JNK in regulating A2AAR expression.

Adenosine receptors are G-protein coupled receptors that mediate important physiological processes in both the central and peripheral nervous system, including vasodilation, respiratory depression, wakefulness, and spontaneous locomotor activity. There are four major adenosine receptor subtypes, A 1 , A 2A , A 2B , and A 3 ; each is encoded by a distinct gene, and each has unique affinities for adenosine analogs and methylxanthine derivatives (1)(2)(3). In the developing rat brain, adenosine A 2A receptor (A 2A AR) 1 mRNA is expressed transiently in var-ious regions (4). Moreover, a severalfold increase in A 2A AR protein levels occurs during early postnatal development in a number of brain regions, whereas a decrease in A 2A AR mRNA is observed in other regions (5).
PC12 cells, derived from a rat pheochromocytoma, have been used extensively as a model for neuronal differentiation and development (6). In response to NGF, these cells differentiate into sympathetic-like neurons and extend neurites (6). The signal transduction pathways activated by NGF originate at both high (TrkA) and low (p75) affinity receptors, and downstream targets of each receptor have been implicated in regulating expression of genes involved in differentiation, neurotransmission, and neuronal function (6 -13). Stimulation of the receptor tyrosine kinase TrkA results in the activation of Ras, Src, phospholipase C-␥, SNT, and phosphoinositide 3-OH kinase (14 -17). In PC12 cells, active Ras triggers a cascade of phosphorylation events leading to activation of ERK1/ERK2 via Raf-1 (18 -20) or p38 MAP kinase and stress-activated protein kinase/c-Jun NH 2 -terminal kinase (SAPK/JNK) via MAP kinase kinase kinase (21)(22)(23). p75 activation increases ceramide production (24) and activates NFB (25).
Gene products regulated by NGF in PC12 cells include several G-protein coupled receptors, such as the M 4 muscarinic, secretin, and adenosine A 2A receptors (7,26,12). Using gene expression profiling (expressed sequence tags) coupled with Northern analysis, a decrease in A 2A AR mRNA could be demonstrated as early as 3 days (3 d) posttreatment with NGF and levels remained depressed for up to 12 d (7). In situ hybridization with an A 2A AR oligonucleotide probe detected a 50% decrease in the number of grains per cell in NGF-differentiated PC12 cells, confirming that NGF decreases A 2A AR mRNA levels (26). Corresponding to the changes in mRNA levels, immunoreactive A 2A AR protein declines by more than half after 7 d of NGF treatment, and the number of binding sites for the A 2A AR selective antagonist, [ 3 H]SCH 58261, decreases by 3-fold (26). When PC12 cells are treated with A 2A AR agonists, a transient down-regulation of A 2A AR mRNA and protein occurs (27). Despite these observations, the specific cellular mechanisms regulating A 2A AR mRNA levels have not been thoroughly delineated. In this study, we provide the first insights into the downstream pathways employed by NGF to control A 2A AR expression in PC12 cells. Such pathways may likewise play an important role in the regulation of A 2A AR expression during brain development.

EXPERIMENTAL PROCEDURES
Cell Culture-Rat pheochromocytoma cells (PC12) were obtained from the American Type Tissue Culture collection (Manassas, VA). PC12nnr5, clone 106, and srcDN2 cells were a generous gift from Gordon Guroff at NICHD, National Institutes of Health (Bethesda, MD). The dominant-negative Ras cell line, 17.26, was obtained from Robert Maue at Dartmouth Medical Center (Hanover, NH). PC12 cell lines were maintained on rat tail collagen, Type IV (Upstate Biotechnology, Saranac Lake, NY) as described previously (7). SrcDN2, 17.26, and clone 106 were cultured in the presence of 300 g/ml Geneticin (Life Technologies, Inc.). Cells were treated with PD98059 or SB203580 (Calbiochem, San Diego, CA) and mouse 2.5S NGF (Promega, Madison, WI) as described below.
Northern Blot Analysis-Poly(A ϩ ) RNA was isolated, fractionated through a denaturing agarose gel, and transferred to Hybond N ϩ membranes (NEN Life Science Products) essentially as described previously (12). A 32 P-labeled 2.3-kilobase pair SstI/XhoI fragment from a rat A 2A cDNA clone and a 1.2-kilobase pair EcoRI/XhoI fragment from a rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA were used as probes (7). Expression levels of A 2A AR mRNA were normalized to GAPDH mRNA levels. Blots were analyzed on a Molecular Dynamics PhosphorImager.
Data are expressed as the mean Ϯ S.E. of n independent experiments.
Ribonuclease Protection Assay--At the times indicated, total RNA was isolated from PC12nnr5 cells. An RNase protection assay was performed essentially as described by Lee et al. (28). Construction of plasmid 118GAPDHpSP73 for generating an antisense riboprobe of the GAPDH mRNA was described previously (12). Plasmid 296A2ApSP72 for generating an antisense riboprobe template of the A 2A AR mRNA was constructed by subcloning an EcoRI/PvuII restriction fragment of 296 nucleotides into pSP72 (Promega). Riboprobes transcribed from EcoRI-linearized 296A2ApSP72 correspond to nucleotides 596 -891 of the A 2A AR cDNA clone in pBluescript (Stratagene, La Jolla, CA) (7). Authenticity of the plasmid construct was verified by dideoxy sequencing.
Transfection of PC12 Cells with SAPK␤(K55A)-PC12 cells were plated onto 100-mm dishes, grown to 70 -80% confluency, and transiently transfected with 5 g of SAPK␤(K55A) in expression vector pCMV5 (29) or an empty expression vector with LipofectAMINE 2000 (Life Technologies, Inc.). At 18 -24 h after transfection, cells were treated with PD98059 and/or NGF, and they were harvested for total RNA at the indicated times.
SDS-PAGE and Immunoblotting-PC12 cells were plated onto 35-mm collagen-coated tissue culture dishes at 90% confluency and incubated 24 h. The medium was then replaced with Dulbecco's modified Eagle's medium lacking serum, and the cells were incubated for an additional 24 h. Serum starved monolayers were washed with cold phosphate-buffered saline, and cells were harvested/lysed in 95°C SDS-PAGE sample buffer containing 50 mM Tris, pH 6.8, 10% glycerol, 0.1% bromphenol blue, 2% SDS, 0.7 M ␤-mercaptoethanol, 50 mM sodium fluoride, 2 mM sodium orthovanadate, 1 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 10.8 g/ml aprotinin, and 10 g/ml leupeptin. Lysates were placed on ice during sonication and reheated to 95°C for 5 min. Proteins were separated on a 7.5% SDS-PAGE gel and transferred to a polyvinylidene difluoride membrane (Amersham Pharmacia Biotech). Membranes were processed for Western analysis as described by the New England Biolabs (Beverly, MA) protocol supplied with the antibodies. Phospho-p44/42 MAP kinase (ERK1/ERK2) antibody, phospho-p38 MAP kinase antibody, p44/42 MAP kinase antibody, SAPK/ JNK antibody, and p38 MAP kinase antibody were from New England Biolabs. Phospho-JNK (G-7) monoclonal antibody was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-TrkA (Ab-1) monoclonal antibody was from Calbiochem. Goat anti-rabbit or goat anti-mouse horseradish peroxidase-conjugated secondary antibodies (Upstate Biotechnology) allowed detection of proteins by the ECLϩPlus detection system (Amersham Pharmacia Biotech).
Radioligand Binding-Crude membrane preparations were obtained as described recently (30). 125 I-ZM241358, an A 2A AR-specific antagonist was used to measure specific binding to A 2A ARs in crude membrane preparations. Data are expressed as the mean Ϯ S.E. of n independent determinations.

RESULTS AND DISCUSSION
To assess the contribution of individual components of the NGF signal transduction pathway leading to regulation of A 2A AR mRNA and protein, PC12 sublines lacking functional TrkA or overexpressing dominant-inhibitory forms of Ras or Src were used. The role of the three MAP kinase family members p38, ERK1/ERK2, and SAPK/JNK in NGF-mediated regulation of A 2A AR mRNA was also determined.
TrkA Is Required for Down-regulation of A 2A AR mRNA and Protein-Earlier experiments have concentrated on the effects of long term (7-12 days) NGF treatment on A 2A AR expression in PC12 cells (7,26). In the present study, shorter periods of treatment were examined to define initial pathways responsible for NGF regulation. Steady state A 2A AR mRNA declined to 62 and 43% of untreated control cells following 1 and 3 d of NGF treatment, respectively (Fig. 1A). Correspondingly, binding of the A 2A AR antagonist 125 I-ZM241358 to PC12 cells decreased by 50 and 45% following 1 and 3 d of NGF treatment, respectively (Fig. 1B). Thus, NGF-induced down-regulation of A 2A AR mRNA and protein is apparent as early as 24 h following NGF treatment, with a further decrease in A 2A AR mRNA occurring at 3 d.
The contribution of TrkA and p75 to the regulation of A 2A AR mRNA and protein were examined in PC12nnr5 cells, which are a PC12 subline that expresses p75, lacks functional TrkA receptors, and does not differentiate in response to NGF (31). As the basal steady-state level of A 2A AR mRNA is reduced in PC12nnr5 cells, a ribonuclease protection assay was performed to quantitate A 2A AR mRNA. As shown in Fig. 1, NGF failed to down-regulate both A 2A AR mRNA and protein levels. For comparison, the effects of NGF on A 2A AR expression were studied in clonal cell line 106. Clone 106, derived from PC12nnr5 cells stably transfected with TrkA, has levels of 125 I-NGF binding similar to those in native PC12 cells and differentiates in response to NGF (32). When cultures of clone 106 were treated with NGF for 1 or 3 d, both A 2A AR mRNA and protein levels were down-regulated to the same extent as native PC12 cells (Fig. 1). Because TrkA is implicated in A 2A AR mRNA and protein regulation, potential roles for TrkA-associated signaling components, Src and Ras, were examined.
NGF-mediated Down-regulation of A 2A AR mRNA and A 2A AR Protein Is Impaired by Dominant-negative Src-As oncogenic Src mimics NGF-induced neurite outgrowth and phosphorylation of a similar set of cellular substrates, a role for Src in the signal transduction pathway initiated by NGF has been implicated (33). Therefore, the effects of NGF on A 2A AR mRNA and protein were examined in srcDN2 cells that overexpress a dominant-negative, kinase-defective Src mutant (34). Upon treatment of srcDN2 cells with NGF for 1 or 3 d, down-regulation of steady-state mRNA was not observed (92 and 100% of untreated cells, respectively) (Fig. 2). Likewise, A 2A AR protein levels remained near control levels following 1 and 3 d of NGF treatment (110 and 81%, respectively). Thus, Src appears to be critical for NGF-induced down-regulation of both A 2A AR mRNA and protein.
Ras Mediates NGF-stimulated Down-regulation of A 2A AR mRNA and Protein-PC12 cells undergo a Ras-dependent transient induction of several immediate-early genes within minutes of NGF treatment that precedes neurite outgrowth (35).
The delayed response genes are transcriptionally active hours to days following NGF treatment (6,7), and the induction of several genes, such as agrin, tau, transin, and SCG10, has been shown to be Ras-dependent (8,10,11). Furthermore, transcriptional down-regulation of the epidermal growth factor receptor requires Ras activity (36). The data shown in Fig. 3 demonstrate that NGF can also mediate down-regulation of A 2A AR mRNA in a Ras-dependent manner. In 17.26 cells expressing a dominant-negative Ras mutant (35), steady state mRNA levels following 1 and 3 d of NGF treatment were 88 and 81% of untreated control cells, respectively, which is a less dramatic decrease than that seen in native PC12 (compare Figs. 1A and 3A). was comparable to that found in native PC12 cells (data not shown). In agreement with our findings, srcDN2, 17.26, and native PC12 cells exhibited similar levels of 125 I-NGF binding to TrkA (32). Taken together, these results indicate that both Src and Ras are necessary TrkA-signaling components that regulate A 2A AR mRNA and protein levels. Ras has multiple downstream effectors that activate divergent signaling pathways (reviewed in Ref. 37), such as Raf-1 and MAP kinase kinase kinase (22). The most thoroughly described Ras activated pathway is the Raf-dependent activation of ERK1/ERK2 (38,39). More recently, Raf-independent Ras-activated MAP kinase pathways have been identified. For example, the Ras effector MAP kinase kinase kinase 1 activates SEK, which in turn activates SAPK/JNK and p38 (40 -43). As such, the activity of p38, ERK1/ERK2, and SAPK/JNK was examined in PC12, srcDN2, and 17.26 cells.

MAP Kinase Family Members Are Activated in PC12 Cells but Not in Sublines Expressing Dominant-inhibitory Src or Ras
Mutants-NGF activation of p38, ERK1/ERK2, and SAPK/ JNK was assessed with phosphorylation state-specific antibodies (Fig. 4). In native PC12 cells, NGF activated p38 and ERK1/ ERK2 at early time points (15 and 30 min), whereas SAPK/ JNK was not activated until 3 d. In agreement with previous reports, expression of dominant-negative Ras in 17.26 cells inhibited NGF activation of ERK1/ERK2 (18,20) and SAPK/ JNK (22), confirming that growth factor activation of these two MAP kinases is Ras-dependent. Likewise, activation of p38 above basal levels in NGF-treated 17.26 cells was not observed. As demonstrated previously (34), ERK1/ERK2 was activated in srcDN2 cells albeit to a slightly lesser extent than native PC12 cells. Moreover, NGF did not effectively activate p38 or SAPK/ JNK in srcDN2 cells (Fig. 4). These findings demonstrate that Src is required for the full activation of the MAP kinases. As these kinases are distal components of the NGF signal transduction pathway, and their activation was impaired or inhibited by dominant-negative signaling components more proximal to TrkA, it is possible that one or more of these MAP kinase family members plays a role in mediating the effects of NGF on A 2A AR mRNA. Notably, a recent report demonstrated that two MAP kinase family members, JNK and ERK1/ERK2, had opposing effects on tau promoter activity and affected promoter activity over different time frames (10).
Role of ERK1/ERK2, p38, and SAPK/JNK in Regulating A 2A AR mRNA and Protein-To examine potential roles of the individual MAP kinases in affecting down-regulation of A 2A AR mRNA by NGF, the synthetic compound SB203580 was used to inhibit p38 (44). Although p38 activation was impaired in srcDN2 and 17.26 cells, this kinase does not appear to be involved in A 2A AR mRNA regulation as SB203580 failed to inhibit NGF-induced down-regulation of A 2A AR mRNA in native PC12 cells (Fig. 5A). Similarly, another p38 inhibitor, SB202190, also failed to inhibit A 2A AR mRNA down-regulation (data not shown). Inhibition of p38 activity in PC12 cells was verified by an in vitro immunocomplex kinase assay. Whereas p38 from lysates of NGF-stimulated cells phosphorylated glutathione S-transferase-ATF-2 (4-fold above basal levels), p38 from NGF-stimulated cells pretreated with SB203580 did not appreciably phosphorylate its substrate (data not shown). ERK1/ERK2 activity was inhibited by employing the MAP kinase kinase inhibitor PD98059 (45). Cells co-treated with NGF and PD98059 demonstrated significantly less down-regulation of A 2A AR mRNA than cells treated with NGF alone, suggesting that ERK1/ERK2 plays at least a partial role in controlling A 2A AR mRNA steady state levels (compare Figs. 5B and 1A). As reported previously (45), PC12 cells pretreated with PD98059 did not extend neurites following 3 d NGF as did FIG. 4. NGF activation of MAP kinases in native PC12 cells and srcDN2 or 17.26 sublines. To assess activation of p38 MAP kinase and ERK1/ERK2, 24 h serum-starved PC12 (WT), srcDN2, and 17.26 cells were left untreated or incubated for 15 or 30 min with 100 ng/ml NGF. To assess SAPK/JNK activation, cells were untreated or treated with 100 ng/ml NGF for 1 or 3 d in Dulbecco's modified Eagle's medium containing 1% serum. Representative Western blot are shown from experiments that were repeated at least twice. Blots were probed with phosphorylation state-specific antibodies, stripped, and reprobed with an antibody that recognized the respective MAP kinase subfamily member, independent of phosphorylation state. Left, p38 MAP kinase; center, ERK1/ERK2; right, SAPK/JNK. cells treated with NGF alone (data not shown). Western analysis also confirmed that PD98059 inhibited ERK1/ERK2, but not p38 and SAPK/JNK, activity in PC12 cells treated with NGF (data not shown).
To inhibit SAPK/JNK, PC12 cells were transiently transfected with SAPK␤(K55A), a kinase-defective construct of SAPK␤/JNK3. Following 1 and 3 d of NGF treatment, PC12 cells expressing SAPK␤(K55A) did not extend neurites, whereas empty vector-transfected PC12 cells extended neurites to the same extent as nontransfected cells (data not shown). These findings are in agreement with studies demonstrating that differentiation of PC12 cells requires the SAPK/ JNK signal transduction pathway (46). As shown in Fig. 5C, PC12 cells expressing SAPK␤(K55A) did not undergo NGFinduced down-regulation of A 2A AR mRNA following 1 d of treatment (102% of untreated cells). After 3 d of NGF incubation, SAPK␤(K55A) transfectants had a slight decrease in A 2A AR mRNA (79% of untreated cells) that was not as great as that of empty vector-transfected cells. When kinase-defective SAPK␤ transfectants were co-incubated with PD98059 and NGF for 3 d, down-regulation of A 2A AR mRNA was completely blocked (Fig. 5D). In contrast, when cells transfected with an empty vector were treated for 3 d with NGF alone, A 2A AR mRNA was down-regulated to the same extent (42%) as wildtype PC12 cells (Fig. 1A). As inhibition of either ERK1/ERK2 or SAPK/JNK individually results in partial inhibition of A 2A AR mRNA regulation, these results indicate that ERK1/ERK2 and SAPK/JNK are both required for complete down-regulation of A 2A AR mRNA.
The capacity of NGF to down-regulate A 2A AR mRNA following inhibition of the different MAP kinases was mimicked at the protein level (Fig. 6). It will be of interest in the future to determine whether NGF utilizes other mechanisms (besides mRNA regulation) to down-regulate A 2A AR protein (e.g. ubiquitin-mediated protein degradation).
Concluding Remarks-To summarize, the data presented here demonstrate that NGF-induced down-regulation of A 2A AR mRNA and protein levels is TrkA-, Src-, and Ras-dependent. Furthermore, the MAP kinase family members ERK1/ERK2 and SAPK/JNK are distal signal transduction components activated by NGF and are implicated here as having important roles in mediating regulation of A 2A AR mRNA. Recent reports have demonstrated a role for mitogen-activated protein kinase family members in regulating mRNA stability. NGF-induced stabilization of the M 4 muscarinic receptor mRNA requires ERK1/ERK2 (12) and p38 plays a role in stabilizing cyclooxygenase-2 mRNA (47,48). As NGF destabilizes A 2A AR mRNA transcripts, 2 we are currently examining the role of ERK1/ ERK2 and SAPK/JNK in NGF-mediated A 2A AR mRNA destabilization.