Our first goal was to find a second model system that was similar to the BE(2)-C human neuroblastoma cell line (expressing modest levels of α_2A-AR), but that didn't express β-ARs. Kazmi and Mishra previously identified the SH-SY5Y cell line as expressing two α₂-AR binding sites 24, while Parsley et al.25 reported that it expressed a single AR subtype, α_2C, based upon functional and molecular studies. Since receptor expression varies depending on differentiation state and passage number, it was necessary to determine which α₂-AR subtypes were expressed in our population of SH-SY5Y cells, using a combination of binding, functional, and molecular approaches.

SH-SY5Y cells expressed α₂-AR levels slightly greater than the level detectable in BE(2)-C cells (B_max: SH-SY5Y, 67.6 ± 8.2 ; BE(2)-C, 40.8 ± 7.0 fmol/mg protein). According to nonlinear and linear regression analysis of saturation binding, the data best fit a single-site model in SH-SY5Y cells, as observed previously in BE(2)-C cells. Rauwolscine and yohimbine competed for specific [³H]rauwolscine binding to SH-SY5Y cell membranes with higher affinity than prazosin, the α_2B/C-selective antagonist (Table 1; 24). Apparent K_i values of agonists and antagonists against [³H]rauwolscine binding were determined for comparison with previously reported values in cells natively expressing α_2A-, α_2B, or α_2C-ARs (HT29 and BE(2)-C, NG108-15, OK; 152627) or cell lines expressing cloned α₂C10, α₂C2, and α₂C4 28. Values obtained from binding studies in SH-SY5Y cells correlated only to values from BE(2)-C cells and showed the greatest similarity with those derived from native and cloned α_2A-AR-containing cell membranes (Table 2). These results are consistent with binding of [³H]rauwolscine to an α_2A-AR in SH-SY5Y cells.

Functional studies were performed by measuring the ability of various α₂-AR agonists to inhibit forskolin (10 μM)-stimulated cAMP accumulation in intact cells. All α₂-AR agonists inhibited forskolin-stimulated cAMP accumulation in a concentration-dependent manner; no stimulation of cAMP accumulation was noted in the absence of forskolin. Inhibition of cAMP accumulation by the α₂-AR agonist UK14,304 (30 nM; Fig. 1) was completely reversed by 10 nM yohimbine, whereas the α_2B/C-selective antagonist ARC-239, at a concentration over 30-fold higher than that of the agonist, failed to reverse the actions of UK14,304. Thus, both binding and functional data support the classification of the α₂-AR subtype in this neuroblastoma cell line as α_2A.

Since Parsley et al.25 were unable to detect α_2A-AR RNA by performing RT-PCR with total RNA extract, we optimized our chances for detecting α_2A-AR RNA by generating RT-PCR products from SH-SY5Y mRNA using primer pairs selective for individual α₂-AR subtypes (Table 3; 2930) or a primer pair that recognizes two α₂-AR receptor subtypes distinguished by their restriction nuclease digestion products (Table 3; 30). RT-PCR with α₂C10/C4 primers gave a 233 bp product specific for α_2A- and α_2C-ARs; restriction digestion of this fragment with BglII, that would specifically cleave α_2A-AR, resulted in two fragments of 117 bp and thereby established expression of α_2A-AR mRNA in SH-SY5Y cells. RT-PCR with α₂C4 primers gave a 630 bp fragment, which was successfully digested by BstXI to produce three fragments of 271, 225, and 78 bp, consistent with the presence of an α_2C-AR gene product (Fig. 2). RT-PCR products were neither noted in samples lacking reverse transcriptase (-), nor were they produced with primers selective for α₂C2 (α_2B-AR; data not shown). While SH-SY5Y cells express mRNA for both α_2A- and α_2C-ARs, it appears that the predominant functional α₂-AR in our cell line is the α_2A-AR.

Since these cells appear to express α_2A-ARs with properties similar to those in BE(2)-C cells 15 but lack a β-AR, pcDNA 3.0 plasmid vector containing the human β₂-AR gene was transfected into SH-SY5Y cells. Colonies of stable transfectants were selected and maintained by their resistance to G418 (600 μg/mL) and subsequently clonal populations of β₂-AR-expressing SH-SY5Y cells (SHβ₂AR) were screened for β-AR expression using [³H]CGP-12177 for binding studies as described in Methods. Since BE(2)-C cells express very low levels of β₂-AR (B_max: 18.5 ± 6.2 fmol/mg protein), the SHβ₂AR4 cell line that expressed 14.78 ± 4.19 fmol/mg protein of the β₂-AR was selected for the subsequent studies. To ensure that the β-ARs were functional, the ability of isoproteranol (ISO) to stimulate cAMP accumulation was assessed (Table 1). The α_2A-AR responses were also tested in this new cell line to confirm that α_2A-AR function had not been altered by the expression of the β₂-AR (Table 1).

To determine whether the presence of the β-AR influences α_2A-AR signaling, the ability UK14,304 to inhibit forskolin-stimulated cAMP accumulation was evaluated after wildtype (wt) and SHβ₂AR4 cells were exposed to vehicle or the indicated concentration of agonist for 16–24 hr. Wt SH-SY5Y cells (Fig 3A) require a 30-fold higher concentration of NE (30 μM) to desensitize the α_2A-AR signal than SHβ₂AR4 cells (1 μM; Fig 3B). Both the potency (-Log EC₅₀ (M): 5.2 ± 0.1) and efficacy (I_max (%): 17.0 ± 1.6; P < 0.05 Fig. 3A) of UK14,304 were reduced by 30 μM NE compared to vehicle treatment in wt cells (-7.6 ± 0.2 M and 43.2 ± 6.8%); modest concentrations of NE (1 μM) and EPI (300 nM) are insufficient to alter the α_2A-AR signal in the wt SH-SY5Y cell line. In contrast, chronic treatment of the β-AR-expressing SHβ₂AR4 cells with 300 nM EPI desensitized the α_2A-AR signal causing loss of UK14,304 potency (-Log EC₅₀ (M): Vehicle 6.9 ± 0.2; EPI 6.3 ± 0.2) and efficacy (I_max (%): Vehicle 68.2 ± 5.4; EPI 49.3 ± 10.4; p < 0.05; Fig. 3B). Unlike EPI, which co-activates both ARs, NE, at the concentrations employed activates only α_2A-ARs and does not alter α_2A-AR signaling. We concluded that the difference in α₂-AR signaling following EPI treatment between the transfected and wt SH-SY5Y was attributable to the presence of functional β₂-ARs, respectively. To ensure that the vector was not responsible for the observed difference between the wt and SHβ₂AR4 cells, similar experiments were conducted in SH-SY5Y cells transfected with the vector alone (minus the β₂-AR gene). These vector only-expressing clones responded to EPI (300 nM) and NE (1 μM) pretreatments as the parent SH-SY5Y cells did (Fig. 3C).

To validate the importance of the β₂-AR in the desensitization of the α₂-AR signal, we included the β-AR selective antagonist propranolol (30 nM) with the chronic 300 nM EPI treatment. Addition of propranolol blocks EPI-induced α_2A-AR desensitization resulting in UK14,304 concentration-response curves indistinguishable from control (-Log EC₅₀ (M) for EPI + Prop 6.7 ± 0.1; I_max (%) for EPI + Prop 67.9 ± 0.4; p < 0.05; Fig. 3D). Propranolol treatment alone did not alter UK14,304 potency or efficacy.

To ensure that the β₂-AR is functioning properly following catecholamine treatment, we evaluated the ability of ISO to stimulate cAMP accumulation over basal in SHβ₂AR4 cells. The β₂-AR signal is desensitized following chronic EPI but not NE treatment, consistent with the fact that NE has a low affinity for the β₂-AR. Inclusion of propranolol (30 nM) inhibited EPI-induced β₂-AR desensitization (p < 0.05; Fig. 4), but had no effect in the absence of EPI.

Our study in BE(2)-C cells suggests that β₂-AR-induced α_2A-AR desensitization following long-term EPI exposure is due in part to down-regulation of the α₂-ARs. To determine if the same mechanism is responsible for the EPI-induced α_2A-AR desensitization in SHβ₂AR4 cells, changes in α_2A-AR expression following catecholamine treatment were evaluated. Specific binding was measured with a single concentration of radioligand. We, and others, have shown that this is sufficient for accurate assessment of changes in receptor number for the α_2A-AR 915. Chronic exposure of SHβ₂AR4 cells to 300 nM EPI down-regulates the α_2A-ARs by 20% (p < 0.05; Fig. 5). The α_2A-AR down-regulation in this cell line, as in BE(2)-C cells, requires β₂-AR co-activation since loss of α_2A-ARs is prevented when 30 nM propranolol is included with EPI. Down-regulation of the α_2A-AR is not observed following chronic activation of α_2A-AR alone by 1 μM NE. Further, 300 nM EPI does not alter the expression of α_2A-AR in wt SH-SY5Y cells as compared to vehicle-treated cells (% of vehicle: 88.6 ± 25.9; n = 2) consistent with a lack of α_2A-AR desensitization. Hence, it can be concluded that chronic EPI treatment induces a loss of α_2A-AR response via β₂-AR-dependent down-regulation of α_2A-ARs in SHβ₂AR4, but not in wt SH-SY5Y cells.

We previously established that EPI-induced α_2A-AR desensitization and down-regulation in BE(2)-C cells is mediated via β₂-AR-dependent GRK3 up-regulation 15. Therefore, GRK3 levels in whole cell SHβ₂AR4 lysates were evaluated following 24 hr catecholamine treatments. Chronic EPI exposure altered neither GRK3 nor GRK2 levels in the transfected SH-SY5Y cell line (Table 4). Therefore, unlike results in BE(2)-C cells, increases in whole cell GRK3 levels do not contribute to the modest α_2A-AR desensitization or down-regulation observed in the SHβ₂AR4 cells.

Although GRK3 levels in whole cell lysates remain unaltered in SHβ₂AR4 cells, it is not known whether GRK3 recruitment to the membrane is regulated via chronic EPI treatment in that cell line. Since GRK2 and GRK3 have been shown to regulate α_2A-AR signaling 10, we wanted to determine whether the membrane recruitment of either GRK isoform was changed following chronic EPI exposure in SHβ₂AR4 cells. GRK2 and GRK3 are cytosolic proteins that anchor to the membrane via interaction with free Gβγ subunits; thus both kinases translocate from the cytosol to the membrane to regulate receptor signaling upon activation. Taking this characteristic of GRK2 and GRK3 into account, the levels of both kinases in membrane fractions following chronic EPI exposure were evaluated. SHβ₂AR4 cells exhibit an increase in membrane-associated GRK2 and GRK3 with 24 hr EPI treatment compared to vehicle (P < 0.05; Fig. 6). In SHβ₂AR4 cells, the same propranolol concentration (30 nM) that inhibited EPI-induced α_2A-AR desensitization and down-regulation also attenuated EPI-induced increase in GRK2 and GRK3 content in the membrane fraction (P < 0.05; Fig. 6). In contrast, no increased translocation of GRKs by EPI treatment was observed in wt SH-SY5Y cells that do not express β₂-ARs. Therefore, this increased GRK2 and GRK3 translocation to the membrane following prolonged EPI treatment in SHβ₂AR4 cells is β₂-AR dependent.

The following drugs were purchased or obtained from the indicated sources: (-) epinephrine (EPI), (±)norepinephrine (NE), sodium ascorbate, UK14,304 (Sigma-Aldrich, St. Louis, MO.); cell culture media (Gibco, Grand Island, NY); fetal bovine serum (Atlanta Biologicals, Norcross, GA); and antibiotics (Mediatech, Inc., Herndon, VA). GRK2 (C-15) and GRK3 (C-14) primary antibodies and horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA); anti-glyceraldehyde-3-phosphate dehydrogenase (GADPH, Research Diagnostics, Inc., Flanders, NJ).

SH-SY5Y (passages 37–55) human neuroblastoma cells (Dr. Robert A. Ross, Fordham University, Bronx, NY) were maintained in a humidified atmosphere (6% CO₂:94% air) in a 1:1 mixture of Eagle's minimum essential medium with non-essential amino acids and Ham's F-12 that contains 10% fetal bovine serum, 100 U/ml penicillin G and 0.1 mg/ml streptomycin sulfate. Plates of cells greater than 60% confluence were used throughout the study.

Plasmid cDNA with the human β₂-AR gene (provided by Dr. Brian Knoll; University of Houston, Houston, TX) or vector alone was stably transfected into SH-SY5Y cells with the fuGENE 6 Transfecting Reagent (Roche). Ten positive clones were isolated by their resistance to 800 μg/mL of G418 and maintained in media containing 600 μg/mL of G418. SHβ₂AR4 was selected for use in all experiments because it expressed similar levels of β₂-ARs as that expressed natively in BE(2)-C cells; SHβ₂AR4 expressed 14.78 ± 4.19 fmol/mg protein while BE(2)-C express 18.5 ± 6.2 fmol/mg protein 15. This β₂-AR level remained consistent to passage 12 in SHβ₂AR4 cells. After passage 12, SHβ₂ARs neither expressed β₂-ARs nor maintained resistance to G418, suggesting that the cells no longer expressed the transfected plasmid.

Total RNA was isolated from several different passages of freshly harvested SH-SY5Y cells by the guanidinium isothiocyanate/phenol-chloroform extraction method 37. Total RNA concentrations were determined by UV spectroscopy; integrity of each isolate was determined by electrophoresis through a 1% agarose gel in the presence of 0.01 M sodium phosphate buffer. Poly(A) mRNA was isolated using a Dynabead oligo(dT)₂₅ Kit (Dynal, Oslo, Norway) and was used for RT-PCR reactions. Each RT reaction (20 μL) contained 5–10 μg total or poly(A) RNA preincubated with 5 ng/μL oligo(dT)_12–18, for 10 min at 70°C. The reaction mixture contained 80 μM each of deoxynucleotides (dATP, dCTP, dGTP and dTTP), RT buffer (50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl₂), and 5 mM dithiothreitol, and was preincubated for 2 min at 42°C before the addition of Moloney Murine Leukemia Virus reverse transcriptase (200 U/μl) for 60 min at 42°C; a 5 min incubation at 95°C terminated the reactions.

ODNs 2930 corresponded to sequences for the various human α₂-AR (α_2A antisense: 5'-AGA CGA GCT CTC CTC CAG GT-3'; sense: 5'-AAA CCT CTT CCT GGT GTC TC-3'), α_2A/2C-(antisense: 5'-GTG CGC TTC AGG TTG TAC TC-3'; sense: 5'-AAA CCT CTT CCT GGT GTC TC-3'), or α_2C-AR (antisense: 5'-CGT TTT CGG TAG TCG GGG AC-3'; sense: 5'-GTG GTG ATC GCC GTG CTG AC-3'). The contents of each RT reaction tube were diluted to a final volume of 50 μL with 10% DMSO, 80 μM each of dATP, dCTP, dGTP and dTTP, 8 μM each of the appropriate sense/antisense primer pair, 1.5 mM MgCl₂, and magnesium free buffer [containing 10 mM Tris-HCl (pH 9.0), 0.1% Triton X-100, 50 mM KCl] in sterile distilled water. Reaction mixtures were overlayed with mineral oil and subjected to a hot start for 5 min at 95°C. DNA polymerase (2.5 U taq, 5 U/μl, Promega, Madison, WI) was added to each reaction tube after the hot start, and the tubes were subjected to a PCR reaction of 30 cycles in a thermal cycler (MJ Research Inc., Watertown, MA) for 1 min at 94°C, 1.5 min at 55°C, and 2 min at 72°C with a final elongation step at 72°C for 7 min. Reaction products were separated by electrophoresis through 2% agarose gels and visualized by ethidium bromide staining. PCR products were isolated from the gel using a DNA extraction kit (Amicon Inc., Bedford, MA). Identity of the purified PCR products was confirmed by their susceptibility to digestion with restriction enzymes specific for each reaction product (see Table 1; 30).

To determine the effects of α₂-AR agonists on forskolin-induced cAMP accumulation, intact cells were incubated for 5 minutes at 37°C in HBSS buffer (in mM): NaCl (137), KCl (5), Na₂HPO₄ (0.6), KH₂PO₄ (0.4), NaHCO₃ (4), D-glucose (6), MgCl₂ (0.5), MgSO₄ (0.4) and CaCl₂ (1), containing the phosphodiesterase inhibitor IBMX (0.5 mM). In some experiments, antagonists also were included in this step. To prohibit oxidation, sodium ascorbate (0.11 mM) was included when assaying catecholamines. Upon addition of forskolin (10 μM) and agonist, assay tubes were incubated for an additional 10 min at 37°C. Removing the tubes to a boiling water bath for 5 min terminated the assay. All assays were performed in duplicate in a total volume of 0.5 ml. After boiling, samples were centrifuged for 5 min at 14000 × g, and cAMP levels from the supernatant fractions were determined in a [³H]cAMP (0.8 pmol) binding assay as previously described 38. β-AR-mediated stimulation of cAMP accumulation was performed in the same manner except that forskolin was not included in the assay mixture. Forskolin (10 μM) stimulated cAMP accumulation to 587 ± 88 pmol/mg protein (n = 46), 15-fold over basal levels (40.5 ± 2 pmol/mg protein).

Bovine serum albumin was used as a standard in the determination of protein levels in intact cells and cell membranes as described 39.

K_d, B_max, IC₅₀ and LogEC₅₀ values were determined by nonlinear regression analysis using GraphPad Prism (GraphPad Software http://www.graphpad.com). The K_i values were calculated according to the Cheng-Prusoff equation 40 in which K_i = (IC₅₀)/(1+S), where S = [concentration of radioligand]/[K_D of radioligand]. Comparisons between groups were made by two-way Student's t-tests or ANOVA and Tukey's or Dunnett's post hoc test (where appropriate; GraphPad Software, San Diego, CA), and groups were considered significantly different if p ≤ 0.05.