Desensitization of α_2A-adrenoceptor signalling by modest levels of adrenaline is facilitated by β₂-adrenoceptor-dependent GRK3 up-regulation

Abstract

1. Adrenaline (ADR) and noradrenaline (NA) can simultaneously activate inhibitory α₂- and stimulatory β-adrenoceptors (AR). However, ADR and NA differ significantly in that ADR is a potent β₂-AR agonist while NA is not. Only recently has the interaction resulting from the simultaneous activation of α₂- and β₂-AR been examined at the cellular level to determine the mechanisms of α₂-AR regulation following concomitant activation of both α₂- and β₂-ARs by chronic ADR.

2. This study evaluates β₂-AR regulation of α_2A-AR signalling following chronic ADR (300 nM) and NA (1 and 30 μM) treatments of BE(2)-C human neuroblastoma cells that natively express both β₂- and α_2A-ARs.

3. Chronic (24 h) treatment with ADR (300 nM) desensitized the response to the α_2A-AR agonist, brimonidine, in BE(2)-C cells. Addition of the β-AR antagonist, propranolol, blocked the ADR-induced α_2A-AR desensitization. Unlike ADR, chronic NA (1 μM) treatment had no effect on the α_2A-AR response. However if NA was increased to 30 μM for 24 h, α_2A-AR desensitization was observed; this desensitization was partially reversed by propranolol.

4. Chronic ADR (300 nM) treatment reduced α_2A-AR binding levels, contributing to the α_2A-AR desensitization. This decrease was prevented by addition of propranolol during ADR treatment. Chronic NA (30 μM), like ADR, treatment lowered specific binding, whereas 1 μM NA treatment was without effect.

5. Chronic ADR treatment produced a significant increase in GRK3 levels and this was blocked by propranolol or GRK2/3 antisense DNA treatment. This antisense DNA, common to both GRK2 and GRK3, also blocked chronic ADR-induced α_2A-AR desensitization and down-regulation.

6. Acute (1 h) ADR (300 nM) or NA treatment (1 μM) produced α_2A-AR desensitization. The desensitization produced by acute treatment was β-AR independent, as it was not blocked by propranolol.

7. We conclude that chronic treatment with modest levels of ADR produces α_2A-AR desensitization by mechanisms that involve up-regulation of GRK3 and down-regulation of α_2A-AR levels through interactions with the β₂-AR.

Introduction

Adrenaline (ADR) and noradrenaline (NA) are endogenous catecholamines that activate three different families of adrenoceptors (ARs): α₁-ARs (α_1A, _1B, and _1D) that activate phospholipase C, α₂-ARs (α_2A, _2B, and _2C) that inhibit adenylate cyclase, and β-ARs (β₁, β₂, and β₃) that activate adenylate cyclase. ADR and NA share similar efficacies and potencies for all these receptors except the β₂-AR, where ADR has a higher affinity than NA.

The interaction between β- and α₂-ARs is especially interesting because these ARs mediate opposing actions on adenylate cyclase. The interplay between these two ARs is important for maintaining homeostasis of various physiological mechanisms and, in some cases, causing the manifestation of pathophysiological conditions. In peripheral tissues, α₂ and β-ARs play important roles in modulating uterine smooth muscle tone (Lécrivain et al., 1998), insulin release from pancreatic islet cells (John et al., 1990), cytokine release (Kalinichenko et al., 1999) and vasomotor tone. In addition, neuronal α₂- and β₂-ARs play an important role in the modulation of noradrenergic neurotransmission in the human CNS and PNS. Neuronal α₂-ARs are considered potential therapeutic targets for the treatment of major cardiovascular disorders, such as hypertension and cerebral ischaemia, and CNS disorders, such as pain and depression (Lanier et al., 1996). Hence the interaction between α₂- and β-ARs in a single cell after exposure to ADR or NA has potentially great physiological and pathophysiological significance (Meana et al., 1992; Shi et al., 1996; Kalinichenko et al., 1999).

The influence of cross talk between receptors that exert reciprocal effects on a common signal transduction pathway within a single cell has been examined previously. For example, Malbon and co-workers determined that chronic (24 h) activation of adenylyl cyclase by β₂-ARs resulted in a 3 fold increase in the expression of the inhibitory G protein G_αi2 in S49 lymphoma cells (Hadcock et al., 1990). Moreover, maximal inhibition of forskolin-stimulated adenylyl cyclase activity by somatostatin was increased from 35 to 65% after 24 h isoprenaline (ISO) treatment. In a subsequent report, this same group described cross-talk in the opposite direction (Hadcock et al., 1991): chronic stimulation of the inhibitory pathway by somatostatin resulted in an increased expression of the stimulatory G protein G_αs and an increased efficacy of ISO on adenylyl cyclase activity. In summary, this suggests that stimulation of one signalling pathway results in sensitization of the opposing pathway.

The results described above clearly illustrate the potential for cross talk between inhibitory and stimulatory receptors in a single cell. However, there is a paucity of information regarding the consequences of chronic simultaneous activation of both the stimulatory and inhibitory pathways in the same cell. This information is important, since in vivo studies suggest that α₂-AR desensitization may occur more readily when the inhibitory α₂-AR and stimulatory β₂-AR are chronically and simultaneously activated by ADR (Schwartz & Eikenburg, 1988; Apparsundaram & Eikenburg, 1996). In contrast, chronic NA treatment that activates only the α₂-AR in vivo does not produce α₂-AR desensitization (Eikenburg, 1990). Studies utilizing β₂-AR antagonists during chronic ADR treatment suggest a role for the β₂-AR in ADR-induced α₂-AR desensitization (Apparsundaram & Eikenburg, 1996). These observations are particularly noteworthy when considered with in vitro evidence that the α₂-AR is more resistant to desensitization, requiring much higher agonist concentrations for desensitization than other ARs, such as the β₂-AR (Eason & Ligget, 1992; Lanier et al., 1996). Therefore, the present study was undertaken to examine the differential regulation of α₂-AR signalling by modest concentrations of ADR (300 nM) and 1 μM NA in a neuronal cell line (BE(2)-C) that endogenously expresses both α_2A- and β₂-ARs. The objective was to determine why long-term simultaneous α₂- and β₂-AR activation by ADR (300 nM) produces α₂-AR desensitization, while α₂-AR activation alone by 1 μM NA fails to do so. For the purpose of this study, desensitization is defined as simply a loss of agonist response.

Methods

Materials

The following drugs were purchased or obtained from the indicated sources: (−)adrenaline (ADR), isoprenaline bitartrate (ISO), p-aminoclonidine (PAC), guanabenz, oxymetazoline, phenylephrine, yohimbine (RBI, Natick, MA, U.S.A.); ARC-239 (Tocris Cookson, St. Louis, MO, U.S.A.); (±)noradrenaline (NA), agmatine, idazoxan, sodium ascorbate, adrenal cortex acetone powder, IBMX, cAMP, forskolin, hydroxyapatite (Sigma Chemical Co., St. Louis, MO, U.S.A.); oligodeoxynucleotides (ODNs; Sigma-Genosys, Woodlands, TX, U.S.A.); [³H]-rauwolscine (60–80 Ci mmol⁻¹), [³H]-RX821002 (56–60 Ci mmol⁻¹), [³H]-cAMP (35–45 Ci mmol⁻¹) and [¹²⁵I]-CYP (2000 Ci mmol⁻¹; Amersham Corp., Arlington Heights, IL, U.S.A.); Liquiscint scintillation fluor, (National Diagnostics, Atlanta, GA, U.S.A.); cell culture media (Gibco, Grand Island, NY, U.S.A.); foetal bovine serum (Atlanta Biologicals, Norcross, GA, U.S.A.); antibiotics (Mediatech, Inc., Herndon, VA, U.S.A.); G_αi1/2 (AS/7), G_αi3 (EC/2, New England Nuclear Corp., Boston MA, U.S.A.), G_αo, GRK2 and GRK3 (H-222), GRK2 (C-15) GRK3 (C-14) and GRK5 (C-20) primary antibodies and horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, U.S.A.); and glyceraldehyde-3-phosphate dehydrogenase antibody (GADPH, Research Diagnostics, Inc., Flanders, NJ, U.S.A.).

Brimonidine was provided by RBI as part of the Chemical Synthesis Program of the National Institute of Mental Health, Contract N01MH30003. Rauwolscine was a generous gift from CarlRoth (GmbH, Karlsruhe, Germany).

Cell culture

BE(2)-C (passages 16–58) human neuroblastoma cells (Dr Robert A. Ross, Fordham University, Bronx, NY, U.S.A.) 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 containing 10% foetal bovine serum, 100 U ml⁻¹ penicillin G and 0.1 mg ml⁻¹ streptomycin sulphate. Plates of cells greater than 60% confluence were used throughout the study, and lifted after incubation for 5 min at room temperature with phosphate buffered saline containing 1 mM EGTA. Intact cells were sedimented by centrifugation at 1000×g for 5 min, and resuspended in the appropriate buffer or reagent, as described below.

Receptor binding

Preparation of cell membranes

Cells were homogenized with a polytron (setting 3, 10 s) in 20 volumes of Tris-HCl buffer (50 mM, pH 7.4) containing (mM) NaCl 100, Na₂ EDTA 1 and PMSF 0.1, incubated for 15 min at 25°C, and the membranes sedimented by centrifugation (30 min, 34,000×g) at 4°C. Pellets were resuspended in 0.32 M sucrose, and aliquots of the membrane fractions were stored frozen (−80°C) until use.

Saturation binding

Levels of α₂-ARs in BE(2)-C cell membranes (0.5 mg ml⁻¹) were determined with various concentrations of either [³H]-rauwolscine (0.3–12 nM) or [³H]-RX821002 (0.1–10 nM) in a total volume of 1–2 ml in potassium phosphate buffer (50 mM, pH 7.4) containing MgSO₄ (5 mM) at 37°C for 45 min. Thereafter, 2 ml Tris-HCl (5 mM, pH 7.4, 4°C) was added to the mixture to terminate the binding reaction before filtration over no.32 glass fibre filter strips (Schleicher & Schuell, Keene, NH, U.S.A.) using a PHD cell harvester (Cambridge Technology, Cambridge, MA, U.S.A.). The reaction tubes and the filter strips were rinsed twice with a further 2–3 ml of buffer. Levels of radioactivity were determined by scintillation spectroscopy in a Beckman LS6000 liquid scintillation counter. Assays were performed in triplicate, and specific binding was determined by subtracting the binding in the presence of yohimbine or phentolamine (10 μM; non-specific) from the binding in its absence; specific binding comprised 50–75% of total binding. Saturation studies indicated that agonist treatments did not alter the K_d of the radioligand for the α₂-AR, so a single concentration (2 nM) of either [³H]-rauwolscine or [³H]-RX821002 was used to determine levels of binding following catecholamine treatment.

In some studies, plated cells were washed (4×5 ml) with serum-free media and treated with an unmodified phosphodiester antisense ODN common to both GRK2 and GRK3: 5′-ACC GCC TCC AGG TCC GCC AT-3′ (20 μM; Shih & Malbon, 1994) for 4 h, before adding sera back for the remainder of the treatment period. Cells treated with ODNs corresponding to the complementary common GRK sense (20 μM) served as controls in those experiments.

β-AR levels were determined as described (Standifer et al., 1989) with [¹²⁵I]-CYP (2-300 pM) and 0.3 mg ml⁻¹ of membrane protein at 37°C for 60 min in a total volume of 250 μl. Non-specific binding was determined in the presence of 10 μM alprenolol, while specific binding was determined by subtracting the binding in the presence of alprenolol from the binding in its absence.

Competitive binding

Cell membrane fractions were incubated as described above, except that the concentration of [³H]-rauwolscine was fixed (2 nM), and various (5–9) concentrations of unlabelled drugs were included. For β-AR displacement studies, multiple concentrations of competitor were incubated with 100 pM [¹²⁵I]-CYP.

cAMP accumulation

To determine the effects of α₂-AR agonists on forskolin-induced cAMP accumulation, intact cells were incubated for 5 min 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 were also 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 14,000×g, and cAMP levels from the supernatant fractions were determined in a [³H]-cAMP (0.8 pmol) binding assay as previously described (Standifer et al., 1994). β-AR-mediated stimulation of cAMP accumulation was performed in the same manner, except that forskolin was not included in the assay mixture.

Immunoblotting

Membrane proteins (30 μg) were resolved by SDS–PAGE through 10% gels. Proteins were transferred to PVDF membrane, blocked with 5% non-fat dried milk in PBS containing 0.1% Tween (PBST) and incubated overnight at 4°C with dilutions of a rabbit polyclonal antibody directed against G_αi3 (1 : 1000), G_αi1/2 (1 : 500), G_αo (1 : 500), GRK2 and GRK3 (1 : 1000), GRK2 (1 : 1000), GRK3 (1 : 1000) or GRK5 (1 : 500). Blots were subjected to four washes before incubating for 60 min at room temperature with a goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (1 : 2000) in PBST. Immunoreactive bands were visualized by enhanced chemiluminescence (Amersham Corp., Arlington Heights, IL, U.S.A.). The intensity of each immunoreactive band was determined using a Nucleovision Imaging Workstation (Nucleotech Corp., San Carlos, CA, U.S.A.), and normalized to the GAPDH loading control (1 : 8000).

Protein determination

Bovine serum albumin was used as a standard in the determination of protein levels in intact cells and cell membranes as described (Lowry et al., 1951).

Data analysis

K_d, B_max, IC₅₀ and LogEC₅₀ values were determined by nonlinear regression analysis using GraphPad Prism version 3.0 for Windows 95 and 98 (GraphPad Software, San Diego, CA, U.S.A., www.graphpad.com). K_i values were calculated according to the Cheng–Prusoff equation (Cheng & Prusoff, 1973) in which K_i=(IC₅₀) [(1+S)]⁻¹, where S=[concentration of radioligand] [K_D of radioligand]⁻¹. 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), and groups were considered significantly different if P⩽0.05.

Results

Chronic ADR induces α_2A-AR desensitization and down-regulation

To test the hypothesis that α₂-AR regulation by chronic ADR is β-AR-dependent, α₂-AR function following chronic 300 nM ADR or 1 μM NA pretreatment (16–24 h) was assayed in BE(2)-C cells (passage <50) by measuring the ability of the α₂-AR agonist, brimonidine, to inhibit forskolin (10 μM)-stimulated cAMP accumulation. Chronic 300 nM ADR exposure desensitized the α₂-AR, significantly reducing brimonidine efficacy (P<0.0001) while reducing brimonidine potency only 2 fold (Figure 1A). The decrease in brimonidine efficacy was prevented when the nonselective β-AR antagonist propranolol (30 nM) was added to the chronic ADR treatment (Figure 1A); 30 nM propranolol alone had no effect. To confirm that 300 nM ADR was sufficient to modulate β-AR activity, the ability of a maximally stimulating concentration of ISO (250 nM) to increase cAMP accumulation after chronic ADR treatment was also determined. Pretreatment with 1 μM ISO desensitized the β-AR response and was included as a positive control. Chronic 300 nM ADR pretreatment desensitized the β-AR response (Figure 2; P<0.05), indicating that ADR was activating both α₂- and β-ARs.

Figure 1.

Chronic 300 nM ADR and 30 μM NA desensitizes the α₂-AR response in BE(2)-C cells. The α₂-AR response, as determined by the ability of brimonidine to inhibit forskolin (10 μM)-stimulated cAMP accumulation in intact BE(2)-C human neuroblastoma cells, was measured 16–24 h after pretreatment with any or all of the following: ADR (300 nM), propranolol (30 nM or 1 μM), ADR+30 nM propranolol, 1 μM NA, 30 μM NA, 30 μM NA+1 μM propranolol, ISO (1 μM) or vehicle (0.1 mM ascorbate). Forskolin increased cAMP levels over 10 fold (287±109 pmol mg⁻¹ protein) compared to the basal state (21.8±3.5 pmol mg ⁻¹protein), and levels of both were unaltered by any pretreatments. (A) ADR (300 nM) treatment reduced brimonidine efficacy (n=7; ^*P<0.05 by ANOVA) compared to vehicle in BE(2)-C cells cultured up to passage 50. Alpha₂-AR desensitization was not noted if the β-AR antagonist propranolol (30 nM) was included during chronic ADR exposure (n=6; #P<0.05 as compared to ADR by ANOVA). Propranolol (30 nM) alone (n=5) had no effect on α₂-AR signalling. (B) Chronic 1 μM NA treatment did not alter the α₂-AR signal as compared to vehicle (n=6–9) but 30 μM NA treatment reduced brimonidine potency and efficacy (n=3; ^*P<0.05). Inclusion of a higher propranolol concentration (1 μM) blocked 30 μM NA-induced loss of α₂-AR potency, but not efficacy. Neither 1 μM ISO nor 1 μM propranolol alone altered the α₂-AR response (data not shown). (C) To better visualize the ability of different concentrations of ADR and NA to activate α₂-ARs in BE(2)-C cells, the concentration-response curves from which the values in Table 2 were derived for each agonist are shown. Each point represents the mean±s.e.mean of three (ADR) or five (NA) independent observations, performed in duplicate. (D) Chronic 300 nM ADR treatment does not produce a significant reduction of the brimonidine response as compared to vehicle (n=3) in higher passage cells in which β-AR binding and responsiveness has dropped off (Table 1).

Figure 2.

Chronic 300 nM ADR, 30 μM NA, and 1 μM ISO desensitize the β-AR response in BE(2)-C cells. BE(2)-C cells were incubated for 16–24 h with vehicle (0.1 mM ascorbate), 300 nM ADR, 1 μM NA, 30 μM NA or 1 μM ISO. After washing, intact cells were assessed for ISO-stimulated (250 nM) cAMP accumulation. Chronic 300 nM ADR, 30 μM NA, and 1 μM ISO (n=2–9; ^*P<0.05), but not 1 μM NA (n=6), pretreatment desensitized the β-AR response to ISO compared to the corresponding vehicle-treated control. None of the pretreatments altered basal cAMP levels (27.6±6.1 pmol mg⁻¹ protein). Data represent mean±s.e.mean; comparisons were made by ANOVA with Dunnett's post-hoc test.

Unlike 300 nM ADR, 1 μM NA pretreatment had no effect on brimonidine potency or efficacy (Figure 1B). This concentration of NA also failed to desensitize the β-AR response (Figure 2), suggesting that a concentration of NA producing only activation of the α₂-AR alone was not sufficient to produce α₂-AR desensitization in this cell line. However, while 300 nMM ADR produces a near maximal α₂-AR response (72% inhibition of forskolin-stimulated cAMP accumulation; Figure 1C), 1 μM NA inhibits forskolin-stimulated cAMP accumulation a bit less (53.8%). Also, if the population of β-AR in the cells were primarily of the β₂-AR subtype, NA would be much less effective than ADR at activating those receptors. To achieve equivalent α₂-AR activation as ADR (Figure 1C) and ensure activation of β-ARs, chronic pretreatment was also performed with a higher concentration of NA (30 μM; Figure 1B). This concentration reduced brimonidine efficacy to levels produced by 300 nM ADR treatment, produced a 10 fold loss of brimonidine potency, and desensitized the β-AR response to the same extent as ADR pretreatment (Figure 2). A higher propranolol concentration (1 μM) was only partially effective in preventing the α₂-AR desensitization induced by 30 μM NA pretreatment (Figure 1B), as it reversed the loss of brimonidine potency but not the loss of efficacy. Whether this reflects (1) a significantly higher α₂-AR occupancy than 300 nM ADR, (2) the fact that NA was present in 30 fold excess compared to propranolol, so that complete β-AR blockade was not achieved, or (3) a combination of both is not clear. It is unlikely that differential α₂-AR occupancy played a big role because 300 nM ADR produces an α₂-AR response that is 90% of maximal (Figure 1C, Table 2). It is more likely that there was an insufficient concentration of propranolol to completely antagonize the actions of NA at the β-AR. Unfortunately, higher concentrations of propranolol could not be used without producing nonspecific effects. While both ADR and NA also can activate α₁-ARs, it is unlikely that their ability to desensitize that α₂-AR response in this cell line involves α₁-ARs: the α₁-AR agonist phenylephrine failed to stimulate IP₃ formation in BE(2)-C cells at concentrations up to 100 μM (data not shown) confirming the lack of functional α₁-ARs in this cell line.

Table 2.

Characteristics of α₂-ARs in BE(2)-C cells

To verify the involvement of the β-AR in the process of α₂-AR desensitization by 300 nM chronic ADR, we took advantage of the fact that the β-AR number in our cells significantly declines with increasing passage number while α₂-AR levels and responsiveness are maintained. After passage 50, β-AR responsiveness wanes. By passage 55, both β-AR responsiveness and number are reduced over 80% (Table 1). In BE(2)C cells>passage 50, chronic ADR treatment does not alter brimonidine efficacy, underscoring the contribution of the β-AR to α₂-AR desensitization by ADR (Figure 1D).

Table 1.

Effect of cell passage number on α₂- and β-ARs in BE(2)-C cells

To establish that α₂-AR desensitization was not produced by activation of the β-AR alone, and that it required simultaneous activation of both α₂- and β-ARs, cells were pretreated with the β-AR-specific agonist, ISO. ISO pretreatment (1 μM) had no effect on brimonidine potency (log M EC₅₀ vehicle −7.44±0.14 vs ISO −7.59±0.23) or efficacy (I_max value: ISO 70.2±2.65; vehicle 69.7±7.3; n=2–6). None of the pretreatments altered basal or forskolin-stimulated levels of cAMP accumulation compared to vehicle-treated controls. Together, these results confirm previous reports that treatment with very high concentrations of catecholamines are required to desensitize the α₂-AR response (Eason & Ligget, 1992; Lanier et al., 1996) if no β-ARs are present or activated, but that much lower concentrations of catecholamines can produce α₂-AR desensitization if activating both α₂-AR and β-ARs simultaneously.

To confirm the β₂-AR classification, the β-AR response to 100 nM ISO was dose-dependently reduced by the β₂-selective antagonist, ICI 118,551, while the β1-selective antagonist, CGP 20712A was inactive at the highest concentration tested (1 μM; Figure 3). These results indicate that ISO increases cAMP accumulation through β₂-ARs in BE(2)-C cells. In addition, propranolol (30 nM) also blocked β-AR stimulation of cAMP accumulation, supporting its effectiveness at blocking the β-AR in the desensitization studies (Figure 3). Binding studies (Altememi et al., 1997; data not shown) further supported the β₂-AR subtype distinction.

Figure 3.

The ISO response in BE(2)-C cells is mediated through β₂-ARs. Cells were assayed for ISO-stimulated accumulation of cAMP in the presence and absence of three different β-AR antagonists. The non-selective antagonist propranolol and the β₂-selective antagonist ICI 118,551 antagonized the stimulatory effect of ISO (100 nM), while the β₁-selective antagonist CGP 20712A was without effect (^*P<0.05, ^**P<0.01, ^***P<0.001 by ANOVA; n=3–13). Antagonists alone had no effect on basal cAMP levels at these concentrations (data not shown).

To determine whether ADR-induced desensitization of α₂-ARs was a result of receptor down-regulation, changes in levels of specific α₂-AR binding were measured using a single concentration of radioligand. This was sufficient for accurate assessment of changes in receptor number since α₂-AR down-regulation does not change radioligand affinity as observed by our own (K_d values: vehicle 0.95±.03; ADR 0.71±.01 nM) and others' studies (Eason & Liggett, 1992). Alpha₂-AR levels were reduced 60% following chronic 300 nM ADR treatment (Figure 4; P<0.05), consistent with the functional deficits noted (Figure 1). Furthermore, inclusion of 30 nM propranolol to block activation of β₂-ARs during chronic ADR treatment also prevented the α₂-AR down-regulation; propranolol treatment alone had no effect. Alpha₂-AR levels were also reduced following 24 h pretreatment with 30 μM NA (Figure 4; P<0.05), but were unaltered by 1 μM NA or ISO treatments that also failed to produce α₂-AR desensitization.

Figure 4.

Chronic ADR treatment down-regulates α₂-ARs. BE(2)-C cells were incubated for 16–24 h with vehicle (0.1 mM ascorbate), 300 nM ADR, 30 nM propranolol, ADR+30 nM propranolol, 1 μM NA, 30 μM NA or 1 μM ISO. Specific binding to cell membrane homogenates was calculated by subtracting the binding of a single concentration of radioligand (2 nM) in the presence of phentolamine (10 μM) from the binding in its absence. ADR profoundly down-regulated α₂-AR levels, but this down-regulation was blocked by the β-AR antagonist, propranolol (^*P<0.05 by ANOVA). Chronic 30 μM NA treatment also reduced α₂-AR levels while ISO, 1 μM NA or propranolol alone had no effect. Data are presented as mean±s.e.mean of 3–5 independent experiments; α₂-AR binding in control cells was 1143±283 c.p.m. mg⁻¹ protein.

Since down-regulation of the α_2A-AR subtype requires pretreatment with higher agonist concentrations than that of α_2B- and α_2C-ARs (Heck & Bylund, 1997; Deupree et al., 2002), various binding and antagonism studies were conducted to determine the α₂-AR subtype expressed in BE(2)-C cells. The nonselective α₂-AR antagonists, rauwolscine and yohimbine, competed for specific [³H]-rauwolscine binding to membranes with high affinity (Table 2; Figure 5A); the two α_2B/C-selective antagonists, ARC-239 and prazosin, displaced [³H]-rauwolscine from binding sites with 100–400 fold lower affinity than rauwolscine. Hill slope values of antagonist binding did not differ from unity, consistent with binding to a single site (Table 2). The various α-AR agonists showed significant differences in their ability to competitively displace [³H]-rauwolscine from its binding site (Table 2), with the more α_2A-AR-preferring agonists (oxymetazoline, brimonidine, guanabenz) and nonselective agonists (ADR, p-aminoclonidine) exhibiting higher affinities than the α_2B/C-AR-preferring agonist, agmatine or the α₁-AR agonist, phenylephrine. Unlike the α₂-AR agonists, phenylephrine exhibited only weak, partial agonist actions in cyclase assays (Table 2). To further distinguish between the different α₂-AR subtypes that could be expressed in BE(2)-C cells, another process was employed comparing the affinity ratios of prazosin to OXY or OXY to yohimbine (Hieble et al., 1997; Table 2). Prazosin : OXY (141) and OXY : yohimbine (1.54) ratios were within the range reported for native and recombinant α_2A-ARs, and differ by at least 10 fold from α_2C-AR values (from Hieble et al., 1997). Apparent K_i values were generated for comparison with previously reported values from cells natively expressing α_2A-, α_2B, or α_2C-ARs (HT29, NG108-15, OK; Gleason & Hieble, 1991; Blaxall et al., 1991) or cell lines expressing cloned α₂C10, α₂C2, and α₂C4 (Jansson et al., 1994; Jasper et al., 1998). Values obtained from our binding studies correlated significantly with those derived from native and cloned α_2A-AR-containing cell membranes only (Table 3), with no correlation in cells or membranes containing α_2B or α_2C-ARs.

Figure 5.

Alpha_2A-AR antagonists displace [³H]-rauwolscine binding and block brimonidine-mediated inhibition of cAMP accumulation. (A) Binding assays were performed as described in ‘Methods' using [³H]-rauwolscine in the presence of 6–8 concentrations of rauwolscine, yohimbine, ARC 239, or prazosin in BE(2)-C cell membrane homogenates. Curves were best fit by nonlinear regression analysis (GraphPad Prism, San Diego, CA, U.S.A.). Each point is the mean±s.e.mean of triplicate determinations from 3–8 experiments. (B) Rauwolscine dose-dependently and significantly antagonized the ability of brimonidine (100 nM) to inhibit cAMP accumulation, while the α_2B/C-AR-selective antagonist, ARC-239, had no effect. The results represent the mean±s.e. of 3–9 experiments, performed in duplicates.

Table 3.

Correlation of human neuroblastoma cell α₂-AR pK_i values with native and cloned α₂-AR subtypes

The classification of α₂-AR is further borne out by functional antagonism studies. Inhibition of cAMP accumulation by brimonidine (100 nM; Figure 5B) in BE(2)-C cells was reversed by rauwolscine in a concentration-dependent manner, whereas the α_2B/C-selective antagonist ARC-239 failed to reverse the actions of the agonist. This further supports the classification of the functional α₂-AR subtype in this neuroblastoma cell line as α_2A.

Previous studies have shown that chronic activation of the α₂-AR leads to down-regulation of inhibitory G-proteins (G_αi; Jewell-Motz et al., 1998; Eason & Liggett, 1992). We observed a small reduction in G_αi3 levels following chronic 300 nM ADR treatment, but this effect never quite achieved significance (Table 4).

Table 4.

G-protein levels following treatment in BE(2)-C cells

Chronic ADR up-regulates GRK3 facilitating α_2A-AR desensitization and down-regulation

β-AR-dependent α₂-AR desensitization could occur via β₂-AR activation of protein kinase A (PKA; Reutter et al., 1998), G-protein-coupled receptor kinase 2 or 3 (GRK2/3; Benovic et al., 1987) or PKC (Shih & Malbon, 1994). Cyclic AMP levels in the cell would need to be increased by ADR treatment in order to activate the cAMP-dependent enzyme, PKA. In the absence of forskolin, ADR (300 nM) alone has little effect on cAMP levels (13.8±10.5% inhibition of basal; n=3), even though β₂-, as well as α₂-ARs, are activated. This suggests that α₂-AR desensitization is not PKA-dependent.

Agonist-occupied α_2A-AR is a substrate for GRK2 and GRK3, but not for GRK5 or GRK6 (Jewell-Motz & Liggett, 1996). To examine the role of GRKs in β₂-AR-faciliated α_2A-AR desensitization, we utilized antisense DNA that is common to a coding sequence found in both GRK2 and GRK3 (GRK 2/3), but not present in GRK5 or GRK6. Cells were pretreated with either the GRK2/3 antisense or the corresponding sense DNA for 4 h in serum free media (Shih & Malbon, 1994), before adding sera back for the remainder of the treatment period (24 h±ADR). Under these conditions, 300 nM ADR-induced α₂-AR desensitization was blocked by GRK2/3 antisense, but not sense, treatment (Figure 6A). Antisense and sense treatments alone were without effect on the α₂-AR response. Pretreatment with GRK2/3 antisense also ablated chronic 300 nM ADR-induced α₂-AR down-regulation (Figure 6B), suggesting that GRK regulates both desensitization and downregulation of the α₂-AR following prolonged agonist treatment. This is consistent with the idea that α₂-AR desensitization results from α₂-AR down-regulation. Cell lysates from these experiments were resolved on SDS–PAGE to insure that the GRK 2/3 antisense ODN reduced GRK 2/3 protein levels. Using antibodies recognizing both GRK2 and GRK3, we observed that chronic 300 nM ADR pretreatment alone up-regulated GRK 2/3 (Figure 6A, insert) and GRK 2/3 antisense blocked the up-regulation. Utilizing antibodies specific for GRK2, GRK3, or GRK5 isoforms, we found that chronic ADR preferentially up-regulated GRK3 (Figure 7A); levels of GRK6 in these cells are too low to consistently detect by Western blotting. Inclusion of propranolol (30 nM; Figure 7A) blocked this up-regulation. Neither chronic 1 μM NA nor 250 nM ISO treatment altered GRK3 levels, but after increasing the NA concentration 30 fold (Figure 7B), GRK3 levels were up-regulated as noted with ADR pretreatment. Chronic 30 μM NA-induced GRK3 up-regulation was partially ablated by the addition of 1 μM propranolol, similar to the desensitization results (Figure 1B), though this ablation failed to achieve significance by ANOVA. Levels of GRK5 were neither up-regulated by chronic treatments (Figure 7), nor altered by GRK2/3 antisense treatment (data not shown).

Figure 6.

GRK antisense ODN treatment inhibits chronic 300 nM ADR-induced α₂-AR desensitization and down-regulation. BE(2)-C cells were washed with serum-free media and incubated in the presence or absence of an antisense (AS) or sense ODN recognizing sequences common to both GRK2 and 3 (20 μM) for 4 h. Serum was returned to the media for an additional 24 h with ADR (300 nM) or vehicle. (A) Again, 300 nM ADR treatment reduced brimonidine potency and efficacy (n=8; ^*P<0.001 by ANOVA) as compared to vehicle-treated cells. Treatment with GRK antisense (n=5), but not GRK sense (n=3) ODNs reversed these effects and prevented ADR-induced α₂-AR desensitization (P<0.05). Treatment with GRK antisense (n=3) or sense (n=4) ODNs alone had no effect (data not shown). Insert: 30 μg of whole cell lysate from each treatment group was resolved by SDS–PAGE through a 10% gel as indicated by the representative blot. Chronic 300 nM ADR increased GRK levels using antisera common to GRK2 and GRK3 (n=4–5; ^*P⩽0.05) compared to vehicle-treated controls. GRK antisense ODN treatment blocked ADR-induced increases in GRK expression as determined by densitometric analysis of immunoblots normalized to a GAPDH loading control. (B) Specific binding to cell membrane homogenates was calculated by subtracting the binding of [³H]-rauwolscine (2 nM) in the presence of phentolamine (10 μM) from the binding in its absence. GRK antisense ODN ablated chronic ADR-induced α₂-AR down-regulation (P<0.05 by ANOVA) while GRK sense ODN had no effect. Neither GRK antisense nor sense ODN alone altered α₂-AR levels. Data are presented as mean±s.e.mean of four independent experiments; α₂-AR binding in control cells was 2815±525 c.p.m. mg⁻¹ protein.

Figure 7.

β₂-AR-dependent GRK3 up-regulation. BE(2)-C cells were treated with vehicle (0.1 mM ascorbate), 300 nM ADR, propranolol (30 nM or 1 μM), ADR+30 nM propranolol, 1 μM NA, 30 μM NA, 30 μM NA+1 μM propranolol, or 250 nM ISO for 24 h. Thirty μg of whole cell lysate from each treatment group was resolved by SDS–PAGE through a 10% gel. Immunoreactive bands were normalized to the GAPDH loading control. (A) GRK isoform-specific antibodies indicated that the GRK3 isoform, but not GRK2 or GRK5, was significantly up-regulated following chronic 300 nM ADR treatment (n=4; ^*P⩽0.05) compared to vehicle-treated controls. Addition of 30 nM propranolol ablated ADR-induced GRK3 up-regulation while 30 nM propranolol alone had no effect. (B) Chronic 30 μM NA treatment up-regulated GRK3 (^*P⩽0.05) similar to ADR treatment while treatment with 1 μM NA, 250 nM ISO, or 1 μM propranolol alone had no effect on GRK3 levels. Though inclusion of 1 μM propranolol partially prevented 30 μM NA-induced GRK3 up-regulation, it was not significantly different from 30 μM NA by ANOVA with Tukey's post-hoc test.

Acute 300 nM ADR and 1 μM NA induce α₂-AR desensitization

In the present study, chronic co-activation of α₂- and β₂-AR by 300 nM ADR induces α₂-AR desensitization while activation of α₂-AR alone by 1 μM NA has no effect. To determine whether the same mechanism regulates α₂-AR signalling with acute agonist treatment, cells were treated with ADR (300 nM) or NA (1 μM) for only 60 min. Unlike chronic treatments, both acute 300 nM ADR and 1 μM NA pretreatments reduced brimonidine efficacy as compared to vehicle (P<0.05; Figure 8) in BE(2)-C cells. As noted with chronic ADR treatment, brimonidine potency was reduced by only 2 fold. Addition of 30 nM propranolol to either acute 300 nM ADR- (Figure 8) or 1 μM NA-treated cells (data not shown) failed to block α₂-AR desensitization. Therefore, α₂-ARs in these cells can desensitize following acute exposure to both 300 nM ADR and 1 μM NA suggesting that co-activation of the β₂-AR is not necessary for acute α₂-AR desensitization by submaximal concentrations of catecholamines. These results also indicate the 1 μM NA is sufficient to produce α₂-AR desensitization, but that mechanisms of α₂-AR desensitization vary with different times of agonist exposure.

Figure 8.

Acute 300 nM ADR and 1 μM NA induce β₂-AR-independent α₂-AR desensitization. The α₂-AR response as determined by the ability of brimonidine to inhibit forskolin (10 μM)-stimulated cAMP accumulation in intact BE(2)-C human neuroblastoma cells was measured 1 h after pretreatment with any or all of the following: vehicle (0.1 mM ascorbate), 300 nM ADR, 30 nM propranolol, ADR+30 nM propranolol, or 1 μM NA. Both 300 nM ADR and 1 μM NA treatment reduced brimonidine potency and efficacy (n=5–6; ^*P<0.05 by ANOVA) compared to vehicle in BE(2)-C cells cultured up to passage 50. Addition of 30 nM propranolol to either ADR or NA (data not shown) treatment failed to block α₂-AR desensitization.

Discussion

Cross-regulation of two opposing signalling cascades activated by the same endogenous neurotransmitter is a complex physiological event. Several in vivo studies suggested that α₂-AR desensitization following prolonged agonist exposure was regulated by β₂-AR activation (Schwartz & Eikenburg, 1988; Apparsundaram & Eikenburg, 1996), however the mechanism of β₂-AR-faciliated α₂-AR desensitization was not examined. Further elucidation of the cellular mechanisms responsible for this cross-talk required the study of this process in a cell system expressing both β₂- and α₂-ARs, preferably at physiological levels. Therefore, we set out to determine how simultaneous activation of natively expressed β₂- and α₂-ARs by chronic 300 nM ADR in BE(2)-C cells produces α₂-AR desensitization, while activation of the α₂-AR alone by 1 μM NA fails to do so.

β₂-AR involvement in α₂-AR desensitization is clearly implicated by several results in the present study. First, α₂-AR desensitization occurred following chronic 300 nM ADR and 30 μM NA, but not 1 μM NA or ISO treatments. Lands et al. (1967) established that ADR has a higher potency at β₂-ARs than NA, while ADR and NA have equal potencies at each α₂-AR subtype (Lanier et al., 1996; Jasper et al., 1998; Deupree et al., 2002). Therefore, chronic 300 nM ADR desensitized the α₂-AR response by activating both α₂- and β₂-ARs simultaneously, while 1 μM NA or ISO activated only α₂- or β₂-ARs, respectively. Only when a higher concentration of NA was used (30 μM), a concentration that activates both α₂- and β-ARs, was α₂-AR desensitization noted. Second, α₂-AR desensitization produced by modest levels of ADR was observed only when the cells expressed both α₂- and β₂-ARs. When BE(2)-C cells are cultured beyond 50 passages, β-AR expression is lost and 300 nM ADR fails to produce α₂-AR desensitization. Finally, the β-AR antagonist, propranolol prevents chronic ADR (300 nM)-promoted α₂-AR desensitization and partially antagonizes α₂-AR desensitization produced by the higher NA concentration (30 μM). Collectively, these observations illustrate that chronic simultaneous activation of α₂- and β₂-ARs enables α₂-AR desensitization at lower concentrations of ADR than NA. Moreover, since the degree of α₂-AR desensitization produced by the higher concentration of NA is partially reduced by β-AR blockade, it suggests that the β-AR may also enhance α₂-AR desensitization at higher concentrations of NA that also activate both α₂- and β-AR.

Acute 300 nM ADR treatment also desensitized the α₂-AR, but unlike chronic treatment, the desensitization was independent of β₂-AR activation as indicated by several results. First, both acute 300 nM ADR (co-activating α₂- and β₂-ARs) and 1 μM NA (activating only α₂-ARs) treatments induced α₂-AR desensitization to a similar extent. Second, inclusion of 30 nM propranolol during 300 nM ADR or 1 μM NA treatments failed to block α₂-AR desensitization. The fact that 1 μM NA desensitizes the α₂-AR after acute, but not chronic, treatment underscores the difference in mechanisms responsible for α₂-AR regulation following short- and long-term agonist exposure.

The present study is the first to provide evidence for the mechanisms that contribute to β₂- AR-dependent α₂-AR desensitization after chronic ADR exposure. When recombinant α₂-ARs were expressed in the absence of β-ARs, long-term (24 h) agonist-induced α₂-AR desensitization was attributed to α₂-AR (Pleus et al., 1993; Heck & Bylund, 1997; Jewell-Motz et al., 1997) and G_αi/o down-regulation (Eason & Liggett, 1992; Liggett et al., 1992; Jewell-Motz et al., 1998). β₂-AR-dependent α₂-AR desensitization appears to involve α₂- AR down-regulation more than reduction of G_αi/o. Since the three α₂-AR subtypes have different propensities to down-regulate (Pleus et al., 1993), binding and functional studies were utilized to establish the α₂-AR subtype expressed in BE(2)-C cells. Several results support the α_2A-AR designation. First, rauwolscine and yohimbine compete for [³H]-rauwolscine binding with much higher affinity than the α_2B/C-AR selective antagonists ARC-239 and prazosin, that displayed only a weak displacement capability. Second, when apparent pKi values for various α₂-AR agonists and antagonists binding to BE(2)-C membrane homogenates were compared with previously reported values in cell lines expressing a single α₂-AR subtype, a significant correlation was observed only with those cells that expressed α_2A-AR. Finally, the inhibitory effect of α₂-AR agonists on cAMP production is readily reversed in a concentration-dependent fashion by the antagonist rauwolscine. The failure of the selective α_2B/C-AR antagonist ARC-239 to produce any antagonism of brimonidine is consistent with activation of α_2A-AR in BE(2)-C cells. Alpha_2A-ARs require exposure to higher agonist concentration to induce receptor down-regulation than α_2B or α_2C-ARs (Deupree et al., 2002). In the present study, we report down-regulation of the α_2A-AR by ADR after exposure to much lower agonist concentrations than previously reported for this subtype (Eason & Liggett, 1992; Pleus et al., 1993; Jewell-Motz et al., 1997). Whether this is due to more physiological AR levels in our system, a more physiological concentration of ADR (300 nM), a β₂-AR-facilitated effect permitting down-regulation of the α₂-AR at lower ADR concentrations, or a combination of these effects is not clear. Nevertheless, our results suggest the β₂-ARs on BE(2)-C cells regulate the α₂-AR response, in part, by directly modulating the level of α_2A-ARs.

The results of the present study differ from previous reports where the effect of β-AR activation on inhibitory receptor number/function has been studied in other cells/tissues. For example, in HT-29 cells, 24 h stimulation of adenylyl cyclase with forskolin or cAMP analogues resulted in an increase in inhibitory α_2A-AR expression (Sakaue & Hoffman, 1991). That response was attributed, in part, to increased α_2A-AR gene expression by cAMP response elements in the AR gene. A major difference between our study and these previous studies is that we simultaneously activated both the stimulatory (β₂-AR) and inhibitory (α₂-AR) receptors during chronic treatment. This difference is significant for two reasons. First, chronic simultaneous activation of α₂- and β₂-ARs in BE(2)-C cells by ADR resulted in α_2A-AR down-regulation. Second ADR treatment does not induce cAMP accumulation even though both α₂- and β₂-AR are simultaneously activated. Therefore it seems unlikely that cAMP or PKA contributed to the response in the present study. The mechanism most likely to contribute to the β₂-AR-dependent α₂-AR desensitization is G protein coupled receptor kinases (GRKs).

Recently several reports have implicated GRKs in regulation of receptor signalling following prolonged agonist treatment (McGraw et al., 1998; Thakker & Standifer, 2002). Several lines of evidence support a role for GRKs in the present study. First, chronic ADR treatment of BE(2)-C cells resulted in up-regulation of GRK3 expression. Since the α_2A-AR is a substrate for GRK3 when agonist-occupied (Benovic et al., 1987; Jewell-Motz & Liggett, 1996), an increase in the expression of GRK3 could render the α₂-AR more readily desensitized upon re-challenge with the α₂-AR agonist (brimonidine). In support of this suggestion, injection of recombinant GRK3 into isolated chick dorsal root ganglion neurons increased GRK3 levels in those cells and rendered the α₂-AR more susceptible to homologous desensitization (Diverse-Pierluissi et al., 1996). This effect was specific for GRK3 as increasing levels of GRK2 failed to facilitate α₂-AR desensitization. Second, β-AR blockade eliminates α₂-AR desensitization as well as the up-regulation of GRK3 by ADR. Moreover, antisense ODNs, which block GRK3 up-regulation, also prevent ADR-induced α₂-AR desensitization and down-regulation. A recent study by Deupree et al. (2002) reports that intact GRK phosphorylation sites are required for the down-regulation of the α_2C-AR, suggesting that GRK phosphorylation regulates α_2C-AR down-regulation. Similarly in the present study up-regulation of GRK3 is required for the profound α₂-AR down-regulation induced by modest ADR concentrations. In fact, in the present study, 300 nM ADR produced desensitization and down-regulation of the α_2A-AR at least as marked as that reported at 30–300 fold higher concentrations of ADR by other investigators (Eason & Liggett, 1992; Jewell-Motz et al., 1997). Therefore, we suggest that chronic simultaneous activation of α₂- and β₂-ARs by ADR renders the α₂-AR more sensitive to homologous desensitization and α₂-AR down-regulation via GRK3 up-regulation.

Abbreviations

AR

adrenoceptor

ARC-239

2-(2,4-(O-methoxyphenyl)-piperazin-1-yl)-ethyl- 4,4dimethyl-1,3-(2H,4H)-isoquinolindione

brimonidine

5-Bromo-N- (4,5-dihydro-1H-imidazole-2-yl)-6-quinoxalinamine (or UK14,304)

G_αi

alpha subunit of inhibitory guanine nucleotide binding proteins

G_αs

alpha subunit of stimulatory guanine nucleotide binding proteins

G_βγ

beta and gamma subunits of guanine nucleotide binding proteins

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

GRK

G protein coupled receptor kinase

HBSS

Hank's balanced salt solution

IBMX

3-isobutyl-1- methylxanthine

I-CYP

iodo-cyanopindolol

ODN

oligodeoxynucleotide

PKA

cAMPdependent protein kinase

PKC

phospholipase C-dependent protein kinase

PMSF

phenylmethyl sulfonylflouride

PNS

peripheral nervous system

PVDF

polyvinylidene diflouride

SDS–PAGE

sodium dodecyl sulphate-polyacrylamide gel electrophoresis

TBS/T

tris-buffered saline/tween