Muscarinic Acetylcholine Receptor Modulation of Mu (μ) Opioid Receptors in Adult Rat Sphenopalatine Ganglion Neurons

Abstract

The sphenopalatine ganglion (SPG) neurons represent the parasympathetic branch of the autonomic nervous system involved in controlling cerebral blood flow. In the present study, we examined the coupling mechanism between mu (μ) opioid receptors (MOR) and muscarinic acetylcholine receptors (mAChR) with Ca²⁺ channels in acutely dissociated adult rat SPG neurons. Successful MOR activation was recorded in ∼40–45% of SPG neurons employing the whole cell variant of the patch-clamp technique. In addition, immunofluorescence assays indicated that MOR are not expressed in all SPG neurons while M₂ mAChR staining was evident in all neurons. The concentration-response relationships generated with morphine and [d-Ala2-N-Me-Phe4-Glycol5]-enkephalin (DAMGO) showed IC₅₀ values of 15.2 and 56.1 nM and maximal Ca²⁺ current inhibition of 26.0 and 38.7%, respectively. Activation of MOR or M₂ mAChR with morphine or oxotremorine-methiodide (Oxo-M), respectively, resulted in voltage-dependent inhibition of Ca²⁺ currents via coupling with Gα_i/o protein subunits. The acute prolonged exposure (10 min) of neurons to morphine or Oxo-M led to the homologous desensitization of MOR and M₂ mAChR, respectively. The prolonged stimulation of M₂ mAChR with Oxo-M resulted in heterologous desensitization of morphine-mediated Ca²⁺ current inhibition, and was sensitive to the M₂ mAChR blocker methoctramine. On the other hand, when the neurons were exposed to morphine or DAMGO for 10 min, heterologous desensitization of M₂ mAChR was not observed. These results suggest that in rat SPG neurons activation of M₂ mAChR likely modulates opioid transmission in the brain vasculature to adequately maintain cerebral blood flow.

INTRODUCTION

Sphenopalatine ganglion (SPG) neurons are well known to play an important role in cerebral blood flow regulation as well as lacrimal and nasal gland secretion (Hara et al. 1993; Lee et al. 2001; Smith et al. 2002). SPG neurons are a major source of vasoactive substances including nitric oxide (NO) and vasoactive intestinal peptide (VIP) (Chen and Lee 1995; Gibbins 1990; Hara et al. 1985; Leblanc et al. 1987; Shimizu et al. 2001). The release of these transmitters by activated SPG neurons leads to an increase in cerebral blood flow. Some studies have presented evidence indicating that acetylcholine (ACh) and NO are co-released from nerve terminals innervating the brain vasculature (Chen and Lee 1993; Toda and Okamura 2003). It is thought that released ACh binds to muscarinic acetylcholine receptors (mAChR) leading to inhibition of NO release and subsequently a reduction of the NO-mediated neurogenic vasodilation (Lee et al. 2001).

SPG blockade has been employed clinically for the treatment of migraine headaches, cluster headaches, and other types of facial pain, including trigeminal neuralgia (Felisati et al. 2006; Obah and Fine 2006). One commonly used method of blocking SPG neurons is the local application of topical anesthetics. Employing this approach, one study found that patients undergoing treatment of trigeminal neuralgia reported a significant relief in pain when treated with carbamazepine along with local application of the opiate, buprenorphine, to either SPG or superior cervical ganglion (SCG) neurons (Spacek et al. 1997). The authors speculated that the presence of mu (μ) opioid receptors (MOR) in SPG or SCG could exert a modulatory role in neuronal pathways observed in trigeminal neuralgia. Opioid administration has been shown to inhibit the nociceptive neurotransmission within the trigeminal neurons in rats (Williamson et al. 2001). Morphine, the most common opiate employed in the treatment of chronic pain, exhibits a high potential for both tolerance and abuse. The morphine-mediated desensitization of MOR is generally believed to be responsible for the tolerance that is observed. However, the exact mechanism underlying the tolerance that develops is poorly understood.

Previous studies have shown that rat SPG neurons express the M₂ mAChR subtype (Liu et al. 2002; Margas and Ruiz-Velasco 2007). Activation of M₂ mAChR leads to voltage-dependent inhibition of Ca²⁺ channel currents, and N-type Ca²⁺ channels are the main carriers of Ca²⁺ ions in SPG neurons (Liu et al. 2000; Margas and Ruiz-Velasco 2007). Given that local application of the opiate analgesic buprenorphine to SPG resulted in greater pain relief, the purpose of the present study was to examine whether acutely isolated rat SPG neurons express MOR that couple to Ca²⁺ channels. In addition, we wanted to determine if M₂ mAChR and MOR employ the same signal transduction components, and how acute desensitization of either receptor would affect the subsequent coupling of the other receptor to Ca²⁺ channels. Electrophysiological and immunofluorescence techniques were employed to examine the interplay between both receptor signaling elements.

METHODS

Sphenopalatine ganglion (SPG) neuron isolation

The experiments performed were approved by the Penn State College of Medicine Institutional Animal Care and Use Committee (IACUC). Single SPG neurons were isolated from adult rats employing the method previously described (Margas and Ruiz-Velasco 2007). Briefly, male Wistar rats (175–250 g) were initially anesthetized with CO₂ and decapitated with a laboratory guillotine. Thereafter, the SPG was removed and cleared of connective tissue in cold Hank's balanced salt solution (Sigma Chemical, St. Louis, MO). The ganglia were enzymatically dissociated in modified Earle's balanced salt solution containing 0.6 mg/ml collagenase (Roche Pharmaceuticals, Switzerland) and 0.4 mg/ml trypsin (Worthington Biochemical, Lakewood, NJ) for 40 min at 35°C in a shaking water bath. The neurons were next dispersed by vigorous shaking and then centrifuged twice for 6 min at 130 × g. Finally, the cells were resuspended in minimal essential medium (MEM), supplemented with 10% fetal calf serum, 1% glutamine, and 1% penicillin-streptomycin (all from Invitrogen, Carlsbad, CA) and plated onto 35 mm poly-l-lysine coated dishes. The dissociated neurons were stored in a humidified incubator (5% CO₂-95% air) at 37°C.

Electrophysiology and data analysis

Ca²⁺ currents were recorded using the whole cell patch-clamp technique. The patch pipettes were pulled from borosilicate glass (Garner Glass, Claremont, CA) and pulled on a P-97 micropipette puller (Sutter Instrument, Novato, CA). SPG whole cell Ca²⁺ currents were acquired with the Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA), analog filtered at frequency 2–10 kHz (–3 dB, 4-pole low-pass Bessel filter), digitized with custom-designed S5 software (Stephen R. Ikeda, National Institutes of Health/NIAAA) equipped with an 18-bit AD converter board (Instrutech, Elmont, NY). Cell membrane capacitance and pipette series resistance were electronically compensated (80–85%).

Solutions and drugs

The external solution, referred to in the text as “TEA external,” contained (in mM): 145 tetraethylammonium hydroxide (TEA-OH), 140 methanesulphonic acid, 10 HEPES, 15 glucose, 10 CaCl_2, and 0.003 TTX, to pH 7.40 with TEA-OH, and an osmolality of 325 mosM/kg. In some experiments, the external solution, referred to as “Tris external,” contained (in mM): 155 tris hydroxymethyl aminomethane (Tris), 20 HEPES, 10 CaCl₂, 10 glucose, and 0.0003 tetrodotoxin (TTX), to pH 7.40 with methanesulphonic acid, and an osmolality of 318 mosM/kg. The pipette solution contained (in mM) 120 N-methyl-d-glucamine, 20 TEA-OH, 20 HCl, 11 EGTA, 10 HEPES, 1 CaCl₂, 4 MgATP, 0.3 Na₂GTP, and 14 Na-creatine phosphate, to pH 7.20 and an osmolality of 305 mosM/kg. Unless specified, the Ca²⁺ current recordings were performed with “TEA external” solution.

Stock solutions of morphine, [d-Ala2-N-Me-Phe4-Glycol5]-enkephalin (DAMGO), N-ethylmaleimide (NEM), methoctramine (all from Sigma Chemical), and oxotremorine-methiodide (Oxo-M; Research Biochemicals International, Natick, MA) were prepared in H₂O and diluted in the external solution to their final concentration prior to use. In some experiments, SPG neurons were pretreated overnight (12–14 h) with pertussis toxin (PTX; List Biological Laboratories, Campbell, CA; 500 ng/ml) present in the culture medium.

Morphine, DAMGO, and Oxo-M concentration-response relationships

The concentration-response relationships were determined by the sequential application of the receptor agonist in increasing concentrations. Two to three different concentrations were used with each cell to minimize desensitization. The results were then pooled and the concentration-response curves were fit to the Hill equation: I = I_MAX/[1 + (IC₅₀/[ligand])^nH], where I is the percentage inhibition, I_MAX is the maximum inhibition of the Ca²⁺ current, IC₅₀ is the half-inhibition concentration, [ligand] is the agonist concentration and nH is the Hill coefficient. These parameters were obtained with Prism 4.0 software (GraphPad Software, San Diego, CA). All data are presented as means ± SE.

Statistical analysis

Data and statistical analyses were performed with Igor Pro (Lake Oswego, OR) and Prism 4.0 (GraphPad Software) software packages, respectively with P < 0.05 considered statistically significant. Graphs and current traces were generated with Igor Pro and Canvas 8.0 (Deneba Software, Miami, FL) software packages. The effect of specific agonists on homologous and heterologous desensitization of M₂ mAChR and MOR receptors was estimated with repeated-measures ANOVA test, followed by the Newman-Keuls test employing Prism 4.0.

Immunofluorescence staining

Immunofluorescence procedures were performed as previously described (Margas and Ruiz-Velasco 2007; Yang et al. 2006). Briefly, SPG neurons were initially rinsed five times with 1X PBS and then fixed for 20 min with 2% formaldehyde/2% sucrose. Next the cells were permeabilized with 0.05% Tween 20 in 1× PBS and 5% goat serum (Vector Labs, Burlingame, CA) for 10 min at 37°C. The cells were rinsed five times and then incubated for 60 min in 1× PBS supplemented with 5% goat serum. Thereafter the primary antibody was added to the neurons for 60 min. The primary antibodies employed were rabbit anti-MOR (1:250) and mouse anti-M₂ mAChR (1:200; both from Abcam, Cambridge, MA). The cells were then rinsed and incubated for 45 min in cold 1× PBS and 5% goat serum containing the secondary, goat anti-rabbit or goat anti-mouse antibody conjugated with Alexa Fluor 568 IgG at a final concentration of 3 μg/ml. Finally, the neurons were rinsed three times with 1× PBS. In some experiments, the cells were stained with the secondary antibody without exposing the neurons to the primary antibody. Phase contrast and immunofluorescence images were acquired with an Orca-ER 1394 digital CCD camera (Hamamatsu Photonics) and Nikon TE2000 microscope (Nikon) equipped with filter set (G-2E/C; Nikon) containing an excitation filter at 540 ± 15 nm, a dichroic beam splitter of 585 nm [long-pass (LP)] and emission filter at 620 ± 30 nm. The collected images were analyzed with iVision software (BioVision Technologies, Exton, PA). Optical sections of labeled cells were collected with a ×40 objective and taken in 2-μm steps covering the Z-axis field using the Pro Scan II (Prior Scientific, Cambridge, UK) motorized stage. The acquired images were deconvolved with the Huygens Essential (Scientific Volume Imaging BV, NL) software and pseudocolored with iVision software.

RESULTS

MOR are natively expressed in a subpopulation of rat SPG neurons

In the first set of experiments, the concentration-response relationships for morphine and the high-affinity MOR agonist, DAMGO, were determined. Figure 1, A and C, shows the time courses of peak Ca²⁺ current amplitude of prepulse (•) and postpulse (○) obtained before and during application of morphine and DAMGO, respectively. Ca²⁺ currents were evoked every 10 s with the double-pulse voltage paradigm (shown on Fig. 1B, top) and peak current amplitude was measured isochronally 10 ms after the initiation of the prepulse and postpulse in the absence or presence of the agonist. The voltage protocol consists of a 20-ms test pulse to +10 mV (prepulse), that is followed by a conditioning pulse to +80 mV, and a second 20-ms test pulse (postpulse) to +10 mV before returning to the holding potential of −90 mV. Figure 1B shows the numbered Ca²⁺ current traces evoked with this paradigm and correspond to those plotted in A. Exposure of the cell to 0.003 μM morphine (trace 3) resulted in ∼3% inhibition of the prepulse current. Following a recovery period, 0.3 μM morphine was applied to the neuron and led to a 29% inhibition of the prepulse Ca²⁺ current (trace 5). After a recovery period, application of 3 μM morphine resulted in a 32% inhibition of the current (trace 7). Note also that with 0.3 or 3 μM morphine, the Ca²⁺ current inhibition was greater during the prepulse than the postpulse (traces 6 and 8), indicating a voltage-dependent inhibition of the Ca²⁺ currents that is characterized by the “kinetic slowing” of the prepulse current and an enhanced postpulse current amplitude (Ikeda 1991). The time course in Fig. 1C shows the effect of 0.003, 0.03, and 10 μM DAMGO on the peak Ca²⁺ currents, and the corresponding numbered current traces are shown to the right (Fig. 1D). Although a small inhibitory effect (∼3%) was observed with 0.003 and 0.03 μM DAMGO (traces 3 and 5), exposure to 10 μM DAMGO (trace 7) resulted in a 44% current inhibition that was also voltage-dependent (Fig. 1D). The concentration-response curves for morphine (○) and DAMGO (•) are plotted in Fig. 1E. The data points for both agonists were fit to the Hill equation. The IC₅₀ (nM), pIC₅₀ (±SE), maximum current inhibition (%, ±SE), and Hill coefficient (±SE) obtained were 15.2 and 56.1, 1.82 ± 0.33 and 1.25 ± 0.13, 26.0 ± 2.5 and 38.7 ± 1.9, and 0.88 ± 0.61 and 2.43 ± 0.99 for morphine and DAMGO, respectively. The plots show that morphine exhibited just over a threefold greater potency than DAMGO. On the other hand, DAMGO had a higher efficacy than morphine, indicating the latter displayed partial agonist properties. It should be noted that not all SPG neurons tested exhibited coupling of agonist-activated MOR and Ca²⁺ channels. Thus we defined coupling of MOR with Ca²⁺ channels when the current inhibition was >15% and voltage-dependent during exposure to either morphine or DAMGO (both at 10 μM). Overall, ∼40% of all neurons examined with this agonist concentration fulfilled the criteria. Figure 1F shows the scatter plot with the median response of the Ca²⁺ current inhibition for the neurons that displayed coupling of MOR and Ca²⁺ channels following exposure to either 10 μM morphine or DAMGO.

Fig. 1.

[d-Ala2-N-Me-Phe4-Glycol5]-enkephalin (DAMGO) and morphine concentration-response relationships in rat sphenopalatine ganglion (SPG) neurons. A and C: time courses of Ca²⁺ current amplitude inhibition acquired from the sequential application of 0.003, 0.3, and 3 μM morphine (A) and 0.003, 0.03, and 10 μM morphine (B) in SPG neurons. The Ca²⁺ currents were evoked every 10 s with the double-pulse voltage protocol (shown in B, top) and current amplitude was measured isochronally 10 ms after initiation of both prepulse (•) and postpulse (○). B and D: superimposed current traces from the SPG neurons in A and C indicate those before and after agonist application. E: concentration-response curves in SPG neurons exposed to DAMGO (•) and morphine (○), respectively. Each point represents the mean (±SE) prepulse Ca²⁺ current inhibition. The smooth curves were obtained by fitting the points to the Hill equation. F: scatter plot showing the Ca²⁺ current inhibition with either 10 μM morphine or DAMGO of all neurons that displayed coupling between MOR and Ca²⁺ channels. —, the median; numbers in parentheses, the number of cells tested. G: summary graph showing mean (±SE) prepulse Ca²⁺ current inhibition in the presence of 10 μM DAMGO (control) and following NEM pretreatment with re-exposure to 10 μM DAMGO [DAMGO + N-ethylmaleimide (NEM)]. *, P < 0.0001 compared with control, paired t-test, numbers in parentheses indicate the number of cells.

MOR are well known to couple to effectors via pertussis toxin (PTX)-sensitive Gα_i/o protein subunits (Bailey and Connor 2005). Thus in the next set of experiments the coupling of MOR and N-type Ca²⁺ channels was examined. The use of PTX as a tool to determine Gα subunit coupling specificity was not possible because not all SPG neurons exhibited MOR expression, making negative results difficult to interpret. Therefore the sulfhydryl alkylating agent NEM was employed to uncouple Gα_i/o protein subunits acutely from MOR. This approach has been successfully employed in other neuron types (Jeong and Ikeda 1998; Jeong and Wurster 1997; Shapiro et al. 1994; Wollmuth et al. 1995). Initially, SPG neurons were briefly (<20 s) exposed to 10 μM DAMGO. If the voltage-dependent inhibition (>15%) of Ca²⁺ current was observed, then the cells were treated with NEM (100 μM) for 2 min. Thereafter, DAMGO was reapplied again. In all six SPG neurons that exhibited an initial response to DAMGO, treatment with NEM significantly (P < 0.0001) decreased the DAMGO-mediated Ca²⁺ current inhibition.

In a separate set of experiments, expression of other “classical” opioid receptors (i.e., δ and κ) and the opioid receptor-like 1 (NOP) receptor in SPG neurons was determined. Acutely isolated neurons were exposed to the κ opioid receptor agonist, U-50488 (10 μM), the δ receptor agonist, deltorphin II (2 μM), and NOP receptor agonist, nociceptin (10 μM). In 11–12 neurons tested, application of the three receptor agonists failed to elicit coupling between Ca²⁺ channels and their respective receptor subtype (data not shown). Therefore the observed effects of the opioid agonists described in this report were mediated via MOR activation.

Muscarinic acetylcholine receptor (mAChR) signaling in SPG neurons

The next set of experiments were undertaken to study the modulation of Ca²⁺ channels following the activation of M₂ mAChR with the mAChR agonist, Oxo-M. This is the sole mAChR subtype expressed in SPG neurons that is coupled to the voltage-dependent inhibition of Ca²⁺ currents (Liu et al. 2002; Margas and Ruiz-Velasco 2007). In the present study, we observed that both morphine- and DAMGO-mediated Ca²⁺ current inhibition was more robust when the external recording solution contained TEA (TEA external). However, it is well established that TEA-containing solutions cause a rightward shift in ACh and Oxo-M concentration-response relationships with regard to Ca²⁺ channel currents (Caulfield 1991). To circumvent this effect, Tris-containing solutions are typically employed (Caulfield et al. 1994; Mathie et al. 1992). Figure 2 illustrates results from experiments comparing the coupling of activated M₂ mAChR to Ca²⁺ channels in TEA- and Tris-containing external solutions. Figure 2, A and C, shows the time courses of peak Ca²⁺ current inhibition in SPG neurons before and after exposure to the mAChR agonist, Oxo-M, in Tris- and TEA-containing solutions, respectively. The traces shown in Fig. 2, B and D, correspond to the numbers for Tris- and TEA-containing solutions, respectively. As with DAMGO and morphine, application of 10 μM Oxo-M led to a voltage-dependent inhibition of the Ca²⁺ currents (trace 7, Fig. 2, B and D). Figure 2E is a plot that compares the concentration-response relationships for both solutions. The data points were also fit to the Hill equation as described in the preceding text. As expected, the presence of TEA in the external solution resulted in a rightward-shift (∼100 fold) of the Oxo-M IC₅₀ from 35.5 to 3250 nM. The maximum current inhibition and Hill coefficient values were 63.1 ± 5.2 and 58.1 ± 4.2, and 0.85 ± 0.15 and 0.95 ± 0.29%, respectively. It can be seen that the Oxo-M-mediated Ca²⁺ current inhibition in solutions containing TEA were lower when compared with Tris-containing solutions at concentrations ranging from 0.003 to 10 μM. The data recorded in TEA external solutions shows that exposure of 10 μM Oxo-M resulted in a 44.6 ± 3.3% inhibition of the Ca²⁺ currents, which was lower than that observed in Tris-containing solutions (58.6 ± 1.6%, Fig. 2E). Nevertheless, the effect of 10 μM Oxo-M obtained in TEA-containing solutions appeared rather robust (trace 7, Fig. 2D; trace 3, Fig. 4B; trace 3, Fig. 5B; trace 3, Fig. 6B) and the results described in the following text (unless specified) were obtained in the presence of TEA.

Fig. 2.

Oxo-M concentration-response relationships of rat SPG neurons. A and C: time courses of Ca²⁺ current amplitude inhibition acquired from the sequential application of 0.001, 0.01, and 10 μM in Tris-containing external solution (A) and 0.01, 1, and 10 μM Oxo-M in TEA-containing external solution (C). Ca²⁺ currents were evoked every 10 s with the double-pulse voltage protocol (shown in B, top) and current amplitude was measured isochronally 10 ms after initiation of both prepulse (•) and postpulse (○). B and D: superimposed current traces from the SPG neurons in A and C indicate those before and after Oxo-M application. E: concentration-response curves in SPG neurons exposed to Oxo-M in Tris (○)- and TEA-containing (•) external solutions, respectively. Each point represents the mean (±SE) prepulse Ca²⁺ current inhibition. The smooth curves were obtained by fitting the points to the Hill equation. F: summary graph showing mean (±SE) prepulse Ca²⁺ current inhibition mediated by 30 μM Oxo-M in control and PTX-treated (500 ng/ml) SPG neurons. *, P < 0.0001 compared with control, Student's t-test, numbers in parentheses indicate the number of experiments.

Fig. 4.

Oxo-M-mediated homologous desensitization of M₂ mAChR in MOR-expressing SPG neurons. A: time course of Ca²⁺ current amplitude inhibition acquired from the application of 10 μM morphine for 30 s, followed by a 10-min and 30-s exposure to 10 μM Oxo-M. The Ca²⁺ currents were evoked every 10 s with the double-pulse voltage protocol (shown in B, top) and the current amplitude was measured isochronally 10 ms after initiation of both prepulse (•) and postpulse (○). ■, application of morphine for 30 s, Oxo-M for 10 min and 30 s. The 2nd application of Oxo-M was performed following a 2-min interval. B: superimposed Ca²⁺ current traces from the neuron in A before (1 and 2) exposure to Oxo-M, 20 s (3 and 4) and 10 min (5 and 6) in the continued presence of Oxo-M, and 20 s (9 and 10) during the 2nd exposure to Oxo-M. C: summary graph showing the mean (±SE) Ca²⁺ current inhibition produced by Oxo-M at 20 s and 10 min and at 20 s during the 2nd application (re-exposure) of the agonist. Numbers in parenthesis indicate the number of experiments, *, P < 0.001 employing 1-way ANOVA followed by the Newman-Keuls test.

Fig. 5.

Oxo-M-mediated heterologous desensitization of MOR in SPG neurons. A: time course of Ca²⁺ current amplitude inhibition acquired from an initial application of 10 μM morphine for 30 s (Mor 1st), followed by a 10 min to 10 μM Oxo-M and a 30 s (Mor 2nd) exposure to morphine. The Ca²⁺ currents were evoked every 10 s with the double-pulse voltage protocol (shown in B, top) and the current amplitude was measured isochronally 10 ms after initiation of both prepulse (•) and postpulse (○). ■, application of morphine for 30 s (Mor 1st and Mor 2nd) and Oxo-M for 10 min. B: superimposed Ca²⁺ current traces from the neuron in A before exposure to morphine (1 and 2, 11 and 12) and Oxo-M (5 and 6), 20 s in the presence of morphine (3 and 4, 13 and 14), and 10 min in the continued presence of Oxo-M (9 and 10). C: summary graph showing the mean (±SE) Ca²⁺ current inhibition produced by the 1st and 2nd application of morphine and in the continued presence of Oxo-M at 20 s and 10 min. Numbers in parenthesis indicate the number of experiments, *, P < 0.02 employing paired t-test.

Fig. 6.

Heterologous desensitization of M₂ mAChR following prolonged MOR activation is time-dependent. A: time course of Ca²⁺ current amplitude inhibition acquired from an initial application of 10 μM Oxo-M for 30 s (Oxo-M 1^st), followed by a 10 min to 10 μM morphine and a 30 s (Oxo-M 2nd) exposure to Oxo-M. The Ca²⁺ currents were evoked every 10 s with the double-pulse voltage protocol (shown in B, top) and the current amplitude was measured isochronally 10 ms after initiation of both prepulse (•) and postpulse (○). ■, application of Oxo-M for 30 s (Oxo-M 1st and Oxo-M 2nd) and morphine for 10 min. B: superimposed Ca²⁺ current traces from the neuron in A before exposure to Oxo-M (1 and 2, 11 and 12) and morphine (5 and 6), 20 s in the presence of Oxo-M (3 and 4, 13 and 14), and 10 min in the continued presence of morphine (9 and 10). C and D: summary graphs showing the mean (±SE) Ca²⁺ current inhibition produced by the 1st and 2nd application of Oxo-M and in the continued presence of 10 μM morphine or DAMGO at 20 s and 10 min. Numbers in parenthesis indicate the number of experiments, *, P < 0.005 employing paired t-test.

Next the coupling specificity of M₂ mAChR with Gα protein subunits was examined in SPG neurons pretreated overnight with PTX. Figure 2F is a summary graph showing that PTX pretreatment significantly (P < 0.0001) decreased the Oxo-M (10 μM)-mediated Ca²⁺ current inhibition when compared with control neurons. Thus these results and those described in the preceding text (Fig. 1G) indicate that MOR and M₂ mAChR employ Gα_i/o protein subunits that are coupled to Ca²⁺ channels in SPG neurons.

Morphine- and Oxo-M-mediated homologous desensitization of MOR and M₂ mAChR, respectively, in SPG neurons

In the next set of experiments, we examined whether prolonged (i.e., 10 min) exposure of an opioid receptor agonist to SPG neurons would result in rapid desensitization of MOR. These experiments were conducted with morphine, one of the most widely employed opiate drugs. Figure 3A shows a time course of peak Ca²⁺ currents evoked every 10 s with the double-pulse protocol described in the preceding text. The numbers indicate the current amplitudes before morphine application (traces 1, 2, 7, and 8), 20 s (traces 3 and 4) and 10 min (traces 5 and 6) into the exposure of 10 μM morphine and 30 s (traces 9 and 10) during the second exposure to morphine. A 2-min recovery period between the first and second morphine application was allowed to elapse. Unless otherwise noted, this protocol was performed in all subsequent experiments. The plot in Fig. 3A shows that 20 s following the first morphine application, Ca²⁺ currents were inhibited ∼40% (see trace 3, Fig. 3B), whereas at the end of 10 min exposure, the current inhibition decreased to 15% (trace 5, Fig. 3B). Re-exposure to morphine following the 2-min recovery period resulted in a 28% inhibition of the Ca²⁺ currents (trace 9, Fig. 3B). The summary plot in Fig. 3C shows that within 10 min of morphine exposure, Ca²⁺ channel current was significantly lower than that observed at 20 s, representing a 75 ± 5% change of the response. In addition, the recovery period was insufficient to regain the full modulation of the current inhibition as the morphine-mediated Ca²⁺ current inhibition was also significantly lower (33 ± 6% change in response) during the second exposure when compared with the initial application at 20 s. Therefore the gradual decrease of coupling (i.e., homologous desensitization) between MOR and Ca²⁺ channels resulted from tonic MOR activation.

Fig. 3.

Morphine-mediated homologous desensitization of mu (μ) opioid receptors (MOR) in SPG neurons. A: time course of Ca²⁺ current amplitude inhibition acquired from the application of 10 μM morphine for 10 min and 30 s following a 2-min interval. The Ca²⁺ currents were evoked every 10 s with the double-pulse voltage protocol (shown in B, top) and the current amplitude was measured isochronally 10 ms after initiation of both prepulse (•) and postpulse (○). ■, application of morphine for 10 min and 30 s. The 2nd application of morphine was performed following a 2-min interval. B: superimposed Ca²⁺ current traces from the neuron in A before (1 and 2) exposure to morphine, 20 s (3 and 4) and 10 min (5 and 6) in the continued presence of morphine, and 20 s (9 and 10) during the 2nd exposure to morphine. C: summary graph showing the mean (±SE) Ca²⁺ current inhibition produced by morphine at 20 s and 10 min and at 20 s during the 2nd application (re-exposure) of the agonist. Numbers in parenthesis indicate the number of experiments, * and **, P < 0.001 and P < 0.05, respectively, employing 1-way ANOVA followed by the Newman-Keuls test.

A similar paradigm just described was followed in the next set of experiments aimed to determine whether chronic M₂ mAChR activation with Oxo-M (10 μM) would lead to acute receptor desensitization. In the following experiments, all neurons received a brief application (30 s) of 10 μM morphine to identify SPG neurons that coexpressed MOR and M₂ mAChR. The time course in Fig. 4A shows peak Ca²⁺ current evoked every 10 s with the double-pulse protocol (Fig. 4B, top) from an SPG neuron that was briefly exposed to morphine. Following the recovery period, Oxo-M was applied for 10 min. The initial (20 s) and prolonged (10 min) Oxo-M-mediated Ca²⁺ current inhibition were 54% (trace 3) and 26% (trace 5), respectively. The numbered Ca²⁺ current traces are shown to the right (Fig. 4B). After the 2-min recovery period, re-exposure to Oxo-M led to an inhibition of 55% (trace 9). As observed with morphine in the preceding text, the chronic exposure to Oxo-M also significantly (P < 0.05) decreased the Ca²⁺ current inhibition when compared with either the initial 20 s exposure to Oxo-M (resulting in a 53 ± 7% change) or during the second application of the agonist (Fig. 4C). On the other hand, unlike morphine in the preceding text, the 2-min recovery period was sufficiently long to allow almost a full recovery (a 14 ± 5% change in response) of the Oxo-M-mediated Ca²⁺ current inhibition when compared with that measured at 20 s. These results suggest that the 10 min exposure of either morphine or Oxo-M (both at 10 μM) will lead to homologous desensitization of MOR and M₂ mAChR, respectively, in SPG neurons.

Prolonged exposure of SPG neurons to Oxo-M leads to heterologous desensitization of MOR, whereas prolonged morphine or DAMGO application does not desensitize M₂ mAChR

In the next series of experiments, the functional interaction of M₂ mAChR and MOR was investigated by initially desensitizing either receptor with its respective agonist, followed by measuring the agonist-mediated Ca²⁺ current inhibition produced by the other agonist. Figure 5A is the time course recorded from an SPG neuron coexpressing both receptors. Morphine (10 μM) was first applied to the neuron and the Ca²⁺ current inhibition was measured at 20 s. In this cell, morphine caused a 40% inhibition of the Ca²⁺ currents (trace 3). Following a 2-min interval, Oxo-M (10 μM) was applied for 10 min. The initial (20 s, trace 7) Oxo-M-mediated Ca²⁺ current inhibition was 55% (trace 7), and the response at the end of the 10-min exposure was 14% (trace 9), indicating homologous desensitization of M₂ mAChR. Morphine was then reapplied after the 2-min recovery period, and Ca²⁺ currents were inhibited less (14%) than the initial morphine exposure (trace 13). The summary plot in Fig. 5C shows that mAChR desensitization led to significantly lower Oxo-M (P = 0.0015)- and morphine-mediated (P = 0.012) Ca²⁺ current inhibition. This corresponded to a 66 ± 13 and 32 ± 6% change in effect for Oxo-M and morphine, respectively. As a negative control, in one group of SPG neurons, (n = 6), morphine (10 μM) was applied briefly first, followed by a 12-min interval in which no agonist was applied, and then re-exposed to morphine. Under these conditions, the mean Ca²⁺ current inhibition were 22.9 ± 2.2 and 24.1 ± 3.2% during the first and second morphine application, respectively. Overall, these results suggest that acute prolonged exposure of SPG to Oxo-M leads to both homologous (e.g., M₂ mAChR) and heterologous (e.g., MOR) receptor desensitization.

Next, we wanted to determine whether the heterologous desensitization of MOR by Oxo-M was mediated specifically by M₂ mAChR. We have previously reported that exposure of SPG to the M₂ mAChR blocker, methoctramine, will remove the Oxo-M-mediated Ca²⁺ current inhibition (Margas and Ruiz-Velasco 2007). SPG neurons were first exposed to morphine (10 μM), and this was followed by a 5-min incubation with methoctramine (10 μM). Thereafter, the neurons were exposed to 10 μM Oxo-M and methoctramine for 10 min. Following a 2-min recovery period, morphine was reapplied. Under these conditions, the initial morphine-mediated Ca²⁺ current inhibition was 38.9 ± 7.3% (n = 4), whereas it was 31.0 ± 7.1% (n = 4), P = 0.095, after the exposure to Oxo-M plus methoctramine (data not shown), a 14% change of the response. In control experiments, a similar paradigm was performed except that after the initial morphine application, a 5-min recording period elapsed before the 10-min exposure to Oxo-M. Following the Oxo-M application and a 2-min recovery period, morphine was applied again. The Ca²⁺ current inhibition for the first and second morphine applications were 39.3 ± 5.2 and 24.9 ± 5% (n = 6), respectively, P = 0.003 (data not shown), a 40% change of the response. These results show that the block of M₂ mAChR prevented the Oxo-M-mediated heterologous desensitization of MOR.

Figure 6 shows experiments aimed to examine whether the acute desensitization of MOR would lead to heterologous desensitization of M₂ mAChR following a 2-min recovery period (as shown in Fig. 5A). The time course shown in Fig. 6A shows the peak Ca²⁺ current evoked every 10 s with the paradigm described in the preceding text (top, Fig. 6B). During the first Oxo-M application, the current was inhibited by 51% (trace 3). When the neuron was exposed next to morphine (10 μM) for 10 min, the inhibition of Ca²⁺ currents mediated by MOR was completely desensitized (cf. traces 5 and 9). However, after the 2-min recovery period, the re-application of Oxo-M lead to similar inhibition (50%) of the Ca²⁺ channel currents (trace 13). Figure 6C illustrates the summary plot of the Ca²⁺ current inhibition mediated by morphine and the twice-fold application of Oxo-M before and after opiate exposure. Although morphine caused a significant (P = 0.019) decrease in Ca²⁺ current inhibition following a 10-min exposure (a 58 ± 19% change in inhibition), no statistically significant effect (P = 0.064) was observed when comparing the first and second application of Oxo-M (a 16 ± 7% change of the inhibitory response). The failure of morphine to heterologously desensitize the Oxo-M-mediated response may have been a result of the inability of the partial agonist to induce MOR internalization. Thus similar experiments were performed with DAMGO, a compound known to internalize MOR, to determine whether heterologous desensitization would be observed following a 10-min incubation with the agonist. The summary plot in Fig. 6D shows that the Oxo-M-mediated Ca²⁺ current inhibition did not change greatly (P = 0.89), whereas a significant (P = 0.009) decrease in coupling of MOR and Ca²⁺ channels (a 58 ± 12% change in response) was observed following a 10-min DAMGO (10 μM) application. In the negative control group (n = 6), Oxo-M was applied twice with a 12-min interval in which no agonist was employed. Under these conditions, the Oxo-M-mediated Ca²⁺ current inhibition (%) during the first and second application were 32 ± 5 and 35 ± 6%, respectively. These results suggest that the absence of heterologous desensitization of M₂ mAChR was not a result of employing either a partial or strong MOR agonist.

Alternatively, the lack of heterologous desensitization may have been that the interval between the 10-min morphine application, and the re-exposure to Oxo-M was sufficiently long enough to resensitize the muscarinic receptors. Therefore in the next set of experiments, we measured the recovery from desensitization of M₂ mAChR in neurons that had been exposed initially to Oxo-M, followed by a 10-min morphine application and finally re-exposed to Oxo-M 30, 60, and 90 s after removal of morphine. Figure 6E shows a plot of the Oxo-M-mediated Ca²⁺ current inhibition during the first and second Oxo-M (10 μM) applications. When the time interval for Oxo-M re-exposure was shortened to 30 or 60 s, M₂ mAChR desensitization was observed. That is, the Oxo-M-mediated Ca²⁺ current inhibition was significantly lower during the second application at 30 s (P = 0.01) and 60 s (P = 0.001) with 16 ± 4 and 25 ± 5% changes in responses, respectively. When the recovery interval following morphine removal was 90 s, the Ca²⁺ current inhibition during the second Oxo-M application was also lower than the initial Oxo-M exposure but was not significantly different (P = 0.060). These results demonstrate that homologous desensitization of MOR with a 10-min morphine exposure does not appear to significantly alter the coupling of Ca²⁺ channels with M₂ mAChR when the recovery period is >90 s.

Immunofluorescence staining with MOR and M₂ mAChR antibodies

Finally, immunofluorescence assays were performed to image the cell surface expression of MOR and M₂ mAChR on SPG neurons. To obtain a better resolution, deconvolved fluorescence imaging was employed to collect images of section planes spanning the z axis. Figure 7 shows phase and fluorescence images of acutely isolated SPG neurons immunostained with antibodies specific for MOR (Fig. 7Aii) and M₂ mAChR (Bii). We have previously employed the latter antibody to detect M₂ mAChR expression in rat stellate ganglion neurons (Yang et al. 2006). The fluorescence images for both receptor subtypes show staining of MOR and M₂ mAChR was primarily observed along the cell membrane, suggesting that both receptors are located on the surface of SPG neurons. As mentioned in the preceding text, in ∼40% of SPG neurons application of morphine resulted in coupling of MOR to Ca²⁺ channels. Consequently, when the neurons were stained with the MOR antibody, not all cells displayed immunofluorescence to the secondary antibody (data not shown). On the other hand, M₂ mAChR appeared to be expressed in all immunostained neurons.

Fig. 7.

Immunofluorescence. Phase (Ai and Bi) and deconvolution fluorescence (Aii and Bii) images of adult rat SPG neurons. Primary rabbit antibody to MOR or mouse antibody to M₂ mAChR was employed and followed by Alexa Fluor 568-conjugated IgG secondary antibodies. Neurons were imaged at ×40 and sections were collected in 2-μm steps with a filter set containing an excitation filter at 540 ± 15 nm, a dichroic beam splitter of 585 nm (LP) and an emission filter at 620 ± 30 nm. Images were deconvolved and pseudocolored; scale bar represents 20 μm.

DISCUSSION

A clinical study that showed improvement in pain relief following the local application of an opiate analgesic to either the SPG or SCG (Spacek et al. 1997) prompted us to investigate signaling events that couple opioid receptors with Ca²⁺ channels in acutely isolated rat SPG neurons. Of the three classical opioid receptor subtypes, our results show that MOR are expressed in SPG neurons because exposure to morphine and DAMGO inhibited Ca²⁺ channel currents in ∼40–45% of all cells tested. Although morphine and DAMGO can activate κ and δ opioid receptors (Williams et al. 2001), it is likely that the agonist-mediated Ca²⁺ current inhibition occurred as a result of MOR activation because the κ and δ receptor agonists, U-50488 and deltorphin II, respectively, failed to effect the Ca²⁺ currents. The results of concentration-response relationships measured for both morphine and DAMGO were also consistent with morphine's profile as a weak partial agonist (Connor et al. 2004; Williams et al. 2001) when compared with latter. However, morphine did exhibit a threefold leftward shift in IC₅₀ value when compared with DAMGO. The expression of the opioid receptor-like 1 (NOP or ORL1) opioid receptor in SPG neurons was also tested, and no evidence was obtained supporting the coupling of this opioid receptor subtype to Ca²⁺ channels. Thus these findings indicate that the μ opioid receptor subtype plays an important role in SPG neuron function.

Previously it has been shown that of the five mAChR subtypes, in SPG neurons only, the M₂ mAChR couples to Ca²⁺ channels (Liu et al. 2002; Margas and Ruiz-Velasco 2007). In SPG neurons, M₂ mAChR have been shown to modulate the NO release and NO-mediated vasodilation via N-type Ca²⁺ channel inhibition (Lee et al. 2001). In the present study, the interaction of MOR and M₂ mAChR was also examined. Under our recording conditions, a robust inhibition of Ca²⁺ channel currents by morphine or DAMGO (both at 10 μM) was observed in the presence of TEA-containing solution. As expected from previous work (Caulfield 1991), the presence of TEA shifted the Oxo-M concentration response relationship rightward. Nevertheless the concentration of Oxo-M (10 μM) employed in this study was sufficient to obtain approximately a 50% inhibition of the Ca²⁺ currents, a robust value near that measured in Tris-containing solutions. Employing these recording conditions, the coupling of the specific Gα protein subunits to MOR and M₂ mAChR was shown to occur via Gα_i/o subunits because pretreatment with NEM or PTX resulted in a loss of coupling for each receptor, respectively. Rat intracardiac parasympathetic neurons have been shown to employ similar signaling components with regard to coupling of Ca²⁺ channels and MOR (Adams and Trequattrini 1998), M₂ mAChR (Jeong and Wurster 1997), and M₄ mAChR (Cuevas and Adams 1997). Thus in rat SPG neurons MOR and M₂ mAChR activate a common signal transduction pathway to effect voltage-dependent inhibition of Ca²⁺ channel currents.

The effect of prolonged exposure of morphine and Oxo-M on SPG Ca²⁺ channel currents was also investigated in this study. Our results show that in the continued (10 min) exposure of morphine there was a significant decrease in Ca²⁺ channel inhibition (i.e., homologous desensitization) when compared with the initial (20 s) effect of the agonist. Furthermore, when the cells were allowed to recover for 2 min, the morphine-mediated Ca²⁺ current inhibition was also significantly lower than that observed at 20 s. That is, a partial recovery (∼65%) of coupling between MOR and Ca²⁺ channels was evident after the recovery period. Although the ability of morphine to cause MOR internalization is known to be poor, it shows a tendency to cause receptor desensitization and eventual tolerance (Arttamangkul et al. 2008; Borgland et al. 2003; Connor et al. 2004; Dang and Williams 2005; Samoriski and Gross 2000; Yu et al. 2009). The incomplete recovery may be explained by the receptor activity versus endocytosis (RAVE) model, which proposes that in the continued presence of morphine, cells are capable of eliciting adaptive mechanisms that oppose the actions of the activated MOR (Waldhover et al. 2004; Whistler et al. 1999). Our observations are consistent with this theory, and it thus seems very likely MOR were not internalized but rather remained on the surface of the cell membrane. Nevertheless, MOR endocytosis cannot be ruled out because studies employing epitope-tagged receptors heterologously expressed in nucleus accumbens and striatal neurons (Haberstock-Debic et al. 2003, 2005) have provided evidence suggesting that morphine can initiate receptor internalization. However, the rapid recovery of coupling between MOR and Ca²⁺ channels that we observed supports the RAVE model. Ultimately, it will be essential to examine MOR trafficking in the continued presence of morphine to attempt to differentiate the mechanism responsible.

Similar to morphine, the 10-min exposure of SPG neurons to Oxo-M resulted in homologous desensitization of M₂ mAChR. Yet, the 2-min recovery period was sufficient for almost a full recovery (∼86%) of the M₂ mAChR and Ca²⁺ channel coupling observed during the first application of Oxo-M. The rapid reversibility of M₂ mAChR-mediated signaling has been previously described to be a consequence of rapid receptor internalization, followed by a slow recovery period that requires protein synthesis-dependent and -independent components (Rosenberry and Hosey 1999). In that study, M₂ mAChR were heterologously expressed in HEK cells and then exposed for 10 or 30 min to the mAChR agonist carbachol. The recovery of mAChR to the surface was measured with binding assays performed 30 min to 6 h after agonist exposure (Rosenberry and Hosey 1999). Unlike these observations, our results suggest that in SPG neurons, M₂ mAChR are capable of undergoing rapid desensitization as well as rapid (i.e., within 2 min) resensitization. Whether the recovery of M₂ mAChR is ligand-dependent (i.e., carbachol vs. Oxo-M) remains to be determined.

Our results, showing that the acute prolonged activation of M₂ mAChR significantly decreased the MOR-mediated Ca²⁺ current inhibition (heterologous desensitization), demonstrate another G-protein-coupled receptor that is capable of modulating MOR. For example, recently it was shown that neurokinin I receptor (NK1R) activation in primary striatal and amygdala neurons resulted in both opioid-mediated endocytosis and desensitization (Yu et al. 2009). The authors found that these processes were dependent on the sequestration of β-arrestin 2. Similarly, Förster resonance energy transfer techniques were employed to examine signaling between MOR and α_2A-adrenergic receptors heterologously expressed in HEK 293 cells. The MOR- α_2A-adrenergic heterodimers were shown to effect the signaling of each receptor type (Vilardaga et al. 2008). In the present study, prolonged activation of M₂ mAChR affected MOR coupling with Ca²⁺ channels. On the other hand, chronic MOR stimulation with either morphine (partial agonist) or DAMGO (full agonist) did not lead to heterologous desensitization of M₂ mAChR unless the recovery period was <60 s. Whether the effects of these agonists occur at the cell membrane (i.e., morphine) or intracellularly via internalization (i.e., DAMGO), our results suggest that coupling of M₂ mAChR with Ca²⁺ channels is dependent on recovery from resensitization. Given that ACh is the principal neurotransmitter of parasympathetic neurons, it may be crucial for signaling via mAChR to remain intact and unimpeded from the influence of other G-protein-coupled receptors such as MOR. It is possible that M₂ mAChR employ multiple signaling pathways that serve as “reserves” in the event that another G-protein-coupled receptor is tonically activated. Our results do show, however, that MOR can exert a small influence on M₂ mAChR function because Oxo-M-mediated Ca²⁺ current inhibition did not fully return to control levels following the 10-min morphine exposure.

The release of ACh and NO by SPG is essential in maintaining cerebral blood flow. However, ACh released from SPG neurons can have untoward effects such as plasma protein extravasation in the dura mater and may lead to the release of proinflammatory and chemotactic proteins. This was demonstrated in a study that observed the mAChR-mediated neurogenic inflammation following electrical stimulation of rat SPG (Delepine and Aubineau 1997). Moreover, some forms of migraine and facial pain are associated with an enhanced parasympathetic activity. Sensory neurons such as the trigeminal ganglion (TG) also play a role in controlling cortical blood flow. Electrical stimulation of both SPG and TG neurons has been shown to increase cerebral blood flow via release of NO and CGRP by these neurons, respectively (Ayajiki et al. 2005). SPG and TG blocks are performed to treat several forms of head and face neuralgias. One study found that intranasal application of lidocaine in patients with migraine resulted in relief of pain intensity, yet the cutaneous allodynia remained (Yarnitsky et al. 2003). The authors concluded that the increased parasympathetic activity sensitized peripheral and/or central nociceptive neurons but that the allodynia was maintained by central nociceptors. In anesthetized rats, dural vasodilation mediated by either electrical stimulation of the dura mater or calcitonin gene-related peptide (CGRP) was blocked by morphine or DAMGO (Williamson et al. 2001). These results suggested that opioids influence the trigeminal nucleus caudalis peripherally by acting on MOR located on trigeminal fibers and inhibiting CGRP release. Evidence that MOR are expressed in TG neurons has been previously provided (Borgland et al. 2001). With these results and our findings, it is tempting to speculate on the potential advantages of MOR expression in SPG with regard to the treatment of migraine and other facial neuralgias. First, the local application of opiates would exclude systemic or toxic effects that can occur with other routes of administration. Second, the actions of administered opiates would most likely remain local with minimal risk for tolerance and addiction. Third, MOR activation would not significantly perturb the mAChR signaling pathway in SPG neurons and potentially diminish the release of other vasoactive compounds to limit their contribution to the inflammatory response.

In conclusion, we have identified a subpopulation of acutely isolated rat SPG neurons that express MOR. Activation of MOR by either morphine or DAMGO and M₂ mAChR by Oxo-M led to voltage-dependent inhibition of Ca²⁺ channel currents via Gα_i/o subunits. The acute prolonged activation of MOR with either morphine or DAMGO and M₂ mAChR with Oxo-M resulted in homologous desensitization of the receptors. Further, acute activation of M₂ mAChR led to heterologous desensitization of MOR, but the opposite was not observed. These results suggest that MOR expressed in human SPG neurons can serve as therapeutic targets in patients suffering from some forms of migraine and facial pain.

GRANTS

This study was supported by National Institute of Health Grants HL-074311 and DA-025574 to V. Ruiz-Velasco.