Activation of a Gq-coupled membrane estrogen receptor rapidly attenuates α₂-adrenoceptor-induced antinociception via an ERK I/II-dependent, non-genomic mechanism in the female rat

Abstract

Though sex differences in pain and analgesia are known, underlying mechanisms remain elusive. This study addresses the selective contribution of membrane estrogen receptors (mER) and mER-initiated non-genomic signaling mechanisms in our previously reported estrogen-induced attenuation of α₂- adrenoceptor-mediated antinociception. By selectively targeting spinal mERs in ovariectomized female rats using E2BSA (membrane impermeant estradiol analogue), and ERα selective agonist PPT, ERβ selective agonist DPN, GPR 30 agonist G1 and Gq-coupled mER (Gq-mER) agonist STX, we provide strong evidence that Gq-mER activation may solely contribute to suppressing clonidine (an α₂- adrenoceptor agonist)-induced antinociception, using the nociceptive tail flick test. Increased tail flick latencies (TFL) by intrathecal (i.t.) clonidine were not significantly altered by i.t. PPT, DPN, or G1. In contrast, E2BSA or STX rapidly and dose-dependently attenuated clonidine-induced increase in TFL. ICI 182,780, the ER antagonist, blocked this effect. Consistent with findings with the lack of effect of ERα and ERβ agonists that modulate receptor-regulated transcription, inhibition of de novo protein synthesis using anisomycin also failed to alter the effect of E2BSA or STX, arguing against a contribution of genomic mechanisms. Immunoblotting of spinal tissue revealed that mER activation increased levels of phosphorylated extracellular signal regulated kinase (ERK) but not of protein kinase A (PKA) or C (PKC). In vivo inhibition of ERK with U0126 blocked the effect of STX and restored clonidine antinociception. Although estrogen-induced delayed genomic mechanisms may still exist, data presented here indicate that Gq-mER may solely mediate estradiol-induced attenuation of clonidine antinociception via a rapid, reversible, and ERK-dependent, non-genomic mechanism, suggesting that Gq-mER blockade might provide improved analgesia in females.

Numerous studies have reported sex-related differences in the perception of pain and analgesic response to drugs highlighting higher suffering and prevalence of several pain disorders in women, e.g. temporomandibular joint disorder (TMD), migraines, fibromyalgia, rheumatoid arthritis, irritable bowel syndrome, etc. (Fillingim and Maixner, 1995; Unruh, 1996; Berkley, 1997; LeResche, 1997; Lipton et al., 2001; LeResche et al., 2003; Fillingim and Gear, 2004; Wiesenfeld-Hallin, 2005; Greenspan et al., 2007; Fillingim et al., 2009; Ruau et al., 2012). However, mechanisms underlying these sex differences have not been clearly understood. While psychosocial factors may exacerbate pain in women, important biological mechanisms, e.g. pain facilitating effect of the gonadal hormone, estrogen is reported in several studies (Craft, 2007; Tang et al., 2008; Ji et al., 2011). We and others have shown that estrogen attenuates G-protein-coupled receptor (GPCR)-mediated antinociception (Flores et al., 2001; Stoffel et al., 2005; Claiborne et al., 2006; Nag and Mokha, 2004; 2006; Thompson et al., 2008). α₂-Adrenoceptor, a GPCR, is present in the spinal dorsal horn and mediates descending inhibition of pain arising from noradrenergic pontine nuclei A5, A6 (Locus coeruleus) and A7 (Köliker Fuse) (Westlund et al., 1983; Jones and Gebhart, 1986; Fairbanks et al., 2009). Among three receptor subtypes, α_{2A, 2B} and _2C, the α_2A subtype predominantly mediates the antinociceptive effect of α₂-adrenoceptor agonists, e.g. clonidine (Lakhlani et al., 1997; Stone et al., 1997; Wang et al., 2002).

We previously reported that antinociception produced by selective activation of the α₂-adrenoceptor is sex-specific and attenuated by estradiol injected subcutaneously, 48 hr before nociceptive testing in ovariectomized (OVX) rats or in females at proestrous stage (with highest endogenous estradiol level) of the estrous cycle (Nag and Mokha, 2006; Thompson et al., 2008). There are four known estrogen receptors - ERα, ERβ, G-protein coupled estrogen receptor 30 (GPR30; also known as GPER), and Gqcoupled membrane estrogen receptor (Gq-mER). Classical cytosolic ERα and ERβ, upon activation, translocate to the nucleus and act as transcription factors to induce genomic changes. Recent evidence indicates rapid, non-genomic effects of post transcriptionally modified membrane bound ERα and ERβ (Levin, 2009; Micevych and Dominguez, 2009). GPR30 (Revankar et al., 2005; Thomas et al., 2005) and a recently discovered Gq-mER (Qiu et al., 2003) are membrane bound receptors. Selective contribution of these membrane estrogen receptors (mER) to estradiol-induced attenuation of the α₂-adrenoceptor-mediated antinociception is not known. Hence, this study targeted spinal mERs using intrathecal administration of E2BSA (a membrane impermeable estradiol derivative), and PPT, DPN, G1 and STX (selective agonists at ERα, ERβ, GPR30 and Gq-mER respectively). In addition, since protein kinases C (PKC), A (PKA), and extracellular signal regulated kinase I,II (ERKI/II) constitute intracellular signaling that impairs potassium channel function leading to increased neuronal excitability (Ji and Woolf, 2001; Hu and Gereau, 2003; Hu et al., 2003), we investigated whether the effect of mER activation on clonidine-induced antinociception was dependent on increased spinal activation of PKA, PKC, and/or ERKI/II, and/or on genomic mechanisms requiring de novo protein synthesis.

1. Experimental Procedures

1.1 Subjects

Sprague–Dawley OVX female rats (250–274 g; 3–4 months old; Harlan, Indianapolis, IN) were housed in the animal care facility at Meharry Medical College certified by the American Association for the Accreditation of Laboratory Animal Care (AAALAC) under a 12-h light/dark cycle (lights on: 7:00 AM) and had free access to food and water. Acclimation period was a minimum of 5 days. All experimental protocols were approved by the Institutional Animal Care and Use Committee of Meharry Medical College and conformed to guidelines established by the National Research Council Guide for the Care and Use of Laboratory Animals and the International Association for the Study of Pain (IASP). We made all efforts to minimize the number of rats and the level of stress in this study.

1.2 Intrathecal cannulation

At least two weeks following ovariectomy, a stretched PE-10 cannula (Intramedic, Sparks, MD; dead volume 10 µl) was implanted under ketamine and xylazine anesthesia (72 and 4 mg/kg i.p., respectively) as described (Yaksh and Rudy, 1976; Lawson et al., 2010). Briefly, the head of the rat was shaved and secured in a stereotaxic frame (David Kopf, Tujunga, CA), skin above the head/neck region incised and the atlanto-occipital membrane was cleared to expose the dura. The cannula was inserted into the subarachnoid space through a small slit in the dura and gently pushed to a length of 8.5 cm to reach the lumbosacral enlargement. After securing the cannula to the skull with dental cement, the wound was sutured and the animal was placed on a heating blanket until it regained consciousness. Animals recovered for at least 7 days before undergoing nociceptive testing. Animals were killed with sodium pentobarbital or Beuthanasia (150 mg/kg; i.p.; Schering-Plough, Union, NJ) at the end of testing or when showing any sign of neurological impairment after surgery (<10%). Position of the cannula was confirmed functionally by administering 10 µl of 2% lidocaine (i.t.) which temporarily paralyzed the animal’s hind limbs and anatomically by injecting Chicago sky blue dye (Sigma) and accessing dye spread on the dorsal lumbar spinal cord.

1.3 Tail flick test

Nociceptive heat-induced tail-flick assay was conducted using a tail-flick analgesia meter (Model 33T, IITC Life Science, Woodland Hills, CA) which automatically measured tail flick latency (TFL). Since this nociceptive test has been utilized by several investigators in the pain field, we decided to choose the same for better comparison of data obtained in this study. A trigger temperature was used to bring tail temperature of all rats to 32°C before intense heating began for consistency since tail temperature may fluctuate between individual rats and with minor variations in the room temperature and time of testing. A cut off time of 15 sec was set to prevent tissue damage. Animals were loosely restrained in a Plexiglas cylinder and habituated for at least 15 min. The radiant heat stimulus was applied 3–7 cm from the tip of the rat’s tail at three separate spots (~ 1 cm apart) to prevent sensitization by repeated stimulation. Determination of main drug effects was carried out using coded drug solutions to which the experimenter was blind; further, animals were loosely restrained in Plexiglas tubes covered with dark cloth and TFL was recorded automatically to avoid experimenter bias.

1.4 Drugs

Clonidine (1.75 µg; Sigma, St. Louis, MO; Nag and Mokha 2006) or vehicle was injected in separate groups at time ‘0’. E2BSA, a membrane impermeable estradiol analogue (0.09, 0.37 nM, and 0.5 mM corresponding respectively to 0.9, 3.7 fmoles, and 5 nmoles / 10 µl; Sigma), 0.1 µM – 2 mM (or 1 pmole - 20 nmoles in 10 µl) propylpyrazole-triol (PPT; an ERα agonist; Tocris, Minneapolis, MN), diarylpropionitrile (DPN; an ERβ agonist; Tocris), G1 (a GPR30 agonist; Tocris) or STX (a Gq-mER agonist provided by Dr. MJ Kelly) was injected 5 min before clonidine. Proestrous (0.37 nM) and diestrous (0.09 nM) equivalent doses of estradiol (in E2BSA) were calculated from previously reported plasma levels of estradiol in normally cycling rats during proestrous (~100 pg/mL) and diestrous stages (~25 pg/ml; Butcher et al., 1974; Claiborne et al., 2006). These w/v concentrations convert directly to the molar solutions used in the present study using 272.39 as the formula weight of E2BSA. Molarities used are of E2 in E2BSA. The highest dose of E2BSA (0.5 mM) is in fact a commonly used dose of estradiol and was derived from previously reported intrathecally applied estradiol that enhanced baroreflex function and autonomic tone and evoked action potentials in spinal neurons (Saleh et al.,2000; Zhang et al., 2012). All four estrogen receptor subtype selective agonists used in this study have several fold higher potency towards their target receptor subtype as compared to other estrogen receptor subtypes. EC50 value of PPT for ERα is ~ 0.2 nM, binding affinity is 400-fold higher than that for ERβ and it is devoid of activity on ERβ in gene transcription assays (Stauffer et al., 2000). EC50 vale of DPN for ERβ is 0.85 nM as compared to 66 nM for ERα (78 fold higher potency; Meyers et al., 2001). EC 50 value of G1 is 2 nM for GPR30 and it is devoid of activity at ERα and ERβ up to 10 µM (Bologa et al., 2006). Similarly, EC50 value of STX is 2.6 nM for Gq-mER with a 17 fold higher potency than estradiol (EC50 = 46 nM) and it is fully efficacious in ERα and/or β knockout mice (Qiu et al., 2006) as well as in GPR30 siRNA-treated neurons (Kenealy et al., 2011). It is noteworthy that while above EC50 values are determined through in vitro experimentation, effective doses of various agonists delivered systemically in vivo are much higher and are also higher than those employed in the current study (up to 20 nmoles / agonist); e.g. up to ~13 µmoles PPT / day for 8 days (Harris et al., 2002) and ~6.4 µmoles STX /day for 7 days (Smith et al., 2013). ICI 182,780 (an ER antagonist; 10 µM; Tocris) was co-injected with E2BSA or STX. Anisomycin (a protein synthesis inhibitor; 125 µg; Sigma; Miletic et al., 2010) was injected 15 min before clonidine. U0126 (20 µg/10 µl; Wang et al., 2011) that inhibits ERK I/II via MEK inhibition, or U0124, an inactive analogue of U0126 was injected 30 min before clonidine. PPT and DPN were dissolved in ethanol; G1, STX, anisomycin, and ERK I/II inhibitors in DMSO solution but the diluted working solutions contained <1% ethanol or <10% DMSO. E2BSA solution was prepared in phosphate buffered saline and was filtered using centrifuge filters to separate free E2 from bound before use as described (Stevis et al., 1999). Each drug was administered in a 10 µl volume except for the co-administration of PPT+DPN+G1 which were prepared as a cocktail in 10 µl. We have conducted pilot experiments with intrathecal injection of 40 µl saline and did not observe any volume related effects on the TFLs for up to 90 min. Appropriate vehicles were used as controls.

1.5 Immunoblotting

Lumbosacral spinal cord of anesthetized (150 mg/kg pentobarbital) rats were harvested ~10 min following in vivo i.t. clonidine, E2BSA or STX treatment and quick confirmation of drug effects on TFLs. Tissues were stored in 0.5 ml RNAlater (Ambion, Austin, TX) at −80°C till further analysis. Tissue homogenates were prepared in 0.5 ml RIPA lysis buffer (Santa Cruz Biotech, Dallas, TX) containing trisbuffered saline (TBS), 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 0.004% sodium azide. PMSF, sodium orthovanadate and protease inhibitor cocktail was added to RIPA (10 µl/ml) immediately before use. Total protein contents were analyzed using a Lowry assay (Lowry et al., 1951) based DC reagent kit (Bio-Rad, Hercules, CA). SDS PAGE was run using NuPAGE gel system (Life Technologies, Grand Island, NY); samples were prepared as per manufacturer’s instructions, heated at 65°C for 10 min and loaded (40 µg) onto the gel. Proteins were transferred onto nylon membrane and processed for immunoblotting using selective primary antibodies against PKA, pPKA (Upstate, Lake Placid, NY), PKC, pPKC (Pierce, Rockford, IL), ERK I/II, pERK I/II (Cell Signaling Technology Inc., Danvers, MA) and actin (Sigma). All incubations were carried out in closed containers on Belly Dancer orbital shakers (Stovall, Greensboro, NC). Blots were first blocked with 5% non-fat dairy milk in tris-buffered saline containing 0.05% Tween 20 (TBST; Santa Cruz) for 1 hr, then incubated with primary antibody for 12–48 hr at 4°C. After washing in TBST (6×5 min), blots were incubated in compatible HRP-tagged secondary antibody (bovine anti-rabbit IgG-HRP, Santa Cruz) for 1 hr, washed and reacted with SuperSignal West Dura Chemiluminescent Substrate (Pierce) for 5 min. Immunopositive bands were visualized with Gel Doc System (UVP, LLC, Upland, CA) and images stored for densitometry analysis using LabWorks 4.6 (UVP) software. The data was normalized against actin and presented as normalized phospho-protein/total protein.

1.6 Data analysis

Behavioral data were analyzed by ANOVA with appropriate within (Time) and between subject (Group) factors, corrected for repeated measures, using SPSS (SPSS Inc., Chicago, IL). The Fisher’s post-hoc test was used for intergroup comparisons only where ANOVA yielded a significant main effect. Area under the curve (AUC) was calculated by trapezoid method using Prism (Graphpad Software, Inc., San Diego, CA) on time course plots to obtain a single measure of the overall drug response and was analyzed using ANOVA and Fisher’s post-hoc test. Densitometry data were analyzed similarly. Significance level (p) was set at 0.05. Data are plotted as mean±SEM using SigmaPlot (Systat Software, San Jose, CA).

2. Results

2.1 Basal nociception was comparable across all groups

Figures 1A–4A and 6A reveal that basal nociception did not significantly differ between any of the groups studied. Thus, the baseline TFL before i.t. clonidine administration in all groups employed in the behavioral part of this study were comparable, and a repeated measures ANOVA did not yield significant main effects of Group [F_(23-61)=1.54; p=0.10] or Time [F_(2,122)=2.11; p=0.13].

Figure 1.

Spinal activation of mERs rapidly attenuates α₂-adrenoceptor-mediated antinociception and ER antagonist (ICI 182,780) blocks it. A, Intrathecal injection of clonidine induced significant increase in TFLs that persisted through the 90 min time course as compared to baselines (veh + clon). This increase was significantly attenuated by a high (0.5 mM) and a proestrous-equivalent (0.37 nM) dose of E2BSA but not by a diestrous-equivalent dose (0.09 nM). ICI 182,780 abolished the effect of E2BSA while E2BSA by itself did not affect TFLs. Yohimbine abolished the effect of clonidine. B, AUCs analysis of the data revealed significantly decreased AUCs in 0.5 mM and 0.37 nM E2BSA-treated groups but not in 0.09 nM E2BSA or ICI-treated groups. [*p < 0.01 compared to veh + clon group, #p < 0.01 compared to higher doses of E2BSA (0.5 mM and 0.37 nM)]. This data provides evidence of mER-induced rapid, presumably non-genomic, mechanism that leads to attenuation of clonidine antinociception.

Figure 4.

E2BSA or STX-induced attenuation of clonidine antinociception does not respond to blockade of de-novo protein synthesis. A, Anisomycin, a protein synthesis inhibitor, pretreatment did not alter E2BSA or STX-induced attenuation of clonidine antinociception. Further, anisomycin injection alone also did not change TFLs. B, AUC of all anisomycin-treated groups, with or without mER activation, remained significantly lower as compared to the clonidine only control group. (*p < 0.01). This data provides firm evidence in support of our hypothesis that mER activation-induced genomic changes are not required to attenuate clonidine antinociception.

Figure 6.

In vivo disruption of ERK I/II signaling allows clonidine antinociception to resurface following Gq-mER activation. A, Intrathecal administration of U0126, an ERK I/II phosphorylation inhibitor, resulted in restoration of clonidine-induced increase in the TFL following STX injection which was indistinguishable from that of clonidine only control group. U0124, an inactive analogue of U0126, failed to block STX-induced attenuation of clonidine antinociception; and U0126 treatment alone did not significantly change TFLs. B., STX failed to reduce clonidine-induced increase in AUC in animals with U0126 pretreatment. However, it successfully reduced AUC in animals with U0124 pretreatment. C, intrathecal co-administration of PPT, DPN, and G1 failed to change whereas STX significantly increased pERK I/II levels. Further, U0126 significantly reduced spinal level of pERK I/II. D, preadsorption of ERK I/II and pERK I/II antibodies with respective control peptides almost completely eliminated the staining. (*p < 0.01 compared to clonidine only control group). This data provides strong evidence that Gq-mER activation induced rapid attenuation of clonidine antinociception is spinal ERK I/II signaling-dependent.

2.2 E2BSA-induced activation of mERs rapidly and dose-dependently attenuates α₂-adrenoceptor-mediated antinociception

To assess the impact of mER-mediated effects on α₂-adrenoceptor antinociception, we examined the effect of i.t. E2BSA administration on clonidine-induced antinociception using the tail flick test. ANOVA yielded significant main effects of Group [F_(6,21)=51.14; p < 0.01], Time [F_(12,252)=38.69; p < 0.01] and an interaction between Group × Time [F_(72,252)=8.10; p < 0.01]. This interaction stemmed from the fact that after three comparable baselines at first three time points, veh+clon, 0.09 nM E2BSA+clon and ICI-treated groups had significantly higher TFLs as compared to the rest of the groups though the rest of the time course. Post hoc comparisons revealed that intrathecal injection of clonidine significantly increased TFL as compared to baselines at time 5 min through the rest of the 90 min time course [n=3; p < 0.05; Figure 1A]. E2BSA at 0.5 mM or 0.37 nM (equivalent to high estradiol concentration at the proestrous stage of the estrous cycle) doses rapidly attenuated clonidine-induced increase in TFL starting at the 5 min time point and this effect persisted through the 90 min time course (n=5/group; p < 0.05; Figure 1A). In contrast, a lower dose of E2BSA (0.09 nM, equivalent to low estradiol concentration at the diestrous stage) did not significantly change clonidine-induced increase in TFLs (n=5).

The effect of E2BSA (0.5 mM) was blocked by the broad spectrum ER antagonist, ICI 182,780 and restored clonidine antinociception (n=3; p < 0.05). Similarly, the effect of clonidine was fully blocked by yohimbine, an α₂-adrenoceptor blocker (n=4; p < 0.05). E2BSA (0.5 mM; n=3) itself did not produce a pronociceptive effect as indicated by no significant change in TFLs as compared to baselines. AUC analysis further confirmed these effects [F_(6,27)=36.62; p < 0.01]: AUCs of 0.5 mM or 0.37 nM E2BSA + clonidine (clon), 0.5 mM E2BSA + vehicle (veh), and yohimbine (yo) + clon groups were significantly lower than those of veh + clon, ICI 182,780 + E2BSA + clon, and 0.09 mM E2BSA + clon groups (p < 0.01; Figure 1B). In a separate experiment, we tested the effect of intrathecally administered membrane permeable estradiol on clonidine antinociception which produced an effect identical to E2BSA shown in Figure 1 (data not shown). These data are consistent with the interpretation that E2BSA-induced spinal activation of mERs leads to rapid attenuation of α₂-adrenoceptor-mediated antinociception and that E2BSA, by itself, does not produce any pro-nociceptive effects.

2.3 Activation of Gq-mER, but not ERα, ERβ, or GPR30 attenuates α₂-adrenoceptor-mediated antinociception

To determine the contribution of individual mER subtypes to the observed effect of E2BSA on clonidine antinociception, we exploited a variety of mER-selective agonists. Selective activation of the Gq-mER using STX significantly and dose-dependently attenuated clonidine-induced increase in TFL (Figure 2A). ANOVA yielded significant main effects of Group [F_(3,9)=32.66; p < 0.01], Time [F_(12,108)=23.22; p < 0.01] and an interaction between Group × Time [F_(36,108)=9.74; p < 0.01]. This interaction stemmed from the fact that after three comparable baselines at first three time points, veh+clon-treated group had the highest TFLs followed by 1 mM STX + clon -treated group with no significant increase in TFLs of 2 mM STX-treated groups. Post hoc comparisons revealed that at 2 mM dose, STX rapidly abolished clonidine-induced increase in TFL starting at time point 5 min and this effect persisted through the 90 min time course (n=4; p < 0.02). However, at 1 mM dose, STX partially reduced the effect of clonidine (n=3) with TFL significantly lower than veh + clon group at time points 15–25, 45 and 60 min whereas significantly higher than 2 mM STX + clon group at time points 5–20 and 30 min (p < 0.05). Injection of 2 mM STX alone did not affect the TFL as compared to the baselines (n=3; Figure 2A). AUC analysis further confirmed these impressions (Figure 2B). AUCs of STX-treated groups were significantly decreased in a dose-dependent manner [F_(3,12)=13.69; p < 0.01] with AUC of 2 mM STX-treated group significantly lower (p < 0.05) than those of 1 mM STX-treated and clonidine only control groups.

Figure 2.

Selective activation of the Gq-mER rapidly, and dose-dependently attenuates clonidine antinociception. A, Intrathecal pretreatment with STX, a Gq-mER selective agonist, at 1 mM and 2 mM doses significantly and progressively attenuated clonidine-induced increase in TFLs. By itself, 2 mM STX did not change TFLs. B, AUCs of all STX-treated groups were significantly lower as compared to the clonidine only control group. (*p < 0.05 compared to veh + clon group). This data provides evidence of an exclusive role played by the Gq-mER in rapidly attenuating clonidine antinociception.

In contrast, selective activation of ERα, ERβ or GPR30 using PPT, DPN or G1, respectively, failed to attenuate clonidine-induced increase in TFL (Figure 3A). ANOVA yielded significant main effect of only Time [F_(12,132)=62.60; p < 0.01], but not the Group [F(4,11) =1.92; p=0.18]. The interaction between Group × Time [F_(48,132)<1; p=0.70] also failed to reach significance. Data from only the highest dose (2 mM; n=3/group) of ER agonists are shown; lower doses in nM to µM range too were ineffective. Even co-administration of all three agonists (2 mM equimolar; n=4) did not significantly change clonidine’s effect (Figure 3A). AUCs also did not significantly differ between groups [F_(4,15)=1.17; p=0.38; Figure 3B]. These data suggest that activation of the Gq-mER solely and rapidly attenuates clonidine antinociception whereas individual or concurrent activation of the remaining mERs - ERα, ERβ and/or GPR30 does not.

Figure 3.

Selective activation of ERα ERβ or GPR30, individually or concurrently, fails to alter clonidine antinociception. A, Intrathecal pretreatment with PPT, DPN, or G1 (ERα, ERβ and GPR30 selective agonists, respectively) did not attenuate clonidine-induced increase in TFL. B, Consistent with the time course data, AUCs did not significantly differ between any groups. This data provides evidence of an absence of any role played by ERα, ERβ and GPR30 in E2BSA-induced rapid attenuation of clonidine antinociception.

2.4 Blockade of de novo protein synthesis using anisomycin fails to alter E2-BSA- or STX-induced attenuation of clonidine antinociception

To determine if Gq-mER activation-induced de novo protein synthesis contributed to the suppression of clonidine-induced nociception, we pretreated rats intrathecally with anisomycin, a protein synthesis inhibitor, as outlined in methods. ANOVA yielded significant main effects of Group [F_(3,12)=67.68; p < 0.01], Time [F_(12,144)=19.34; p < 0.01] and an interaction between Group × Time [F_(36,144)=11.61; p < 0.01]. This interaction merely stemmed from the fact that after three comparable baselines at first three time points, only veh+clon had significantly higher TFLs as compared to the rest of the groups. Post hoc comparisons revealed that anisomycin pretreatment did not significantly alter the lack of clonidine-induced increase in TFL in E2BSA (n=5) or STX (n=4) treated rats (Figure 4A). In addition, anisomycin itself did not change TFL which remained at the level of baseline through the 90 min time course (n=4). Further, all anisomycin-treated groups still had significantly lower AUC as compared to the clonidine alone control group [F_(3,15)=120.94; p < 0.01; Figure 4B]. These data are consistent with the interpretation that de novo protein synthesis, which would be expected to be required for genomic responses to estradiol, does not underlie Gq-mER-mediated rapid attenuation of clonidine-induced antinociception.

2.5 E2BSA or STX treatment significantly increases spinal activation of ERK I/II but not PKC or PKA

To explore the role ERK I/II, PKC or PKA played in mER-induced rapid and reversible suppression of clonidine-induced antinociception, we harvested lumbosacral spinal cord tissue from animals treated intrathecally with E2BSA (0.5 mM) or STX (2 mM) and/or clonidine and quantitated the relative amount of these enzymes that were phosphorylated in the absence and present of E2BSA or STX. Densitometry analyses of immunoblots revealed that E2BSA or STX treatment significantly increased levels of pERK I/II in all groups as compared to the clonidine only control group [F_(4,14)=11.89; p < 0.01; Figure 5A]. STX induced as much increase in pERK I/II level as did E2BSA. However, pPKC [F_(4,14)=1.79; p=0.21; Figure 5B] or pPKA [F_(4,14)=2.97; p=0.07; Figure 5C] levels did not significantly change with E2BSA or STX treatment (n=3/group). These data provide evidence that Gq-mER-induced cellular cascades that lead to attenuation of clonidine antinociception parallel the activation of ERK I/II but not of PKC and PKA.

Figure 5.

Spinal activation of ERK I/II but not PKC or PKA significantly increases following mER activation. A, Significantly increased levels of phosphorylated ERK I/II were detected by immunoblotting of the spinal cord tissue collected after in vivo mER activation using E2BSA or STX. Representative qualitative data from pERK I/II, ERK I/II and actin positive immunoblots is presented under each column. B, C, In contrast, mER activation did not significantly alter spinal activation of PKC or PKA (*p < 0.05 compared to clonidine only control group). This data provide evidence that mER-activation-induced rapid attenuation of clonidine antinociception involves ERK I/II signaling.

2.6 Behavioral reversal of the effect of STX by spinal blockade of ERK I/II

To determine whether Gq-mER –mediated activation of ERK I/II was causal in suppressing clonidine-induced antinociception, we inhibited ERK I/II phosphorylation using intrathecal administration of U0126 in STX and clonidine treated rats. ANOVA yielded significant main effects of Group [F_(3,8)=91.13; p < 0.01], Time [F_(12,96)=36.98; p < 0.01] and an interaction between Group × Time [F_(36,96)=8.88; p < 0.01]. This interaction stemmed from the fact that after three comparable baselines at first three time points, only veh+clon and U0126+STX+clon-treated groups had significantly higher TFLs as compared other groups though the rest of the time course. Post hoc comparisons revealed that the active ERK I/II inhibitor,

U0126 completely abolished the effect of STX and restored clonidine-induced increase in TFL (p < 0.01; Figure 6A) whereas an inactive ERK I/II inhibitor, U0124 failed to alter the effect of STX. U0126 injection itself did not significantly change TFLs as compared to the baseline TFL (n=3/group) (Figure 6A), nor was there a significant difference between AUC of U0126 + STX + clon group and clonidine only control group (Figure 6B). In a separate control experiment, intrathecal co-administration (2 mM) of PPT, DPN, and G1 that was behaviorally ineffective in altering clonidine antinociception, failed to significantly change the level of pERK I/II whereas STX again significantly increased the same in the lumbar spinal cord [F_(3,14)=11.72; p < 0.01; Figure 6C]. Further, U0126 treated rats showed significantly lower spinal levels of pERK I/II as compared to that of STX-treated rats (p < 0.01; Figure 6C). Specificity of ERK I/II and pERK I/II antibodies were determined by preadsorbing them with the respective control peptides (Cell Signaling Technology) for 1–2 hr at room temperature which almost completely eliminated the staining (Figure 6D). These data provide strong evidence that Gq-mER-induced rapid attenuation of clonidine antinociception is causally dependent on Gq-mER activation of the ERK I/II cascade.

3. Discussion

The present studies provide novel evidence that estradiol-induced attenuation of clonidine antinociception is mediated primarily if not exclusively by the newly discovered Gq-mER, and is secondary to activation of ERK I/II by this receptor. Although technical limitation does not allow us to rule out a genomic component, our results are consistent with the interpretation that non-genomic mechanisms underlie this action of Gq-mER since i) it quickly attenuated clonidine-induced increase in the TFL (within 5 min; too short a time for mER-induced genomic changes to occur and alter clonidine’s effect), and ii) anisomycin failed to alter the effect of E2BSA or STX. The dose of anisomycin employed in this study (125 µg) has been previously shown to almost completely (>90%; Grollman, 1967; Rosenblum et al., 1993) block de novo protein synthesis. Since genomic responses to estradiol would depend on synthesis of new proteins, ability of E2BSA or STX to attenuate clonidine antinociception despite anisomycin pretreatment provides evidence that a genomic mechanism does not underlie Gq-mER-mediated rapid attenuation of clonidine-induced antinociception.

The effect of E2BSA or STX was blocked by ICI182,780, a broad spectrum antagonist at ERα, βµand Gq-mER suggesting estrogen receptor-mediated effects. On the contrary, a reported agonistic activity of ICI182,780 at GPR30 (Thomas et al., 2005) is not likely to play a role in findings presented since i) clonidine antinociception was fully restored following ICI182,780 pretreatment, which blocks ERα, β, and Gq-mER but not GPR30, ii) ICI182,780 treatment alone did not change TFL, and iii) a GPR30 selective agonist, G1, did not affect clonidine antinociception. The data presented also do not support involvement of multiple estrogen receptor-mediated mechanisms since concurrent activation of ERα, ERβ and GPR30 by co-administration of PPT, DPN and G1 was as ineffective in altering clonidine antinociception as that of each of these three specific ER subtypes.

We reported previously that estrogen-induced changes in the spinal expression of α₂-adrenoceptor protein or mRNA did not correlate with the observed behavioral attenuation of clonidine antinociception (Thompson et al., 2008). These findings, together with our current data, would suggest that the Gq-mER-mediated rapid and reversible suppression of clonidine-mediated antinociception depends primarily on non-genomic mechanisms.

The α₂-adrenoceptor is widely distributed in the brain and spinal cord, including the dorsal horn (Unnerstall et al., 1984; Sullivan et al., 1987; Stone et al., 1998). Consequently, epidural or intrathecal clonidine treatment has been utilized to alleviate severe neuropathic or cancer pain even in patients with limited relief by and development of tolerance to opiates (Eisenach et al., 1995; Hassenbusch et al., 1999; Rauck et al., 1993; Dobrydnjov et al., 2005). Activated α₂-adrenoceptors couple to pertussis toxin-sensitive Gi/Go proteins that couple to inhibition of adenylyl cyclase and attenuation of cyclic-AMP synthesis, to suppression of voltage-gated Ca⁺⁺ channel activities, and to activation of outward currents via K⁺⁺ channels including G-protein coupled inwardly rectifying K⁺⁺ (GIRK) channels (Bylund et al., 1992; Lipscombe et al., 1989; Sonohata et al., 2004; see Figure 7). Resulting neuronal hyperpolarization hinders pain signaling. GIRK channels are present in the superficial dorsal horn and most brain regions (Karschin et al., 1996; Marker et al., 2005). Although a direct co-localization of GIRK with α₂-adrenoceptor in the spinal cord has not been shown, a significant loss of clonidine-induced antinociception by deletion of GIRK2 channels in mice suggests a predominantly post-synaptic, GIRK2- mediated, α₂-adrenoceptor-induced antinociception (Blednov et al., 2003; Mitrovic et al., 2003). Hence, mER-induced disruption of α₂-adrenoceptor coupling to the GIRK channel, or direct effects of Gq-mER on GIRK , could represent mechanisms by which these receptors impair α₂-adrenoceptor-mediated antinociception (Figure 7).

Figure 7.

Hypothesized Gq-mER signaling: Upon agonist binding, Gi-coupled α₂-adrenoceptor inhibits calcium channels and activates GIRK channels, resulting in reduced neuronal excitability, thus producing antinociception. Gq-coupled mER activates PLC and results in IP3-mediated intracellular Ca⁺⁺ mobilization, as well as DAG-mediated activation of PKC and in turn, PKA. Present results demonstrate Gq-mER-induced, MEK I/II-mediated activation of ERK I/II. This activation is expected to inhibit GIRK channels (and possibly other K⁺ channels), resulting in increased neuronal excitability and thus attenuating α₂ adrenoceptor-mediated antinociception. Alternatively, ERKI/II activation may have previously unreported effects on the coupling of the α₂-adrenoceptor to Gi/Go, another formal possibility by which Gq-mER might mediate suppression of the antinociceptive effects of clonidine. Abbreviations: ER, endoplasmic reticulum; DAG, diacylglycerol; ERK I/II, extracellular signal regulated kinases I/II; MEK I/II, MAPK/ERK kinases I/II; PIP2, phosphatidylinositol 4,5-biphosphate; PLCβ, Phospholipase C-β; IP3, inositol (1,4,5) trisphosphate

All estrogen receptors (ERα, β, GPR30 and Gq-mER) are shown to be localized in spinal dorsal horn neurons (Shughrue et al., 1997; Papka et al., 2001; Dun et al., 2009; Zhang et al., 2012). Gq-mER, as the name suggests, is coupled to Gα_q protein that induces cellular cascades (Qiu et al., 2003; Figure 7) beginning by activation of phospholipase C β (PLCβ), DAG-induced activation of PKCδ and in turn, PKA which may phosphorylate, and thus inhibit potassium channel (including GIRK) function. IP3-induced mobilization of intracellular Ca⁺⁺ may further increase neuronal excitability. Based on this evidence derived from hypothalamic neurons, increased pPKA and pPKC levels would be predicted following spinal Gq-mER activation. However, this was not the case following E2BSA or STX treatment in the present study; yet spinal ERK I/II phosphorylation was significantly increased following STX. Although a definitive evidence of ERK activation by PKC or PKA in the dorsal horn is lacking, there is strong evidence supporting the view that PKC and PKA may directly (MEK-mediated) or indirectly (Raf-mediated) activate ERK (Gutkind, 2000; Hu and Gereau, 2003; Hu et al., 2003). A downstream activation of ERK is suggested since ERK inhibition prevented PKC- and PKA-mediated modulation of A-type K⁺ current (Hu and Gereau, 2003) as well as action potentials in the superficial dorsal horn neurons (Hu et al., 2003). A recent study has also reported estradiol-induced phosphorylation of spinal PKA and ERK (Zhang et al., 2012). In addition, Gs coupling of membrane ERα and GPR30 is expected to induce adenylyl cyclase-mediated activation of cAMP-dependent PKA (Razandi et al., 1999; Filardo et al., 2007). Activated PKA may lead to K⁺ channel phosphorylation via ERK and thus attenuate clonidine antinociception. It is likely that ERK I/II activation was mediated in this study by kinases other than PKC and PKA, nevertheless, there may be subtle, cell-specific changes in PKC and PKA levels which could not be detected due to sensitivity limits of Immunoblotting performed on the lumbar spinal tissue in toto. In contrary, PKC or PKA-independent Gq-mER-induced cellular cascades may exist comprising other kinases including ERK I/II.

Agonist bound ERβ, on the other hand couples to Gαi (Kumar et al., 2007) and has been shown to produce visceral antinociception (Cao et al., 2012) in comparison to the pronociceptive effect of ERα shown by the same group (Ji et al., 2011). Such opposing effects of mERs were not observed in the present study due likely to a participation of genomic mechanisms in the above studies vs. rapid, non-genomic mechanisms investigated in the present study.

Finally, this study employed a model for acute pain testing, i.e. tail flick test; thus estrogenic mechanisms that attenuate clonidine antinociception investigated in this study may differ under chronic pain condition due to involvement of a variety of inflammatory mediators and sensitization of nociceptive signaling pathways. Further studies utilizing chronic pain models will be required to determine those mechanisms.

3.1 Conclusions

Although estrogen-induced delayed genomic mechanisms may still exist, results of the present study led us to conclude that estradiol rapidly attenuates clonidine-induced antinociception via a Gq-mER-mediated, ERK I/II activation-dependent, non-genomic signaling mechanism presumably impairing K⁺ channel function (Figure 7). Development of a Gq-mER antagonist would appear to prove useful for improving α₂-adrenoceptor-mediated analgesia in women. Relatively widespread distribution of α₂-adrenoceptor and Gq-mER and their diverse roles across multiple physiological functions outside the pain system e.g., cardiovascular, reproductive, learning and memory, suggest broader implications of our novel findings.

Highlights.

• Estradiol rapidly attenuates α₂-adrenoceptor antinociception via non-genomic mechanism

• Activation of Gq-mER but not ERα, ERβ and/or GPR30 mediates estradiol’s effect

• ERK activation is required for estradiol attenuation of clonidine antinociception

Abbreviations

AUC

area under the curve

DAG

diacylglycerol

DPN

2,3-bis(4-Hydroxyphenyl)-propionitrile

E2BSA

β-Estradiol 6-(O-carboxy-methyl)oxime bovine serum albumin

ERK

extracellular signal regulated kinase

GIRK

G-protein coupled inwardly rectifying potassium channels

GPCR

G-protein-coupled receptor

GPR30

G-protein coupled estrogen receptor 30

Gq-mER

Gq-coupled mER

i.t.

intrathecal

IASP

International Association for the Study of Pain

IP3

inositol (1,4,5) trisphosphate

mER

membrane estrogen receptors

OVX

ovariectomized

PIP2

phosphatidylinositol 4,5-biphosphate

PKA

protein kinase A

PKC

protein kinase C

PLCβ

Phospholipase C-β

PPT

4,4',4"-(4-Propyl-[1H]-pyrazole-1,3,5-triyl)trisphenol

SEM

standard error of the mean

TBS

tris-buffered saline

TBST

tris-buffered saline containing 0.05% Tween 20

TFL

tail flick latency

TMD

temporomandibular joint disorder