DRGs from all spinal levels were harvested from early postnatal (p0-1) or adult mice (4-6 weeks old), which contained both (β-arr2+/+) or neither (β-arr2-/-) of the β-arr2 alleles in the C57BL/6 background. DRGs were dissociated in trypsin (Invitrogen, Carlsbad, CA) for the early postnatal neurons, or Collagenase (Liberase TH and TM, Roche, Indianapolis, IN) for the adult neurons. 1 × 10⁴ cells were plated on each poly-D-lysine (Sigma, MO) and laminin (Invitrogen, CA) coated coverslip of MatTek dishes, (Ashland, MA) as previously described 17.

After 24-48 h in vitro, the whole-cell patch-clamp (Axopatch 200A amplifier, Axon Instruments Inc., CA) technique was used to record VDCC activity from small DRG neurons (capacitance less than 15 pF). The external solution contained (in mM): 130 TEA-Cl, 10 CaCl₂, 5 HEPES, 25 D-glucose and 0.25 tetrodotoxin (pH 7.2). Recording electrodes contained (in mM): 105 CsCl, 40 HEPES, 5 D-glucose, 2.5 MgCl₂, 10 EGTA, 2 Mg²⁺-ATP and 0.5 Na⁺-GTP (pH 7.2). VDCCs were activated using voltage-steps from -80 mV to 10 mV at 20 s intervals. Such steps were carried out in the absence or presence of a prepulse to 80 mV, a 2 step-protocol. Voltage-activated currents were low-pass filtered at 2 KHz, digitized (Digidata, Axon Instruments Inc., CA) at 10 KHz, and stored on a PC. Leak currents were nulled using the P/4 subtraction method. All ligands were diluted in the external solution on the day of the experiment and applied through the perfusion system at ~10 ml/min. Mean current amplitudes were measured (pCLAMP 9.0, Molecular Devices, CA) between 5 and 10 ms after initiating the depolarizing step. Once a stable recording was obtained, Ca²⁺ in the external solution was replaced by Ba²⁺ (10 mM) to minimize Ca²⁺ current inactivation.

For behavioral experiments male and female 2-3 month old progeny of heterozygous matings of mice fully backcrossed into the C57Bl/6 background were used to test thermal or mechanical analgesia or place conditioning, each mouse undergoing a single test. Thermal analgesia was also tested in double knockout mice lacking both μ receptors and β-arrestin 2 generated from heterozygous matings of both lines, both of which were fully back-crossed into the C57Bl/6 background. All animal research was conducted in accordance with the federal regulations as set forth in the Animal Welfare Act (AWA), the 1996 Guide for the Care and Use of Laboratory Animals, PHS Policy for the Humane Care and Use of Laboratory Animals, as well as UCLA's policies and procedures as set forth in the UCLA Animal Care and Use Training Manual.

Thermal nociception was measured by either the tail immersion 19 or the Hargreaves 20 test and mechanical nociception by von Frey filaments 21. For all tests animals were habituated to the test environment for 2 days prior to testing which took place between 8.00 and 16.00 h. For the tail immersion test the last 2.5 cm of the tail was immersed in 48°C water and the latency for tail withdrawal measured. The maximum immersion time was 15 s. After 3 basal measurements, saline or the test drug was injected subcutaneously (sc) at a volume of 10 μl/g body weight (bw) and the time to respond measured 30, 60 and 90 min thereafter. The compounds tested and doses used were (in mg/kg): naloxone (0.5), 6α-naloxol (1.0), 6β-naloxol (10), naltrexone (0.5) and 6β-naltrexol (10). Naloxone and naltrexone were obtained from Sigma (MO) and the hydroxyl derivatives from the National Institute on Drug Abuse. For the Hargreaves test each mouse was habituated to the 10 cm × 10 cm × 15 cm enclosure on a heated (22°C) glass plate for 20 min prior to testing. A medium intensity infrared beam was shone on the rear foot pad of the hind paw and the latency to lift the paw was measured. The average responses of both left and right hind paws were obtained for each data-point with a minimum interval of 5 min between testing either paw. After the baseline measurements were obtained naloxone was injected (0.5 mg/kg) and the response measured 30 min later. For mechanical nociception the response threshold to punctate mechanical stimuli was tested according to the up-and-down method 21 as previously described 22. In this case, the plantar surface of one hind paw was stimulated with a series of eight von Frey filaments (bending force ranging from 0.01 to 2 g). A response was defined as a lifting or shaking of the paw upon stimulation.

The conditioning apparatus (Coulbourn Instruments, Allentown, PA, USA) and experimental protocol have been described previously 23. Briefly, arenas were divided into three distinctively patterned chambers: a neutral start chamber and two conditioning chambers, which were discernable by pattern and odor. The drug-paired chamber was randomized across subjects and treatments. On day 1 (habituation session), mice were placed in the start chamber, permitted access to the entire apparatus for 15 minutes, and time spent in each chamber was recorded. On the mornings of days 2-4 (conditioning sessions), animals received saline (10 μl/g bw sc) and were confined to the 'vehicle-paired' chamber for 30 minutes. Four hours later, the same mice received a single naloxone injection (0.25, 0.5, 1.0, or 10 mg/kg sc) or vehicle and were confined to the 'drug-paired' chamber for 30 min. On day 5 (test session), animals, in a drug-free state, were placed in the neutral chamber and permitted to explore the entire apparatus for 15 minutes, with time spent in each chamber recorded. Separate groups of β-arr2-/- and β-arr2+/+ mice were used for each dose of naloxone.

Nociception assays involved sample sizes of 9-22 mice per group. Data from these and the electrophysiological assays were analyzed by one way ANOVA for genotype, and the post hoc Scheffe test. Data are presented as the mean ± SEM of the time to respond. For place conditioning analysis of initial bias between future drug-paired versus future vehicle-paired chamber was conducted by three way ANOVA with factors of genotype, dose, and the repeated-measures factor of chamber. Statistical comparisons of times spent in the drug-paired chamber between habituation and test day were made using three way ANOVA with factors of genotype, dose, and the repeated-measures factor of day (habituation versus test).

VDCC inhibition by μ receptors is mediated by Gβγ subunits which bind to VDCCs in a voltage-dependent manner 24. Inhibition mediated by Gβγ subunits can be reversed by strongly depolarizing pre-pulses leading to facilitation of current mediated by VDCCs 15. We previously demonstrated that facilitation can be used to monitor the level of inhibitory opioid receptor coupling to VDCCs in the absence of agonist 17. We used a voltage protocol to generate a current (activated by stepping from -80 to 10 mV) either in the absence (P1) or in the presence (P2) of a depolarizing pre-pulse from -80 to 80 mV (Figure 1A). In our previous study we used small to medium sized neurons 17. In the current study we measured facilitation in small (classified as having a whole-cell capacitance of < 15 pF) DRG neurons, by determining the P2/P1 ratio (Figure 1A). In agreement with our previous report, the pre-pulse-evoked facilitation was significantly larger (p < 0.005) in β-arr2-/- (1.10 ± 0.02) compared to β-arr2+/+ DRG neurons (1.02 ± 0.01) from early postnatal mice and was augmented when the non-hydrolyzable GTP analog, GTP-γ-S (100 μM) was included in the intracellular recording solution (Figure 1A). An increased facilitation was not seen in DRG neurons from postnatal mice lacking both μ receptors and β-arr2 (1.03 ± 0.02, p < 0.05 Figure 1A), indicating that the μ receptor is required for this phenomenon. Similar to early postnatal DRG neurons, a pre-pulse evoked facilitation was also seen in adult DRG neurons from β-arr2-/- (1.10 ± 0.02) but not β-arr2+/+ (1.03 ± 0.02) mice (p < 0.05, Figure 1B). Again, facilitation was enhanced by the presence of intracellular GTP-γ-S (Figure 1B). Early postnatal DRG neurons were used for all subsequent electrophysiological experiments.

We previously demonstrated that facilitation in β-arr2-/- neurons is mediated by constitutively active μ receptors; it is inhibited by the inverse agonist naltrexone, but unaffected by the neutral antagonist CTAP 17. The relatively poor bioavailability of the peptide CTAP compared to alkaloids 25 prompted us to use voltage-dependent facilitation of VDCC activity in β-arr2-/- neurons as an assay to identify alkaloids that lack negative efficacy for subsequent use in vivo. We examined naloxone and naltrexone, previously classified as inverse agonists and 6α-naloxol, 6β-naloxol and 6β-naltrexol classified as neutral antagonists in the reward pathway in vivo 11 and in assays of adenylyl cyclase activity in vitro 9. Naloxone (1 μM) reduced facilitation recorded from β-arr2-/- DRG neurons (Figure 2A). By contrast, 6α-naloxol (1 μM), 6β-naloxol (1 μM) and 6β-naltrexol (1 and 10 μM) did not influence facilitation in β-arr2-/- neurons. In agreement with our previous study 17 naltrexone (1 μM), like naloxone, inhibited facilitation. The naloxone-evoked reduction in pre-pulse evoked facilitation in β-arr2-/- neurons was prevented by the co-application of the neutral antagonist, 6α-naloxol (Figure 2A). None of the alkaloids tested affected facilitation in β-arr2+/+ neurons.

We also examined the effect of the inverse agonists on currents activated by depolarizing neurons from -80 to 10 mV in the absence of a depolarizing pre-pulse. This approach reveals the amplitude of current recovery by inverse agonist-evoked inhibition of constitutive μ receptor activity. Naloxone (1 μM) and naltrexone (1 μM) caused small enhancements of VDCC mediated currents recorded from β-arr2-/- neurons (5.3 ± 1.0% and 4.3 ± 1.1% of control amplitude) not seen in recordings from β-arr2+/+ neurons (Figure 2B). By contrast, the neutral antagonists 6α-naloxol (1 μM), 6β-naloxol (1 μM) and 6β-naltrexol (1 and 10 μM), had no significant effect on the amplitude of currents recorded from DRG neurons, regardless of genotype. The enhancement of VDCC mediate currents by naloxone (1 μM) was inhibited by co-application of the neutral antagonist, 6α-naloxol (1 μM; Figure 2B).

We tested the ability of the neutral antagonists to inhibit agonist induced μ receptor coupling to VDCCs. The specific μ receptor agonist DAMGO (1 μM) inhibited currents recorded from β-arr2+/+ neurons to 28 ± 3% of control current amplitude. By contrast, DAMGO (1 μM) co-applied with the neutral antagonists (1 μM of each) 6α-naloxol, 6β-naloxol, or 6β-naltrexol, barely inhibited currents (2.6 ± 1.0%, 0.1 ± 0.5% and 2.3 ± 1.6%, respectively, n = 4-10). These data confirm that, at the concentrations used, the neutral antagonists bind to μ receptors and inhibit VDCC coupling.

Having determined the relative intrinsic efficacies of naloxone, naltrexone, 6α-naloxol, 6β-naloxol, and 6β-naltrexol in vitro we used these bioavailable compounds to investigate the behavioral significance of constitutive μ receptor activity in vivo.

Mice lacking β-arr2 have previously been shown to exhibit a delayed thermal response in the tail withdrawal assay 26. We similarly found an attenuated tail withdrawal response in β-arr2-/- mice that was absent in their β-arr2+/+ littermates (Figure 3A). This enhanced tail withdrawal latency was not seen in mice lacking both μ receptors and β-arr2 (μ-/-/β-arr2-/-: 2.8 ± 0.2 s, μ+/+/β-arr2-/-: 4.6 ± 0.4 s, μ+/+/β-arr2+/+: 2.5 ± 0.3 s, p < 0.05 μ-/-/β-arr2-/- vs μ+/+/β-arr2-/-) indicating that the μ receptor is necessary for the prolonged response to noxious heat.

We examined whether constitutive activity of the μ receptor is responsible for the prolonged withdrawal latency in β-arr2-/- mice. Initially we used the inverse agonist, naloxone (Figure 3B) and the neutral antagonists, 6α-naloxol (Figure 3C) and 6β-naloxol (Figure 3D) using doses of 0.5, 1 and 10 mg/kg for naloxone, 6α- and 6β-naloxol respectively, previously shown to be effective in vivo 912132728. We examined a time course of up to 90 min after each injection, to allow sufficient time for different rates of receptor activation. Naloxone, 30 min after administration, reduced tail withdrawal latency of β-arr2-/- mice from 4.6 ± 0.4s to 3.0 ± 0.3s (Figure 3B), but had no effect in β-arr2+/- or β-arr2+/+ mice (Figure 3B). The mean withdrawal latencies in naloxone treated β-arr2-/- mice at the 30 and 60 min time points were similar to those of β-arr2+/+ mice with or without naloxone administration. Unlike naloxone, neither the neutral antagonist 6α-naloxol nor 6β-naloxol significantly affected tail withdrawal latencies in β-arr2+/+, β-arr2+/- or β-arr2-/- mice (Figure 3C and 3D). However, the co-administration of 6α-naloxol with naloxone abolished the reduction in tail withdrawal latency seen in β-arr2-/- mice treated with naloxone alone, indicating that both drugs have access to μ receptors in vivo (Figure 3E). Consistent with these effects being mediated by μ receptors, naloxone had no effect when administered to μ-/-/β-arr2-/- mice 30 min after injection (untreated: 2.81 ± 0.24 s, Nal: 3.03 ± 0.22 s, p = 0.5).

Although naloxone is an inverse agonist at μ receptors, it also binds to δ and κ receptor subtypes 29. We therefore tested the effect of naltrexone (0.5 mg/kg), an inverse agonist with a similar negative efficacy to that of naloxone 30, but with higher μ receptor selectivity 29 as well as the structurally related neutral antagonist 6β-naltrexol. Since 6β-naltrexol may have a lower affinity for the μ receptor and a slower rate of activation, a higher dose of 6β-naltrexol, effective in vivo, was administered 2731. We found that similar to naloxone, naltrexone reduced tail withdrawal latency in β-arr2-/- mice, but was without effect in β-arr2+/+ or β-arr2+/- mice 30 min after administration (Figure 3F). By contrast, the neutral antagonist 6β-naltrexol (10 mg/kg) had no effect on tail withdrawal latencies, irrespective of genotype (Figure 3G).

Similar to the tail-immersion response, β-arr2-/- mice also exhibited a delayed response to thermal pain assessed by the Hargreaves test (β-arr2+/+: 2.17 ± 0.11 s, β-arr2-/-: 3.55 ± 0.03 s, p < 0.05; Figure 4A). The response latency was reduced by naloxone (0.5 mg/kg; β-arr2+/+: 2.30 ± 0.34 s, β-arr2-/-: 2.75 ± 0.35 s, p < 0.05 vs untreated β-arr2-/-). However, there was no effect of genotype in the response to mechanical pain assessed by von Frey filaments (β-arr2+/+: 1.31 ± 0.31 g, β-arr2-/-, 1.29 ± 0.15 g; Figure 4B). These data suggest that the enhanced analgesic tone mediated by constitutively active μ receptors in β-arr2-/- mice is specific to the thermal nociceptive pathway.

In addition to nociception, μ receptors regulate reward and hedonic tone 2332. Agonists of the μ receptor induce a conditioned place preference, a measure of reward, while antagonists and inverse agonists produce a conditioned place aversion (CPA), a measure of μ receptor-mediated hedonic tone 12. As enhanced constitutive activity of the μ receptor induced by morphine treatment leads to increased naloxone aversion 12 we examined whether the enhanced μ receptor constitutive activity seen in β-arr2-/- mice would similarly affect basal hedonic tone. Although naloxone administration (0.25-10 mg/kg sc) produced an aversion to the drug-paired chamber (F_1,61 = 33.75, p < 0.0001) in a dose-dependent manner (dose × day interaction: F_4,61 = 6.41, p < 0.001, Figure 4C), there was no effect of genotype. This was evident from the lack of interaction between genotype and day (F_1.,61 = 0.62, p = 0.44), or between genotype, day and dose (F_4,61 = 0.15, p = 0.96). There was no overall initial bias between the future drug-paired versus vehicle-paired chamber on habituation day (F_1,61 = 0.10, p = 0.75), no interaction with dose (F_4,69 = 0.58, p = 0.68) or genotype (F_1,61 = 1.29, p = 0.26), and no three-way interaction (F_4,61 = 0.59, p = 0.67).