Contrasting cardiovascular properties of the µ-opioid agonists morphine and methadone in the rat

European Journal of Pharmacology 762 (2015) 372–381

Contents lists available at ScienceDirect

European Journal of Pharmacology journal homepage: www.elsevier.com/locate/ejphar

Cardiovascular pharmacology

Contrasting cardiovascular properties of the m-opioid agonists morphine and methadone in the rat Kenneth H. Tung, James A. Angus, Christine E. Wright n Cardiovascular Therapeutics Unit, Department of Pharmacology and Therapeutics, University of Melbourne, Victoria 3010, Australia

art ic l e i nf o

a b s t r a c t

Article history: Received 23 April 2015 Received in revised form 10 June 2015 Accepted 10 June 2015 Available online 20 June 2015 Keywords: Arterial blood pressure Heart rate Mesenteric artery Papillary muscle Right atria Left atria Methadone Morphine Naloxone QT interval Rat Chemical compounds studied in this article: Morphine sulphate (PubChem CID: 5464280) Methadone hydrochloride (PubChem CID: 4095) Mepyramine maleate (PubChem CID: 6039) Cimetidine (PubChem CID: 50963) Endothelin-1 (PubChem CID: 44284481) Naloxone hydrochloride (PubChem CID: 5464092) Nω-nitro-l-arginine (PubChem CID: 440005) Indomethacin (PubChem CID: 3715) Nifedipine (PubChem CID: 4485)

Morphine and methadone share the property of μ-opioid receptor agonism yet have markedly different cardiovascular actions suggesting additional properties are at play. We investigated the i.v. dose–response relationships of the opioids on cardiovascular metameters in anaesthetised rats in the absence or presence of H1- and H2-receptor antagonism and the μ-opioid antagonist naloxone. In vitro tissue assays were employed to define more clearly cardiac and vascular mechanisms of action. Morphine (9, 30, 90 mg/kg i.v.) decreased heart rate (HR) and mean arterial pressure (MAP) – responses that were blocked by naloxone pretreatment (10 mg/kg i.v.). In contrast, methadone (3, 10, 30 mg/kg i.v.) caused dramatic short-lived (1–3 min) bradycardia, hypotension and lengthening of the QT interval before stabilising 5 min after i.v. dosing. Only the steady-state responses of HR and MAP were blocked by naloxone. Mepyramine (10 mg/kg i.v.) and cimetidine (100 mg/kg i.v.) also blocked the naloxone-sensitive components. In isolated small mesenteric arteries precontracted by K þ 62 mM or endothelin-1, methadone (1– 30 μM) relaxed vessels while morphine (1–100 μM) had no effect. Pretreatment with naloxone (10 μM), indomethacin (30 μM) or nitro-L-arginine (100 μM) did not affect the relaxation to methadone. In rat isolated left atria, morphine and methadone inhibited inotropic responses at high concentrations (100 μM). In rat papillary muscle and right atria, methadone was more than 30 times more potent at lengthening the refractory period and slowing the atrial rate than morphine. We conclude that methadone is a potent vasodilator agent, possibly through blocking L-type calcium channels. & 2015 Elsevier B.V. All rights reserved.

1. Introduction Morphine and methadone are

μ-opioid receptor agonists. The

Abbreviations: ECG, electrocardiogram; HR, heart rate; KPSS, isotonic potassium physiological salt solution; MAP, mean arterial pressure; NO, nitric oxide; NOS, nitric oxide synthase; pEC50, the negative log10 of agonist concentration that caused 50% of the maximum response; PSS, physiological salt solution; QTC, QT interval of electrocardiogram corrected for heart rate n Corresponding author. E-mail addresses: [email protected] (K.H. Tung), [email protected] (J.A. Angus), [email protected] (C.E. Wright). http://dx.doi.org/10.1016/j.ejphar.2015.06.016 0014-2999/& 2015 Elsevier B.V. All rights reserved.

μ-opioid receptors are subclassified into μ1 and μ2 for high and low affinity sites, respectively (Pasternak and Wood, 1986; Thompson et al., 1993; Wolozin and Pasternak, 1981). A third μ3-subtype has been postulated to occur on endothelial cells to release nitric oxide (NO) that is sensitive to the μ-opioid receptor antagonist naloxone (Stefano et al., 1995). The opioid agonists cause varying degrees of hypotension and bradycardia caused by CNS actions, and direct and indirect actions on the heart and vasculature. A major concern for the use of methadone to suppress the physical dependence in the heroin or morphine addict has been its association with potentially fatal ventricular arrhythmias, Torsades

K.H. Tung et al. / European Journal of Pharmacology 762 (2015) 372–381

de Pointes (TdP). This arrhythmia is caused by blockade of the rapid-rectifying K þ channel that is encoded by the human ether-ago-go-related gene, the hERG channel (Katchman et al., 2002). This channel is crucially responsible for the repolarization of the cardiac action potential (Zhou et al., 1998). In this study, the dose-related hypotensive, bradycardic and ECG actions of morphine and methadone in anaesthetised rats were investigated. These opioid agonists were tested in the absence or presence of the μ-opioid antagonist naloxone or H1- and H2-histamine receptor antagonists to determine the role of μopioid receptors and potential for histamine release in vivo. To analyse the direct actions of these opioids, their activity was tested in isolated vascular rings of mesenteric small arteries, papillary and right and left atrial muscle. These in vitro assays provided a reference for the likely concentration-dependent actions in vivo. Methadone was consistently 30 fold more potent than morphine in slowing the atrial rate and increasing the refractory period, accounting for the dramatic acute fall in heart rate and increase in QTc in vivo that were not sensitive to the μ-opioid receptor antagonist naloxone. The work supports the notion that methadone has significant L-type calcium channel blocking activity – an action not shared with morphine.

2. Materials and methods Sprague Dawley rats (male; 285 7 3 g) were used. Experiments were approved by the University of Melbourne Animal Ethics Committee in accordance with the Australian code for the care and use of animals for scientific purposes (8th edition, 2013, National Health and Medical Research Council, Australian Government, Canberra). 2.1. In vivo experiments Rats were initially lightly anaesthetised with 5% halothane (Veterinary Companies of Australia; Artarmon, NSW, Australia) in O2 and air, then surgical anaesthesia was induced with pentobarbitone (60 mg/kg i.p.). Atropine (1 mg/kg s.c.) was given to inhibit bronchial secretions. After infiltration with a long-acting local anaesthetic (0.5% ropivacaine HCl; Naropin, AstraZeneca, Sydney, Australia), a tracheotomy was performed for mechanical ventilation (O2 and room air; initial settings 6.5 ml/kg stroke volume and 75 breaths/min; Rodent ventilator 7025, Ugo Basile, Comerio, Italy) and a carotid artery cannulated for the measurement of phasic and mean arterial pressure (MAP) via a calibrated blood pressure transducer (Cobe, Argon Medical, Athens, TX, USA) connected to a Powerlab data acquisition system (8SP; ADInstruments, Sydney, Australia). Heart rate (HR) was computed from phasic blood pressure. The right jugular vein was cannulated for drug administration. Electrocardiogram (ECG) leads (lead II), connected to a BIO Amp (ML136; ADInstruments), were inserted under the skin; parameters were analysed using the Chart ECG Analysis Module (v2.0; ADInstruments). Regular arterial blood gas analyses were performed (ABL5, Radiometer Medical A/S, Copenhagen, Denmark) to monitor blood pH (7.43 7 0.01), CO2 and O2 levels prior to the start of the experiment; optimum blood gas parameters were maintained by adjusting ventilator stroke volume and rate. Body temperature was maintained at 38 °C by a homoeothermic blanket with a rectal probe (Harvard Apparatus, Holliston, MA, USA). Stable cardiovascular parameters and anaesthetic state were maintained for 10 min prior to the start of the 62 min experiment protocol. Thereafter, the anaesthetic agent (s) was not readministered to obviate its effects on cardiovascular variables (depth of anaesthesia was adequate for the typically 72 min experimental period).

373

2.1.1. In vivo protocols 2.1.1.1. Effects of μ-opioid agonist alone. Rats were injected (i.v.) with 3 volumes of saline (each 2.5 ml/kg) or cumulative doses of morphine (9, 30 and 90 mg/kg) or methadone (3, 10 and 30 mg/ kg) at 21 min intervals. This allowed sufficient time for MAP and HR to stabilise after each dose, according to observations from pilot studies. Since the plasma half-life for morphine is 1.5–2 h and for methadone more than 20 h, doses were given in a cumulative manner; doses shown are the total in each case. To control for effects of bolus volumes, all drug doses were standardized to 2.5 ml/kg as an i.v. infusion over 3.5 min. 2.1.1.2. Effects of μ-opioid receptor antagonism. Rats were pretreated with the μ-opioid receptor antagonist naloxone (10 or 20 mg/kg i.p.) 20 min prior to the administration of the 3 doses of saline, morphine or methadone, as described above. 2.1.1.3. Effects of histamine H1 7H2 receptor antagonism. Rats were pre-treated with the histamine H1-receptor antagonist mepyramine (10 mg/kg i.p.) or with the histamine H1- and H2-receptor antagonists mepyramine and cimetidine (10 and 100 mg/kg, respectively, i.p.) 20 min prior to the administration of the 3 doses of saline, morphine or methadone, as above. These doses of mepyramine and cimetidine were shown to inhibit responses to histamine (10–100 μg/kg i.v.). 2.1.1.4. ECG analyses. During the protocols described above, the ECGs of 4 consecutive heartbeats were acquired and averaged at each time point. The QT interval – time (ms) between the start of the QRS complex and the end of the T wave – was measured. Since the QT interval is highly dependent on the heart rate (RR interval), it must be corrected, giving the QTc interval. For this purpose, we used the formula of Fridericia (QTc ¼RR  QT  0.33) (Fridericia, 1920), considered to be the more suitable choice than the Bazett correction (EMEA, 2006). 2.2. In vitro experiments Male Sprague Dawley rats were anaesthetised with 5% halothane in O2 and killed by decapitation. The heart and a loop of intestine with mesenteric arteries attached were immediately excised and placed in cold physiological salt solution (PSS) of the following composition (in mM): NaCl 119; KCl 4.69; MgSO4 1.17; KH2PO4 1.18; glucose 11 (5.5 for arteries); NaHCO3 25; EDTA 0.026; and CaCl2 2.5. The PSS was continuously saturated with carbogen (95% O2; 5% CO2) at pH 7.4. 2.2.1. Rat mesenteric arteries Rat second or third order small mesenteric arteries were carefully dissected from surrounding tissues under a microscope; the endothelium remained intact. Each vessel segment (2 mm long) was mounted on 40 μm diameter wires in an isometric myograph (Danish Myo Technology, Aarhus, Denmark) containing oxygenated PSS at 37 °C. One wire was connected to a force transducer and the other to a micrometre; force was recorded on a Powerlab data acquisition system. From a computer-fitted curve, parameters were determined to stretch each vessel to 0.9L100, where L100 is the circumference of the artery when distended at a transmural pressure of 100 mmHg (Angus and Wright, 2000; Mulvany and Halpern, 1977). After an equilibration period of 20 min, each vessel was exposed to potassium-depolarising solution (KPSS; PSS with an equimolar substitution of KCl for NaCl; K þ 124 mM) for 2 min to obtain a reference maximum contraction (100% KPSS tone). The myographs were then washed out and vessels pre-contracted with endothelin-1 (1–3 nM) to approximately 70–100% KPSS tone. In another series of experiments,

K.H. Tung et al. / European Journal of Pharmacology 762 (2015) 372–381

2.2.2. Rat right and left atria and papillary muscles Right atria were dissected from the ventricles and surrounding tissue. Two stainless steel S-shaped hooks were then attached to the right atrium. Papillary muscles were isolated from right ventricles and a silk suture (5.0, Genzyme Biosurgery, Fall Creek, MA, USA) was secured to the chordae tendineae. A stainless steel S-shaped hook was then attached at the junction of the papillary and septum. Both cardiac tissue preparations were then placed vertically in organ baths and attached to a force transducer (Grass FT03, Grass Instruments, Quincy, MA, USA) via the S-shaped hooks and silk suture (for papillary muscle). A passive force of 9.8 mN for right atria and 4.9 mN for papillary muscles was then applied. Platinum field electrodes, on either side of the tissue, were then used to field stimulate the papillary muscle (Grass S88). Force signals were amplified (Baker Medical Research Institute amplifier Model 108, Prahran, VIC, Australia) and recorded with a Powerlab data acquisition system (8SP, ADInstruments). Right atrial tissue was allowed to beat spontaneously and the rate was computed. Papillary muscle was paced continuously at 1 Hz (S1), which was intermitted by an extra pulse (S2). The S1–S2 interval was shortened gradually until the tissue failed to initiate a contraction. This S1–S2 interval was defined as the absolute refractory period (ARP). Both S1 and S2 pulses were of 0.3 ms duration and delivered at 1.2 times the threshold voltage; this was to prevent autonomic neurotransmitter release (Angus and Harvey, 1981). The tissues were equilibrated for 60 min with washes every 15 min. Several measurements were then made at 10 min intervals to ensure a stable baseline before drug treatment began. Cumulative concentration–response curves were constructed to morphine (10  7–10  3 M) or methadone (10  7–10  4 M). Rat left atria were mounted as per right atria and pre-stretched to 0.5 g force for 10 min before readjusting to 0.5 g. Through platinum field electrodes, the atria were stimulated to contract at 1 Hz, with pulses set at 0.3 ms duration and voltage 50% above threshold. The maximum active ionotropic response was then determined with the addition of isoprenaline 0.1 μM. After returning to baseline, the left atria were stimulated with a submaximum isoprenaline concentration (0.01 μM) and tissues allowed to stabilise before adding cumulative concentrations of morphine, methadone or nifedipine. 2.3. Drugs Drugs used in this study were dissolved or diluted in 0.9% saline (unless otherwise stated): atropine sulphate (Sigma; St. Louis, MO, USA); cimetidine (GSK; Boronia, VIC, Australia); endothelin-1 (Auspep, Parkville, VIC, Australia); heparin sodium (Pharmacia & Upjohn; Rydalmere, NSW, Australia); indomethacin (Sigma); (  ) isoprenaline bitartrate (Sigma); mepyramine maleate (Tocris Cookson; Bristol, UK); racemic methadone hydrochloride s (10 mg/ml, Physeptone , GSK); morphine sulphate (Hospira; Melbourne, VIC, Australia); naloxone hydrochloride (Sigma); nifedipine (Sigma Aldrich, St Louis, Missouri, USA); Nω-nitro-L-

Heart rate Dose 1

Dose 2

Dose 3

400

Heart rate beats/min

vessels were precontracted with K þ 62 mM. Single morphine (10  7–10  3.5 M) or methadone (10  7–10  3 M) cumulative concentration–response curves were then constructed in the absence (control) or presence of (i) Nωnitro-L-arginine (10  4 M), to inhibit nitric oxide synthase; (ii) indomethacin (10  5.5 M), to inhibit cyclooxygenase-1 and -2; or (iii) naloxone (10  5 M). Tissues were equilibrated for 30 min with the respective antagonist before opioid agonist addition. For comparison with an L-type voltage-operated calcium channel (Cav1.2) antagonist, we tested nifedipine concentration–response curves in K þ 62 mM precontracted arteries.

350

* 300

250

* Saline 2.5 ml/kg n = 6 Morphine 9, 30, 90 mg/kg n = 6 Methadone 3, 10, 30 mg/kg n = 6

200 -10

0

10

20

30

40

50

60

Mean arterial pressure

120

100

MAP mmHg

374

80

* *

60

40 -10

0

10

20

30

40

50

60

Time min Fig. 1. Effects of opioid agonists morphine and methadone on heart rate (top panel) and mean arterial pressure (MAP; lower panel) in anaesthetised rats. Vehicle volumes (0.9% saline) or cumulative doses of opioid agonist were administered i.v. at 0, 21 and 42 min (Doses 1, 2 and 3, respectively). Error bars are average S.E.M. from repeated measures analysis of variance (RM ANOVA; see Methods). nPo 0.05 compared to saline (vehicle) group (RM ANOVA). n, number of rats.

arginine (Sigma); and sodium pentobarbitone (Sigma). Nifedipine was dissolved in dimethylsulfoxide and diluted in water. 2.4. Statistics and analyses Data are presented as mean 7S.E.M. The vertical error bars shown on Figs. 1–3 and 5–8 are average S.E.M. (except where indicated) calculated from repeated measured ANOVA (RM ANOVA). Average S.E.M. was calculated as [error mean square/number of animals (or tissues)]0.5 after subtracting the sums of the squares “between animals (or tissues)” and “between times (or concentrations)” from the “total” sum of squares for each treatment (Wright et al., 1987, 2002). Results were analysed by RM ANOVA with Greenhouse–Geisser correction (Ludbrook, 1994), with post-hoc contrast where appropriate, calculated using the statistical programme SuperANOVA™ 1.11 for Macintosh. Baseline values of cardiovascular variables were compared between groups by one-way ANOVA, with Dunnett’s post-hoc test (InStat, GraphPad Software, San Diego, USA). In addition to analysing the absolute values for HR and MAP the delta values were computed taking the zero baseline for all treatment groups and saline as the value of HR and MAP at time 0 min. In the saline control group, the HR and MAP increased over the 60 min by 197 7 beats/min and 14 75 mmHg, thus the

K.H. Tung et al. / European Journal of Pharmacology 762 (2015) 372–381

To compare the potency of morphine and methadone in cardiac or vascular tissues in vitro, the negative log10 of each agonist concentration that caused 50% of the maximum response, the pEC50, was analysed (where applicable). In papillary muscle, the pEC50 could only be estimated, as there was no clear upper plateau in each agonist concentration–ARP response curve. pEC50 values were compared between groups by unpaired Student’s t-test. Responses were calculated as % of the maximum contraction to KPSS 100% in each artery or of the maximum inotropic response to isoprenaline (0.1 μM) in the left atrium. The right atrial rate (beats/ min) and papillary muscle refractory period (ms) were measured directly without normalisation. P values o0.05 were considered statistically significant in all cases.

QTc interval - Fridericia's correction 0.06

Saline 2.5 ml/kg n = 5

QTc interval

Morphine 9, 30, 90 mg/kg n = 4

0.05

Methadone 3, 10, 30 mg/kg n = 6

0.04

Methadone 3, 10, 30 mg/kg + naloxone 10 mg/kg n = 4

* *

Naloxone 10 mg/kg n = 3

0.03 0.02 0.01 0.00 -0.01 -10

0

10

20

30

40

50

60

3. Results

Time min Fig. 2. Effects of opioid agonists on QTc interval calculated using Fridericia’s correction (QTc ¼ QT  RR  0.33) shown as change (Δ) from baseline (0 min) in anaesthetised rats. Vehicle volumes (0.9% saline; alone or with naloxone 10 mg/kg pretreatment) or cumulative doses of morphine or methadone (alone or with naloxone 10 mg/kg pre-treatment) were administered i.v. at 0, 21 and 42 min (indicated by arrows). Error bars are average S.E.M. from RM ANOVA. nPo 0.05 compared to vehicle group (RM ANOVA). n, number of rats.

3.1. In vivo experiments 3.1.1. Effects of opioid agonists on MAP, HR and QTc in rats Prior to administration of any treatment or vehicle, baseline (time 0 min) values for HR, MAP and QTc interval were similar in all rat treatment groups (P 40.05, one-way ANOVA). In the salinetreated rats, there was no significant change in HR (P40.05), however, there was a gradual small increase in MAP over the 62 min experimental period (P ¼0.03, RM ANOVA; Fig. 1), presumably due to the waning depressor effect of pentobarbitone. Morphine decreased HR in a dose-dependent manner (P o0.0001, RM ANOVA; Fig. 1); maximum changes (from the 0 min baseline)

delta increase or decrease for HR and MAP in the saline control group was deducted at the same time point from the plateau responses to each of the 3 doses of morphine and methadone (see Figs. 4 and 5). These plateau time points were taken 20 min after the i.v. dose of agonist to avoid the transient effects of the bolus.

Heart rate

Heart rate 450

450

Heart rate beats/min

Dose 1

Dose 2

Dose 3

Dose 1

Dose 2

Dose 3

400

400

* 350

350

300

300

250

-10

0

10

20

30

* *

250

Saline 2.5 ml/kg n = 6 Saline + naloxone 10 mg/kg n = 4 Morphine 9, 30, 90 mg/kg n = 6 Morphine + naloxone 10 mg/kg n = 3

200

Methadone 3, 10, 30 mg/kg n = 6 Methadone + naloxone 10 mg/kg n = 4 Methadone + naloxone 20 mg/kg n = 3

200 40

50

-10

60

0

10

20

30

40

50

60

Mean arterial pressure

Mean arterial pressure 120

120

100

MAP mmHg

375

*

100

80

80

60

60

* *

40

40 -10

0

10

20

30

Time min

40

50

60

-10

0

10

20

30

40

50

60

Time min

Fig. 3. Effects of saline (2.5 ml/kg), morphine (9, 30 and 90 mg/kg) (left panels) or methadone (3, 10 and 30 mg/kg) (right panels) on heart rate (top panels) and mean arterial pressure (MAP; lower panels) in the absence or presence of the μ-opioid receptor antagonist naloxone ( þ naloxone, 10 or 20 mg/kg) in anaesthetised rats. Vehicle volumes (0.9% saline) or cumulative doses of opioid agonist were administered i.v. at 0, 21 and 42 min (Doses 1, 2 and 3, respectively). Error bars are average S.E.M. from RM ANOVA. n P o0.05 compared to respective no pre-treatment group (RM ANOVA). n, number of rats.

376

K.H. Tung et al. / European Journal of Pharmacology 762 (2015) 372–381

HR beats/min

Plateau

in heart rate

0 -40 Morphine alone n = 6 Morphine + naloxone 10 mg/kg n = 3

-80

Methadone alone n = 6 Methadone + naloxone 10 mg/kg n = 4

-120

Methadone + naloxone 20 mg/kg n = 3

-160 1

Plateau

2

3

in mean arterial pressure

MAP mmHg

15 0 -15 -30 -45 -60 1

2

3

Agonist dose number i.v. Fig. 4. Plateau changes (approx. 20 min post-dose) in heart rate (top) and mean arterial pressure (bottom) from time 0 min for each treatment group to explore the effect of naloxone (10 or 20 mg/kg) μ-opioid receptor antagonism on the responses to the 3 cumulative doses of morphine (9, 30 and 90 mg/kg) and methadone (3, 10 and 30 mg/kg). The average changes 7 1 S.E.M. (those not shown are contained within the symbol) were corrected for the contemporaneous change in respective saline or saline plus naloxone 10 mg/kg group.

were  317 8, 46 717 and  647 18 beats/min with 9, 30 and 90 mg/kg, respectively (n ¼6). Methadone caused a much larger dose-dependent bradycardia than morphine, with peak effects of 53 78, 96 77 and  1457 6 beats/min with 3, 10 and 30 mg/ kg, respectively (n ¼6), occurring 4 min after each dose (P o0.0001, RM ANOVA; Fig. 1). Morphine caused falls in MAP (P o0.0001, RM ANOVA), however the maximum decrease of  2975 mm Hg (n ¼6) was observed with the first dose, 9 mg/kg, and did not change with additional doses (Fig. 1). Similar to the pattern observed with HR, methadone caused initial large dose-dependent falls in MAP (within 3–4 min) of  40 75,  517 5 and  58 76 mm Hg with 3, 10 and 30 mg/kg, respectively (n ¼6; P o0.0001), followed by a partial recovery to a plateau response which appeared to be doseindependent. This plateau response compared closely to the effects of morphine on MAP (Fig. 1). Further, methadone (3–30 mg/ kg) widened the pulse pressure, i.e. the difference between systolic and diastolic blood pressure, notably due to a decrease in the latter (data not shown). The QTc interval was not affected by either saline or morphine (P 40.05, RM ANOVA), but was significantly and dose-dependently prolonged by methadone (Po0.0001, RM ANOVA; Fig. 2). The maximum QTc interval increase was þ0.044 70.011 ms (n ¼6), an approximately 60% increase from baseline (0.073 70.004 ms).

3.1.2. Effects of μ-opioid antagonist pre-treatment in rats Following pre-treatment with naloxone 10 mg/kg i.p. in the saline-treated rats, HR and MAP slowly increased over the experimental period as noted for saline alone (Fig. 3). When these effects were corrected for the plateau changes in HR and MAP following doses of morphine and methadone, there is clear evidence that naloxone at 10 mg/kg inhibited the agonist-induced bradycardia and hypotension for doses one and two (Fig. 4). However, at the third respective agonist dose, naloxone 10 mg/kg did not prevent the falls in HR or MAP (Po0.05 each, RM ANOVA; Figs. 3 and 4). Doubling the naloxone dose (20 mg/kg) still did not inhibit the third dose (30 mg/kg) of methadone. For the acute falls in both HR and MAP following doses 1–3 (3– 30 mg/kg) of methadone, naloxone appeared to have little effect. Similarly, naloxone 10 mg/kg (Fig. 2) or 20 mg/kg (data not shown) had no effect on the methadone-induced QTc prolongation. 3.1.3. Effects of H1- and H2-histamine receptor antagonist pretreatment in rats Mepyramine, a selective H1-histamine receptor antagonist, did not affect HR, MAP or QTc interval in rats treated with saline (P 40.05; data not shown). Mepyramine attenuated the plateau phase of morphine-induced bradycardia, but did not affect the peak fall in HR or MAP (Fig. 5). With the combined administration of cimetidine and mepyramine, both peak and plateau HR

K.H. Tung et al. / European Journal of Pharmacology 762 (2015) 372–381

Heart rate

Heart rate Dose 1

Heart rate beats/min

400

Dose 2

Dose 3

Dose 1

400

Dose 2

Dose 3

* * 350

350

300

300

250

250 Methadone 3, 10, 30 mg/kg n = 6

Morphine 9, 30, 90 mg/kg n = 6

Methadone + mepyramine 10 mg/kg n = 4

Morphine + mepyramine 10 mg/kg n = 3 Morphine + mepyramine + cimetidine 100 mg/kg n = 3

200 -10

0

10

20

30

40

Methadone + mepyramine + cimetidine 100 mg/kg n = 4

200 50

60

Mean arterial pressure

120

MAP mmHg

377

-10

0

100

80

80

60

60

40

20

30

40

50

60

50

60

Mean arterial pressure

120

100

10

40

-10

0

10

20

30

40

50

60

Time min

-10

0

10

20

30

40

Time min

Fig. 5. Pre-treatment with histamine H1 (mepyramine) or H1 þH2 (mepyramineþ cimetidine) receptor antagonists on effects of morphine (9, 30 and 90 mg/kg; left panels) or methadone (3, 10 and 30 mg/kg; right panels) on heart rate (top panels) and mean arterial pressure (MAP; lower panels) in anaesthetised rats. The histamine receptor antagonist(s) was administered i.p. at time -20 min. Cumulative doses of opioid agonist were administered i.v. at 0, 21 and 42 min (Doses 1, 2 and 3, respectively). Error bars are average S.E.M. from RM ANOVA. nP o0.05 compared to morphine alone group (RM ANOVA). n, number of rats.

responses to morphine were attenuated (P o0.05; RM ANOVA; Fig. 5). Neither the H1-histamine antagonist alone nor in combination with the H2-receptor antagonist affected the bradycardia and hypotension peak or plateau responses to methadone (P 40.05, RM ANOVA; Fig. 5). The peak QTc prolongation caused by the highest cumulative dose of methadone (30 mg/kg) tended to be augmented in the presence of mepyramine or the combination of mepyramine and cimetidine (þ 0.057 70.004 and þ0.060 70.005, respectively), however this was not statistically significant (P 40.05, 1-way ANOVA; data not shown). 3.2. Effects of opioid agonists in vascular and cardiac tissues in vitro 3.2.1. Effects of opioid agonists on rat mesenteric arterial tone Methadone caused concentration-dependent relaxation of endothelium-intact, endothelin-1 pre-contracted, rat mesenteric arteries with a pEC50 4.5 70.1 (n ¼4; P o0.0001, RM ANOVA; Fig. 6A). In contrast, morphine did not cause relaxation, but instead a slight but significant contraction (pEC50 5.1 70.3; P o0.0001). Methadone-induced vascular relaxation was unaffected by pre-incubation of tissues with the nitric oxide (NO) synthase inhibitor Nω-nitro-L-arginine 10  4 M (pEC50 4.6 7 0.2), the cyclooxygenase (COX) inhibitor indomethacin 10  5.5 M (pEC50 4.4 70.1) or naloxone 10  5 M (pEC50 4.4 70.1) (each n ¼4; P 40.05, RM ANOVA; Fig. 6B–D). These findings indicate that this relaxation to methadone was not due to μ-opioid receptor activation, or release of NO or prostaglandins. In mesenteric arteries, depolarised and contracted by K þ 62 mM, again morphine caused a slight rise in force up to 91 710% KPSS maximum (n¼ 5; P ¼0.002, RM ANOVA), while methadone

was 5.8 times more potent at relaxing the K þ -precontracted arteries (pEC50 5.26 70.06, n¼ 5; Fig. 7A) than when arteries were precontracted by endothelin-1 (Fig. 6A). 3.2.2. Effects of opioid agonists on rat atria and papillary muscle Methadone and morphine decreased spontaneous right atrial rate in a concentration-dependent manner (n ¼4 each; P o0.0001, RM ANOVA; Fig. 8A), with pEC50 values of 4.86 70.07 and 3.54 70.10, respectively. Atrial rate fell from a baseline of 285 714 to 12278 beats/min with methadone 10  4 M, and from 257 715 to 184 79 beats/min with morphine 10  3 M. Methadone was at least 20 fold more potent than morphine at causing bradycardia (n ¼4; P ¼0.02, unpaired t-test). In separate right atria, the effects of methadone or morphine were tested in the presence of naloxone 10 μM. The μ-opioid receptor antagonist had no effect on the bradycardic responses to either methadone (Fig. 8A) or morphine (data not shown). Methadone or morphine (1–30 μM) did not cause any significant change in the ionotropic response to isoprenaline (0.01 μM; P ¼0.05 and 0.06 respectively, RM ANOVA; Fig. 7C). For comparison, nifedipine was E270 times more potent at relaxing the K þ -depolarised rat mesenteric artery (pEC50 9.15 70.08, n ¼4; Fig. 7B) compared with its negative inotropic action in rat left atria (pEC50 6.72 70.16, n ¼3; Fig. 7D). Methadone markedly prolonged the absolute refractory period (ARP) in rat isolated papillary muscle with a threshold of 10  4.5 M and an estimated pEC50 of 4.04 70.13 (n ¼4; Fig. 8B). The ARP increased from the baseline of 6576 to 230 78 ms with methadone 10  3.5 M. In contrast, morphine had no effect until 10  3 M, when the ARP was prolonged to 179 716 ms; the estimated pEC50 was 3.29 70.06 (n ¼4; Fig. 8B). Methadone was significantly more

378

K.H. Tung et al. / European Journal of Pharmacology 762 (2015) 372–381

No pretreatment

% KPSS tone

120

120

100

100

80

80

60

60

40

40

20

Nw-nitro-L-arginine pretreatment

20 Control n = 4

Morphine n = 4 Methadone n = 4

0 -7

-6

-5

Nw-nitro-L-arginine 10-4 M n = 4

0

-4

-3

-7

Indomethacin pretreatment

% KPSS tone

120

-6

-5

-4

-3

Methadone log [M]

Agonist log [M]

Naloxone pretreatment

120

100

100

80

80

60

60

40

40 20

20 0 -7

Control n = 4

Control n = 4

Indomethacin 10-5.5 M n = 4

Naloxone 10-5 M n = 4

-6

-5

-4

Methadone log [M]

0 -3

-7

-6

-5

-4

-3

Methadone log [M]

Fig. 6. (A) Effects of morphine or methadone on vascular tone (expressed as % KPSS tone) in rat isolated small mesenteric arteries precontracted with endothelin-1. (B–D) Single vascular relaxant effects of methadone in the absence (Control) or presence of the (B) nitric oxide synthase inhibitor Nω-nitro-L-arginine 10  4 M; (C) COX inhibitor indomethacin 10  5.5 M; or (D) μ-opioid antagonist naloxone 10  5 M. Each artery was precontracted with endothelin-1 to a level of 70–100% of the response to KPSS before the start of the morphine or methadone concentration–response curve (only a single curve was constructed per tissue). Vertical error bars are average S.E.M. from RM ANOVA; horizontal error bars represent the EC50 7 S.E.M. n, number of vessels from separate rats.

potent than morphine at ARP prolongation based on these estimates of pEC50 (P ¼0.002, unpaired t-test).

4. Discussion The primary goal of this work was to define the mechanisms of the in vivo cardiovascular actions of morphine and methadone in the anaesthetised rat by using selective receptor antagonists and secondary experiments in isolated tissues. Our central finding is that i.v. bolus doses of morphine and methadone that effectively lower HR and MAP at steady state are antagonised by naloxone. Thus the steady state hypotension and bradycardia actions are likely to be μ-opioid-dependent. Our design of using 3 cumulative doses of opioids allowed for the analysis of surmountable

antagonism as the doses of naloxone at 10 and 20 mg/kg i.p. were insufficient to block the effects of morphine at 90 mg/kg or methadone at 30 mg/kg. Naloxone doses of 5 mg/kg i.p. in rats block the antiallodynic effects of morphine 10–30 mg/kg i.p. (Bian et al., 1995). Thus the range of doses of 5–20 mg/kg of naloxone should indicate a μ-opioid-sensitive mechanism. The careful correction for the time-dependent increases in resting MAP and HR in the control saline-injected rats over the 60 min test period allowed more accurate assessment of the opioids’ effects as shown by the delta changes in MAP and HR at plateau after each dose. Of note, the acute (3–4 min) bradycardic and hypotensive effects caused by each dose of methadone were not affected by naloxone and thus unlikely to be μ-opioid receptor-dependent. This rapid acute response to methadone may be due to L-type voltage-operated calcium channel antagonism, an action not shared with morphine.

K.H. Tung et al. / European Journal of Pharmacology 762 (2015) 372–381

% KPSS tone

100

Mesenteric artery: K+ 62 mM tone

100

80

80

60

60

40

40

20

20 Morphine n = 5 Methadone n = 5

0 -7.0

-6.5

-6.0

Mesenteric artery: K+ 62 mM tone

Nifedipine n = 4

0 -5.5

-5.0

-4.5

-10.0

Left atria

% Maximum response

80

60

60

40

40

20

20

Morphine n = 3 Methadone n = 3 -6.0

-5.5

-9.0

-8.0

-7.5

Nifedipine n = 3

0

-5.0

-8.5

Left atria

100

80

0

-9.5

Nifedipine log [M]

Agonist log [M]

100

379

-4.5

Agonist log [M]

-8.0

-7.5

-7.0

-6.5

-6.0

Nifedipine log [M]

Fig. 7. Effects of morphine or methadone (A) or nifedipine (B) on vascular tone (expressed as % KPSS tone) in rat isolated small mesenteric arteries. Each artery was precontracted with K þ 62 mM to a level of 70–80% of the response to KPSS before the start of the morphine, methadone or nifedipine concentration–response curve (only a single curve was constructed per tissue). (C) Effects of morphine or methadone or (D) nifedipine on inotropy in rat isolated left atria (expressed as % maximum response to isoprenaline 0.1 μM). Each atrium was stimulated with a submaximal concentration of isoprenaline (0.01 μM) before cumulative concentration–response curves were constructed to morphine, methadone or nifedipine (one curve per atrium). Vertical error bars are average S.E.M. from RM ANOVA; horizontal error bars represent the EC50 7 S.E.M. (where applicable). n, number of tissues from separate rats.

Further in vivo experiments with the combined histamine H1and H2-receptor antagonists revealed that only the sustained bradycardia following morphine was significantly attenuated, while the hypotension following morphine and both the bradycardia and acute/plateau hypotension responses following methadone were largely unaffected. The sustained hypotensive and bradycardia responses in vivo to morphine and methadone are most likely to be mediated by μopioid-sensitive actions given the effects of naloxone. There is evidence that in dogs morphine suppresses central sympathetic outflow (Muldoon et al., 1987) and this may be caused by μopioid-induced histamine release (McGrath and Shepherd, 1976). There is also evidence of the role of μ2-opioid receptors on the vagus causing bradycardia in conscious rats in response to morphine (Paakkari et al., 1992). Similarly, in dogs (Atanackovic and de Jongh, 1950) and in man (Eckenhoff and Oech, 1960) morphine

enhanced vagal tone. In our experiments rats were anaesthetised with pentobarbitone, a known vagolytic anaesthetic agent, and following atropine (1 mg/kg s.c.) there would be little if any vagal μ-opioid effect. Nevertheless the 40–60 beats/min fall in HR for morphine and methadone were completely prevented by naloxone (Fig. 4). To determine if this bradycardia was directly at the sinoatrial node, we used rat isolated right atria to show that methadone was indeed a potent direct-acting bradycardic agent some 21-fold more potent than morphine (Fig. 8A). This bradycardia was insensitive to naloxone and may again be due to L-type voltageoperated calcium channel inhibition. There is substantial evidence that μ-opioid agonists stimulate histamine release in vivo (Brashear et al., 1974; Rosow et al., 1982; Schurig et al., 1978; Thompson and Walton, 1964). While direct measurements of increases in histamine in blood following intravenous morphine in humans was difficult to show until 1982

380

K.H. Tung et al. / European Journal of Pharmacology 762 (2015) 372–381

Right atrium

Papillary muscle

300

240

Refractory period ms

Atrial rate beats/min

210

250

200

150 Morphine n = 4 Methadone + naloxone 10 M n = 3 BL -7

-6

150 120 90

Methadone n = 4 100

180

-5

-4

60

-3

Agonist log [M]

BL -7

-6

-5

-4

-3

Agonist log [M]

Fig. 8. Effects of morphine or methadone on (A) rate in isolated right atria and (B) absolute refractory period in isolated papillary muscle. In (A), methadone was also tested in the presence of naloxone 10 μM. BL, baseline values ( 7 S.E.M.; error bars not visible are contained within the symbol) prior to addition of agonist. Vertical error bars on lines are average S.E.M. from RM ANOVA. Horizontal error bars represent estimated EC50 7 S.E.M. (where applicable). n, number of tissues from separate rats.

(Rosow et al., 1982), the rat isolated peritoneal mast cell preparation first described by Uvnas and Thon (1959) has been a reliable qualitative rather than quantitative in vitro assay to examine histamine release. Moss and Rosow (1983) noted that relatively high concentrations of test drugs are required to demonstrate release in rat isolated mast cells compared with clinical practise. For example, Ellis et al. (1970) showed threshold histamine-releasing concentrations of 3  10  3 M for morphine and Grosman (1981) found that methadone (3  10  4 M) was 10 fold more sensitive compared with morphine. Histamine release from basophils and platelets would be additional sources of histamine in vivo. Our findings that combined histamine H1- and H2-receptor antagonism was required to inhibit the HR falls to morphine in the rat were in accord with i.v. morphine results in patients undergoing coronary bypass surgery (Philbin et al., 1981). In contrast, there are reports that methadone does not consistently release histamine in humans. At 20 mg i.v. (0.3 mg/kg), a dose that causes prolonged postoperative analgesia, only 2 of 23 patients had marked increases in plasma histamine and no obvious haemodynamic effects were observed (Bowdle et al., 2004). Our results confirm that histamine H1- and H2-receptor antagonists do not significantly attenuate the falls in heart rate and blood pressure to methadone in the rat. The role of μ-opioid receptor activation by morphine in causing local release of mast cell histamine, tryptase and protein as measured by microdialysis was questioned when locally coadministered naloxone did not block the autacoid release or vasodilatation (Blunk et al., 2004). However, only a single dose of naloxone was tested at a typical clinical dose. Again in vivo in the rat, QTc measurement revealed that the rapid onset methadone-induced prolongation of the QT interval was highly significant at the 3 i.v. doses and unaffected by the μopioid antagonist naloxone. Moreover the isolated papillary muscle experiments that were used to measure the absolute refractory period, showed a 30 fold higher sensitivity to methadone than morphine for lengthening refractory period, and 21 fold in isolated atria, slowing spontaneous atrial rate. The 21–30 fold higher sensitivity to methadone compared with morphine on atrial refractoriness and slowing atrial rate was significantly higher than the 3 fold difference in equivalent doses causing hypotension.

These differences may suggest that methadone’s “direct” atrial slowing is a different action to that causing QT prolongation and refractoriness in the rat. The isolated mesentery artery experiments revealed that morphine and methadone have quite different ‘direct’ effects on vascular tone induced by endothelin-1. The isolated vessels would have few mast cells attached in this preparation. In agreement, Lee and Berkowitz (1977), using rat aortic strips, showed morphine 10  5–3  10  4 M contracted the mesenteric small artery while methadone relaxed the vessel over the same concentration range. Lee and Berkowitz (1977) concluded that methadone had a ‘calcium antagonist’ action to cause this relaxation. Stefano et al. (1995) concluded that morphine activated a novel μ3 opiate receptor on endothelial cells to release NO and cause relaxation of rat aortic rings precontracted by phenylephrine. They measured NO release directly from rat aortic endothelial cells and showed that naloxone (5  10  6 M) abolished the NO signal and relaxation stimulated by morphine (1 μM). In our hands, the mesenteric artery relaxation by methadone was not mediated by opioid receptors, NO or prostaglandins (i.e. insensitive to naloxone 10  5 M, Nω-nitro-L-arginine or indomethacin, respectively). We chose this concentration of naloxone (10  5 M) to test the opioid mechanism of this relaxation as there is very strong evidence that low concentrations (0.03–0.1 μM) enhance the release of acetylcholine from the myenteric plexus of the guinea pig ileum through a specific block of an endogenous ligand of the opioid receptor (Waterfield and Kosterlitz, 1975). Moreover, augmentation of H þ production in rat isolated parietal cells by methionine-enkephalin is specifically blocked by 1–10 μM naloxone stereoselectively and the authors caution using concentrations greater than 10 μM when demonstrating specific opioid antagonism (Schepp et al., 1986). We cannot confirm any role for NO, or other autacoid, in mediating the vascular relaxation by methadone. However, our work with isolated mesenteric arteries suggests that the relaxation of the K þ precontracted vessels following methadone is due to L-type voltage-operated calcium channel antagonism. In supporting this conclusion, in the endothelin precontracted arteries the IC50 for methadone was 5.8 times greater than when the arteries were precontracted by K þ depolarisation. This finding favours the notion that L-type voltage-operated calcium channel antagonists

K.H. Tung et al. / European Journal of Pharmacology 762 (2015) 372–381

are more potent against K þ depolarised precontracted arteries than arteries precontracted by receptor-operated calcium channel agonists (Angus et al., 1983). In addition, as with nifedipine, methadone was vascular to cardiac selective in these assays of mesenteric arteries and left atria stimulated with isoprenaline. These in vitro experiments raise the important question as to the relative range of blood concentration that would be reached in the in vivo experiments. Our estimations, assuming a blood volume in the rat of 60 ml/kg body weight and plasma half life of more than 2 h, suggest for 9, 30, 90 mg/kg i.v. morphine sulphate, a peak blood concentration of 3.95  10  4–9.95  10  3 M. This range would be decreased by 35% for protein binding, i.e. to 2.6  10  4–2.6  10  3 M. For methadone hydrochloride, the 3– 30 mg/kg dose range could theoretically reach 1.62  10  4–1.62  10  3 M and with 80% plasma protein binding (range 71–87% human plasma protein in vitro (Olsen, 1973)), the unbound range would be 3.2  10  5–3.2  10  4 M. The estimations of free morphine and methadone concentration in blood correlate very well with the in vitro range of concentration–response curves assessed here. Thus, putting the in vitro and in vivo data together shows that the acute bradycardia, QT prolongation and hypotension observed with methadone are probably caused by the non μ-opioid receptor-mediated actions of calcium channel antagonism, as the high blood concentration would exceed 10  5 M. In conclusion, i.v. morphine and methadone cause μ-opioidsensitive sustained hypotension in the rat. But in addition, methadone has acute and dramatic short-lived effects to prolong the QT interval, slow atrial rate and cause vasodilatation. This vasodilatation is most likely due to L-type voltage-operated calcium channel inhibition.

Conflicts of Interest None.

Acknowledgements We thank Mr Mark Ross-Smith for assistance with some of the myography and isolated atria experiments.

References Angus, J.A., Brazenor, R.M., Le Duc, M.A., 1983. Responses of dog large coronary arteries to constrictor and dilator substances: implications for the cause and treatment of variant angina pectoris. Am. J. Cardiol. 52, 52A–60A. Angus, J.A., Harvey, K., 1981. Refractory period field stimulation of right atria: a method for studying presynaptic receptors in cardiac autonomic transmission. J. Pharmacol. Methods 6, 51–64. Angus, J.A., Wright, C.E., 2000. Techniques to study the pharmacodynamics of isolated large and small blood vessels. J. Pharmacol. Toxicol. Methods 44, 395–407. Atanackovic, D., de Jongh, D., 1950. Action of methadone on blood pressure regulating mechanisms. Arch. Int. Pharmacodyn. Ther. 82, 494–499. Bian, D., Nichols, M.L., Ossipov, M.H., Lai, J., Porreca, F., 1995. Characterization of the antiallodynic efficacy of morphine in a model of neuropathic pain in rats. Neuroreport 6, 1981–1984. Blunk, J.A., Schmelz, M., Zeck, S., Skov, P., Likar, R., Koppert, W., 2004. Opioid-induced mast cell activation and vascular responses is not mediated by m-opioid receptors: an in vivo microdialysis study in human skin. Anesth. Analg. 98, 364–370. Bowdle, T.A., Even, A., Shen, D.D., Swardstrom, M., 2004. Methadone for the induction of anesthesia: plasma histamine concentration, arterial blood pressure, and heart rate. Anesth. Analg. 98, 1692–1697.

381

Brashear, R.E., Kelly, M.T., White, A.C., 1974. Elevated plasma histamine after heroin and morphine. J. Lab. Clin. Med. 83, 451–457. Eckenhoff, J.E., Oech, S.R., 1960. The effects of narcotics and antagonists upon respiration and circulation in man. A review. Clin. Pharmacol. Ther. 1, 483–524. Ellis 3rd, H.V., Johnson, A.R., Moran, N.C., 1970. Selective release of histamine from rat mast cells by several drugs. J. Pharmacol. Exp. Ther. 175, 627–631. EMEA, 2006. S7B The Nonclinical Evaluation of the Potential for Delayed Ventricular Repolarization (QT interval prolongation) by Human Pharmaceuticals. European Medicines Agency, London. Fridericia, L.S., 1920. Die systolendauer im elektrokardiogramm bei normalen menschen und bei herzkranken. Acta Med. Scand. 57, 469–486. Grosman, N., 1981. Histamine release from isolated rat mast cells: effect of morphine and related drugs and their interaction with compound 48/80. Agents Actions 11, 196–203. Katchman, A.N., McGroary, K.A., Kilborn, M.J., Kornick, C.A., Manfredi, P.L., Woosley, R.L., Ebert, S.N., 2002. Influence of opioid agonists on cardiac human ether-ago-go-related gene K(þ ) currents. J. Pharmacol. Exp. Ther. 303, 688–694. Lee, C.H., Berkowitz, B.A., 1977. Calcium antagonist activity of methadone, L-acetylmethadol and L-pentazocine in the rat aortic strip. J. Pharmacol. Exp. Ther. 202, 646–653. Ludbrook, J., 1994. Repeated measurements and multiple comparisons in cardiovascular research. Cardiovasc. Res. 28, 303–311. McGrath, M.A., Shepherd, J.T., 1976. Inhibition of adrenergic neurotransmission in canine vascular smooth muscle by histamine: mediation by H2-receptors. Circ. Res. 39, 566–573. Moss, J., Rosow, C.E., 1983. Histamine release by narcotics and muscle relaxants in humans. Anesthesiology 59, 330–339. Muldoon, S.M., Freas, W., Mahla, M.E., Donlon, M.A., 1987. Plasma histamine and catecholamine levels during hypotension induced by morphine and compound 48/80. J. Cardiovasc. Pharmacol. 9, 578–583. Mulvany, M.J., Halpern, W., 1977. Contractile properties of small arterial resistance vessels in spontaneously hypertensive and normotensive rats. Circ. Res. 41, 19–26. Olsen, G.D., 1973. Methadone binding to human plasma proteins. Clin. Pharmacol. Ther. 14, 338–343. Paakkari, P., Paakkari, I., Feuerstein, G., Siren, A.L., 1992. Evidence for differential opioid m 1- and m 2-receptor-mediated regulation of heart rate in the conscious rat. Neuropharmacology 31, 777–782. Pasternak, G.W., Wood, P.J., 1986. Multiple m opiate receptors. Life Sci. 38, 1889–1898. Philbin, D.M., Moss, J., Akins, C.W., Rosow, C.E., Kono, K., Schneider, R.C., VerLee, T.R., Savarese, J.J., 1981. The use of H1 and H2 histamine antagonists with morphine anesthesia: a double-blind study. Anesthesiology 55, 292–296. Rosow, C.E., Moss, J., Philbin, D.M., Savarese, J.J., 1982. Histamine release during morphine and fentanyl anesthesia. Anesthesiology 56, 93–96. Schepp, W., Schneider, J., Schusdziarra, V., Classen, M., 1986. Naturally occurring opioid peptides modulate H þ -production by isolated rat parietal cells. Peptides 7, 885–890. Schurig, J.E., Cavanagh, R.L., Buyniski, J.P., 1978. Effect of butorphanol and morphine on pulmonary mechanics, arterial blood pressure and venous plasma histamine in the anesthetized dog. Arch. Int. Pharmacodyn. Ther. 233, 296–304. Stefano, G.B., Hartman, A., Bilfinger, T.V., Magazine, H.I., Liu, Y., Casares, F., Goligorsky, M.S., 1995. Presence of the m3 opiate receptor in endothelial cells. Coupling to nitric oxide production and vasodilation. J. Biol. Chem. 270, 30290–30293. Thompson, R.C., Mansour, A., Akil, H., Watson, S.J., 1993. Cloning and pharmacological characterization of a rat m opioid receptor. Neuron 11, 903–913. Thompson, W.L., Walton, R.P., 1964. Elevation of plasma histamine levels in the dog following administration of muscle relaxants, opiates and macromolecular polymers. J. Pharmacol. Exp. Ther. 143, 131–136. Uvnas, B., Thon, I.L., 1959. Isolation of “biologically intact” mast cells. Exp. Cell Res. 18, 512–520. Waterfield, A.A., Kosterlitz, H.W., 1975. Stereospecific increase by narcotic antagonists of evoked acetylcholine output in guinea-pig ileum. Life Sci. 16, 1787–1792. Wolozin, B.L., Pasternak, G.W., 1981. Classification of multiple morphine and enkephalin binding sites in the central nervous system. Proc. Natl. Acad. Sci. USA 78, 6181–6185. Wright, C.E., Angus, J.A., Korner, P.I., 1987. Vascular amplifier properties in renovascular hypertension in conscious rabbits. Hindquarter responses to constrictor and dilator stimuli. Hypertension 9, 122–131. Wright, C.E., Angus, J.A., Korner, P.I., 2002. Structural factors increase blood pressure through the interaction of resistance vessel geometry with neurohumoral and local factors: estimates in rabbits with renal cellophane-wrap hypertension with intact effectors and during neurohumoral blockade. J. Hypertens. 20, 471–483. Zhou, Z., Gong, Q., Ye, B., Fan, Z., Makielski, J.C., Robertson, G.A., January, C.T., 1998. Properties of HERG channels stably expressed in HEK 293 cells studied at physiological temperature. Biophys. J. 74, 230–241.