Modulation of mean arterial pressure and diuresis by renomedullary infusion of a selective inhibitor of fatty acid amide hydrolase

Abstract

The kidneys contribute to the control of body fluid and electrolytes and the long-term regulation of blood pressure through various systems, including the endocannabinoid system. Previously, we showed that inhibition of the two major endocannabinoid-hydrolyzing enzymes, fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase, in the renal medulla increased the rate of urine excretion (UV) and salt excretion without affecting mean arterial pressure (MAP). The present study evaluated the effects of a selective FAAH inhibitor, N-3-pyridinyl-4-[[3-[[5-(trifluoromethyl)-2-pyridinyl]oxy]phenyl]methyl]-1-piperidine carboxamide (PF-3845) on MAP and renal functions. Infusion of PF-3845 into the renal medulla of C57BL/6J mice reduced MAP during the posttreatment phases and increased UV at 15 and 30 nmol/min per gram kidney weight (g kwt), relative to the pretreatment control phase. Intravenous PF-3845 administration reduced MAP at the 7.5, 15, and 30 doses and increased UV at the 15 and 30 nmol⋅min⁻¹⋅g⁻¹ kwt doses. PF-3845 treatment elevated sodium and potassium urinary excretion and medullary blood flow. Homozygous FAAH knockout mice were refractory to intramedullary PF-3845-induced changes in MAP, but UV was increased. Both MAP and UV responses to intramedullary PF-3845 in C57BL/6J mice were diminished by pretreatment with the cannabinoid type 1 receptor-selective antagonist, rimonabant (3 mg/kg, ip) but not the cyclooxygenase 2-selective inhibitor, celecoxib (15 mg/kg, iv). Liquid chromatography-tandem mass spectrometry analyses showed increased anandamide in kidney tissue and 2-arachidonoyl glycerol in plasma after intramedullary PF-3845. These data suggest that inhibition of FAAH in the renal medulla leads to both a diuretic and blood pressure-lowering response mediated by elevated anandamide in kidney tissue or 2-arachidonoyl glycerol in plasma.

INTRODUCTION

The kidneys play a vital role in the long-term regulation of blood pressure by control of body fluid volume and electrolytes. Therefore, it is not surprising that disorders affecting the kidney or renal vasculature commonly lead to secondary form of hypertension (34). Kidneys modulate pressure and diuresis through multiple systems, including the renin angiotensin system (35), the sympathetic nervous system (6), and the endocannabinoid (EC) system (1). The roles of the EC system have been well studied in gastrointestinal tract disorders (36), peripheral neuropathic pain management (25), immunoregulatory role in infectious diseases (15), and the brain (31). However, little is known about specific mechanisms of EC regulation of renal functions and its role in hypertension.

The EC system consists of the two main cannabinoid receptor agonists, anandamide (AEA) and 2-arachidonylglycerol (2-AG), their hydrolyzing enzymes, fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL), and the cannabinoid receptors, cannabinoid type 1 (CB-1) and type 2 (CB-2). AEA is synthesized mostly by release from N-arachidonoyl phosphatidylethanolamine, mediated by N-arachidonoyl phosphatidylethanolamine-specific phospholipase D, and its agonist effect on CB receptors is controlled by FAAH-mediated metabolism to arachidonic acid and ethanolamine (23). In contrast, 2-AG is synthesized from membrane phospholipids by phospholipase C, β, and diacylglycerol lipase, and it undergoes hydrolysis by MAGL to form arachidonic acid and glycerol (12). Both EC agonists, AEA and 2-AG, have been a subject of great interest because of their capacities to activate cannabinoid receptors (10).

In the cardiovascular system, AEA is reported to have a vasodepressor activity and to cause bradycardia by inhibition of the baroreceptor reflex and sympathetic tone (33). In the kidney, intramedullary infusion of AEA was reported by our laboratory to stimulate diuretic and natriuretic effects with little or no effect on MAP in C57BL/6J mice (29). Furthermore, we also observed that diuretic and natriuretic effects of AEA were dependent on cyclooxygenase (COX)-2 (29). It has been observed that increasing the tissue concentrations of AEA by pharmacological inhibition or genetic ablation of FAAH resulted in augmented hypotensive action of exogenous AEA (26). It can be assumed that inhibiting the EC-hydrolyzing enzymes may produce similar effects as exogenous administration of ECs. There has been much interest in FAAH as a therapeutic target in the treatment of pain and central nervous system disorders (2), inflammation and pain (3), and hypertension (4, 5). More recently, we reported that administration of isopropyl dodecyl fluorophosphate (IDFP), a potent inhibitor of both FAAH and MAGL, into the renal medulla of C57BL/6J mice stimulated diuresis and natriuresis accompanied by increased medullary blood flow (MBF) but did not affect mean arterial pressure (MAP; 1). These effects were dependent on COX-2 but independent of COX-1, suggesting that the effect was dependent on conversion of AEA to COX-2-derived products [i.e., prostamide E₂ (PE₂)].

In this work we tested the hypothesis that infusion of a FAAH-selective inhibitor into the renal medulla would stimulate diuresis and natriuresis. For this purpose, we used N-3-pyridinyl-4-[[3-[[5-(trifluoromethyl)-2-pyridinyl]oxy]phenyl]methyl]-1-piperidine carboxamide (PF-3845), a highly selective FAAH inhibitor (3) and evaluated its effects on MAP, MBF, and urine formation rate (UV) after intramedullary or intravenous infusion. To determine the role of FAAH in the mechanism of PF-3845 effects, the responses were also evaluated in mice carrying homozygous knockout (KO) mutations in the FAAH gene (FAAH KO mice). We also tested the hypotheses that the responses produced by PF-3845 were dependent on metabolism of AEA by COX-2 and/or are mediated through cannabinoid receptors. For this purpose, the effects of PF-3845 were also assessed after pretreatment with celecoxib, a COX-2 inhibitor or rimonabant (SR141716), a CB-1 receptor antagonist. To correlate any observed functional changes elicited by PF-3845 with concentrations of possible EC mediators, AEA, 2-AG, and the COX-2 metabolite of AEA PE₂ were measured in kidney tissue and serum after intramedullary infusion of PF-3845.

MATERIALS AND METHODS

PF-3845 and rimonabant were purchased from Apex BioTech (San Jose, CA) and Cayman Chemical (Ann Arbor, MI), respectively. Celecoxib was from Sigma-Aldirch(St. Louis, MO). AEA, 2-AG, prostaglandin E₂, prostaglandin E₂-ethanolamide, and arachidonic acid, and their deuterated internal standards (AEA-d₈, 2-AG-d₅, prostaglandin E₂-d₄P, prostaglandin E₂-ethanolamide-d₄, and AA-d₈) were purchased from Cayman Chemical. Ammonium acetate, formic acid, sodium chloride, chloroform, HPLC-grade methanol, HPLC-grade acetonitrile, and HPLC-grade water were purchased from Fisher Scientific (Hanover Park, IL). All other reagents for in vivo use were of the highest grade available.

Animals.

Two- to four-month-old male C57BL/6J mice were obtained from Jackson Laboratory (Bar Harbor, ME), and male and female FAAH^+/+ (homozygous wild-type) and FAAH^−/− (homozygous gene KO) mice were obtained from a colony established at Virginia Commonwealth University by Dr. Aron Lichtman. The FAAH^−/− mice were maintained by backcrossing onto a C57BL/6 background for 20 generations. All mice used in experiments weighed 25–35 g and were housed 4–5 per cage in a temperature- (20°C–22°C) and humidity-controlled (50%–55%) facility with a 12:12-h light/dark cycle and ad libitum food and water. All animal protocols were approved by the Institutional Animal Care and Use Committee of Virginia Commonwealth University and were in concordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Acute surgical preparation.

Mice were prepared for acute renal function studies as described (1). Briefly, anesthesia was induced by intraperitoneal administration of ketamine (ketathesia, 100 mg/kg, Harry Schein Animal Health, Dublin, OH) and thiobutabarbital (Inactin, 75 mg/kg, Sigma-Aldrich). The body temperature of the mice was maintained by placing them on a surgical table that was thermally stable at 37°C. A tracheotomy was performed to facilitate breathing, which was followed by cannulation of the left carotid artery with a blood pressure probe for continuous monitoring of blood pressure and of the right jugular vein for constant rate infusion of vehicle and vehicle containing drugs into the systemic circulation. The blood pressure probe was attached to a pressure transducer connected to a data acquisition system (Windaq, DATAQ Instruments, Akron, OH).

The infusion of drugs into the renal medulla was through a catheter with a tapered tip (2.5 mm in length) inserted into the outer medulla from the dorsal side of the right kidney and anchored on the surface of the kidney using VetBond tissue adhesive (3M Co., Minneapolis, MN) as described (18, 29, 38). The infusion solution contained phosphate-buffered saline (205 mM NaCl, 40.5 mM Na₂HPO₄, and 9.5 mM NaH₂PO₄ (pH 7.4, 610 mOsM) and 10% ethanol at a rate of 2 μl/min. The left kidney ureter was ligated and cut proximal to the kidney. The bladder was cannulated to enable the collection of urine into preweighed tubes for gravimetric determination of UV. Urine sodium and potassium concentrations were measured by flame photometry to permit determination of the urinary sodium (U_Na) and potassium (U_K) excretion rates. All urinary parameters were calculated on a per gram kidney weight (g kwt) basis. Mice were infused through the jugular vein with a vehicle containing 0.9% NaCl solution and 2% bovine serum albumin at the rate of 5 μl/min per 25 g body weight to maintain fluid homeostasis in the body. MBF was measured using a laser Doppler flow probe (OxyFlo Probe, MNP 100XP, Oxford, UK) placed on the lower lateral surface of the kidney. After a 1-h equilibration period and establishment of a stable baseline, urine was collected every 10 min during the experimental period. The two 10-min samples collected immediately before the start of drug treatments, representing the pretreatment control periods, were designated C1 and C2. PF-3845 was administered into the renal medulla in the infusion solution at rates of 3.75, 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹ for 30 min at each dose. The mean value for each treatment was determined from the average between 10- and 30-min of the respective 30-min infusion period. The effects of rimonabant (3 mg/kg, ip bolus) and celecoxib (15 mg/kg, iv) on the responses to PF-3845 were determined using the same strategy. Sham control mice were treated identically to drug-treated mice except that they received intravenous and intramedullary vehicle solutions only for the entire experimental duration. At the termination of the experiment, blood was drawn from the carotid artery into a heparinized tube and centrifuged. The plasma and the right and left kidneys were collected, weighed, and stored at −80°C for later analyses. The position of the medullary catheter was confirmed after sagittal sectioning of the right kidney.

Analysis of AEA, 2-AG, and PE₂ by ultrahigh performance liquid chromatography-tandem mass spectrometry.

Internal standard solution (10 µl) containing 10 ng each of AEA-d₈, 100 ng 2-AG-d₅, and 0.1 ng PE₂-d₄ was added to each plasma (100 µl), kidney tissue (90–100 mg), or calibrator samples. Chloroform/methanol [3 ml of 2:1 (vol/vol)] and 0.73% sodium chloride (200 µl) were added to all samples and calibrators. The tissue samples were then homogenized for 30 s using a Brinkmann Polytron PT 3000 homogenizer and centrifuged at 3500 revolutions/min for 5 min. Plasma samples and calibrators were mixed for 5 min then centrifuged at 3500 revolutions/min for 5 min. The organic phases were collected, and the aqueous phases were extracted twice more with chloroform (1 ml). The organic phases were combined and evaporated to dryness under nitrogen, reconstituted in 100 µl 60:40 water/acetonitrile, and placed in autosampler vials for ultrahigh performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) analysis.

The UPLC-MS/MS analysis of AEA, 2-AG, and PE₂ was performed on a Sciex 6500 QTRAP system with an IonDrive Turbo V source for TurbolonSpray (Ontario, Canada) attached to a Shimadzu UPLC system (Kyoto, Japan) controlled by Analyst 1.6.3 software (Ontario, Canada). Chromatographic separation was performed on a Discovery HS C18 Column 15 cm × 2.1 mm, 3 µm (Supelco, Bellefonte, PA) kept at 25°C and 2 µl of sample was injected. The mobile phase consisted of A: acetonitrile and B: water with 1 g/l ammonium acetate and 0.1% formic acid. The following gradient was used: 0.0 to 2.4 min at 40% A, 2.5 to 6.0 min at 60% A, hold for 2.1 min at 60% A, then 8.1 to 9 min 100% A, hold at 100% A for 3.1 min and return to 40% A at 12.1 min. The flow rate was 1.0 ml/min and total run time was 14 min. The acquisition mode used was multiple reaction monitoring. The transition ions (mass-to-charge ratio), deprotonation potentials (V), and corresponding collision energies (V) for all of the compounds can be found in Table 1. A calibration curve was constructed for the assay by linear regression using the peak area ratios of the calibrators and internal standards. The standard curve ranged from 1 to 100 ng for AEA, 10–1,000 ng for 2-AG, and 0.01–1 ng for PE₂. Calibration curves had a correlation (r²) of 0.9988 or better.

Table 1.

MRM parameters for all compounds

Compound	Transition ions, m/z	DP, V	CE, V
AEA	348.2 > 62.2	26	13
	348.2 > 91.2	26	60
AEA-d8	356.2 > 63.2	26	13
AEA	348.2 > 62.2	26	13
2-AG	379.2 > 287.2	45	26
	379.2 > 269.2	45	28
2-AG-d5	384.2 > 287.2	45	26
PE2	396.2 > 378.0	60	9
	396.2 > 360.0	60	13
PE2-d4	400.2 > 382.0	60	9

AEA, anandamide; 2-AG, 2-arachidonoylglycerol; CE, collision energy; DP, deprotonation potential; MRM, multiple reaction monitoring; m/z, mass-to-charge ratio; PE₂, prostamide E₂.

Statistical analyses.

Data are presented as the mean ± SE. For multiple group comparisons, one- or two-way ANOVA was performed using a Dunnett post hoc test when significant differences were found, using the first preinfusion control phase, C1, as the control. For cross-group comparisons of intramedullary PF-3845 treatment alone to intramedullary PF-3845 treatment after rimonabant or celecoxib pretreatment, one-way ANOVA was performed using a Fisher’s least-significant difference post hoc test. Data were considered statistically significant when P ≤ 0.05.

RESULTS

Effects of intramedullary and systemic infusions of PF-3845 on MAP and UV in C57BL/6J mice.

The effect of PF-3845 infusion into the right renal medulla of C57BL6/J mice on MAP and UV are presented in Fig. 1, A and B. The baseline MAP of C57BL/6J mice during the pretreatment control phases, C1 and C2, was 112 and 116 mmHg, respectively. Although intramedullary infusion of PF-3845 at increasing sequential doses of 3.75, 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹ did not significantly decrease MAP, a downward trend in MAP that was dose-dependent was evident. MAP further declined during the posttreatment control phases, P1 and P2, to 82 and 79 mmHg, respectively, which were significantly lower compared with the C1 pretreatment control phase (P < 0.05). The baseline UV in C57BL/6J mice was 11 and 12 µl⋅min⁻¹⋅g⁻¹ kwt for the C1 and C2 control phases, respectively. Intramedullary infusion of PF-3845 increased UV (P < 0.05) to 22 and 28 µl⋅min⁻¹⋅g⁻¹ kwt at the two highest dose rates, 15 and 30 nmol⋅min⁻¹⋅kg⁻¹, respectively, compared with the C1 phase. During the posttreatment phases, UV remained significantly elevated at P1 (P < 0.05) but not P2. There were no significant changes in either MAP or UV of sham-treated control C57BL/6J mice, which received only vehicle infusion during the entire experimental period.

Fig. 1.

The effects of intramedullary and systemic infusion of PF-3845 on MAP and UV in C57 BL/6J mice. Control infusion periods with vehicle alone (C); 3.75, 7.5, 15, and 30 indicate dose rates for PF-3845 infusion (in nmol⋅kg⁻¹⋅min⁻¹) and posttreatment control infusion periods (P). Data represent the mean ± the standard error of each group. *Significant difference vs. the C1 control group (P < 0.05; n = 5–7 per group). kwt, kidney weight; MAP, mean arterial pressure; PF, PF-3845; PF-3845, N-3-pyridinyl-4-[[3-[[5-(trifluoromethyl)-2-pyridinyl]oxy]phenyl]methyl]-1-piperidine carboxamide; UV, urine formation rate; n, sample size.

For comparison, the effects of intravenous PF-3845 on MAP and UV at the same dose rates was determined (Fig. 1, C and D). The baseline MAP of C57BL/6J mice during the C1 and C2 control phases was 116 and 113 mmHg, respectively, whereas intravenous administration of PF-3845 dropped MAP significantly (P < 0.05) by 20, 22, and 25 mmHg at 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹ doses, respectively, relative to C1. MAP declined further during the posttreatment phases, 35 and 38 mmHg for P1 and P2.

In contrast, intravenous PF-3845 significantly increased UV from 11 and 12 µl⋅min⁻¹⋅g⁻¹ kwt for C1 and C2 to 21 and 28 µl⋅min⁻¹⋅g⁻¹ kwt at the two highest doses (P < 0.05). The peak increase in UV (35 µl⋅min⁻¹⋅g⁻¹ kwt) was reached during the P1 posttreatment phase, declining during P2.

Effects of intramedullary and systemic infusions of PF-3845 on MAP and UV in FAAH KO mice.

Corresponding experiments were conducted to characterize the effect of intramedullary and intravenous PF-3845 in FAAH KO mice (Fig. 2, A–D). The baseline MAP of FAAH KO mice following the 1-h equilibration period was 95 and 96 mmHg during C1 and C2, respectively. Intramedullary infusion of PF-3845 did not significantly change MAP at any of the four doses tested or during the posttreatment phases (Fig. 2A). The baseline UV in FAAH KO mice before intramedullary PF-3845 infusion was 4 and 3 µl⋅min⁻¹⋅g⁻¹ kwt for C1 and C2, respectively. Intramedullary administration of the lowest PF-3845 dose tested (3.75 nmol⋅min⁻¹⋅kg⁻¹) did not change UV. However, UV was significantly increased (P < 0.05) after the three highest doses (18, 20, and 24 µl⋅min⁻¹⋅g⁻¹ kwt after 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹, respectively when compared with the C1 phase (Fig. 2B). UV remained elevated during both posttreatment phases. There were no significant changes in either MAP or UV of sham-treated control FAAH KO mice, which received only vehicle infusion during the entire experimental period.

Fig. 2.

The effect of intramedullary and systemic infusion of PF-3845 on MAP and UV in FAAH KO mice. Control infusion periods with vehicle alone (C); 3.75, 7.5, 15, and 30 indicate dose rates for PF-3845 infusion (in units of nmol⋅kg⁻¹⋅min⁻¹), and posttreatment control infusion periods (P). Data represent the mean ± the standard error of each group. *Significant difference vs. the C1 control group (P < 0.05; n = 5–7 per group). FAAH, fatty acid amide hydrolase; FAAH KO, FAAH homozygous knockout; kwt, kidney weight; MAP, mean arterial pressure; PF, PF-3845; PF-3845, N-3-pyridinyl-4-[[3-[[5-(trifluoromethyl)-2-pyridinyl]oxy]phenyl]methyl]-1-piperidine carboxamide; UV, urine formation rate; n, sample size.

In the experiment to assess intravenously administered PF-3845 on MAP in FAAH KO mice (Fig. 2, C and D), the average MAP during the C1 and C2 control phases was 98 and 96 mmHg, respectively. MAP was not significantly different from the C1 control group at any of the four intravenous PF-3845 doses tested (3.75, 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹). Interestingly, a significant drop in MAP occurred in both posttreatment groups, P1 and P2: 80 and 77 mmHg, respectively (Fig. 2C). The baseline UV of FAAH KO mice measured before intravenous PF-3845 was 4 and 2 µl⋅min⁻¹⋅g⁻¹ kwt during C1 and C2, respectively. Whereas intravenous administration of PF-3845 at the lowest dose (3.75 nmol⋅min⁻¹⋅kg⁻¹) did not significantly increase UV, the 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹ groups were increased to 23, 23, and 26, respectively, and this increase in UV was significant (all P < 0.05) compared with the C1 group. After shifting back to the vehicle, UV remained elevated during the P1 period but not P2.

Effects of intramedullary and intravenous infusion of PF-3845 on U_Na, U_K, and MBF in C57BL/6J mice.

Intramedullary infusion of PF-3845 into the medulla of the right kidney produced significant elevations in both U_Na and U_K (1.2 and 2.2 µl⋅min⁻¹⋅g⁻¹ kwt at 30 nmol⋅min⁻¹⋅kg⁻¹, P < 0.05, in comparison with C1 (Fig. 3, A and B). MBF was also increased (P < 0.05), but the effect was significant after administration of 7.5, 15 and 30 nmol⋅min⁻¹⋅kg⁻¹. MBF increased from 0.9 and 0.9 V in C1 and C2 to 1.4, 1.5, and 1.8 V at the three highest doses (Fig. 3C). Intravenous infusion of PF-3845 at increasing doses elevated U_Na from 0.16 to 1.28 and U_K from 0.32 to 1.45 µl⋅min⁻¹⋅g⁻¹ kwt, respectively at 30 nmol⋅min⁻¹⋅kg⁻¹ versus the C1 control phase (Fig. 3, D and E). MBF increased significantly (P < 0.05) at the 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹ doses, respectively (Fig. 3F). No effects on either salt excretion rate or MBF were observed in control mice infused with vehicle the entire experimental period.

Fig. 3.

The effect of intramedullary and systemic infusions of PF-3845 on U_Na and U_K and MBF in C57BL/6J. Control infusion periods with vehicle alone (C); 3.75, 7.5, 15, and 30 indicate dose rates for PF-3845 infusion (in units of nmol⋅kg⁻¹⋅min⁻¹), and posttreatment control infusion periods (P). Data represent the mean ± the standard error of each group. *Significant difference vs. the respective C1 control group (P < 0.05; n 5–7 for PF-38455 and n = 3 for Sham control). kwt, kidney weight; MBF, medullary blood flow; PF, PF-3845; PF-3845, N-3-pyridinyl-4-[[3-[[5-(trifluoromethyl)-2-pyridinyl]oxy]phenyl]methyl]-1-piperidine carboxamide; U_K, urine potassium excretion rate; U_Na, urine sodium excretion rate; n, sample size.

Effect of a COX-2 inhibitor on the blood pressure-lowering and diuresis-stimulating responses to intramedullary PF-3845.

The baseline MAP of C57BL/6J mice during the C1 and C2 phases was 112 and 116 mmHg, respectively, whereas the MAP after treatment with celecoxib (15 mg/kg, iv, over 30 min) was not significantly changed (102 mmHg, P > 0.05). Subsequent administration of PF-3845 decreased MAP significantly compared with the C1 phase but only during the two posttreatment phases, P1 and P2 (90 and 90 mmHg, respectively, Fig. 4A). The baseline UV of C57BL/6J mice during the control phases, C1 and C2, was 16 and 20 µl⋅min⁻¹⋅g⁻¹ kwt, respectively, whereas after the treatment with celecoxib, it was 15 µl⋅min⁻¹⋅g⁻¹ kwt. Subsequent administration of PF-3845 at 3.75, 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹ into the medulla of the right kidney increased UV significantly to 30, 32, 34, and 44 µl⋅min⁻¹⋅g⁻¹ kwt when compared with the C1 control group (Fig. 4B). UV subsequently declined during the posttreatment phases in the celecoxib-pretreated and PF-3845-treated mice to 33 and 25 µl⋅min⁻¹⋅g⁻¹ kwt (P > 0.05).

Fig. 4.

The effect of celecoxib pretreatment on MAP and UV responses to intramedullary infusion of PF-3845 in C57BL/6J and FAAH KO mice. Control infusion periods with vehicle alone (C); 3.75, 7.5, 15, and 30 indicate dose rates for intramedullary PF-3845 infusion (nmol⋅kg⁻¹⋅min⁻¹). Period of celecoxib pretreatment (Cel; 15 mg/kg, iv), and posttreatment control infusion periods (P). A and B: C57BL/6J; C and D: FAAH KO. Data represent the mean ± the standard error of each group. *Significant difference vs. the respective C1 control group (P < 0.05; n = 5–7 per group). FAAH, fatty acid amide hydrolase; FAAH KO, FAAH homozygous knockout; kwt, kidney weight; MAP, mean arterial pressure; PF, PF-3845; PF-3845, N-3-pyridinyl-4-[[3-[[5-(trifluoromethyl)-2-pyridinyl]oxy]phenyl]methyl]-1-piperidine carboxamide; UV, urine formation rate; n, sample size.

The corresponding MAP and UV responses of FAAH KO mice to celecoxib followed by PF-3845 are shown in Fig. 4, C and D. MAP was not significantly changed by celecoxib pretreatment (104 mmHg) compared with 100 and 100 mmHg for C1 and C2, respectively. Celecoxib also did not block the reduction in MAP by PF-3845 treatment. MAP was reduced to 79, 79, 84, and 85 mmHg at the 15 and 30 nmol⋅min⁻¹⋅kg⁻¹ doses of PF-3845 and during the P1 and P2 posttreatment phases, respectively (P < 0.05, Fig. 4C). The UV of FAAH KO mice were not significantly changed after either the celecoxib pretreatment period (2 µl⋅min⁻¹⋅g⁻¹ kwt vs. 4 for the C1 control group) or after dosing with intramedullary PF-3845 (Fig. 4D).

Effect of a CB-1 receptor antagonist on the blood pressure-lowering and diuresis stimulating responses to intramedullary PF-3845.

The baseline MAP of C57BL/6J mice before rimonabant treatment was 112 and 109 mmHg during the C1 and C2 phases, respectively (Fig. 5A). Administration of rimonabant (3 mg/kg, ip) decreased MAP significantly (P < 0.05) to 91 mmHg, compared with C1. MAP was not changed with subsequent intramedullary administration of PF-3845, compared with the rimonabant pretreatment group. MAP remained at 79, 81, 80, and 80 mmHg at the 3.75, 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹ dose rates. The baseline UV of the C57BL/6J mice was 15 and 16 µl⋅min⁻¹⋅g⁻¹ kwt during C1 and C2, which was not significantly changed by the rimonabant pretreatment (P > 0.05, Fig. 5B. PF-3845 subsequently administered into the medulla of the right kidney did not significantly stimulate diuresis at any dose rate.

Fig. 5.

The effect of rimonabant pretreatment on the MAP and UV responses to intramedullary infusion of PF-3845 in C57BL/6J and FAAH KO mice. Control infusion periods with vehicle alone (C); 3.75, 7.5, 15, and 30 indicate dose rates for intramedullary PF-3845 infusion (nmol⋅kg⁻¹⋅min⁻¹). Period of rimonabant (SR141716A) pretreatment (Rmbt; 3 mg/kg, ip) and posttreatment control infusion periods (P). Data represent the mean ± the standard error of each group. *Significant difference vs. the respective C1 control group (P < 0.05; n = 5–7 per group). FAAH, fatty acid amide hydrolase; FAAH KO, FAAH homozygous knockout; kwt, kidney weight; MAP, mean arterial pressure; PF, PF-3845; PF-3845, N-3-pyridinyl-4-[[3-[[5-(trifluoromethyl)-2-pyridinyl]oxy]phenyl]methyl]-1-piperidine carboxamide; UV, urine formation rate; n, sample size.

In the FAAH KO mice, rimonabant treatment dropped MAP significantly from 117 (for both the C1 and C2 control groups) to 105 mmHg. Intramedullary administration of PF-3845 did not significantly change MAP at any dose level or during the postphases compared with the rimonabant group (P > 0.05, Fig. 5C). Baseline UV of C1 and C2 in FAAH KO mice was 4 and 5 µl⋅min⁻¹⋅g⁻¹ kwt, respectively which was not significantly changed by rimonabant pretreatment (11 µl⋅min⁻¹⋅g⁻¹ kwt, P > 0.05; Fig. 5D). Ensuing infusion of PF-3845 into the right renal medulla produced a significant increase in UV (P < 0.05, compared with the C1 control group) to 12, 20, and 24 µl⋅min⁻¹⋅g⁻¹ kwt at the 7.5, 15, and 30 nmol⋅min⁻¹⋅kg⁻¹ doses, respectively.

Effects of intramedullary infusion of PF-3845 on AEA, 2-AG, and PE₂ concentrations in plasma and kidney tissue of C57BL/6J and FAAH KO mice.

The concentrations of AEA, 2-AG, and related derivatives in kidney tissue and plasma of control and PF-3845-treated mice were determined by UPLC-MS/MS (Fig. 6). In C57BL/6J mice, AEA but not 2-AG was elevated in kidney tissue by PF-3845 treatment (P < 0.05; Fig. 6A). In contrast, plasma 2-AG but not AEA concentrations were significantly increased by PF-3845 (Fig. 6B). In FAAH KO mice, however, kidney AEA was decreased (P < 0.05) but 2-AG was unaffected by intramedullary PF-3845 infusion (Fig. 6C), whereas neither AEA nor 2-AG were changed in plasma of the PF-3845-treated FAAH KO mice (Fig. 6D). PE₂ was not detected in either kidney tissue or plasma by UPLC-MS/MS (10 pg level of detection).

Fig. 6.

Anandamide (AEA) and 2-arachidonylglycerol (2-AG) concentrations in kidney tissue and plasma of sham control and intramedullary PF-3845-infused C57BL/6 and FAAH KO mice. Data represent the mean ± the standard error of the mean of each group. *Significant difference vs. the Sham control group (P < 0.05; n = 5–6 per group) in an unpaired, two-tailed t-test. FAAH, fatty acid amide hydrolase; FAAH KO, FAAH homozygous knockout; kwt, kidney weight; PF, PF-3845; PF-3845, N-3-pyridinyl-4-[[3-[[5-(trifluoromethyl)-2-pyridinyl]oxy]phenyl]methyl]-1-piperidine carboxamide; n, sample size.

DISCUSSION

The present study was designed to investigate the hypothesis that infusion of a FAAH-selective inhibitor into the renal medulla would stimulate diuresis and natriuresis. It also explored the hypotheses that EC receptors and/or COX-2 would be required for PF-3845 to modulate UV. Lastly, the design also enabled us to examine possible interactions between the two main ECs, AEA and 2-AG, by evaluating the effect of pharmacological inhibition of FAAH on AEA and 2-AG concentrations in plasma and kidney tissue. The study resulted in several novel observations. Intramedullary and systemic infusions of PF-3845 increased urinary and salt excretion while decreasing MAP. The data also showed that the mechanism by which PF-3845 increased UV and decreased MAP was independent of COX-2 but dependent on CB-1, effects that were accompanied by elevated kidney AEA and plasma 2-AG. Lastly, the MAP-reducing effect of PF-3845 was not observed in FAAH KO mice, but PF-3845, surprisingly, was still able to stimulate diuresis and natriuresis. The data in our study suggest that multiple mechanisms contribute to the action of intramedullary PF-38455 to lower pressure and stimulate diuresis and natriuresis, including FAAH-dependent and FAAH-independent mechanisms. They also support that crosstalk between the AEA and 2-AG systems in response to inhibition of FAAH in the renal medulla by PF-3845 may also contribute.

Intramedullary infusion of PF-3845 in C57BL/6J mice resulted in increased urinary excretion (Fig. 1B), and the analytical data showed that this effect was accompanied by elevated AEA in kidney tissue (Fig. 6A). These data are consistent with earlier data from our laboratory (29) and others (17) showing that infusion of exogenous AEA into the renal medulla elicits a diuretic effect (17, 29). The current study extends this by showing that elevation of endogenous AEA in response to FAAH inhibition can produce the diuretic effect as well. FAAH is highly expressed in kidney relative to most tissues (19), and in the kidney, it is most highly expressed in renal tubular cells (29). Our recent observation that an inhibitor of both FAAH and MAGL, IDFP, was an effective stimulator of diuresis (1) led us to evaluate a highly selective FAAH inhibitor. The present study found that the diuretic action of PF-3845 is CB1-dependent (Fig. 5B) but COX2-independent (Fig. 4B) in C57BL/6J mice. The finding of CB1-dependence agrees with other studies reporting diuretic effects of CB-1 receptor agonists (8, 27), but it appears contrary to the data of Li et al. (17) who found that the diuretic effect of intramedullary methanandamide, a longer-lasting derivative of AEA, was insensitive to CB-1 blockade (17). The COX-2 independence of the PF-38455-mediated effects on MAP and UV was surprising, given our earlier finding that the diuretic effect of intramedullary AEA was blocked by pretreatment with celecoxib (29). This indicates that the effects of PF-3845 on MAP and UV are not mediated by COX-2-derived metabolites of AEA, such as PE₂ (37). The difference between the results for AEA and PF-3845 may relate to the exogenous versus endogenous nature of AEA. The concentration and disposition of AEA produced indirectly by inhibition of FAAH is likely to differ from when exogenous AEA is infused into the renal medullary interstitium. Resulting differences in the degrees and extents to which AEA versus AEA metabolites contribute to its biologic effects could be responsible. In our study that showed COX-2 dependence of the diuretic effect of intramedullary IDFP (1), it may be noted that the role that MAGL inhibition and accumulation of 2-AG plays in the diuretic mechanism of IDFP remains to be determined. Like AEA, 2-AG is an agonist of CB-1 receptors and undergoes metabolism by COX-2 to bioactive metabolites (16, 30).

Treatment of C57BL/6J mice with PF-3845 resulted in increased urinary excretion of both sodium and potassium. These effects were observed after either intramedullary or intravenous administration of the drug and they were accompanied by increased MBF (Fig. 3). These data support a role of increased MBF in the diuretic mechanism of PF-3845. A washing out of the hyperosmotic environment of the renal medulla is consistent with decreased capacity to reabsorb sodium and fluid from the urinary filtrate. These data have a similarity to those described for intramedullary infusion of the FAAH and MAGL inhibitor, IDFP, including stimulated diuresis, salt excretion, and MBF (1). The data in the current study support that inhibition of FAAH may account for the reported effects of IDFP. The mechanism of increased MBF by intramedullary PF-38455 in C57BL/6J mice may be attributed to an action of AEA to produce medullary nitric oxide (NO), which increases MBF (28).

Surprisingly, administration of PF-3845 either systemically or by intramedullary infusion was still able to produce a diuretic response in FAAH KO mice (Fig. 2, B and C). This observation suggests that a second mechanism independent of FAAH can be activated by PF-3845 treatment in the renal medulla of FAAH KO mice. The analytical data showing that 2-AG concentrations are not significantly changed by intramedullary PF-3845 treatment of FAAH KO mice clearly rule out MAGL as a potential off-target of PF-3845’s inhibitory effects. On the other hand, it is noteworthy that several ethanolamides, including AEA that were elevated in kidney of control FAAH KO mice were decreased upon intramedullary PF-3845. This finding, taken together with the U-shaped diuresis dose-response curves described for cannabinoid agonists (8, 27), suggests that lowering of AEA in kidney tissue from a higher than normal concentration (such as that observed in FAAH KO mice) may have a disinhibiting effect, producing the observed paradoxical stimulation of diuresis. The mechanism by which AEA and other ethanolamides are lowered in kidneys by PF-3845 infusion into the renal medulla of FAAH KO mice is also not known. Mice homozygous for FAAH KO mutations have been shown to exhibit elevated renal AEA concentrations compared with their wild-type counterparts (19), which was also evident in the current study (Fig. 6, A and C). This difference does not translate into either reduced MAPs or increased urine flow or salt excretion rates (data not shown), suggesting the development of underlying compensatory responses in response to chronic loss of FAAH.

Another interesting finding of this study was the drop in MAP after intramedullary or systemic infusion of PF-3845 in C57BL/6J mice (Fig. 1, A and B). This response occurred to a greater degree after intravenous compared with intramedullary dosing. This decrease was dependent on functional FAAH as FAAH KO mice did not show comparable reductions after PF-3845 treatments (Fig. 2, A and B). Similar to the diuretic effect, this MAP-reducing effect in C57BL/6J mice appears to be mediated by CB-1 receptors, based on its blockade by pretreatment with rimonabant and appears to not involve metabolism of ECs by COX-2. The mechanism of the reduced MAP after PF-38455 treatment remains to be determined, but there are several possibilities. The first is that it may be related to the ongoing loss of effective blood volume secondary to increased diuresis and salt excretion. However, the observation that acute intramedullary AEA (29) has apparent similar diuresis-inducing capacity in the absence of any changes in MAP argues against this possibility. The MAP-lowering effect could be mediated by a substance entering the circulation from the kidney after PF-3845 treatment to lower systemic resistance by an action on the peripheral vasculature. AEA has well-known vasodepressor properties mediated by CB-1 receptors (33). However, its plasma level was not found to be altered by intramedullary PF-3845 in this study. The possibility that it is a substance indirectly stimulated by AEA in the kidney and released into the circulation, such as NO (11, 22), may be considered. NO is considered to have a critical role in the renal medulla for the control of MBF and blood pressure (20, 21). Two additional possibilities are that it is mediated by 2-AG, which is increased in plasma of PF-3845-treated mice (Fig. 6A) or that it is due to escape of PF-3845 into the systemic circulation to inhibit FAAH and elevate AEA in the peripheral vasculature.

The data in the present study also suggest that the renomedullary EC system has a profound effect on the pressure-natriuresis relationship. According to the pressure-natriuresis model (14), steady-state blood pressure is controlled primarily by the effective intravascular volume. As blood pressure rises, the renal perfusion pressure rises in parallel, and the kidneys respond by increasing sodium and fluid excretion, restoring blood pressure to its starting level. The data in the present study suggest that activation of the renomedullary EC system by treatment with the FAAH inhibitor can fundamentally alter the relationship between blood pressure and diuresis/natriuresis. Even as MAP fell from the intramedullary treatment with PF-3845 to levels below the starting pressure, the kidneys continued to excrete fluid and sodium. These data support an effect of an activated EC system in the renal medulla on the slope of the pressure natriuresis curve. Other substances including NO (9), renal prostaglandins (7), renal kinins (32), and reactive oxygen species (24) have been proposed to contribute to the mechanism of increased diuresis and natriuresis in response to increased renal perfusion pressure (13). It is interesting that MBF, which is increased by these agents and may be a downstream mediator of the pressure natriuresis mechanism, is also increased by PF-3845 treatment (Fig. 3, C and F).

In summary, the current study supports a role for inhibition of FAAH and elevation of AEA in the renal medulla to stimulate MBF and diuresis and natriuresis and lower MAP through a mechanism involving CB-1 receptors. Studies are ongoing to identify the physiologic response of the renomedullary EC system to increased renal perfusion pressure including hypertensive states and to elucidate the mechanisms by which inhibition of FAAH leads to stimulation of MBF and diuresis and lowering of blood pressure.

GRANTS

This work was supported by National Institute of Digestive and Kidney Disorders Grant, DK102539, National Institutes of Health Center for Drug Abuse Grant P30-DA-033934, and National Institute on Drug Abuse Grant T32DA007027–42.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.A., P.-L.L., and J.K.R. conceived and designed research; A.A., S.K.D., Z.D., and J.L.P. performed experiments; A.A., S.K.D., Z.D., J.L.P., and J.K.R. analyzed data; A.A., S.K.D., Z.D., N.L., P.-L.L., and J.K.R. interpreted results of experiments; A.A., S.K.D., and Z.D. prepared figures; A.A. and S.K.D. drafted manuscript; Z.D., N.L., P.-L.L., and J.K.R. edited and revised manuscript; A.A., S.K.D., Z.D., N.L., J.L.P., P.-L.L., and J.K.R. approved final version of manuscript.