The renal outer medullary potassium (ROMK) channel, which is located at the apical membrane of epithelial cells lining the thick ascending loop of Henle and cortical collecting duct, plays an important role in kidney physiology by regulating salt reabsorption. Loss-of-function mutations in the human ROMK channel are associated with antenatal type II Bartter’s syndrome, an autosomal recessive life-threatening salt-wasting disorder with mild hypokalemia. Similar observations have been reported from studies with ROMK knockout mice and rats. It is noteworthy that heterozygous carriers of Kir1.1 mutations associated with antenatal Bartter’s syndrome have reduced blood pressure and a decreased risk of developing hypertension by age 60. Although selective ROMK inhibitors would be expected to represent a new class of diuretics, this hypothesis has not been pharmacologically tested. Compound A [5-(2-(4-(2-(4-(1H-tetrazol-1-yl)phenyl)acetyl)piperazin-1-yl)ethyl)isobenzofuran-1(3H)-one)], a potent ROMK inhibitor with appropriate selectivity and characteristics for in vivo testing, has been identified. Compound A accesses the channel through the cytoplasmic side and binds to residues lining the pore within the transmembrane region below the selectivity filter. In normotensive rats and dogs, short-term oral administration of compound A caused concentration-dependent diuresis and natriuresis that were comparable to hydrochlorothiazide. Unlike hydrochlorothiazide, however, compound A did not cause any significant urinary potassium losses or changes in plasma electrolyte levels. These data indicate that pharmacologic inhibition of ROMK has the potential for affording diuretic/natriuretic efficacy similar to that of clinically used diuretics but without the dose-limiting hypokalemia associated with the use of loop and thiazide-like diuretics.
The kidneys play a critical role in the long-term regulation of blood pressure. In humans, all identified mutations in genes that cause mendelian forms of hypertension or hypotension act in the kidney to alter net renal salt and water reabsorption (Lifton et al., 2001). Several mechanisms involving ion channels, exchangers, and transporters act in an integrated manner along the nephron to regulate salt and water reabsorption. At the thick ascending limb of Henle (TALH), ∼30% of salt reabsorption occurs through the luminal Na+/K+/2Cl− cotransporter, which is the target of furosemide, a loop diuretic used clinically in the treatment of congestive heart failure. In the distal convoluted tubule (DCT), the Na+/Cl− cotransporter is responsible for ∼7% of salt reabsorption. This cotransporter represents the clinical target of the thiazide-class of diuretics used in the treatment of hypertension. The final step in the regulation of salt reabsorption takes place in the cortical collecting duct (CCD) through the amiloride-sensitive epithelial sodium channel. Because of the tight coupling between sodium reabsorption and potassium secretion at the CCD, the use of loop or thiazide diuretics is clinically associated with hypokalemia, whereas amiloride intervention causes hyperkalemia.
Bartter’s syndrome, characterized by renal salt wasting and polyuria-associated low blood pressure and hypokalemic alkalosis, is caused by recessive loss-of-function mutations in any one of the four genes involved in salt reabsorption at the TALH: the apical Na+/K+/2Cl− cotransporter, the basolateral Cl− (α or β subunits) channel, and the renal outer medullary potassium (ROMK) channel, the product of the KCNJ1 gene present at the apical membrane (Simon et al., 1996; Palmer et al., 1997; Xu et al., 1997; Lu et al., 2002). The ROMK channel has attracted significant interest because of the finding that heterozygous carriers of channel mutations associated with type II Bartter’s syndrome have reduced blood pressure and a decreased risk of developing hypertension by age 60 (Ji et al., 2008). Moreover, it is important to note that ROMK knockout mice and rats recapitulate the phenotype of type II Bartter’s syndrome (Lu et al., 2002; Lorenz et al., 2002; Zhou et al., 2013). These data suggest that ROMK represents a target for the development of novel diuretics for the treatment of hypertension and/or heart failure.
ROMK is a member of the inwardly rectifying family of potassium (Kir) channels (Nichols and Lopatin, 1997). The expression of ROMK (Kir1.1) appears to be almost exclusively restricted to the apical membrane of the epithelial cells lining the TALH and the CCD (Xu et al., 1997; Palmer et al., 1997; Lu et al., 2002). At the TALH, ROMK participates in potassium recycling across the apical membrane that is critical for the proper function of the furosemide-sensitive Na+/K+/2Cl− cotransporter because the K+ concentration in the luminal fluid is much lower than that of Na+ and Cl−. At the CCD, ROMK provides a pathway for potassium secretion that is tightly coupled to sodium reabsorption through the amiloride-sensitive epithelial sodium channel. Because of the presence of ROMK channels at both TALH and CCD, selective inhibitors of this channel would be predicted to provide equivalent or superior diuretic/natriuretic efficacy to loop diuretics such as furosemide, with the added potential benefit of attenuating the hypokalemia associated with the use of loop diuretics or thiazides. Despite all the evidence supporting ROMK as a novel therapeutic target, the development of selective channel inhibitors has only recently been attempted. In addition to the peptide tertiapin, which blocks with high affinity the rat but not the human channel (Jin and Lu, 1998; Felix et al., 2006), two independent groups have reported the identification of small molecule ROMK inhibitors (Lewis et al., 2009; Bhave et al., 2011; Tang et al., 2012).
In our present study, compound A, the discovery of which is described in a separate study (Tang et al., 2013), 5-(2-(4-(2-(4-(1H-tetrazol-1-yl)phenyl)acetyl)piperazin-1-yl)ethyl)isobenzofuran-1(3H)-one, was characterized. This agent inhibits the rat and human Kir1.1 channels with high affinity and displays good selectivity across other ion channel superfamilies as well as pharmacokinetic properties suitable for in vivo testing. Short-term oral dosing of compound A was shown to produce dose-dependent diuresis and natriuresis in normotensive rats and dogs of similar magnitude to that of hydrochlorothiazide but with no significant kaliuresis and no changes in plasma electrolyte levels. Taken together, these data indicate that pharmacologic inhibition of ROMK provides diuretic/natriuretic efficacy similar to that of clinically used diuretics but with the potential benefit of reducing the hypokalemia associated with the use of loop and thiazide class diuretics.
Materials and Methods
Compound A was synthesized at Merck Research Laboratories (Rahway, NJ) as described elsewhere (Tang et al., 2013). The FluxOR thallium detection kit was obtained from Invitrogen (Carlsbad, CA). Probenecid, anhydrous dimethyl sulfoxide (DMSO), and ouabain were obtained from Sigma-Aldrich (St. Louis, MO). 86RbCl was obtained from PerkinElmer Life and Analytical Sciences (Waltham, MA). The QuickChange II site-directed mutagenesis kit was from Stratagene (La Jolla, CA). The pCI-neo vector and the transfection reagents FuGENE 6 and FuGENE HD were from Promega (Madison, WI). Hydrochlorothiazide (HCTZ) was purchased from MP Biomedicals, LLC (Solon, OH). All other reagents were obtained from commercial sources and were of the highest purity commercially available.
All tissue culture reagents were obtained from Invitrogen. All Kir1.1 constructs represent the ROMK1 splice form of Kir1.1. HEK293 cell lines stably transfected with human Kir1.1 (hKir1.1) or rat Kir1.1 (rKir1.1), CHO cell line stably expressing hKir1.1 (CHO-hKir1.1), and MDCKII-Flp cell line stably expressing rKir1.1 (MDCK-rKir1.1) were obtained as previously described elsewhere (Felix et al., 2006, 2012). A CHO cell line stably expressing hKir2.3 (CHO-hKir2.3) was constructed by transfecting cells with hKir2.3-pCIneo expression plasmid using FuGENE 6. Stable pools of CHO cells expressing hKir2.3 were prepared by Geneticin (G418) selection (1000 μg/ml) and analyzed for functional expression of hKir2.3 using a membrane potential, fluorescence resonance energy transfer-based assay, as described elsewhere (Solly et al., 2008). Individual stable cell lines were generated by limiting dilution under continuous selection with G418. HEK293 cells stably expressing either human Kir4.1 (HEK-hKir4.1) or human Kir7.1 (HEK-hKir7.1) were prepared by transfecting cells with hKir4.1-phCMV1 or hKir7.1-phCMV1 expression plasmids using FuGENE HD. Stables pools of cells were selected using 1000 μg/ml G418. Individual stable cell lines were generated by limiting dilution under continuous selection with G418 and were analyzed for functional expression of Kir4.1 or Kir7.1 using a fluorescence-based thallium flux assay. HEK293 cells stably transfected with human Kir2.1 were obtained from EMD Millipore Corporation (Billerica, MA). HEK293 cells were grown in minimum essential-α media (MEM), 10% fetal bovine serum (FBS), 500 µg/ml G418, 1× penicillin/streptomycin/glutamine, and 1× MEM nonessential amino acids. CHO cells were grown in Iscove’s modified Dulbecco’s medium supplemented with HT Supplement Solution with 10% heat inactivated FBS, 500 mg/ml G418, and 1% penicillin/streptomycin. The MDCK-rKir1.1 cell line was grown in DMEM + Glutamax supplemented with penicillin/streptomycin, 200 μg/ml hygromycin B, and 10% FBS. All lines were maintained at 37°C in a 10% CO2 atmosphere.
Kir1.1 Thallium Flux Assay.
Permeation of thallium through open hKir1.1 channels was determined as previously described elsewhere (Felix et al., 2012). In brief, HEK293 cells stably transfected with hKir1.1 were plated using a Thermo Scientific Matrix WellMate (Thermo Scientific, Waltham, MA) at approximately 20,000 cells/well on black-wall, clear bottom, 384-well poly-d-lysine-coated plates (Becton Dickinson, Franklin Lakes, NJ) in 50 µl of growth medium and incubated overnight (16–20 hours) at 37°C in a 10% CO2 atmosphere. All liquid handling was done on a Thermo Scientific Matrix PlateMate 2X3. The cell growth medium was removed, and cells were then incubated with 0.025 ml of a solution containing FluxOR dye loading reagent, prepared according to the manufacturer’s instructions in Hank’s buffered saline solution (HBSS) containing 1.26 mM CaCl2 and 0.49 mM MgCl2 (Invitrogen), pH adjusted to 7.4 by the addition of NaOH. After incubation in the dark for 90 minutes at ambient temperature (22–24°C), cells were washed once with 0.04 ml of HBSS buffer solution and incubated in the dark for 30 minutes at ambient temperature (22–24°C) with 0.025 ml of FluxOR assay solution containing 2.5 mM probenecid and 300 μM ouabain in the absence or presence of the test compound.
At the end of the 30-minute incubation period, the plate was placed in a FLIPRTETRA instrument (Molecular Devices, Sunnyvale, CA), illuminated at 490 nm, and the fluorescence emission was recorded at 525 nm. After an 80-second baseline reading, 0.00625 ml of a 5X solution containing 7.5 mM thallium sulfate, 0.75 mM K2SO4, prepared in the FluxOR chloride-free buffer was added, and fluorescence emission was recorded for an additional 8–9 minutes, with an exposure time of 0.4 seconds and a read interval of 10 seconds. The change in fluorescence emission (F/F0) was calculated by averaging the three readings just before the signal reaching a plateau level, usually from 330 to 360 seconds (F), and the baseline was calculated by averaging the initial four readings, usually from 1 to 40 seconds.
Kir1.1 86Rb+ Flux Assays.
The ability of 86Rb+ to permeate through Kir1.1 channels was evaluated as previously described elsewhere (Felix et al., 2012). In brief, CHO cells stably expressing hKir1.1 or HEK293 cells stably expressing rKir1.1 were seeded at 120,000 cells/well in either 96-well white, opaque bottom tissue culture plates (PerkinElmer Life and Analytical Sciences) or clear bottom poly-d-lysine-coated plates (BioCoat; Becton Dickinson, Franklin Lakes, NJ) in complete growth medium containing 1.5 μCi/ml 86Rb+ and incubated in 10% CO2 at 37°C overnight. On the day of the assay, the 86Rb+-containing medium was removed, and the cells were washed once with low K assay buffer containing (in mM): 126.9 NaCl, 4.6 KCl, 2 CaCl2, 1 MgCl2, 10 Hepes/NaOH, pH 7.4. High K assay buffer (100 μl) containing (in mM) 121.5 NaCl, 10 KCl, 2 CaCl2, 1 MgCl2, 10 Hepes/NaOH, pH 7.4, with or without the test compound, was added, and the cells were incubated at ambient temperature (22–24°C) for 30 minutes. An aliquot (30 μl) of the assay buffer was removed and added to 170 μl of MicroScint 20 scintillation cocktail (PerkinElmer Life and Analytical Sciences) in 96-well plates (Packard OptiPlate-96; EMD Millipore); the remaining assay buffer was discarded. Cells were solubilized in the presence of 1% sodium dodecyl sulfate, and 170 μl of MicroScint 20 was added to each well.
Radioactivity associated with the assay buffer and cells was determined on a TopCount counter (Packard, GMI, Ramsey, MN). The amount of radioactivity in the assay buffer (% efflux) was normalized to the total radioactivity content of the assay buffer and cells. For experiments in which serum was included, compounds were prepared in high K assay buffer in the absence or presence of 10, 30, or 100% human (CHO-Kir1.1) or rat (HEK-rKir1.1) serum. All other steps were performed as described earlier.
For Transwell assays, 80,000 MDCK-rKir1.1 cells in 0.5 ml were seeded in BD Falcon cell culture inserts, 0.4-μm-pore, transparent polyethylene terephthalate (PET) membrane (Becton Dickinson), and 2 ml of growth media was added to the lower chamber (Felix et al., 2012). Cells were allowed to adhere and grow at 37°C in 10% CO2 for 3 days. The Transwell medium (apical compartment) was replaced with 0.5 ml of fresh medium, and the growth medium was removed from the lower chamber (basolateral compartment) and replaced with 2 ml of growth medium containing 1.5 μCi/ml 86Rb+, followed by overnight incubation at 37°C in 10% CO2. The Transwell medium was replaced with 0.4 ml of low K assay buffer, and the Transwell was transferred to a chamber containing 2 ml of low K assay buffer to remove excess of 86Rb+. The Transwell medium was replaced with 0.4 ml of high K assay buffer, with or without 10 μM compound A, and the Transwell was transferred to a chamber containing 2 ml of high K assay buffer, with or without 10 μM compound A. After 30 minutes’ incubation at ambient temperature, the Transwell filter was removed, and radioactivity associated with cells and the apical and basolateral media was determined on a TopCount counter after addition of MicroScint 20.
Blocking of wild-type and mutant hKir1.1 channels by compound A was examined by whole-cell voltage clamp, as previously described elsewhere (Felix et al., 2006). In brief, the experiments were performed at room temperature using an EPC-9 amplifier and Pulse software (HEKA Electronics, Lamprecht, Germany). The data were acquired at 10 kHz and filtered at 2.9 kHz. The internal (pipette) solution contained 130 mM KCl, 5 mM NaCl, 2 mM MgCl2, 5 mM EGTA, 0.2 mM MgATP, 5 mM Na-HEPES, pH 7.4. The bath solutions containing two different concentrations of K+, each composed of (140-x) mM NaCl, x mM KCl, 2.7 mM CaCl2, 0.5 mM MgCl2, 5 mM Na-HEPES, pH 7.4, were used in every experiment to assess the quality of the recording. Only cells with a shift in the reversal potential within 5 mV of the calculated shift for a K+ selective current were used. Kir1.1 currents were recorded using voltage ramps over at least 100 mV; unless noted otherwise, the current corresponding to a membrane potential of −100 mV was used to determine fractional inhibition by compound A.
Site-directed mutagenesis of Kir1.1, cloned into the pCIneo expression plasmid, was performed using the QuikChange II kit according to the manufacturer’s protocol. For all mutants, the entire open reading frame was sequenced to exclude secondary amino acid changes. For characterization by whole-cell voltage clamp, mutant constructs were transiently transfected into TsA-201 cells using FuGENE 6, as previously described elsewhere (Felix et al., 2006).
Other Ion Channel Assays.
The functional activities of Kir2.1 (HEK-hKir2.1), Kir4.1 (HEK-hKir4.1), and Kir7.1 (HEK-Kir7.1) channels were determined in thallium flux assays using identical conditions as those described for hKir1.1 (Felix et al., 2012). The activity of Kir2.3 (CHO-hKir2.3) was evaluated by 86Rb+ flux using identical conditions as those described for Kir1.1 (Felix et al., 2012). All procedures for evaluation of human ether-à-go-go related gene (hERG) (CHO-hERG) by either QPatch automated electrophysiology or by [35S]MK-499 binding (Schmalhofer et al., 2010), the human voltage-gated sodium channel Nav1.5 (HEK-hNav1.5) using a fluorescence resonance energy transfer (FRET)-based membrane potential assay (Felix et al., 2004), and the L-type calcium channel Cav1.2 (HEK-hCav1.2) in a fluorescence calcium influx assay (Abbadie et al., 2010) have been previously described elsewhere.
Ancillary Target Binding Assays.
In vitro binding assay screening for a panel of 166 ancillary targets was performed by MDS Pharma Services (King of Prussia, PA).
All protocols for animal experiments were approved by the Institutional Animal Care and Use Committee of Merck Research Laboratories (Rahway, NJ and West Point, PA) and adhere to the guidelines of the Committee for Research and Ethical Issues.
Renal Excretory Function Studies in Anesthetized Rats.
Twelve-week old male Sprague Dawley (SD) rats (body weight [BW] 300∼350 g) were anesthetized with thiobutabarbital sodium (Inactin, 100–110 mg/kg i.p.; Sigma-Aldrich) and then placed on a heating pad to maintain rectal temperature at 37°C throughout the study. A tracheostomy was performed, and a polyethylene (PE) tube (PE-250) was inserted to facilitate spontaneous breathing. A PE-50 catheter was inserted into the left femoral artery to allow for intermittent blood sampling and continuous monitoring of arterial blood pressure using a digital data acquisition system (EMKA Technologies Inc., Falls Church, VA). The left femoral vein was cannulated with a PE-50 catheter for infusion of a solution of 6% albumin at a rate of 0.4 ml/100 g body weight per hour (BW/h) initially, followed by infusion of a maintenance solution of 1% albumin at a rate of 0.35 ml/100 g BW/h.
The right jugular vein was also cannulated with a PE-50 catheter for infusion of vehicle [10% ethanol/40% polyethylene glycol (PEG) 400/50% water] and either compound A or HCTZ (both compounds were dissolved in the above mentioned vehicle) at a rate of 0.05 ml/100 g BW/h. This vehicle, when tested in a separate study, was shown to have no effect on renal excretory function (data not shown).
The bladder was catheterized for urine collection with PE-100 tubing. After 60-minute stabilization, urine was collected over two consecutive 30-minute periods, with blood samples being drawn at their midpoint to assess control values of renal excretory function and blood electrolytes during vehicle administration. Subsequently, compound A at 1.55 mg/kg/h or HCTZ at 5.0 mg/kg/h were administered by constant intravenous infusion for 1 hour, and two successive 30-minute sample collections were performed as described for the control vehicle period. Blood and urine electrolytes (Na+, Cl−, and K+) were measured with an i-STAT Portable Clinical Analyzer (HESKA Corporation, Loveland, CO) and a Roche Modular Chemistry System (Roche Diagnostics, Indianapolis, IN), respectively.
Rat Diuresis Assay.
Adult male SD rats (275–350 g BW) were acclimated to single housing in metabolism cages with free access to food and water for at least 3 days before the experiments. On the day of the study, animals were transferred from metabolism cages to shoebox cages, and access to food and water was restricted for the entire duration of the study. Vehicle [Imwitor 742:Tween 80 (1:1, v:v)] or compound was administered at a dose volume of 1 ml/kg by oral gavage. After 30 minutes, voiding was induced by giving each animal a saline load (18 ml/kg by oral gavage). Animals were then transferred back to metabolism cages for urine collection over the next 4 hours at room temperature. Volume of urine voided was recorded for each rat; urine samples were then centrifuged, portioned in aliquots, and frozen at −20°C until analyzed. HCTZ was used as a positive control. When needed, blood samples were obtained by jugular vein puncture to determine compound plasma exposure levels.
Dog Diuresis Assay.
Female mongrel dogs were trained to lie quietly on their back. A sterile Foley catheter with lubricant on its tip was inserted into the urinary bladder after a local topical anesthetic, such as Cetacaine spray or lidocaine gel, was applied to the urethra and surrounding tissue for the comfort of the dog. Once inserted into the bladder, a balloon was gently inflated to retain the catheter within the bladder, which remained in place for the duration of the urine collection period (total 3 hours). The animals were then placed in a padded sling (size appropriate, manufactured by Alice King Chatham) during the experiment. Sterile percutaneous catheters were inserted into saphenous and cephalic veins for blood collection for chemistry, hematology, and compound level analysis. Immediately after collection of control blood and urine samples (two 30-minute collection periods), vehicle or compound was administered orally by gavage (feeding tube). Six additional blood and urine collections (30 minutes for each period) were obtained. Urine volume was recorded for each collection period. Urine and blood samples were then centrifuged, measured in aliquots, and frozen at −20°C until analyzed. Dogs were continually observed while in sling restraint. Upon completion of the study, the Foley catheter was gently removed, and the dogs were returned to their home cages.
Plasma, Urine, and Kidney Level Analysis.
Plasma, urine, and homogenized kidney tissue concentrations of compound A were determined by liquid chromatography with tandem mass spectrometry (LC-MS/MS) using an Applied Biosystems/MDS Sciex API 5000 LC-MS/MS mass spectrometer (Applied Biosystems/MDS Sciex, Foster City, CA) operated in positive ion atmospheric pressure chemical ionization mode with multiple-reaction monitoring. Plasma was prepared for analysis by addition of 300 μl of acetonitrile to a 50-μl sample of plasma. The mixture was then vortexed and centrifuged. The clear liquid at the top was pipetted away from the pellet that was formed at the bottom of the tube, and it was injected directly onto the LC-MS/MS. The kidney homogenate was prepared by adding 3 parts of water for every 1 part tissue (volume:weight) and then shaking vigorously with grinding beads. At the end of the process, the homogenate becomes an opaque liquid that can be readily pipetted. To an aliquot of the homogenate, a 6-fold volume of acetonitrile was added, and the sample was vortexed and centrifuged. The liquid at the top (supernatant) was injected directly onto the LC-MS/MS. Extracts were chromatographed using a Phenomenex Kinetex 1.7-μm pentafluorophenyl 50 × 2.1 mm column (Phenomenex, Torrance, CA) and eluted at 0.75 ml/min using a linear gradient of acetonitrile.
IC50 values for inhibition were determined according to the Hill equation from concentration-response curves by a nonlinear regression analysis where all parameters were left unconstrained. Data are presented as either mean ± S.D. or mean ± S.E.M. of n experiments. Statistical analysis was conducted using either analysis of variance (ANOVA) followed by Dunnett’s post hoc test using Prism (version 4.0.3; GraphPad Software, Inc., La Jolla, CA) or Student’s t test, as appropriate. Two-tailed P < 0.05 was considered statistically significant.
Compound A Inhibits Kir1.1 Channels.
The search for potent and selective Kir1.1 inhibitors has led to the identification of compound A (Fig. 1). In HEK cells stably expressing hKir1.1 channels, compound A inhibits thallium flux through these channels with an IC50 value (mean ± S.D.) of 24 ± 7 nM (n = 3) (Fig. 2A). In functional cell-based assays that measure the ability of 86Rb+ to permeate through human or rat Kir1.1 channels, compound A inhibits these channels with IC50 values (mean ± SD) of 89 ± 6 (n = 6) (Fig. 2B) and 135 ± 15 (n = 2) nM (Fig. 2C), respectively, in the absence of serum and 506 ± 23 (n = 2) (Fig. 2B) and 576 ± 164 (n = 2) nM (Fig. 2C), respectively, in the presence of 100% human or rat serum.
The selectivity of compound A was assessed in functional assays of other related Kir channels, such as cardiac Kir2.1 and renal Kir2.3, Kir4.1, and Kir7.1. In the absence of serum and at concentrations of up to 100 μM, compound A had no significant effect on either thallium flux through Kir2.1 (Fig. 2A) or 86Rb+ flux through Kir2.3 channels (Fig. 2B). Thallium flux through either Kir4.1 or Kir7.1 channels was not significantly inhibited by compound A at concentrations of up to 100 μM in the absence of serum (Fig. 1; Supplemental Data).
In electrophysiological recordings of hERG channels, compound A inhibits with an IC50 value (average ± S.E.M.) of 5.6 ± 1.3 μM (n = 8), whereas in a binding assay that measures the interaction of [35S]MK-499 with membranes derived from HEK cells expressing the hERG channel, compound A displays an IC50 value of 5.9 ± 0.4 μM (n = 5), a value that is similar to that determined by electrophysiology.
In functional assays, compound A inhibited the human voltage-gated sodium channel Nav1.5 by 42% at 30 μM and the human voltage-gated calcium channel Cav1.2 by 19% at 100 μM. In a panel of 166 enzyme and radioligand binding assays run by MSD Pharma Services, compound A, when tested at 10 μM, only inhibited the serotonin transporter with an IC50 of 9.1 μM.
These data suggest that compound A is a potent and selective Kir1.1 inhibitor. The pharmacokinetic properties of compound A in rats and dogs (Table 1) indicate that the compound has moderate clearance rates (rat 40 ml/min/kg; dog 36 ml/min/kg) and good oral bioavailability (rat 33%; dog 80%), making it suitable for in vivo evaluation.
Mechanism of Inhibition of Kir1.1 Channels by Compound A.
Inhibition of hKir1.1 channels by compound A was examined in more detail using standard whole-cell voltage clamp protocols and HEK cells stably expressing hKir1.1 channels. Voltage ramps from −80 mV to +20 mV were applied at regular intervals to monitor hKir1.1 currents under control conditions and in the presence of increasing concentrations of compound A (Fig. 3). The amplitude of the inward current at −80 mV was inhibited by compound A with an IC50 (mean ± S.D.) of 92 ± 51 nM (n = 3). Similarly, the outward current at +20 mV was inhibited with an IC50 (mean ± S.D.) of 42 ± 20 nM (n = 3) (Fig. 3C). The potency observed in the electrophysiology study was consistent with data generated in the thallium and 86Rb+ flux assays (29 and 106 nM, respectively). The data further suggest that inhibition of hKir1.1 by compound A is not dependent on the direction of current flow (P = 0.16 for comparing IC50 values at −80 and +20 mV).
To gain information about the binding site of compound A in the hKir1.1 channel, several site-directed mutants were generated. Mutagenesis focused mainly on amino acids that differed between hKir1.1 and hKir2.1 and that were predicted to face the ion conduction path, based on comparisons with the chicken Kir2.2 channel (Tao et al., 2009) Amino acids in hKir1.1 were mutated individually to the corresponding amino acids found in hKir2.1. All mutant channels were transiently expressed in TsA-201 cells and were examined by whole-cell voltage clamp. Results from these mutagenesis studies are summarized in Table 2, and representative current traces are shown in Fig. 4. Two of the mutant constructs (S130A and L166V) did not generate measurable currents. With the exception of mutations at position 171, all mutant channels were inhibited potently by compound A. N171 in hKir1.1 was initially mutated to aspartate found in the homologous position in hKir2.1.
As expected, based on the work by MacKinnon and others (Lu and MacKinnon, 1994), currents generated by hKir1.1-N171D were strongly inwardly rectifying. Bath application of 100 nM compound A had no discernible effect on these currents, whereas 10 μM compound A inhibited the current at −100 mV by 59% (Fig. 4A). Based on the prominent outcome of introducing a negative charge at position 171, a charge-neutral substitution to glutamine was examined. As with wild-type hKir1.1, hKir1.1-N171Q currents were weakly inwardly rectifying; however, blocking with compound A, although more potent than that for N171D, was still shifted to weaker potency by almost an order of magnitude (Fig. 4B; Table 2).
MDCK-rKir1.1 cells grown on permeable Transwell supports provide a polarized system where the apical and basolateral membranes are physically separated by an impermeable barrier due to the formation of tight junctions (Simons and Virta, 2006). Addition of 86Rb+ to the basolateral compartment allows the accumulation of the isotope inside the cell through activity of the ouabain-sensitive Na+/K+-ATPase pump. After removal of 86Rb+ from the basolateral compartment, efflux of the isotope into the apical side can be inhibited by addition of the peptide blocker tertiapin-K12/Q13 to the apical but not the basolateral compartment, suggesting that Kir1.1 is exclusively expressed at the apical surface of the MDCK-rKir1.1 cells and that the two membrane compartments are indeed separated by tight junctions formed by the monolayer of cells. When the same experiment is performed with 10 μM compound A, inhibition of 86Rb+ flux into the apical compartment occurs, regardless of whether compound A was added to the apical or basolateral compartments (Fig. 5). These data are consistent with the idea that channel inhibition results from compound A accessing the channel from the cytoplasmic side.
Renal Excretory Function Studies in Anesthetized SD Rats.
The diuretic activity of compound A was assessed in anesthetized euvolemic rats. After intravenous infusion of either compound A or HCTZ for 1 hour, significant and comparable increases in urine flow and urinary Na+ and Cl− excretion rates were observed with both agents (Table 3). Urinary K+ excretion, however, was only significantly increased by HCTZ, but not by compound A. HCTZ also caused a significant decrease in blood K+ levels. Plasma levels for compound A during the infusion period averaged 1.36 ± 0.13 μM, approximately 3 times above the IC50 value for inhibition of rat Kir1.1 channels in the presence of 100% serum. Neither compound A nor HCTZ caused significant changes in blood pressure or heart rate during the study (Table 3). Vehicle had no effect on renal excretory function (data not shown). These data indicate that even at a greater diuretic and natriuretic dose, compound A induces less urinary K+ loss compared with HCTZ. Despite significant diuresis observed with both agents, blood K+ levels did not significantly change with compound A but did significantly decrease with HCTZ.
Compound A Elicits Diuresis and Natriuresis in Conscious, Volume-Loaded Rats.
Oral administration of compound A evoked dose-dependent increases in urine output in conscious, volume-loaded SD rats (Fig. 6A). Four hours after oral dosing, compound A significantly increased the urine flow, starting at a dose of 3 mg/kg (2.6-fold vs. vehicle); maximal increases were observed at 50 mg/kg (4.3-fold vs. vehicle). The diuretic efficacy of compound A appeared to reach a plateau at doses between 10 and 50 mg/kg under these experimental conditions. In similar experiments, HCTZ significantly and dose-dependently increased urine flow starting at 10 mg/kg (2.5-fold vs. vehicle), whereas maximal increases were observed at 100 mg/kg (4.0-fold vs. vehicle). Thus, the diuretic efficacy of compound A and HCTZ appears to be similar at the highest doses tested.
A similar trend was observed when comparing the natriuretic efficacy of compound A and HCTZ (Fig. 6B). Four hours after dosing, compound A and HCTZ significantly, and dose-dependently, increased urinary Na+ excretion starting at doses of 10 mg/kg. Maximal natriuretic effects were observed at the highest doses tested, with compound A increasing urinary Na+ excretion by 3.8-fold and HCTZ by 3.4-fold compared with vehicle-treated rats. It is noteworthy that, despite robust diuresis and natriuresis associated with compound A administration, urinary K+ excretion did not significantly change with any of the doses tested but was significantly increased by the highest dose of HCTZ (100 mg/kg, 1.4-fold vs. vehicle) (Fig. 6C). These data indicate that compound A at maximally efficacious diuretic and natriuretic doses does not lead to significant urinary K+ losses, an effect that is different from the well-known, and documented HCTZ-induced kaliuresis.
After 4.5 hours after oral dosing, concentrations of compound A in plasma were below 0.2 μM in rats that had been given the 1 and 3 mg/kg doses but were 0.39 and 0.92 μM, respectively, in rats dosed with 10 and 50 mg/kg. Compound A levels were higher in urine samples, ranging from 0.74 μM in rats dosed with 3 mg/kg up to 4.8 μM in rats dosed with the 50 mg/kg dose. In a parallel PK study, concentrations of compound A were determined in plasma and kidney samples from rats dosed with 10 mg/kg compound A (Table 4). Kidney levels of compound A were approximately 3.9- to 6.7-fold larger than those found in plasma.
Compound A Causes Diuresis and Natriuresis in Conscious Dogs.
Oral administration of compound A increased urine output in conscious, euvolemic dogs (Fig. 7A). Compound A dosed at either 10 or 50 mg/kg led to significant dose- and time-dependent increases in urine output. The maximal diuretic effect of compound A was obtained with the 50 mg/kg dose, 3 hours after dosing, with urine output increasing from baseline values of 0.43 ml/min up to 2.68 ml/min (8-fold change vs. baseline), but the values appeared to plateau at 1.5 hours after dosing.
Oral HCTZ dosed at 10 mg/kg also increased urine flow in a time-dependent fashion, and the maximal diuretic effect of HCTZ was seen at 1 hour after dosing. The urine flow values increased from 0.35 ml/min at baseline up to a maximum of 1.62 ml/min at 1 hour, but the values remained constant for the remainder of the study (∼3.5- to 4.8-fold change vs. baseline). Similarly, compound A led to dose- and time-dependent increases in urinary Na+ excretion in dogs (Fig. 7B).
The maximal natriuretic effect was seen with 50 mg/kg compound A at 3 hours after dosing. Urinary Na+ excretion increased from baseline values of 55 μEq/min up to 435 μEq/min (16-fold change vs. baseline), and the values appeared to plateau at 1.5 hours after dosing. Oral HCTZ also increased urinary Na+ excretion in a time-dependent fashion; maximal natriuresis was seen at 1 hour after dosing. Urinary Na+ excretion increased from 23 μEq/min at baseline up to a maximum of 289 μEq/min at 1 hour (16-fold change vs. baseline), and these values slowly decreased to 167 μEq/min after 3 hours (9.1-fold change vs. baseline). Thus, the natriuretic efficacy of the largest dose of compound A and HCTZ were comparable, whereas the extent of diuresis was larger with the 50 mg/kg dose of compound A.
Kaliuresis, on the other hand, was only significantly enhanced by HCTZ (Fig. 7C), particularly during the first hour after dosing. The two doses of compound A tested did not significantly change urinary K+ excretion rates compared with the baseline values, but HCTZ significantly increased urinary K+ excretion values from 36 μEq/min at baseline to 63 μEq/min past 1 hour after dosing (1.7-fold change vs. baseline).
The glomerular filtration rate, determined by calculating creatinine clearance, and effective renal plasma flow, determined by PAH clearance, were significantly decreased by HCTZ by approximately 20% but not by either dose of compound A (data not shown). Plasma sodium and potassium levels, as well as hematocrit values, did not change following oral dosing of compound A at 10 or 50 mg/kg; however, plasma potassium levels in HCTZ-treated dogs tended to decrease toward the end of the study, but the decrease did not reach statistical significance.
Several cardiovascular parameters evaluated in the conscious dogs, such as heart rate, PR interval, QRS duration, QTc interval, and blood pressure, were not significantly changed by either compound A or HCTZ during the course of the experiment (data not shown). The plasma and urine levels of compound A were determined and shown to rise with time (Table 5). Similar to the findings in SD rats, the levels of compound A were found to be significantly higher in urine than in plasma.
Taken together, these data demonstrate that in two different animal species single oral doses of compound A evoke a diuretic and natriuretic response that is comparable to HCTZ but with minimal changes in urinary K+ excretion rate and without any significant cardiovascular liabilities.
The results of this study illustrate the characterization of the first small molecule Kir1.1 inhibitor with appropriate selectivity profile and pharmacokinetic properties for in vivo evaluation. Compound A is a potent inhibitor of Kir1.1 channels stably expressed in heterologous systems, and it accesses the channel through the cytoplasmic side. Residues lining the pore within the transmembrane region of the channel below the selectivity filter appear to contribute to high-affinity inhibition of Kir1.1 by compound A. Pharmacologic inhibition of ROMK upon oral dosing of compound A leads to diuresis and natriuresis in two different animal species, SD rats and dogs. The magnitude of these events is similar to those of the clinically used diuretic HCTZ, but compound A appears to have a more favorable urinary potassium/sodium ratio than the relevant doses of HCTZ in these short-term studies. Taken together, these data suggest that selective inhibitors of ROMK represent a novel mechanism for developing diuretic/natriuretic agents with the potential for enhanced efficacy. Assuming that the short-term effects predict chronic pharmacology, they have a more favorable potassium balance than the diuretics that are currently used in the treatment of hypertension and/or congestive heart failure.
Within the diuretic class, thiazides such as HCTZ are the most widely used as first-line therapy to treat uncomplicated hypertension or as an add-on therapy to other mechanism-of-action drugs such as angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers (Sood et al., 2010). Even when HCTZ is used as part of combination therapy with renin-angiotensin-aldosterone system blockade, patients with resistant hypertension often fail to achieve adequate blood pressure control. ROMK inhibitors may offer potential an added efficacy benefit in these populations.
Hypokalemia (serum potassium concentration <3.5 mEq/l) and elevations in fasting blood glucose are the major liabilities associated with thiazides (Palmer and Naderi, 2007). Loop diuretics such as furosemide are mostly used to treat acute episodes of pulmonary and peripheral edema in patients with congestive heart failure, but their long-term use in treating the disease is not recommended because the drugs lose efficacy with time and cause hypokalemia at higher doses. No new diuretics have been developed within the past 4 decades, but an ideal novel diuretic should be potassium neutral, provide equal or greater efficacy to clinically used diuretics, and be suitable for combination therapy. In this sense, ROMK has emerged as an attractive novel target for the development of such agents (Ji et al., 2008).
ROMK is present in two different regions of the nephron (Xu et al., 1997). Inhibition of ROMK at the TALH should mimic the effect of furosemide in providing natriuresis/diuresis. In addition, inhibition of ROMK at the CCD, where it participates in potassium secretion, may ameliorate the hypokalemia caused by loop and thiazide diuretics. The results presented in this study with the selective, small molecule compound A seem to support these expectations of a ROMK inhibitor. Thus, compound A provides diuresis/natriuresis effects comparable to those of clinically used diuretics but with a more favorable urinary potassium/sodium ratio in two different species, and under different experimental paradigms.
It is interesting to note that inhibition of ROMK at the CCD does not appear to decrease urinary potassium excretion, which could lead to hyperkalemia, despite the fact that ROMK contributes to potassium secretion in that part of the nephron. However, we can speculate that other mechanism(s) present at the CCD, such as high-conductance, calcium-activated potassium channels, are likely to contribute to potassium secretion, in particular under conditions of high luminal flow rates that result from inhibition of salt reuptake at the TALH. Indeed, compensatory mechanisms between ROMK and high-conductance, calcium-activated potassium channels in the distal part of the nephron have been observed in studies with mice lacking either channel (Rieg et al., 2007). Similar to furosemide and thiazide diuretics, ROMK inhibitors are expected to activate the renin-angiotensin system, which could attenuate the extent of natriuresis/diuresis upon long-term dosing. Long-term treatment with compound A will be needed to determine whether diuretic/natriuretic resistance develops with time, which may limit the utility of ROMK inhibitors as monotherapy agents. If resistance develops, it would be important to evaluate whether ROMK inhibitors could be administered in combination with an angiotensin-converting enzyme or an angiotensin II receptor blocker to enhance their efficacy. In addition, long-term treatment will also provide insight into the effects of ROMK inhibitors on plasma potassium levels over time.
The search for Kir1.1 inhibitors has provided a limited number of compounds with appropriate potency, selectivity, and physicochemical and pharmacokinetic properties to be used in proof of concept studies (Lewis et al., 2009; Bhave et al., 2011; Tang et al., 2012). Compound A represents the first Kir1.1 inhibitor that fulfils these criteria. The selectivity of compound A for Kir1.1 versus other Kir channels, such as Kir2.1 and Kir2.3, is especially noteworthy. Part of this selectivity appears to arise from the nature of specific residues that line the channel’s pore within the transmembrane domain below the selectivity filter, although other region(s) of the channel may also contribute to the high-affinity interaction of compound A with Kir1.1 channels. Although more studies need to be done to determine the in vivo efficacy of compound A after long-term dosing and the possibility of combination with other mechanism of action drugs, the results of the our study support the idea that ROMK represents a target of interest for the development of novel diuretics with the potential of having an improved plasma potassium profile.
The authors thank Randal Bugianesi, Rodolfo Haedo, Michael Margulis, and Kashmira Shah for expert technical assistance and Drs. Euan MacIntyre, Sandy G. Mills, Adam Weinglass, and Lihu Yang for important discussions during the course of this work.
Participated in research design: Garcia, Priest, Alonso-Galicia, Zhou, Felix, Owens, Roy, Kaczorowski, Pasternak.
Conducted experiments: Priest, Alonso-Galicia, Zhou, Felix, Brochu, Bailey, Thomas-Fowlkes, Liu, Swensen, Hernandez, Pai, Xiao, Hoagland, Owens.
Contributed new reagents or analytic tools: Tang, de Jesus, Pasternak.
Performed data analysis: Garcia, Priest, Alonso-Galicia, Zhou, Felix, Thomas-Fowlkes, Liu, Swensen, Hernandez, Pai, Hoagland, Owens.
Wrote or contributed to the writing of the manuscript: Garcia, Priest, Alonso-Galicia, Zhou, Kaczorowski, Pasternak.
- Received August 8, 2013.
- Accepted October 18, 2013.
↵1 Current affiliation: Kanalis Consulting, L.L.C., Edison, New Jersey.
↵2 Current affiliation: Lilly Research Laboratories, Eli Lilly & Co., Indianapolis, Indiana.
↵3 Current affiliation: Forest Research institute, Inc., Jersey City, New Jersey.
↵4 Current affiliation: Purdue Pharma LP, Cranbury, New Jersey.
↵5 Current affiliation: Novartis, Cambridge, Massachusetts.
↵6 Current affiliation: ImClone Systems, New York, New York.
↵7 Current affiliation: AbbVie, North Chicago, Illinois.
M.L.G., B.T.P., and M.A.-G. contributed equally to this work.
- body weight
- cortical collecting duct
- Chinese hamster ovary
- compound A
- distal convoluted tubule
- inwardly rectifying potassium channel
- dimethyl sulfoxide
- fetal bovine serum
- Hank’s buffered saline solution
- human embryonic kidney 293 cell line
- human ether-à-go-go related gene
- liquid chromatography with tandem mass spectrometry
- minimum essential medium
- polyethylene glycol
- renal outer medullary potassium channel
- thick ascending loop of Henle
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics