Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The diseases of cystic fibrosis (CF) and chronic obstructive pulmonary disease (COPD) are characterized by mucus-congested airways and submucosal glands, meta- and hyperplasia of the mucus-secreting goblet cells, chronic inflammation of the respiratory tract, and a protracted decline in pulmonary function (Celli et al., 1995; O'byrne and Postma, 1999; Pilewski and Frizzell, 1999). Mutations, of which over 700 have been identified, in the cystic fibrosis transmembrane conductance regulator (CFTR) are the known genetic cause of CF (Pilewski and Frizzell, 1999). Cystic fibrosis is a disease affecting approximately 30,000 individuals in the U.S. with a mean life expectancy of 30 years. In contrast, COPD affects 10 to 14 million individuals in the U.S. and has a later onset, approximately 45 years of age (Celli et al., 1995; O'byrne and Postma, 1999), compared with CF. Cigarette smoking is clearly an important contribution in the development of COPD. However, 10 to 15% of COPD patients are nonsmokers. In addition, only 15% of smokers develop COPD. Therefore, there are a number of unknown risk factors that predispose an individual to develop COPD. Except to curtail respiratory infections with antibiotics and the use of bronchodilators, there are no drugs for the treatment of CF and COPD respiratory diseases. The benzimidazolones would appear to offer some potential in the treatment of these diseases.

The secretion of fluid and electrolytes by the airways and submucosal glands facilitates the clearance of mucus. Agents that promote fluid and electrolyte secretion would, therefore, be of potential therapeutic benefit in CF and COPD. Transepithelial Cl secretion secondarily draws sodium and water into the lumen by electrical and osmotic coupling. To stimulate the secretion of Cl, a secretory agonist must cause the opening of two different ion channels, an apical membrane Cl channel and a basolateral membrane K⁺ channel. The opening of a Cl channel is required to allow for the exit of Cl from the cell across the apical membrane. However, if this were the only channel to be activated, the cell would depolarize, Cl would come into equilibrium, and Cl secretion would stop. Thus, to maintain an adequate driving force for apical membrane Cl exit, basolateral membrane K⁺ channels must also be activated to hyperpolarize the cells and thereby provide the driving force for Cl exit. Endogenous secretory agonists, mediated via intracellular signal transduction cascades, open both Cl and K⁺ channels. Pharmacological modulators of Cl secretion will be required to do the same. Several different types of Cl and K⁺ channels are now thought to contribute to the secretion of Cl, including CFTR and hIK1. The latter is an intermediate conductance inwardly rectified Ca²⁺-activated K⁺ channel that was recently cloned (Ishii et al., 1997; Joiner et al., 1997) and was shown to be expressed in Cl secretory epithelia (Gerlach et al., 2000).

The benzimidazolones, NS004 (5-trifluoromethy-1-(5-chloro-2-hydroxyphenyl)-1,3-dihydro-2H-benzimidazol-2-one) and NS1619 (1-(2'-hydroxy-5'-trifluoromethylphenyl)-5-trifluoromethyl-1,3-dihydro-2H-benzimidazol-2-one), were first reported by Olesen et al. (1994) as activators of maxi conductance K⁺ channels. Subsequent studies by Gribkoff et al. (1994) demonstrated that NS004 activates both wild-type and F508 CFTR, the most common CFTR mutation. Thus, as activators of both CFTR and K⁺ channels, the benzimidazolones appear to be a good class of compounds to investigate as potential stimulators of Cl secretion. Unfortunately, NS004, on its own, does not stimulate Cl secretion in T84 cells (Devor et al., 1996a), a well characterized human colonic Cl secretory cell line. However, we did find that the benzimidazolone, 1-EBIO (1-ethyl-2-benzimidazolinone or 1-ethyl-1,3-dihydro-2H-benzimidazol-2-one) causes a sustained stimulation of Cl secretion (Devor et al., 1996a,b). Nystatin-permeabilized monolayers revealed that 1-EBIO activates both an apical membrane Cl conductance and a basolateral membrane K⁺ conductance. Similar studies revealed NS004 fails to activate a basolateral membrane K⁺ conductance but does activate an apical membrane Cl conductance. The effects of NS004 were seen at concentrations of less than 10 µM, whereas those of 1-EBIO required much higher concentrations with a half-maximal effective concentration of approximately 600 µM. Given these differences, we undertook a series of structure activity studies with the goal of developing benzimidazolones of greater Cl secretory potency. The results reported here document our progress toward this goal.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials

The benzimidazolone derivatives shown in Tables 1 through 5 were synthesized according to the procedures described below. 1-EBIO, compounds 1 and 2 and the precursors for the synthesis of the various derivatives were from Aldrich Chemical Co. (Milwaukee, WI). Forskolin was obtained from Calbiochem (La Jolla, CA). Amiloride and bumetanide were obtained from Sigma Chemical Co. (St. Louis, MO). Forskolin and bumetanide were made as 1000-fold stock solutions in ethanol. Benzimidazolones were prepared as 100- or 1000-fold stock solutions in DMSO.

T84 Cell Culture

T84 cells were grown in Dulbecco's modified Eagle medium and Ham's F-12 (1:1) supplemented with 15 mM HEPES, 14 mM NaHCO₃, and 10% fetal bovine serum (FBS). The cells were incubated in a humidified atmosphere containing 5% CO₂ at 37°C. For measurements of short-circuit current (I_SC) T84 cells were seeded on to Costar Transwell cell culture inserts (0.33 cm²), and the culture media were changed every 48 h. I_SC measurements were performed on filters after 14 to 21 days in culture as previously described (Devor et al., 1996a,b).

I_SC Measurements

Costar Transwell cell culture inserts were mounted in an Ussing chamber (Vertical Diffusion chambers, Costar Corp., Cambridge, MA), and the monolayers were continuously short-circuited with a model C558 voltage clamp (University of Iowa, Department of Bioengineering). Transepithelial resistance was measured by periodically applying a 2-mV bipolar pulse, and the resistance was calculated using Ohm's law. The bath solution contained (in mM): 120 NaCl, 25 NaHCO₃, 3.3 KH₂PO₄, 0.8 K₂HPO₄, 1.2 MgCl₂, 1.2 CaCl₂, and 10 glucose. The pH of this solution was 7.4 when gassed with a mixture of 95% O₂-5% CO₂ at 37°C. Forskolin and the other benzimidazolone derivatives were added to both sides of the monolayers at the indicated concentrations. Bumetanide was added only to the serosal bathing solution. Changes in I_SC are calculated as a difference current between the sustained phase of the response and their respective baseline values.

⁸⁶Rb⁺ Uptake Studies

⁸⁶Rb⁺ uptake was measured using the method of Gasko et al. (1976) as modified by Garty et al. (1983) and us (Bridges et al., 1988). In this method, tracer uptake is driven by a large electrochemical potential gradient by passing K₂SO₄-loaded vesicles down a cation exchange column. The removal of the extravesicular K⁺ creates a chemical gradient for K⁺ loss from the vesicles, and, since the intravesicular counter ion, SO₄², is less permeant than K⁺, an inside negative potential is generated by the outward K⁺ gradient. We estimate the membrane potential to be nearly 200 mV, vesicle interior negative.

HEK-293 cells stably expressing hIK1 were grown on plastic dishes to near confluence, washed three times with uptake buffer (100 mM K₂SO₄, 10 mM 3-(N-morpholino)propanesulfonic acid, 1 mM EGTA, 1 mM EDTA, 1 mM MgCl₂, pH 7.2, with Tris base), scraped, and pelleted at 1000g (SW34 rotor, RC5B Sorvall centrifuge, Newton, CT). The cell pellet was resuspended, centrifuged twice, and resuspended in a small volume of uptake buffer (0.5 ml/plate). The cells were homogenized twice for 15 s with a Polytron homogenizer (Brinkmann). The homogenate was centrifuged at 1000g for 10 min, and the supernatant was collected. The supernatant was centrifuged at 20,000g for 1 h, and the pellet was resuspended in uptake buffer (0.5 ml). The uptake assay was run in a 96-well format using a Quatra 6 robot (Tomtec, Orange, CT). A 96-well plate was prepared with 30 µl of ⁸⁶Rb⁺ in sucrose in each well, and compounds in DMSO or vesicle alone (control) were added to the appropriate wells at various concentrations. Five concentrations of compound in duplicate were typically evaluated together with two control wells and two wells of 1 mM 1-EBIO for each compound tested. This allowed us to test six compounds per plate. A plate of 96 cation exchange resin columns was placed over a rack of 96 collection vials and placed on the Tomtec robot. The 30 µl of ⁸⁶Rb⁺-sucrose plus test compounds were aspirated into pipette tips on the Tomtec robot. A 96-well plate was then prepared with 100 µl of membrane vesicles per well obtained immediately after elution of the vesicles off a cation exchange column. This plate was immediately placed on the Tomtec robot, and the uptake was initiated by pipetting the isotope into the vesicle plate and mixing several times. At 3 min a 100-µl aliquot of the uptake reaction was taken and pipetted onto the cation exchange columns. The vesicles were then eluted into the collection vials with sucrose using a fresh set of pipettes. The collection vials were then placed in a scintillation vial and counted on a liquid scintillation counter (Packard Instruments, Meriden, CT). Control experiments demonstrated ⁸⁶Rb⁺ uptake was linear up to 3 min and that the benzimidazolone-stimulated uptake was completely inhibited by charybdotoxin (CTX), results consistent with the activation of hIK1. In addition, the benzimidazolones had no effect on ⁸⁶Rb⁺ uptake in membrane vesicles derived from control, untransfected HEK-293 cells. In addition ⁸⁶Rb uptake in vesicles from control cells was not blocked by CTX, whereas the benzimidazolone-stimulated uptake in vesicles from hIK1-transfected cells was completely inhibited by CTX.

Channel Expression

Xenopus laevis care and handling procedures were in accordance with University of Pittsburgh guidelines. X. laevis frogs were anesthetized with 3-aminobenzoic acid ethyl ester, ovaries surgically removed, and oocytes dissected in modified Barth's solution containing (in mM) 88 NaCl, 2.4 NaHCO₃, 1 KCl, 0.82 MgSO₄, 0.33 Ca(NO₃), 0.41 CaCl₂, 10 HEPES, penicillin-streptomycin (1%), and defolliculated by digestion in calcium-free ND96 solution containing collagenase (Life Technologies, Inc., Grand Island, NY). Oocytes were incubated at 19°C in modified Barth's solution. pBF plasmid containing the gene for hIK1 (kindly supplied by J. P. Adelman, Oregon Health Sciences University, Portland, OR) was linearized using PvuI restriction enzyme (Boehringer Mannheim), and 5'-capped mRNAs were generated using SP6 polymerase (mMESSAGE mMACHINE in vitro transcription kit, Ambion, Austin, TX). mRNAs were evaluated both spectrophotometrically and by agarose gel electrophoresis with ethidium bromide staining. Oocytes were injected with 5 to 50 ng of mRNA 1 to 7 days before recording.

Solutions

For inside-out patch-clamp recording, the bath contained (in mM) 145 K⁺ gluconate, 5 KCl, 2 MgCl₂, 1 EGTA, and 10 HEPES (pH adjusted to 7.2 with KOH). Sufficient CaCl₂ was added to the bath solution to obtain the desired free [Ca²⁺] (program kindly supplied by D. C. Dawson, University of Michigan, Ann Arbor, MI). The pipette solution contained (in mM) 140 K⁺ gluconate, 5 KCl, 1 CaCl₂, 1 MgCl₂, and 10 HEPES (pH adjusted to 7.2 with KOH).

Single-Channel Recording

Before patch-clamp recording, the vitelline membrane was carefully removed from the oocyte following cell shrinkage with hypertonic solution containing (in mM) 200 K⁺ gluconate, 20 KCl, 1 MgCl₂, 1 EGTA, and 10 HEPES (pH adjusted to 7.4 with KOH). Excised, inside-out single channel currents were recorded using a List EPC-7 amplifier (Medical Systems, Great Neck, NY) and were recorded on videotape for later analysis as described previously (Devor and Frizzell, 1993). Pipettes were fabricated from thin-walled borosilicate glass (World Precision Instruments, Sarasota, FL). All recordings were performed at a holding voltage of 100 mV. The voltage is referenced to the extracellular compartment, as is the standard method for membrane potentials. Inward currents are defined as the movement of positive charge from the extracellular compartment to the intracellular compartment and are presented as downward deflections from the baseline in all recording configurations.

Single-channel analysis was performed on records sampled at 5 kHz after low-pass filtering at 1 kHz. The NP_o (the product of the number of channels and the channel open probability) of the channels was determined using Biopatch software (version 3.11; Molecular Kinetics, Pullman, WA). NP_o was calculated from the mean total current (I) divided by the single-channel current amplitude (i), such that NP_o = I/i. The value of i was determined from the amplitude histogram of the current record.

Chemistry

General Procedure. Melting points were obtained on an Electrothermal melting point apparatus. ¹H NMR spectra were recorded on GE 300-WB FT NMR (300 MHz) spectrometer (General Electric, Cleveland, OH). Gas chromatography/mass spectrometry data were recorded on a HP 5985 spectrometer (Hewlett-Packard, Palo Alto, CA). Thin-layer chromatography (TLC) was performed on Sigma brand silica gel GF plates. Log P values were calculated by Hyperchem (Hyperchem, Inc., Gainesville, FL). Hammet  values and Taft E_S values were obtained from Hansch et al. (1995).

General Procedures for the Synthesis of 1-Alkyl-1,3-dihydro-2H-benzimidazol-2-one and 1,3-Dialkyl-1,3-dihydro-2H-benzimidazol-2-one. Procedure A (Davoll, 1960). Hydroxybenzimidazole (1) (2.61 mmol) was dissolved in 4 ml of hot ethanolic KOH (1.7 g in 25 ml of ethanol), then alkyl halide (1.4 mmol) was added, and the reaction was refluxed for 3 h while it was monitored by TLC. The reaction was cooled and quenched by the addition of 10% NaOH (10 ml) and then extracted with ether (3 × 25 ml) and washed with brine (3 × 15 ml) dried (anhydrous Na₂SO₄) and concentrated to give crude oil. This crude mixture was purified on the preparative TLC (70% ethyl acetate/hexane) to yield the mono- and dialkylated products.

Procedure B (Vernin et al., 1981). A mixture of hydroxybenzimidazole (1) (10 mmol) in benzene (40 ml), 50% aqueous NaOH (15 ml), benzyltriethylammonium chloride (2 mmol), and alkyl halide (30 mmol) was stirred for about 4 to 5 h at 60°C, and the reaction was monitored via TLC (benzene/ethyl acetate 6:1). After the completion of the reaction, the mixture was cooled and the organic layer was separated, washed thoroughly with water (3 × 30 ml), dried over anhydrous MgSO₄, concentrated, and solidified by cooling. The solid was washed with petroleum ether and crystallized.

Procedure C (Johnstone and Rose, 1979). Powdered KOH (4 mmol) was added to DMSO (2 ml) and was stirred for 5 min. After 5 min the benzimidazolone to be alkylated (0.1 g, 1 mmol) was added followed by alkyl bromide (2 mmol), and the reaction mixture was stirred at room temperature for about 15 to 20 min depending on the TLC results (50% ethyl acetate/hexane). The reaction mixture was then poured onto water (20 ml) and extracted with ethyl acetate (3 × 20 ml). The combined organic layer was washed with water (3 × 10 ml), dried (anhydrous Na₂SO₄), and concentrated to dryness on a rotary evaporator to give an oily residue. This was purified by preparative TLC (40% ethyl acetate/hexane) to yield mono- and dialkylated products.

Procedure D (Davoll and Laney, 1960). Orthophenylenediamines (4 mmol) were heated with urea (12 mmol) at 140-150°C for 15 min. Even though the melted mixture solidified within 10 min, the heating was continued for 15 min. The reaction mixture was cooled, and NaOH (2.5 N solution, 15 ml) was added. After most of the solid became soluble in NaOH, it was filtered. Concentrated HCl (15 ml) was then added drop wise to the filtrate till the pH became acidic and solid precipitated out. After standing for some time the solid was filtered, washed with water (2 × 10 ml) and hexane (2 × 5 ml), and air-dried to yield a light brown solid.

Compounds. 1,3-Dipropyl-1,3-dihydro-2H-benzimidazol-2-one (11). ¹H NMR (300 MHz, CDCl₃)  7.07-6.97 (m, 4 H, aromatic), 3.84 (t, 4 H, (CH₂CH₂CH₃)₂), 1.84-1.74 (m, 4 H, (CH₂CH₂CH₃)₂), 0.98 (t, 6 H, (CH₂CH₂CH₃)₂).

1-Ethyl-3-propyl-1,3-dihydro-2H-benzimidazol-2-one (14). ¹H NMR (acetone-d₆, 300 MHz)  7.13-6.9 (m, 4 H, aromatic), 3.9-3.75 (m, 4 H, CH₂CH₃, CH₂CH₂CH₃), 1.72-1.69 (m, 2 H, CH₂CH₂CH₃), 1.27-1.18 (t, 3 H, CH₂CH₃), 0.909 (t, 3 H, CH₂CH₂CH₃).

1-Ethyl-3-isopropyl-1,3-dihydro-2H-benzimidazol-2-one (15). ¹H NMR (300 MHz, acetone-d₆)  7.25-6.99 (m, 4 H, aromatic), 4.69-4.65 (m, 1 H, CH(CH₃)_2, 3.88 (q, 2 H, CH₂CH₃), 1.48 (d, 6 H, CH(CH₃)₂), 1.29 (t, 3 H, CH₂CH₃).

1-Ethyl-3-butyl-1,3-dihydro-2H-benzimidazol-2-one (16). ¹H NMR (300 MHz, CDCl₃)  7.07-6.99 (m, aromatic, 4 H), 3.95-3.88 (m, 4 H, CH₂CH₂CH₂CH_3, CH₂CH₃), 1.77-1.65 (m, 2 H, CH₂CH₂CH₂CH₃), 1.50-1.31 (m, 2 H, CH₂CH₂CH₂CH₃), 0.97-0.90 (m, 6 H, CH₂CH₂, CH₂CH₃, CH₂CH₃).

5,6-Dimethyl-1-ethyl-1,3-dihydro-2H-benzimidazol-2-one (24) and 5,6-dimethyl-1,3-diethyl-1,3-dihydro-2H-benzimidazol-2-one (31). Compound 24: ¹H NMR (300 MHz, acetone-d₆),  6.88 (s, 1 H, aromatic), 6.82 (s, 1 H, aromatic), 3.82 (q, 2 H, CH₂CH₃), 2.25 (s, 3 H, CH₃), 2.22 (s, 3 H, CH₃), 1.26 (t, 3 H, CH₂CH₃). Compound 31: ¹H NMR (300 MHz, acetone-d₆),  6.92 (s, 2 H, aromatic), 3.86 (q, 2 H, CH₂CH₃), 2.27 (s, 6 H, CH₃), 1.25 (t, 3 H, CH₂CH₃).

5,6-Dichloro-1-ethyl-1,3-dihydro-2H-benzimidazol-2-one (26, DCEBIO), and 5,6-dichloro-1,3-diethyl-1,3-dihydro-2H-benzimidazol-2-one (32). Compound 32 was prepared using procedure C except that the mono- and di-products were isolated during work-up. Powdered KOH (4.87 mmol, 0.273 g) was stirred with DMSO (2 ml) for 10 min, and then 5,6-dichloro-1,3-dihydro-2H-benzimidazol-2-one (21) (0.173 g, 0.85 mmol) was added followed by ethylbromide (3.54 mmol, 0.262 ml). The reaction was stirred for only 5 min, water (5 ml) was added, and the solid precipitated was filtered and dried to give the di-product (32): ¹H NMR (300 MHz, CDCl₃)  7.14 (s, 2 H, aromatic), 3.92-3.90 (q, 2 H, CH₂CH₃), 1.36 (t, 3 H, CH₂CH₃).

4-Fluoro-N-ethyl-2-nitroaniline (27a). Compound 27a was prepared by starting from 4-fluoro-2-nitroaniline using procedure C. m.p.: 60-62°C (61.5-62°C) (Yagupolskii et al., 1964). Mass: 184 (M⁺). ¹H NMR (300 MHz, CDCl₃)  7.91-7.88 (m, 1 H, aromatic), 6.85-6.80 (m, 2 H, aromatic), 3.33 (q, 2 H, CH₂CH₃), 1.37 (t, 3 H, CH₂CH₃).

5-Fluoro-2(N-ethylamino)aniline (27b). The same procedure was used as for compound 27a (1 g), ethanol (100 ml), palladium-carbon (1 g) hydrazine hydrate (1 ml). The reaction was complete in 2 h after similar work-up, and the crude mixture (500 mg) was used up as such for the next step.

5-Fluoro-1-ethyl-1,3-dihydro-2H-benzimidazol-2-one (27). ¹H NMR (300 MHz, CDCl₃)  9.20 (bs, 1 H, NH), 6.90-6.85 (m, 3 H, aromatic), 3.92 (q, 2 H, CH₂CH₃), 1.35 (t, 3 H, CH₂CH₃).

5,6-Difluoro-1-ethyl-1,3-dihydro-2H-benzimidazol-2-one (28) and 5,6-difluoro-1,3-diethyl-1,3-dihydro-2H-benzimidazol-2-one (33). Compound 28: ¹H NMR (300 MHz, CDCl₃)  9.05 (bs, 1 H, NH), 6.94-6.83 (m, 2 H, aromatic), 4.60 (q, 2 H, CH₂CH₃), 1.34 (t, 3 H, CH₂CH₃). Compound 33: ¹H NMR (300 MHz, CDCl₃, 300 mHz)  6.82 (m, 2 H, aromatic), 3.92-3.85 (q, 2 H, CH₂CH₃), 1.32 (t, 3 H, CH₂CH₃).

4-Bromo-N-ethyl-2-nitroaniline (29a). Compound 29a was prepared by starting from 4-bromo-2-nitroaniline using procedure C. m.p.: 89-90°C. HRMS, calculated for C₈H₉N₂OBr 243.9847, found 243.9848. ¹H NMR (300 MHz, CDCl₃)  8.30 (m, 1 H, aromatic), 7.95 (bs, 2 H, NH), 7.5 (m, 1 H, aromatic), 6.75 (m, 1 H, aromatic,), 3.35 (q, 2 H, CH₂CH₃), 1.35 (t, 3 H, CH₂CH₃).

5-Bromo-2(N-ethylamino)aniline (29b). Compound 29a (372 mg, 1.51 mmol) was dissolved in ethanol (50 ml), and 2.4 drops of concentrated HCl was added, followed by 10% palladium-carbon (380 mg). The reaction mixture was cooled in ice water, and then hydrazine hydrate (0.472 ml) dissolved in ethanol (5 ml) was added (Singh et al., 1995). The reaction was slowly warmed to room temperature and stirred until the starting material completely disappeared on TLC (20% ethyl acetate/hexane). The reaction was worked up by filtering the catalyst, and the filtrate was concentrated to dryness to yield a brown oily residue (165 mg). This was used as such for the preparation of 1-ethyl-5-bromo-1,3-dihydro-2H-benzimidazol-2-one.

1-Ethyl-5-bromo-1,3-dihydro-2H-benzimidazol-2-one (29). The crude oil (Compound 29b; 165 mg), obtained from the above step, was treated further using procedure C. ¹H NMR (300 MHz, CDCl₃),  8.5 (bs, 1 H, NH), 7.22 (m, 2 H, aromatic), 6.88 (d, 1 H, aromatic), 3.90 (q, 2 H, CH₂CH₃), 1.80 (t, 3 H, CH₂CH₃).

N,N'-Ditoluene-p-sulfonyl-o-phenylene diamine (30a). 1,2-Diaminobenzene (1 g, 5.31 mmol) was dissolved in dry pyridine (10 ml), and then p-toluene sulfonic acid was added and the reaction was left overnight. The reaction was then poured over ice and water, stirred, and the oily material slowly solidified on standing. The solid was filtered, washed with water several times, dried, and weighed to give 3 g of crude solid. The crude solid was crystallized with hot ethanol to give a pure solid. Yield: 1.6 g (72.7%). m.p.: 200-203°C (Cheeseman, 1962). Mass: 416 (M⁺). ¹H NMR (300 MHz, CDCl₃),  7.57 (m, 4 H, aromatic), 7.26 (m, 4 H, aromatic), 7.1-6.9 (m, 2 H, aromatic), 6.76 (s, 2 H, aromatic), 1.55 (s, 6 H, CH₃).

5,6-Dibromo-1-ethyl-1,3-dihydro-2H-benzimidazol-2-one (30) and 5,6-dibromo-1,3-diethyl-1,3-dihydro-2H-benzimidazole-2-one (34). Compound 30: ¹H NMR (300 MHz, CDCl₃),  9.43 (bs, 1 H, NH), 7.30 (s, 1 H, aromatic), 7.24 (s, 1 H, aromatic), 3.90 (q, 2 H, CH₂CH₃), 1.35 (t, 3 H, CH₂CH₃). Compound 34: ¹H NMR (300 MHz, CDCl₃),  7.24 (s, 2 H, aromatic), 3.85 (q, 4 H, CH₂CH₃), 1.30 (t, 6 H, CH₂CH₃).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Chemistry. The 1-substituted and 1,3-disubstituted benzimidazolones were synthesized using various known procedures such as procedure A (Davoll, 1960), procedure B (Vernin et al., 1981), or procedure D (Davoll and Laney, 1960). Procedure A (Davoll, 1960) involved heating the benzimidazolone with the alkyl halide in the presence of ethanolic KOH for several hours. However, the yields obtained by this method were extremely low (3-10%) so only a few compounds (e.g., 5, 6, 7, and 11) were synthesized using this procedure (Fig. 1A). Procedure B (Vernin et al., 1981) was a phase transfer catalysis technique using benzyltriethylammonium chloride (Fig. 1A). This method took almost 24 h, and the yields were between 10 and 28% (e.g., 10, 13, and 16). Hence, another simple and rapid method was employed, which had been previously used for alkylating phenols, alcohols, and amides (Johnstone and Rose, 1979). The reaction used stirring of KOH/DMSO at room temperature for 5 min (Fig. 1A) followed by addition of benzimidazolone and alkylhalide. The reaction was complete within 30 min, and the yields were between 50 and 90% (e.g., 4, 9, 11, 12, 14, and 15). Benzimidazolones with substitutions in the aromatic ring were prepared using procedure D (Davoll and Laney, 1960) (Fig. 1B). Substituted 1,2-diaminobenzene derivatives were heated with urea at 140-150°C for about 15 min and were then alkylated using procedure C. In case of symmetrically substituted benzimidazolones (e.g., 18, 21, and 23), alkylation using procedure C resulted in both mono- and dialkylated products, which were easily separated by preparatory chromatography. In contrast, alkylation of asymmetrically substituted benzimidazolones (e.g., 20, 22, and 29), where the aromatic ring is substituted in an asymmetric manner, resulted in complex mixtures of isomers, which were difficult to separate. Therefore, it was necessary to first form the N-alkylnitroamines by alkylating the nitro amine using procedure C and then reducing the nitro amine to diamine (Fig. 1C), followed by heating with urea to give the monoalkylated benzimidazolone with substituents like chlorine, bromine, or fluorine only at the 5 position in the phenyl ring (procedure D). As for compound 30 the precursor 1,2-diamino-4,5-dibromobenzene (30) was synthesized as shown in Fig. 1D, and the final product was obtained after a series of reactions (Fig. 1D).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Schemes (A-D) for the synthesis of various benzimidazolones.

Structure Activity Studies. 1-EBIO (compound 3, Table 2) was the first benzimidazolone to be shown to stimulate Cl secretion, and for the purposes of presentation here it will serve as our reference compound. The half-maximal effective concentration (EC₅₀) of 1-EBIO stimulation of Cl secretion in T84 cells as measured by I_SC was several hundred micromolar, as previously reported (Devor et al., 1996a,b) and as shown in Fig. 2 for comparison. Transepithelial impedance analysis (Singh et al., 2000) revealed that 1-EBIO activated both an apical membrane Cl conductance and a basolateral membrane K⁺ conductance as originally reported in the studies using nystatin-permeabilized T84 monolayers (Devor et al., 1996a,b). However, impedance analysis dose-response studies for these two effects of 1-EBIO revealed the activation of the apical membrane occurred at much lower concentrations (K_S ~ 30 µM) than did the activation of the basolateral membrane (K_S ~ 600 µM). Thus, the rate-limiting activity of 1-EBIO is the activation of basolateral membrane K⁺ channels, which have since been shown to be hIK1 (Gerlach et al., 2000). Therefore, in an effort to improve the potency of 1-EBIO stimulation of Cl secretion, we focused our attention on improving the activity of hIK1 activation. Newly synthesized compounds were first evaluated for their stimulatory effects on ⁸⁶Rb⁺ uptake into membrane vesicles derived from hIK1-expressing HEK-293 cells. [All compounds were tested a minimum of three times at five concentrations in the ⁸⁶Rb⁺ uptake assay. However, the concentration-dependent effects of the various benzimidazolones on ⁸⁶Rb⁺ uptake displayed complex kinetics and could not be uniformly fit to standard kinetic functions. Therefore, the results were normalized to the effect of 60 µM and 1 mM 1-EBIO at the same two concentrations of the test compound. Sixty micromolar was selected as a basis for comparison, because the most potent compound (26, DCEBIO) produced near maximal stimulatory effect at this concentration. One millimolar was used because of the limited availability of the various derivatives.] In addition, the activity of the most potent compound (compound 26, also referred to as DCEBIO) was verified and compared with 1-EBIO using the I_SC method and the patch-clamp technique.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   86Rb+ uptake into membrane vesicles derived from hIK1-expressing HEK cells. Values are the mean ± S.E.M. for n = 5-6 experiments for uptakes at 3 min at the indicated concentration. Control uptake, without drug, was approximately 75 pmol/100 µg of protein and was subtracted from the drug-treated values. The EC50 values for DCEBIO was 19 µM and for 1-EBIO 210 µM.

1,3-Nitrogen Substitutions. The first series of derivatives were prepared to evaluate the importance of the 1-ethyl group of 1-EBIO. Substitution of the 1-ethyl group with a hydrogen (1) methyl (2), propyl (4), isopropyl (5), butyl (6), t-butyl (7), or phenyl (8) groups all led to a decrease in the activity as hIK1 activators as measured by ⁸⁶Rb⁺ uptake (Table 2). All of the compounds were found to cause some degree of stimulation of hIK1 at 1 mM. At the lower concentration of 60 µM, the hydrogen (1) and isopropyl (5) derivatives were inactive and the t-butyl (7) and phenyl (8) derivatives caused very little stimulation. These results suggest the 1-ethyl group is critical for the activation of hIK1 and that neither smaller substituents (hydrogen or methyl) nor a longer alkyl group (propyl or butyl) improve the activity. Moreover, substitution with a branched alkyl chain (isopropyl or t-butyl) decreased the activity compared with the respective unbranched derivatives.

Mono- and Disubstitutions in the Phenyl Ring at the 5 and 6 Positions. Table 4 summarizes the results with compounds substituted with a methyl, chloro, or fluoro groups at the 5 or 5 and 6 positions of the phenyl ring and with hydrogens at the 1 and 3 nitrogen positions. At the lower test concentration (60 µM), the disubstituted derivatives were more potent than the monosubstituted derivatives. The 5,6-dimethyl (18) and 5,6-dichloro (21) derivatives had a 30 to 35% higher potency compared with 1-EBIO at the lower test concentration, but this improved potency was not observed at the higher test concentration. As with some of the compounds in Table 3, the dose-response curves for compounds 18 and 21 were biphasic (data not shown).

I_SC Measurements. To further document the improved potency of DCEBIO (26) as an hIK1 activator, we also performed I_SC measurements on T84 monolayers. A typical I_SC experiment is shown in Fig. 3A, and the dose responses for DCEBIO and 1-EBIO are compared in Fig. 3B. The T84 monolayers were first stimulated with a maximal stimulatory concentration of forskolin (2 µM) to fully activate the apical membrane Cl conductance and the basolateral membrane cAMP-activated K⁺ channels (K_cAMP). Under these conditions the further activation of the basolateral membrane K⁺ conductance (i.e., hIK1) is rate limiting for Cl secretion and can therefore be measured as a further increase in I_SC. For the monolayer shown in Fig. 3A, forskolin caused the I_SC to increase to approximately 60 µA/cm². DCEBIO (100 µM) further increased the I_SC to over 150 µA/cm², and this increase was entirely inhibited by serosal CTX (50 nM), a blocker of hIK1 K⁺ channels. Bumetanide (20 µM), an inhibitor of the Na⁺:K⁺:2Cl-cotransporter, further inhibited the I_SC, thus providing evidence that the current was a Cl secretory current, as has been repeatedly shown for T84 cells. In six similar experiments the baseline I_SC and R_T were 0.8 ± 0.033 µA/cm² and 1468 ± 244 ohm cm². Forskolin increased the I_SC to 95 ± 10.3 µA/cm² and decreased the R_T to 389 ± 37.9 ohm cm². DCEBIO (100 µM) further increased the I_SC to 163 ± 7.3 µA/cm² and decreased the R_T to 227 ± 31.6 ohm cm². CTX decreased the I_SC to 72 ± 5.9 µA/cm² and increased R_T to 359 ± 27.7 ohm cm². Bumetanide further reduced the I_SC to 13 ± 0.9 µA/cm² and R_T to 451 ± 60.8 ohm cm². The DCEBIO decrease in R_T and the CTX increase in R_T are the expected results for the activation and blockade, respectively, of a basolateral membrane K⁺ conductance. Similar experiments were performed at several DCEBIO and 1-EBIO concentrations as summarized in Fig. 3B. These results illustrate the manyfold improved potency of DCEBIO compared with 1-EBIO.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effects of DCEBIO and EBIO on ISC in forskolin-treated T84 monolayers. A, typical ISC current trace showing the effects of forskolin (2 µM), DCEBIO (100 µM), CTX (50 nM), and bumetanide (20 µM) on ISC. B, dose-response curves for DCEBIO and EBIO from experiments similar to that shown in A. Values are the change in ISC after forskolin treatment, mean ± S.E.M. for n = 6-7 experiments. Vertical deflections are the current response to 2-mV pulses. The EC50 value for DCEBIO was 24.8 µM and for 1-EBIO 280 µM. These values compare quite favorably with the 86Rb uptake results.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Effects of DCEBIO and EBIO on ISC in T84 monolayers. A, typical ISC current trace showing the concentration-dependent effects of DCEBIO on ISC. B, dose-response curves for DCEBIO and 1-EBIO for experiments similar to that shown in A. Values are the change in ISC, mean ± S.E.M. for n = 7 experiments. Vertical deflections are the current response to 2-mV pulses. The EC50 value for DCEBIO was 45 µM and for 1-EBIO 1200 µM. In the absence of forskolin the stimulation of Cl secretion requires the activation of both apical membrane Cl channels and basolateral membrane K+ channels. Therefore, the EC50 values in the absence of forskolin can not be compared with those in the presence of forskolin (Fig. 3) where the increase in ISC depends only on the activation of hIK1.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Effects of 293B and CTX on forskolin-treated (A) and DCEBIO-treated (B) T84 monolayers. A, monolayers were first treated with forskolin (24 µM), then CTX (50 nM), and 293 B (3 µM). B, monolayers were first treated with DCEBIO (100 µM), then 293B (3 µM), and finally CTX (50 nM). Results are representative of six similar experiments.

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Effects of DCEBIO and NS004 on Cl currents in T84 monolayers with nystatin-permeabilized basolateral membranes and a mucosal to serosal Cl gradient. Results are representative of six similar experiments.

Patch-Clamp Studies. To further document the improved potency of DCEBIO and to verify its effects on hIK1, we performed patch-clamp studies. hIK1 was expressed in Xenopus oocytes and studied in excised inside-out patches at a bath Ca²⁺ concentration of 400 nM. Figure 7A shows a current trace of hIK1 channel activity in the absence and presence of 750 nM DCEBIO. DCEBIO is seen to cause a dramatic increase in hIK1 activity. The effects of DCEBIO were rapid in onset, requiring only 30 s to reach a new steady state. The effects of DCEBIO in the patch-clamp studies were observed at much lower concentrations than were required in the ⁸⁶Rb⁺ uptake assay and the I_SC measurements. These differences may reflect the different Ca²⁺ concentrations in these different assays as discussed under Discussion. The effects of DCEBIO on hIK1 channel activity were concentration-dependent as shown in Fig. 7B. The dose responses of both DCEBIO and 1-EBIO could be fit to the Hill equation and yielded K_1/2 values of 840 nM and 84 µM and Hill coefficients of 1.8 and 1.9 for DCEBIO and 1-EBIO, respectively. Thus in the patch-clamp studies, DCEBIO was 100-fold more potent than 1-EBIO. As in our previous studies (Syme et al., 2000) control, water-injected oocytes did not display any benzimidazolone-activated K⁺ channels.

View larger version (28K): [in this window] [in a new window]   Fig. 7.   hIK1 channel activity in excised inside-out membrane patches. A, continuous current trace of hIK1 channel activity from excised membrane patches before and after the addition of DCEBIO. B, concentration-dependent changes in channel activity caused by DCEBIO and 1-EBIO. Data were fit to the Hill equation as described in the text.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The discovery of 1-EBIO as a stimulator of Cl secretion is perhaps an excellent example of the role chance can play in science. Shortly after the publication of Gribkoff et al. (1994), which describes the effects of NS004 on CFTR, we attempted to obtain some NS004 for our own studies. However, NS004 was not commercially available at that time and we instead looked for a commercially available compound of similar structure. The compound we selected was 1-EBIO (Devor et al., 1996a,b). The positive results we obtained with 1-EBIO encouraged us to continue to investigate this class of compound. However, had we purchased the unsubstituted derivative (compound 1), it is almost certain we would not have continued to study the benzimidazolones, since this compound fails to activate hIK1 and does not stimulate Cl secretion. Indeed the structure activity studies with the compounds reported here suggest the ethyl group is an optimal substituent at the 1 nitrogen position and that only a hydrogen is tolerated at the 3 nitrogen position, as in 1-EBIO. The importance of the 3 position hydrogen may relate to the need for tautomere formation between the 3 nitrogen and 2 ketone groups. These results also explain why NS004, with a substituted phenyl group at the 1 nitrogen position failed to stimulate Cl secretion in our earlier studies (Devor et al., 1996a). Substitution at the 5 and 6 positions of the phenyl ring with either methyl or chloro groups appeared to impart activity to the 1,3-nitrogen-unsubstituted compound, at least at low concentrations (Table 4). Similar 5 and 6 position substitutions of 1-EBIO yielded compounds of improved activity compared with 1-EBIO (Table 5). The most potent of this compounds, DCEBIO (compound 26), was 4-fold better than 1-EBIO in the ⁸⁶Rb⁺ uptake assay, 20-fold better in the I_SC studies and 100-fold better in the patch-clamp studies compared to 1-EBIO. The rank order of the 5 and 6 position substituents was chloro > bromo > fluoro > methyl  hydrogen. Unfortunately, due to difficulties in isolating pure compounds, we were unable to synthesize the iodo, nitro, or amine derivatives in this series. Based on this limited series of compounds, it would appear that a chloro group satisfies an optimal molecular size (E_S = 1.14) and electronic character ( = 0.3) compared with the methyl group of similar size (E_S = 1.24) but negative  value (0.37) or the larger bromo group with an E_S = 1.34 and similar  value (0.37). Based on this reasoning, none of the iodo, nitro, or amine substituted derivatives are anticipated to have an improved activity compared with the 5,6-dichloro-1-ethyl compound (DCEBIO, compound 26). Hydrophobicity, as judged by the calculated log P values, also appeared to be of little predictive value in evaluating the potential activity of the various derivatives. Further structure activity studies with derivatives substituted in the 4 and 7 positions of the phenyl ring as well as with compounds replacing the 1 and 3 position nitrogens and the 2 position ketone are needed to ascertain if compounds of better potency than DCEBIO are possible. In this regard, we have shown that the benzoxazoles, chlorzoxazone, and zoxazolamine do stimulate Cl secretion and activate hIK1 (Singh et al., 2000; Syme et al., 2000). These benzoxazolones do not have an ethyl group at the 1 nitrogen position and perhaps their activity could be improved with such a substitution.

Several problems were encountered when attempting to evaluate the potency of the benzimidazolones as activators of hIK1. As already noted, several of the compounds displayed a biphasic dose response in the ⁸⁶Rb⁺ assay. A similar behavior was also observed in the I_SC assay (data not shown). We can offer no explanation for this phenomenon at this time. A biphasic dose response was not observed in the patch-clamp assay. A second problem we encountered especially in the I_SC measurements was a dose response with very steep dose dependence with some of the compounds. Compound 25 was one such compound. At concentrations below 30 µM, there was very little change in the I_SC, but at 100 µM there was a rapid and sustained increase in the I_SC that was not further stimulated at higher concentrations. These results suggest a cooperative kinetic interaction. However, not all of the compounds seemed to behave with the same degree of cooperativity in the I_SC studies. The patch-clamp studies on 1-EBIO and DCEBIO did yield dose-response results that could be fit to appropriate kinetic functions, and in both cases a Hill coefficient of approximately 2 was determined, consistent with a cooperative interaction. Unfortunately, it is not practical to perform patch-clamp studies with all of the derivatives. Finally, as already noted the activation of hIK1 in excised membrane patches required much lower concentrations than were needed in the ⁸⁶Rb⁺ uptake assay or the I_SC studies. One possible explanation for this difference is the influence Ca²⁺ levels may have on the benzimidazolone dose response. hIK1 is a Ca²⁺-activated channel. In the patch-clamp experiments, the Ca²⁺ concentration was set at 400 nM. In the ⁸⁶Rb⁺ uptake assay, the vesicles were filled with 1 mM EGTA, so the free Ca²⁺ concentration should be rather low compared with the patch-clamp experiments. Resting levels of intracellular Ca²⁺ in T84 cells was estimated to be around 100 nM, again lower than in the patch-clamp experiments. 1-EBIO has been shown to not influence the affinity of hIK1 for Ca²⁺ (Syme et al., 2000). However, the possible influence of Ca²⁺ on the benzimidazolone affinity has not been investigated, and is one explanation for the observed differences in the apparent affinities in the different assays. It is our hypothesis that the benzimidazolones act directly on the hIK1 channel protein to cause activation. However, it is possible that the site of action is a channel modulatory protein. If there is such a modulatory protein, then our studies indicate this modulatory protein is present in oocytes and mammalian cells and is retained in membrane vesicles and excised membrane patches.

In contrast to CF patients, COPD patients have functional apical membrane CFTR Cl channels, and there is no evidence for any K⁺ channel pathology in COPD. Thus, compounds like the benzimidazolone derivatives reported here and the previously reported effects of benzoxazoles suggest these types of compounds could be of therapeutic benefit in the treatment of COPD. The sustained Cl secretory responses these compounds elicit support this notion. Endogenous Ca²⁺-mediated agonists (e.g., acetylcholine) tend to cause only a transient secretory response and in the late phase inhibit a cAMP-mediated response (Barrett and Keely, 2000). The benzimidazolones or benzoxazoles acting alone or in combination with cAMP agonists cause a sustained and additive Cl secretory response. If mucus secretion is not stimulated by these compounds or there is a favorable shift in fluid secretion compared with mucus secretion, then these compounds should be of benefit in the treatment of COPD. Since chlorzoxazone and zoxazolamine are already FDA-approved drugs, their off label use for the treatment of COPD may be of immediate benefit. Singh et al. (2000) did observe a positive effect of chlorzoxazone on Cl secretion as measured by changes in nasal potential difference in healthy volunteers. However, appropriate clinical trials are necessary to establish the safety and efficacy of these drugs in COPD patients

The potential of the benzimidazolones in the treatment of CF is less certain. It is generally held that most CF-causing mutations in CFTR lead to a loss or absence of functional Cl channels in the apical membrane. However, this notion has been challenged recently (Kalin et al., 1999), and our own studies on primary cultures of human bronchial epithelial cells derived from F508-CFTR patients suggest some functional protein does reach the apical membrane (Devor et al., 2000). If a small amount of functional CFTR does reach the apical membrane in vivo, then these compounds are of therapeutic potential for CF. As reported by Gribkoff et al. (1994) NS004 and NS1619 activate wild-type and F508 CFTR. DCEBIO also activates an apical membrane Cl channel as shown in permeabilized monolayers (Fig. 6), and the stimulation of Cl secretion in intact monolayers (Figs. 4 and 5). The added benefit of DCEBIO is that, unlike NS004 and NS1619, DCEBIO also activates hIK1 thereby improving the driving force for Cl secretion. Although it is our hypothesis that the benzimidazolones act directly on CFTR, this remains to be demonstrated experimentally. The activation of CFTR by the benzimidazolones has only been observed in cell-attached patches or intact epithelial monolayers. In addition to other researchers, we have been unable to detect an activation of CFTR by the benzimidazolones in excised membrane patches, suggesting perhaps that CFTR is activated by an indirect mechanism. A considerable search has been initiated by the Cystic Fibrosis Foundation in collaboration with Aurora BioSciences (San Diego, CA) for compounds that will facilitate the expression of mutant CFTR at the apical membrane. Should such compounds become available, the benzimidazolones, acting directly or indirectly, would lessen the amount of CFTR that must be delivered to the apical membrane to achieve an equivalent Cl secretory response and thus be of therapeutic benefit in the treatment of CF.