This study investigates the effects of intracellular (pHi) and extracellular pH (pHe) on the efflux of Rb+ and Li+ in isolated rat hearts.87Rb and7Li NMR were used to measure Rb+ and Li+ content, respectively, of hearts, and 31P NMR was used to monitor pHi, pHe, and phosphate levels. After 30-min equilibration with Rb+ or Li+, effluxes were initiated by switching perfusion to a Rb+- or Li+-free, high-K+ (20.7 mM) Krebs-Henseleit buffer with 15 μM bumetanide. Monensin (2 μM) increased pHi from 7.10 ± 0.05 to 7.32 ± 0.07 and resulted in activation of Rb+ efflux; the first-order rate constant (k × 103, in min−1) increased from 42 ± 2 to 116 ± 16. Glibenclamide (4 μM) did not inhibit monensin-activated Rb+ efflux (k = 110 ± 17), whereas quinine (0.2 mM) slightly inhibited it by 19 ± 9%. Infusion of 15 mM NH4Cl during Rb+ washout increasedk for Rb+ efflux by 93% (81 ± 8), which was glibenclamide and quinine insensitive, and caused a transient increase in pHi to 7.25 ± 0.08. Intracellular Li+ inhibited NH4Cl-stimulated Rb+ efflux by 55%. Monensin and NH4Cl stimulated Li+ efflux by 40%, increasingk from 29 ± 3 to 43 ± 7 and 41 ± 3, respectively. The stimulation was not sensitive to 10 μM dimethylamiloride. Intracellular acidosis that resulted from the washout of NH4Cl (pH 6.86 ± 0.2) slightly inhibited Rb+ efflux (k = 36 ± 5), whereas NH4Cl itself in the absence of pHi changes did not markedly affect Rb+ efflux. A moderate increase in pHi (7.17 ± 0.06) produced by washout of 15 mM 2,2-dimethylpropionate (DMP)-Tris from hearts preequilibrated with DMP did not markedly affect Rb+ efflux. Neither global alkalosis (pHi 7.4, pHe 7.55) nor acidosis (pHi ≈ pHe 6.8) produced by 3 mM Tris base or 5 mM MES, respectively, affected Rb+ efflux. We suggest that intracellular alkalosis stimulates Rb+(K+) and Li+ effluxes by activating a nonselective sarcolemmal K+(Li+)/cation exchanger or a K+(Li+)-anion symporter.
- rubidium (potassium) ion efflux
- lithium ion efflux
- intracellular pH
- cation/cation exchanger
intracellular and extracellular pH (pHi and pHe, respectively) are important regulators of cation transport in myocytes. In early studies it was noted that pH affects the distribution of K+ in skeletal muscle such that the ratio of intracellular to extracellular K+ concentration ([K+]i/[K+]e) correlates with the ratio of intracellular to extracellular H+ concentration ([H+]i/[H+]e) (41). Intracellular alkalosis results in a net efflux of K+ from the intracellular space (17, 28, 29). Acidification of the extracellular medium by increasing the ratio of CO2 to concentration ([CO2]/[ ]) produces a similar effect (1). In cardiac muscle similar results were obtained for rabbit atria (35) and isolated rat heart (14) preparations. In contrast, in vivo experiments in dog hearts demonstrated a net uptake of K+ in response to respiratory acidosis ( increase) (38). Radiotracer studies (42K) on perfused rat and rabbit septa showed that this effect was caused by the inhibition of unidirectional K+ efflux by respiratory acidosis (33).
Under ischemic conditions, acidosis is more profound than respiratory acidosis and may affect the observed net efflux of K+ (18, 44, 45) mediated by ATP-sensitive K+(KATP) channels (44, 45) and K+-anion cotransporters (43). Patch-clamp experiments have shown that acidosis can increase (7, 11), decrease (5, 11), or not change (12, 27) the activity of KATP channels when the ATP concentration exceeds 0.1 mM. In addition to pH modulation of the above-mentioned K+ efflux pathways, H+ may exchange with K+ and other monovalent cations through the nonselective cation/H+exchanger. This exchanger has been characterized by Periyasamy et al. (32) in isolated sarcolemmal vesicles from bovine hearts.
Thus K+ efflux from myocytes in cardiac muscle is affected by pH changes under both normal and pathological conditions; however, the data are controversial. Measurements of net fluxes are somewhat difficult to interpret because they depend on both unidirectional influx and efflux rates. The evaluation of unidirectional fluxes using tracers, either radioisotope or magnetic resonance, may provide information on unidirectional influx and efflux rates. Rb+ has been used successfully as an NMR-sensitive tracer (87Rb, 28% natural abundance) for K+ in studies of unidirectional K+ uptake and efflux rates in isolated heart preparations (2, 6, 20, 21).
In this study we used 87Rb and7Li NMR spectroscopy to investigate the effects of changes in pHi and pHe on the unidirectional rates of Rb+ and Li+ efflux from myocytes in isolated rat hearts. It was found that intracellular alkalosis produced by infusion of NH4Cl or monensin results in a two- to threefold activation of Rb+ efflux, which is not glibenclamide sensitive, and 40% activation of dimethylamiloride (DMA)-insensitive Li+ efflux. We suggest that this effect could be caused by activation of either the cation/cation exchanger or the cation-anion cotransporter.
The investigation was performed in accordance with theGuide to the Care and Use of Experimental Animals published by the Canadian Council on Animal Care (2nd ed., Ottawa, ON, Canada, 1993).
Male Sprague-Dawley rats (300–400 g,n = 107) were obtained from the animal facility of the Institute for Biodiagnostics. Isolated hearts were perfused in the Langendorff mode with Krebs-Henseleit (KH) buffer aerated with 95% O2-5% CO2 at a constant flow (∼10 ml ⋅ min−1 ⋅ g wet wt−1) maintained by a peristaltic pump. KH buffer contained (in mM) 118 NaCl, 25 NaHCO3, 4.7 KCl, 1.75 CaCl2, 1.2 MgSO4, 11 glucose, and 0.5 EDTA. Temperature of the perfusate was maintained at 36.0–36.5°C by using water-jacketed reservoirs and perfusion line. Functional parameters such as left ventricular systolic pressure (LVSP), end-diastolic pressure (LVEDP), heart rate (HR), perfusion pressure (PP), and coronary flow were monitored continuously before, during, and after the acquisition of NMR spectra by a Gould four-channel physiological recorder. A water-filled balloon was inserted into the LV and connected to a Statham P23 Db pressure transducer to measure LVSP, LVEDP, and HR, which were 129 ± 17 mmHg, 2.9 ± 3.0 mmHg, and 277 ± 33 beats/min, respectively, after the equilibration period. PP was measured by the separate pressure transducer connected to the aortic line and was 68 ± 8 mmHg after the equilibration period. Cardiac performance was evaluated using the rate-pressure product (RPP) as HR × (LVSP − LVEDP), which was 34,900 ± 4,800 beats ⋅ mmHg ⋅ min−1.
For measurements of Rb+ and Li+ effluxes, perfusion was switched to a RbCl- or LiCl-containing solution ([Rb+] = 2.14 mM, [K+] = 3.76 mM, or [Li+] = 15 mM, [K+] = 4.7 mM) for 30 min. Rb+ or Li+ washout was then initiated by switching the perfusion to a Rb+- and Li+-free, high-K+ (20.7 mM) solution containing 15 μM bumetanide, a Na+-K+-2Cl−cotransport inhibitor (10). A high-K+ medium was used for the following reasons: 1) it causes cardiac arrest and eliminates the dependence of Rb+ and Li+ efflux on HR, which is important because some interventions may significantly affect HR;2) the conductance of K+ channels increases at high [K+]e, which makes the assay more sensitive;3) high [K+]esignificantly dilutes the Rb+released from cardiac cells, preventing its reuptake by Na+-K+-ATPase;4) high K+ dramatically decreases energy consumption [∼15% of that in beating hearts (19)] that prevents potential deleterious effects of drugs increasing [Ca2+]i(monensin, veratridine) by maintaining sufficient energy reserve. Two data points obtained during the first 4 min after Rb+ or Li+ efflux was initiated (0–2 and 2–4 min), which corresponded to Rb+ or Li+ washout from the extracellular space, were excluded from the analyses. The remaining data were presented as semilogarithmic plots and analyzed using linear regression (see Figs. 2 and 4). The slopes of the straight lines yielded pseudo-first-order rate constants (k) that are presented for convenience as k × 103 (in min−1).
Protocol A, which describes effects of monensin and related drugs on Rb+efflux, is shown in Fig.1 A. Pharmacological agents, such as monensin (2 μM), nigericin (1–2 μM), or veratridine (16 μM), were added at the beginning of Rb+ washout to ensure that anticipated changes in [Na+]i(monensin, veratridine), [K+]i(nigericin), and [H+]i(monensin, nigericin) had developed. In some experiments monensin was added after 6 min of Rb+ washout (Fig.1 A; see Fig.2 B) to reveal better kinetics of Rb+-efflux activation. Glibenclamide (4 μM) and quinine (0.2 mM) were added at the beginning of Rb+ washout where indicated (Fig. 1 A; see Fig.2 B and Table 2).
Protocol B (Fig.1 B) describes effects of changes in pHi and pHe on Rb+ efflux produced by other agents, such as NH4Cl, Tris base, and MES. Intracellular alkalosis was induced by infusion of 15 mM NH4Cl at 6 min of Rb+ washout. Intracellular acidosis was induced by washout of NH4Cl at 6 min after Rb+ efflux was initiated from hearts equilibrated with NH4Cl for 16 min. In control experiments NH4Cl was present before (10 min) and after initiation of Rb+efflux. The NH4Cl prepulse and washout method (8) is based on the following principles. Ammonia, which is in equilibrium with , penetrates into the cytoplasm faster than , where it binds H+, resulting in transient alkalinization of the cytoplasm. After NH4Cl reaches equilibrium between compartments, pHi returns to normal. Washout of NH4Cl results in faster removal of NH3 from the cytoplasm and shifts the equilibrium ( → NH3 + H+) toward dissociation and release of H+, which results in transient intracellular acidosis.
Global alkalosis was induced by infusion of Tris base (3 mM). Global acidosis (pHi and pHe) was induced by infusion of 5 mM MES at 6 min of Rb+ washout. Pharmacological agents (see Table 1) such as glibenclamide (4 μM), 4-aminopyridine (4-AP; 1.5 mM), 4,4′-dinitrostilbene-2,2′-sulfonate (DNDS; 0.1 mM), and quinine-HCl (0.1–0.2 mM) were added at the beginning of Rb+ washout. In Li+-inhibition experiments, hearts were loaded with 15 mM LiCl for 30 min simultaneously with Rb+ loading. Li+ washout was started simultaneously with Rb+ washout such that at the moment of NH4Cl infusion (after 6-min washout) Li+was present only in the intracellular space (22).
Protocol C (Fig.1 C) describes effects of pHi changes on Rb+ efflux produced by the weak, nonmetabolizable carboxylic acid 2,2-dimethylpropionate (DMP)-Tris salt (pH 7.4). Its principle of action is similar to that of NH4Cl. Protonated DMP (DMPH) penetrates faster than its anion into the cytoplasm, where DMPH releases H+, resulting in transitory acidification of the cytoplasm (16). Subsequent washout of DMP results in rapid diffusion of its protonated form (pK a ∼5.8) from the cytoplasm to the interstitium and shifts the cytoplasmic equilibrium toward proton binding to the DMP anion, thereby producing an increase in pHi. Intracellular alkalosis was induced by washout of DMP at 6 min after Rb+ efflux was initiated from hearts preequilibrated with 15 mM DMP-Tris for 16 min. Intracellular acidosis was produced by infusion of 15 mM DMP at 6 min of Rb+ washout. In control experiments DMP was present before (10 min) and after Rb+ efflux was initiated.
Protocol D (Fig.1 D) describes effects of monensin and NH4Cl on Li+ efflux. Hearts were loaded with Li+ by 30-min perfusion with KH buffer containing 15 mM LiCl at an unchanged Na+ concentration (143 mM) as described previously (22). Li+washout was initiated by switching perfusion to the buffer used for Rb+ washout (20.7 mM KCl, 15 μM bumetanide). DMA (10 μM), an inhibitor of Na+/H+exchanger (25), was infused for 20 min at the beginning of Li+ washout.
Similar protocols were used in the absence of Rb+ for the measurements of phosphates and pHi. Because the Pi level is very low in KCl-arrested hearts, it is very difficult to measure pHi using the Pi peak. Therefore, before KCl arrest and pHi manipulations, the hearts were treated with 10 mM 2-deoxy-d-glucose (DG) for 8 min, which was then washed out for 6 min. This resulted in the appearance of a small but clearly detectable peak of 2-deoxyglucose 6-phosphate (DG-6-P), which was used as a pHi probe. A short period of treatment with DG resulted in only slight decreases in the phosphocreatine (PCr) and ATP levels to 83 ± 14 and 89 ± 2% of their pretreatment levels, respectively. 2-Aminoethylphosphonic acid (AEPA)-Na salt (3 mM, pH 7.4) was used as a pHe probe in the experiments with monensin (protocol A) and Tris and MES (protocol B) and was added 6 min before the above agents were added.
All NMR spectra were obtained using a Bruker AM-360 WB spectrometer with a 20-mm Morris Instruments broad-band probe placed in a wide-bore, vertical 8.4-T magnet. The 23Na signal (95.2 MHz) from the heart and surrounding bath was used for adjusting the homogeneity (shimming) of the magnetic field. The NMR tube was continuously heated from outside with a flow of air, the temperature of which was maintained at 36°C by the temperature control unit of the NMR spectrometer. The temperature of the heart was maintained mainly by a high flow (14–17 ml/min) of thermostated perfusate. The temperature was checked in some experiments by placing the thermocouple in the right ventricle and was in the range of 36.0–36.5°C.
87Rb NMR spectra were acquired at 117.8 MHz with a 2-min time resolution. The pulse duration was 55 μs (90° flip angle), recycle time was 10 ms, sweep width was 18 kHz, and memory size was 512 data points. A capillary containing 1 M RbCl and 5 M KI was used as a concentration reference (2) in which I− serves as a shift reagent. The reference signal was stable during experiments carried out on any given day (≤4 hearts) such that the standard deviation did not exceed 5%.87Rb NMR spectra were obtained in a “dry” perfusion mode by aspirating the perfusate from the bottom of the 20-mm NMR tube (21). Postprocessing by exponential filtering with a line-broadening factor of 150 Hz was applied. After baseline correction (spline), peak areas were measured using an integration subroutine (Bruker). Because only the kinetics of Rb+ efflux were pursued (rate constant) in this study, no attempt was made to measure absolute concentrations of Rb+.
7Li NMR spectra were acquired as described previously (22) on a Bruker AM-360 spectrometer using a broad-band probe (Morris Instruments) operating at 139.95 MHz.7Li NMR spectra were collected every 2 min (60° pulse, 2.2-s repetition time). A capillary containing 0.5 M LiCl plus 50 mM (Tris)3DyTTHA-3Tris ⋅ HCl (7 μmol of Li; Dy, dysprosium; TTHA, triethylenetetraminehexaacetate) was used as a chemical shift and concentration reference. Before Fourier transformation, free induction decays were exponentially multiplied using a line-broadening factor of 20 Hz. Peak heights were used for spectral quantification using a peak-picking subroutine provided by Bruker.
31P NMR spectra were acquired at 145.8 MHz with a 2-min time resolution using a broad-band Bruker probe. The pulse duration was 35 μs (60° flip angle), recycle time was 2 s, sweep width was 10 kHz, memory size was 4 K data points, and the line-broadening factor was 20 Hz. After baseline correction (spline), the heights and chemical shifts of the PCr, Pi, and β-phosphate of ATP peaks were measured using a peak-picking subroutine (Bruker) and referred to their initial values. Cytoplasmic pH was assessed from the chemical shift of the DG-6-P and Pi peaks using calibration curves, as described previously (20, 21). pHe was measured using dependence of the chemical shift of AEPA on pH as a calibration curve. The curve was obtained at 36°C using a medium identical to KH buffer except for the absence of NaHCO3, which was replaced with 25 mM HEPES/HEPES-Na (pH 6.0–8.0).
First-order rate constants for Rb+and Li+ efflux (k × 103min−1) were calculated as follows. Natural logarithms of Rb+peak intensity (ln Rb) were plotted as a function of time (see Figs. 2and 5). In all experiments, the first two data points (0–2 and 2–4 min) were ignored because they were affected by changes in the extracellular Rb+ or Li+ concentration (20-22). The rest of the data were subjected to the linear regression analysis, and k was calculated as the slope of a straight line (see Figs. 2 and 5). In the experiments in which pHi was changed at 6 min of Rb+ or Li+ washout, the data points for the 6- to 20-min time interval were analyzed (see Figs. 2 and 4). In these experiments, only linear parts of the experimental curves were used for calculations of k (5 data points between 9 and 17 min) because activation of Rb+ efflux developed in time [see Figs. 2, A andB (monensin), and 5 (NH4Cl)].
Monensin, nigericin, veratridine, bumetanide, glibenclamide, 4-AP, RbCl, LiCl, quinine-HCl, Tris base, MES, NH4Cl, DG, α-cyano-4-hydroxycinnamic acid (CNHCA), tetramethylammonium chloride (TMA-Cl), and DMP (pivalic acid) were purchased from Sigma Chemical (St. Louis, MO). DNDS was supplied by Molecular Probes (Eugene, OR), and DMA was supplied by Research Biochemicals International (Nattick, MA). Glucose and salts from BDH (Canada) were reagent and analytic grades. Monensin and nigericin were prepared as stock solutions in 95% EtOH and were diluted with water (6% EtOH vol/vol) to prepare solutions for infusion (120 μM each). The concentration of EtOH in the perfusion solution did not exceed 0.1%. Glibenclamide and veratridine were dissolved in DMSO and then diluted with water to yield 240 μM and 1 mM solutions, respectively, in 6–10% (vol/vol) DMSO. To prevent glibenclamide from precipitating from the DMSO-water solution, it was made alkaline by the addition of 1 mM NaOH. Bumetanide was dissolved in DMSO and added to the solution used for Rb+ washout to bring its concentration to 15 μM and DMSO to 0.05%. The final concentration of DMSO in the perfusion solution did not exceed 0.15%. 4-AP (120 mM), NH4Cl (900 mM), quinine-HCl (12 mM), CNHCA (60 mM), and DNDS (6 mM) were used as aqueous solutions; pH was adjusted to 7.4 using HCl and NaOH. DMP (900 mM) was neutralized with Tris base to bring the pH to 7.4. Tris base (180 mM), MES (300 mM), TMA-Cl (900 mM), and DMA (0.6 mM) were dissolved in water.
Data are presented as means ± SD. Comparisons were made using the StatView computer program for Macintosh. A two-tailed paired Student’st-test was used for comparisons within groups (pH changes), and a one-factor ANOVA was used for comparisons between groups at a significance level of 95%.
Effects of monensin.
The rate constant for Rb+ efflux increased by 180% in response to the infusion of monensin (2 μM), a Na+-specific ionophore (Fig.2, A andB, and Tables1 and 2). The activation developed in time and reached maximum after 8 min of monensin application. This time dependence of the activating effect is clearly seen in Fig. 2 B, where the experiments with different times of monensin addition (0 and 6 min after initiation of Rb washout) are compared. Monensin also caused alkalinization of the intracellular medium such that pH increased from 7.1 to 7.4 (maximum) over an 8-min period and reached a steady-state level of ∼7.35, which was maintained for another 8 min (Fig. 3). The “instantaneous” rate constant calculated as [ln Rb(t) − ln Rb(t + 2)]/2 increased over 8 min and reached a maximal level of 120–140 (Fig.4). Time courses ofk and pHi increases were qualitatively similar, although dependence for k was shifted to the right (Fig. 4). This shift probably reflects a nonlinear dependence of k on pH. The average increase in pH caused by monensin over the last 8 min of its action on Rb+ efflux was 0.21 ± 0.06 units (1.6-fold decrease in [H+]i) (Table 2). The pHe (Fig. 3 and Table 2) and PCr (102 ± 17%, averaged over 14 min) did not change in response to the addition of monensin, whereas ATP decreased slightly to 89 ± 12% of its initial value. Glibenclamide (4 μM), an inhibitor of KATP channels (Table1), did not markedly suppress the activating effect of monensin on Rb+ efflux (Fig. 2 and Table 2). Quinine (0.2 mM), an inhibitor of the mitochondrial and sarcolemmal K+/H+exchanger (9, 32), moderately inhibited Rb+ efflux in the presence of monensin (28%) and did not affect basal Rb+ efflux (not shown). This corresponds to 43% inhibition of the monensin-stimulated fraction of Rb+ efflux. However, treatment with quinine plus monensin depressed the recovery of contractile function (RPP; by 32 ± 17%), measured after 20-min washout of the drugs and high KCl (K+) plus bumetanide. It was caused by some decrease in HR (19 ± 16%) and an increase in LVEDP by 21.0 ± 20.0 mmHg at unchanged LVSP. After treatment with monensin alone, cardiac function (RPP) recovered completely to 97 ± 16% of its pretreatment level. These data, combined with the lack of any significant changes in PCr and ATP, show that monensin did not cause any detectable damage to cardiac cells under our experimental conditions. Thus activation of Rb+efflux by monensin was not associated with any increase in sarcolemmal permeability associated with energy deprivation and cell damage. This conclusion is consistent with the lack of inhibition of monensin-activated Rb+ efflux by glibenclamide, because it was shown previously that this inhibitor completely blocked activation of Rb+ efflux caused by energy deprivation with cyanide (21) and 2,4-dinitrophenol (24).
A series of control experiments was performed to test whether the activation of Rb+ efflux by monensin was caused by 1) the direct transfer of Rb+ across the sarcolemma as a complex with monensin (because of its limited Na+/Rb+selectivity), 2) an increase in [Na+]iand an associated increase in [Ca2+]i, or 3) an increase in pHi. Nigericin (1–2 μM), a K+-specific ionophore (see Table1) structurally related to monensin, produced a modest stimulation of K+ efflux and did not significantly change pHi (Table2). Veratridine (16 μM), an activator of Na+ channels (26) (see Table 1) that raises [Na+]i, caused an increase in Rb+ efflux of ∼50%, which is significantly less than the effect of monensin (180%). In two experiments, ouabain, an inhibitor of Na+-K+- ATPase (see Table 1) that reduces Na+ efflux and increases both [Na+]iand [Ca2+]i, did not affect the Rb+ efflux rate (k = 46 ± 2 and 36 ± 2, respectively). These data support the hypothesis that intracellular alkalosis is the main factor activating Rb+ efflux from rat heart myocytes during monensin treatment. None of the above agents significantly affected the recovery of contractile function after their washout simultaneously with the removal of KCl plus bumetanide.
Effects of NH4Cl infusion and washout.
To eliminate the effects of increases in [Na+]iand [Ca2+]ion Rb+ efflux, a NH4Cl pulse technique (8) was used to increase pHi. The rate of Rb+ efflux and its rate constant increased by 93% in response to the infusion of 15 mM NH4Cl (Fig.5 and Table 2), which increased pHi from 7.1 to 7.3 during the first 4 min of infusion and increased the osmolarity of the perfusate by 30 mM. pHi returned to its prepulse level 6–8 min after reaching its peak (Fig.6). The average increase in pHi over the 8-min period was 0.15 ± 0.07 units. Rb+ efflux kinetics followed pH changes: the level of activation increased in time, reached a maximum, and then decreased (see Fig. 5). High-energy phosphates and Pi remained unchanged (not shown). The rate constant for Rb+ efflux decreased slightly (k = 36 ± 5,n = 3) relative to the control when pHi decreased (6.86 ± 0.2,n = 3,P < 0.05) in response to the washout of NH4Cl (Fig. 6). NH4Cl itself did not markedly affect either Rb+ efflux (k = 49 ± 4,n = 3,P = 0.005) or pHi (7.10 ± 0.07,n = 3), when it was infused 16 min before the rate of Rb+ efflux was measured. TMA-Cl (15 mM) was infused instead of NH4Cl to test whether the activating effect of NH4Cl is caused by increased osmolarity of the intracellular medium and associated cell volume change. TMA-Cl did not significantly affect Rb+ efflux (k = 44 ± 3,n = 3). To probe the nature of the transporter(s) responsible for the observed alkalosis-stimulated Rb+ efflux, a variety of inhibitors (see Table 1) was tested:1) glibenclamide (4 μM), an inhibitor of KATP channels (Table2); 2) 4-AP (1.5 mM,n = 4,k = 83 ± 25), an inhibitor of some voltage-gated K+ channels;3) DNDS (0.1 mM,n = 4,k = 72 ± 10), an inhibitor of anion exchange and cotransport (3); and4) quinine (0.1–0.2 mM) (Fig. 5and Table 2). None of these substances significantly affected alkalosis-stimulated Rb+ efflux. Quinine in combination with NH4Cl depressed the recovery of cardiac function (by 33 ± 5%) measured after 20-min washout of NH4Cl, quinine, and high K+ plus bumetanide. It was mainly caused by a decrease in HR (by 36 ± 5%) and some increase in LVEDP (by 6.0 ± 6.3 mmHg) at unchanged LVSP. Quinacrine, which is a more potent inhibitor of K+/H+exchanger [inhibition constant (K i) = 10 μM] than quinine (K i = 60 μM) (32), did not affect NH4Cl-stimulated Rb+ efflux at low concentrations of 20 and 40 μM and abolished the efflux at 100 μM. However, treatment with this dose produced toxic effects, such as contracture and cardiac arrest, which persisted even after 20 min of drug washout. NH4Cl alone or in combination with drugs other than quinacrine did not produce any changes in cardiac performance measured after their washout.
NH4Cl infusion stimulated Rb+ efflux not only under conditions of KCl arrest but also in beating hearts in the presence of bumetanide. However, the effect was smaller, with the rate constant increasing from 47 ± 4 to 63 ± 2, a 39 ± 14% increase (n = 5,P < 0.05). The drugs were not tested on beating hearts because activation was relatively small and some drugs affected heart rate. Quinine caused cardiac arrest, perhaps blocking Na+ channels in a manner similar to that of its stereoisomer, quinidine (37).
Effects of DMP infusion and washout.
Washout of DMP from preloaded hearts (Fig.7; see Fig.1 C) resulted in a moderate and transient alkalinization of intracellular medium. In the first 2 min, pHi reached a maximum of 7.20 ± 0.03 and returned to normal. The average pH for the 6-min period that followed DMP removal was 7.17 ± 0.06 [Fig. 7, DMP(+/−)], and the change in pH (ΔpH) was 0.07. This modest and transient alkalosis caused no activation of Rb+ efflux (k = 42 ± 3,n = 4) relative to the control in the absence of DMP. Transient acidosis (pH 6.92 ± 0.06) caused by the infusion of DMP resulted in moderate inhibition of Rb+ efflux [k = 35 ± 2,n = 3, DMP(−/+)]. Hearts preequilibrated with DMP for 16 min showed a slight decrease in pHi (7.01 ± 0.05) and Rb+ efflux [k = 33 ± 4,n = 3, DMP(+/+)]. Thus there is little difference (−27%) in kvalue between hearts with DMP-induced alkalosis and DMP control. CNHCA (1 mM) did not affect the response of Rb+ efflux to DMP, indicating a lack of Rb+-DMP cotransport through the monocarboxylate carrier. DMP did not affect levels of high-energy phosphates and recovery of contractile function after washout of DMP and simultaneous removal of KCl and bumetanide.
Effects of global alkalosis and acidosis.
To test whether Rb+ efflux was affected by changes in both intra- and extracellular pH, the effects on Rb+ efflux of a weak base, Tris, and a weak acid, MES, were investigated. The efflux was not affected by global alkalosis or global acidosis produced by the infusion of 3 mM Tris base or 5 mM MES acid, respectively (Table 2). Tris caused a rapid increase in pHe and pHi such that they reached new steady-state levels of 7.55 and 7.4, respectively (Table 2) within 2 min (time course not shown). MES infusion resulted in a rapid decrease in pHe and pHi to 6.8 and 6.7, respectively (Table 2). Fast changes in pHiimply that any pH gradients imposed by a pHe jump do not persist any longer than 2 min. Tris and MES did not affect levels of high-energy phosphates and subsequent recovery of contractile function.
Effect of intracellular alkalosis on Li+ efflux.
To test the ionic specificity of alkalosis-activated Rb+(K+) transporter(s), hearts were loaded with Li+ as described previously (22) (Fig. 1 D). The rate constant of Li+ efflux increased by ∼40% from 29 ± 3 (n = 4) to 43 ± 7 (n = 5,P = 0.008) and 41 ± 3 (n = 5,P = 0.007) in response to NH4Cl and monensin infusions, respectively. An inhibitor of Na+/H+exchanger, DMA (10 μM), inhibited neither NH4Cl-stimulated (k = 45 ± 8,n = 3,P = 0.01) nor monensin-stimulated (k= 49 ± 3,n = 3,P = 0.01) Rb+ efflux. These data indicate that the stimulated fraction of Li+ efflux is not caused by the operation of amiloride-sensitive Na+/H+exchanger and can be mediated by the pH-dependent carrier(s) involved in Rb+ efflux. To test this hypothesis, the effect of intracellular Li+ on Rb+ efflux was investigated. Hearts loaded with Li+([Li+]i= 21.5 ± 3.4 mM) (22) showed a reduced response to NH4Cl stimulation (Fig. 5); the rate constant for Rb+ efflux decreased from 81 ± 11 to 60 ± 9 (n = 4,P = 0.004). Thus NH4Cl-stimulated fraction of Rb+ efflux decreased to ∼45% [(60 − 42)/ (81 − 42)] of maximal NH4Cl effect. This inhibitory effect reflects competition between intracellular ions (Rb+ and K+ with Li+) because Li+ was present only in the intracellular space. This inhibition was not caused by decreased [Rb+]i, because Li+ loading only slightly decreases [Rb+]iplus [K+]i(12%) (22) and Rb+ efflux rate constant does not depend on [Rb+]iin a wide range of its concentrations (24). A lack of deleterious effects of 15 mM LiCl on cardiac energetics and contractile function was demonstrated (22). The addition of monensin, NH4Cl, and DMA during Li+ washout did not change subsequent recovery of contractile function.
Mechanisms of activation of Rb+ efflux by monensin, NH4Cl, and DMP.
Monensin functions as an ionophore, carrying Na+ into the cell (down its concentration gradient) and extruding H+ into the interstitium (31, 34,36). Under similar experimental conditions (hyperkalemia and cardiac arrest by carbachol) we have shown that 2 μM monensin increases [Na+]iby ∼30% in 10 min relative to its level in arrested hearts (23). However, cardiac arrest resulted in a decrease in [Na+]irelative to that in beating hearts (20, 23) such that monensin in effect nearly restored the level of Na+ to that observed in beating hearts. Nonetheless, an increase in [Na+]i(if any) could activate Na+-dependent K+ channels (30) and stimulate Rb+ efflux. However, the moderate stimulatory effect of veratridine on Rb+ efflux (∼50%), which is known to increase [Na+] and [Ca2+] (26), indicates that these ions could contribute only partially to the stimulatory effect of monensin, which was 3.6 times larger (∼180%). Furthermore, the stimulatory effect of veratridine could be associated in part with Rb+ efflux through the Na+ channels, which becomes less selective to Na+ in the presence of veratridine (26).
A fraction of the monensin-stimulated Rb+ efflux could be ascribed to limited selectivity of monensin for Na+ relative to K+(Rb+), which could transfer Rb+ across the sarcolemma in the form of a Rb+-monensin complex. This fraction cannot be large because the stability of monensin complexes in water decreases in the following order: Na+, Li+, K+, Rb+ (526, 286, 77, <10 M−1, respectively) (15). Furthermore, nigericin, a K+-specific ionophore (K:Na selectivity ratio of 24), did not produce any significant stimulation of Rb+ efflux. These selectivity profiles indicate that the fraction of K+/Rb+transferred as a monensin complex cannot exceed that of nigericin. Indeed, nigericin might produce an even greater stimulatory effect than monensin if it were the major mechanism for monensin stimulation. This is to be anticipated because of the higher selectivity of nigericin for K+(Rb+), comparable pK values for both ionophores, and the assumption of similar permeabilities for both ionophore-K+(Rb+) complexes. Theoretically, nigericin should produce intracellular acidosis (25, 34) by exchanging intracellular K+ for extracellular H+. This was not observed in our experiments because of the small activation of K+ efflux and the presence of a functional Na+/H+exchanger, which can reduce intracellular acidosis.
We postulate that the largest contribution to the monensin effect resulted from intracellular alkalosis. Indeed, alkalosis produced by the infusion of NH4Cl resulted in a high degree of stimulation of Rb+ efflux (∼100% increase). In this case, [Na+]iand [Ca2+]iare not expected to increase. The infusion of NH4Cl and DMP also increased the osmolarity of the extracellular medium by 30 mM and possibly caused transient cell shrinkage. However, NH4Cl increased and DMP decreased Rb+ efflux. Furthermore, the addition of 15 mM TMA-Cl instead of NH4Cl did not affect the Rb+ efflux. In contrast, monensin, which increases [Na+]i, should result in transient cell swelling, i.e., an effect opposite to shrinkage in the presence of NH4Cl. However, activation of Rb+ efflux occurred in both cases. In other words, there was no correlation between direction of volume changes and activation of Rb+efflux. Note that the Na+-K+-2Cl−cotransporter, which is activated by volume changes (10), was blocked by 15 μM bumetanide.
It is also possible that extracellular was exchanged for intracellular Rb+ through the K+/cation exchanger, thereby increasing Rb+ efflux. However, stimulation of the hypothetical Rb+/ exchange did not occur when the intra- and extracellular compartments were equilibrated with , which was present on both sides of the sarcolemma. Modest alkalosis induced by washout of the anion of the weak carboxylic acid DMP did not result in stimulation of Rb+ efflux. Contrariwise, pHi decrease produced by DMP infusion or NH4Cl washout slightly inhibited Rb+efflux (∼20%). Interestingly, transport of monocarboxylates into guinea pig myocytes inhibited KATPcurrent, whereas their washout activated the current, possibly caused by intracellular acidosis and alkalosis, respectively (5). Thus the stimulation of Rb+ efflux by monensin and NH4Cl indicates that the major factor activating Rb+efflux is intracellular alkalosis. The logarithm of the Rb+ efflux rate constant depended linearly on pHi in the narrow pH range between 7.1 and 7.35 (Fig. 7), revealing a strong activation of Rb+ efflux, and was pH independent in the range between 6.85 and 7.05. The slope of the straight line is ∼2, which indicates participation of more than one proton or OH− in the transport process. For instance, one H+ or OH− could be transported as a substrate, whereas another H+could play a regulatory role. Alternatively, transport of 2 H+ or 2 OH− could be coupled to the efflux of 2 Rb+. The pK of ionizable group(s) involved in the activation of Rb+ efflux is ∼7.0–7.1.
Mechanisms responsible for alkalosis-activated Rb+ efflux.
The large increase in Rb+ efflux with increasing pHi could be associated with a positive pH dependence (dk/dpH > 0) of various K+ transporters, such as voltage-gated and ATP-, Na+-, and Ca2+-sensitive K+ channels as well as K+-anion cotransporters. However, increasing pH on both sides of the sarcolemma did not produce any such activation. This implies that if the above transporters are involved in intracellular alkalosis-stimulated Rb+ efflux, extracellular alkalosis abolishes this effect. The lack of sensitivity to glibenclamide, 4-AP, and DNDS argues against involvement of KATP, 4-AP-sensitive channels, and /Cl−-dependent cation transporters, respectively. Similarly, the Na+-K+-2Cl−cotransporter can be excluded because all measurements were performed in the presence of 15 μM bumetanide. Patch-clamp experiments have shown that neither global acidosis nor alkalosis affects the activity of KATP channels, whereas intracellular alkalosis inhibits these channels (12). The increase in Rb+ efflux could be caused by the activation of a sarcolemmal K+/H+exchanger (32), which is similar to the mitochondrial exchanger (9,13), by intracellular alkalosis. However, the inhibitor of the sarcolemmal and mitochondrial K+/H+exchanger quinine has some effect on monensin-activated and no effect on NH4Cl-activated Rb+ efflux at moderate concentrations (0.1–0.2 mM), which exceeded theK i for the sarcolemmal exchanger (60 μM) (32). Note, however, that thisK i value was obtained using sarcolemmal vesicles isolated from bovine heart at a low [Rb+] or [K+] of 1 mM (32). In our experiments [Rb+]i+ [K+]iis ∼100 mM and [K+]eis 20.7 mM, which could dramatically reduce the efficiency of potential competitive quinine inhibition. In addition, rat heart exchanger can be less sensitive to quinine than bovine heart exchanger. Furthermore, the drug may inhibit other transport systems, such as Na+ and K+ channels, similar to its stereoisomer, quinidine (37). At the least, cardiac arrest produced by 0.2 mM quinine, which was observed in our experiments with beating hearts, is consistent with this interpretation. In the experiments on KCl-arrested hearts, quinine depressed recovery of contractile function assessed after washout of quinine and high K+. For these reasons the experiments with quinine on the intact hearts are not conclusive.
Other transport systems responsible for activation of Rb+ efflux by alkalosis could be1) a cation-anion cotransporter capable of symporting K+ and OH− or Cl− and2) a hypothetical H+-sensitive K+ channel with pH regulatory sites on both sides of the sarcolemma. Interestingly, there is some resemblance to nonselective ion channels of gap junctions, which can be activated by alkaline pHi (39). However, Findlay (12) did not find any significant effect of NH4Cl-evoked intracellular alkalosis on the background K+current in guinea pig myocytes over a wide range of membrane potentials (from −100 to 45 mV). Moreover, intracellular alkalosis inhibited KATP channels (by ∼35%) activated pharmacologically (12). Therefore, the channel hypothesis seems to be less likely.
If Rb+ efflux is coupled to the movement of acid equivalents (K+/H+exchanger or K+-OH−cotransporter), the Rb+ efflux is not driven by ΔpH. Under the conditions used in the majority of our experiments, the K+ gradient was ∼7 (log [K+]i/[K+]e∼ 0.85) and the H+ gradient changed from negative (outwardly directed, ΔpH = 0.30) to slightly negative (ΔpH = 0.10–0.20) and was much less than the K+ gradient. A similar situation occurs for the K+/H+exchanger in Amphiuma red blood cells, where K+ efflux coupled to H+ uptake is activated by osmotic swelling and is involved in regulatory volume decrease (4). In our experiments the kinetics of K+(Rb+) efflux are controlled by relatively small changes in transmembrane ΔpH, which decreased from 0.35 to 0.10–0.20 during intracellular alkalosis. Furthermore, global alkalosis and acidosis also decreased ΔpH to 0.10–0.15 (see Table 2) without any effect on Rb+ efflux. In other words, it is hardly possible that a small increase in driving force due to a decrease in ΔpH is responsible for very significant activation of Rb+ efflux. Rather, it is a reflection of kinetic regulation of the efflux by [H+]i. Interestingly, it was demonstrated (32) that in sarcolemmal vesicles Rb+/H+exchange was not active under conditions of global acidosis (pHi = pHe ∼6.0) or alkalosis (pHi = pHe ∼8.0).
According to Periyasamy et al. (32) the K+/H+exchanger can also catalyze partial K+/K+, K+/Li+, Na+/Na+, K+/Na+, and Li+/Na+exchanges. In our Rb+ efflux experiments the K+/H+exchanger could exchange intracellular Rb+ for extracellular K+ such that Rb+ efflux would be coupled to H+ and K+ entry. If so, [K+]emay affect Rb+ efflux, depending on the transporter kinetics model. In this study we have found that under normokalemic conditions ([K+]e= 4.7 mM, beating hearts) activation of Rb+ efflux resulting from NH4Cl-induced alkalosis decreased (k = 63 ± 2 vs. 81 ± 8 in KCl-arrested hearts). This can be explained by a higher maximal rate for Rb+/K+exchange relative to Rb+/H+exchange. Alternatively, the inhibition of alkalosis-activated Rb+ efflux by a decrease in [K+]emay result from the more negative plasma membrane potential, implying positive net charge movement during Rb+ efflux from the cytoplasm. In this case a decrease in [K+]erenders the diastolic membrane potential more negative and decreases the driving force that moves Rb+from the cell.
Inhibition of NH4Cl-activated Rb+ efflux by intracellular Li+ and activation of Li+ efflux by intracellular alkalosis imply that Rb+ and Li+ interact with similar pH-sensitive transporter(s). The quantitative difference in the levels of activation of Rb+ efflux (100 and 180%) and Li+ efflux (40%) can be explained by competition between higher [Rb+]i+ [K+]i(∼100 mM) and lower [Li+]i(∼22 mM) (22) as well as by lower maximal rates for Li+ efflux. Note that the differences between absolute values of basal and stimulated Rb+ and Li+ fluxes (J =k[ion]i) are even greater. In fact, Rb+ + K+ flux increased from 4.2 to 8.1–11.6 mM/min (ΔJ = 3.9–7.4 mM/min), whereas Li+flux increased from 0.62 to 0.92 mM/min (ΔJ = 0.3 mM/min). These data show that the carrier(s) responsible for alkalosis-stimulated Rb+ efflux is not absolutely selective to Rb+(K+), corresponding to properties of sarcolemmal monovalent cation/cation antiporter described by Periyasamy (32).
The above mechanisms imply the direct participation of H+ or OH− in Rb+ transport or their direct effects on ion transporters. We can exclude any indirect effects of intracellular alkalosis, such as effects on signaling pathways or affinity of Ca2+-binding proteins to Ca2+, because global alkalosis produced by Tris base did not activate Rb+ efflux.
Intracellular alkalosis is a relatively rare occurrence in cardiomyocytes relative to acidosis, which is characteristic of ischemia. Transient alkalosis can occur early in ischemia due to the breakdown of PCr at unchanged ATP and cessation of aerobic CO2production. If this occurs, the K+release that takes place early in ischemia (0–3 min) (18,45) can be, in part, an alkalosis-activated process. In addition, it has been shown in rat myocytes that stimulation of Na+/H+exchanger by pharmacological activation of protein kinase C results in intracellular alkalosis (42). A parallel decrease in extracellular and CO2 also results in intracellular alkalosis (14). In this study, an increase in extracellular K+ was observed, probably caused by K+ efflux from the cells (17).
In conclusion, intracellular alkalosis stimulates unidirectional effluxes of Rb+ and Li+ from myocytes in isolated rat hearts. We suggest that this phenomenon reflects the activation of a sarcolemmal K+(Li+)/cation exchanger that extrudes K+, Rb+, and Li+ and takes up cations such as H+ and K+. We do not yet have any direct evidence for transport of H+, such as stoichiometric Rb+ efflux and H+ influx, because no selective inhibitor of alkalosis-activated Rb+ efflux has yet been found. If such an inhibitor is identified, it should increase the alkalosis produced by either monensin or NH4Cl and allow calculation of the K+/H+stoichiometry.
This work was supported in part by a grant from the Heart and Stroke Foundation of Manitoba.
Address for reprint requests and other correspondence: V. V. Kupriyanov, Inst. for Biodiagnostics, National Research Council of Canada, 435 Ellice Ave., Winnipeg, MB, Canada R3B 1Y6 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 1999 the American Physiological Society