INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

SKELETAL MUSCLE FIBERS must regulate potassium transport to provide a compensatory response to the ion fluxes that occur during contractile activity. Under resting conditions, steady state is maintained by the active transport of potassium into the muscle fiber to compensate losses due to leakage. However, during contractile activity, potassium efflux can exceed potassium influx (30, 42); this exacerbated potassium loss is thought to contribute to muscle fatigue (4, 14, 26). The potassium lost by the muscle accumulates in the extracellular space and the general circulation (26, 30), but the rapid upregulation of inward transport activity eventually restores intracellular potassium after cessation of contractile activity (12, 37). The increased inward transport activity must eventually be downregulated to maintain steady state and prevent hypokalemia. In addition to its involvement in potassium homeostasis during contractile activity, inward potassium flux in skeletal muscle is also an important buffering mechanism for ingested potassium. After a meal, ingested potassium could increase plasma potassium to dangerous levels if it were not buffered (25, 43). Therefore, the mechanisms responsible for inward potassium flux in skeletal muscle are of great interest.

The best-known potassium influx pathway in skeletal muscle is the Na⁺-K⁺-ATPase. Na⁺-K⁺-ATPase activity can be stimulated by an increase in intracellular sodium (3, 40) and, in some nonmuscle cells, by phosphorylation of its -subunit (2, 3). Contractile activity, increased levels of circulating catecholamines, and increased levels of insulin are known stimulators of Na⁺-K⁺-ATPase activity in skeletal muscle (37). Inwardly rectifying potassium channels may also participate in potassium recovery, but probably only under conditions where locally high extracellular or low intracellular concentrations of potassium shift the equilibrium potential for potassium to values more positive than the resting membrane potential (19, 41). Another potential mechanism for potassium influx is through the activity of sodium-potassium-chloride cotransporters (NKCC). The regulation of skeletal muscle NKCC activity is the topic of this study.

NKCC activity provides an electroneutral, inwardly directed flux at virtually all physiological ion concentrations and membrane potentials encountered by muscle cells. In skeletal muscle at rest, the activity is low and would not contribute appreciably to net potassium flux (46). In fact, it has been argued that typical NKCC activity is not present in rat skeletal muscle (9, 10). However, we have previously reported the expression of NKCC-like proteins in skeletal muscle and found at rest the activity is quite low (46). Therefore, we were interested in whether NKCC activity could be stimulated, especially by stimuli that are known to increase Na⁺-K⁺-ATPase activity in skeletal muscle [e.g., contractile activity, catecholamines, and insulin (7, 15, 37)]. In addition, we were interested in whether the extracellular-signal regulated kinase 1/2 (ERK1/2) cascade mediates stimulation of NKCC activity, as has been reported for heart muscle after ₁-adrenergic stimulation (1, 22). This is a relevant question considering that contractile activity, insulin, and isoproterenol can stimulate the ERK1/2 pathway in muscle (39, 47).

We report here that epinephrine and contractile activity, but not insulin, stimulate skeletal muscle NKCC activity. The epinephrine-stimulated NKCC activity can be mediated via ₁- or -adrenergic stimulation. Furthermore, we demonstrate that the intracellular signaling pathway for stimulation of NKCC activity involves activation of the ERK1/2 arm of the mitogen-activated protein kinase (MAPK) pathways.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal care and muscle preparation. Female Sprague-Dawley rats (50-100 g) were used for all experiments. Animals were housed in light- and temperature-controlled quarters where they received food and water ad libitum. Animals were randomly assigned to experimental groups, and all animals were handled identically. Animals were anesthetized with pentobarbital sodium (45 mg/kg ip) for tissue removal. Soleus (25-45 mg each) and plantaris muscles (50-90 mg each) were removed by carefully dissecting the proximal tendons at the muscle origin and severing the distal tendon. The muscles were placed in oxygenated Krebs-Ringer buffer at 25°C in preparation for further treatment (see NKCC activity in skeletal muscle). The integrity of the muscle preparation was verified by the ability of the muscle to elicit a visible unloaded contractile response to a single electrical impulse and, in randomly chosen muscles, the membrane potential of some muscle fibers. Microelectrode measurement of membrane potential indicated intact preparations with initial potentials in a range of 70 to 75 mV. In selected muscles, membrane potentials changed by 0.8 ± 1.6 (SE) and 1.2 ± 1.7 during 30 min and 2 h of incubation, respectively, in oxygenated Krebs-Ringer solution at 25°C (n = 8 muscles). The Animal Care and Use Committee of the University of Tennessee Health Science Center approved all procedures.

⁸⁶Rb influx rate constant calculation. A first-order Michaelis-Menten rate constant for ⁸⁶Rb influx into muscle can be calculated if the following assumptions are made: 1) [⁸⁶Rb] in the bathing is much less than the Michaelis-Menten constant (K_m) for the transport processes, 2) the transport of ⁸⁶Rb is far from steady state, and 3) steady-state conditions exist for transport intermediates and diffusion processes. In these experiments, the first condition was met by the low concentration of ⁸⁶Rb in the solutions used to bathe the muscles (1 µCi/ml ⁸⁶Rb at 2-10 nmol Rb/µCi). To determine the transport kinetics and obtain an assay time for the second requirement, we measured the time course of ⁸⁶Rb uptake by muscles suspended in oxygenated (95% O₂-5% CO₂) Krebs-Ringer solution (in mM: 120 NaCl, 25.1 NaHCO₃, 4.7 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 1.3 CaCl₂, and 10 D-glucose, pH 7.4) at 30°C (Fig. 1). After incubation, the muscles were rinsed three times, 5 min each in 50 ml Krebs-Ringer solution at 4°C, blotted dry, weighed in tared scintillation vials, and counted as described in NKCC activity in skeletal muscle. The data indicate that a steady state for ⁸⁶Rb transport is obtained after 15 min of incubation. To assess the third requirement for calculation of a transport rate constant (steady-state intermediates and diffusion), potassium transport processes in the muscles were blocked by preincubating for 15 min at 30°C in oxygenated Krebs-Ringer solution containing 1 mM ouabain, 10 µM bumetanide, 5 mM tetraethylammonium, and 1 mM BaCl₂. The muscle was then transferred to an identical tissue bath containing ⁸⁶Rb and allowed to incubate at 30°C for various times, blotted dry, and immediately counted without rinsing (Fig. 1). Similarly, the rinsing procedure was verified by allowing a 10-min incubation of the treated muscle in the ⁸⁶Rb-containing bath followed by rinsing in Krebs-Ringer solution at 4°C for various times before counting (Fig. 1). From these data it is clear that most of the diffusion of ⁸⁶Rb is taking place during the first 2 min of incubation. Thus an incubation time of 10 min in the ⁸⁶Rb-containing solution was chosen for all experiments to minimize the impact of diffusion on the rate constants and avoid being near a steady-state condition for total ⁸⁶Rb uptake.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   A: uptake of 86Rb into the soleus muscle at 30°C approached steady state by 15 min of incubation in 86Rb-containing Krebs-Ringer solution (vehicle, n = 6 muscles per point). The combination of 1 mM ouabain, 10 µM bumetanide, 5 mM tetraethylammonium (TEA), and 1 mM BaCl2 in the incubating solution diminished the uptake of 86Rb () to an amount that is easily washed away within 5 min of washout in ice-cold Krebs-Ringer solution (). CPM, counts per minute; inset: expanded ordinate scale for diffusion () and washout () values. B: 86Rb uptake rate constants calculated from 86Rb uptake data indicate a large influence of the diffusion during the first 2 min of incubation, whereas incubation longer than 10 min was affected by the approach to steady state. On this basis, 10 min was chosen as the uptake period for calculation of the rate constants.

1

1

NKCC activity in skeletal muscle. The soleus and plantaris muscles were removed from anesthetized rats and immediately placed in oxygenated Krebs-Ringer solution. A stock solution of 10³ M bumetanide was prepared by dissolving bumetanide in 50% DMSO in Krebs-Ringer solution. The working bumetanide Krebs-Ringer solution was prepared by further diluting the stock bumetanide with Krebs-Ringer solution to a final concentration of 10⁵ M bumetanide. Krebs-Ringer vehicle was similarly prepared except that bumetanide was omitted. The final concentration of DMSO in the vehicle and bumetanide buffers was 0.5%.

5

₁- And -adrenergic stimulation and blockade. ₁- and -Adrenergic receptors were specifically stimulated by adding 30 µM phenylephrine (₁-adrenergic receptor agonist) or 30 µM isoproterenol (-adrenergic receptor agonist) to the bumetanide and vehicle incubation media. The -adrenergic receptor antagonist propranolol was added to the preincubation and incubation media of phenylephrine-treated muscle to block any effects that phenylephrine might have on the -adrenergic receptors. Likewise, the ₁-adrenergic antagonist prazosin was added to the preincubation and incubation media for isoproterenol-treated muscle. Separate experiments were performed to test the specificity of the phenylephrine stimulation. The ₁-adrenergic antagonist prazosin (50 µM) alone, or in combination with the -adrenergic antagonist propranolol (1 µM), was added to the preincubation and incubation media of phenylephrine-stimulated muscle. Parallel experiments were done for isoproterenol: propranolol, as well as a combination of propranolol and prazosin, were tested with the isoproterenol-treated muscle. Any effects that prazosin or propranolol had on NKCC activity in the presence of agonists were tested by adding only prazosin or propranolol to the preincubation and incubation media of the unstimulated muscle. The bumetanide-sensitive ⁸⁶Rb rate constant was determined as before. In some experiments, muscles were treated with a combination of 30 µM phenylephrine and 30 µM isoproterenol in the incubation medium.

Potassium content measurement. Potassium content was measured by flame photometry. After 1 min of electrical stimulation (see NKCC activity in skeletal muscle), muscles were removed from the incubation bath and placed in ice-cold Krebs-Ringer solution. The muscles were removed from the stimulating electrodes, blotted dry, and clamp-frozen. Tissue was dried under vacuum until weight did not change. The muscle was digested in 10 vol 1:1:1 water-HNO₃-H₂SO₄ for 24-48 h and diluted with water for flame photometry.

MEK1/2 inhibition. Involvement of MAPK intracellular signaling component MAPK kinase 1/2 (MEK1/2) in the activation of NKCC was tested by preincubating muscles for 15 min in the presence of bumetanide or DMSO vehicle, inhibitors of MEK1/2 (10 µM PD-98059 or 0.5 µM U-0126), an inactive analog to U-0126 (0.5 µM U-0124), or vehicle control (DMSO). Muscles were then either electrically stimulated for 1 min at 10 Hz and then incubated for 10 min in incubation media (Krebs-Ringer solutions with 1 µCi/ml ⁸⁶Rb, either the inhibitors or their vehicle control, and bumetanide or DMSO vehicle), or they were stimulated using 30 µM isoproterenol or 30 µM phenylephrine in the incubation media.

Western blotting of phospho-ERK. Whole muscle was preincubated as described before. Incubation time for stimulation lasted for 5 min and included no ⁸⁶Rb in the medium. After incubation, the muscles were placed on ice in ice-cold lysis buffer (10 mM Tris · HCl, pH 7.4, 5 mM EDTA, 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM sodium fluoride, 100 µM Na₃VO₄, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 2 µg/ml aprotinin, 2 µg/ml leupeptin, 2 µg/ml antipain, and 1 µg/ml pepstatin A), homogenized with a Teflon pestle, and centrifuged at 4°C for 5 min at 5,000 g. Protein concentration of the supernatant was measured by the bicinchoninic acid assay (Pierce, Rockford, IL). Twenty-five micrograms of protein were mixed with SDS denaturing buffer, warmed to 95°C for 5 min, and electrophoresed on a 10% SDS-PAGE gel. The gels were electroblotted with the semidry blotter from Buchler Instruments (Fairfield, NJ). The membrane was incubated overnight at 4°C in the Western blocking buffer (1.5 mM NaH₂PO₄, 8 mM Na₂HPO₄, 0.15 M NaCl, 0.3% Triton X-100, pH 7.4) supplemented with 3% BSA. Immunologic reactions were performed at room temperature for 1.5 h in blocking buffer containing 1% BSA and the specific antibody. Either the phospho-ERK antibody or the ERK-2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) were used at a 1:1,000 dilution. The membrane was subsequently washed four times for 5 min each with blocking buffer and incubated for 45 min at room temperature with horseradish peroxidase-conjugated goat anti-mouse IgG or goat anti-rabbit IgG secondary antibody (Sigma, St. Louis) in blocking buffer supplemented with 1% BSA. After extensive washing with blocking buffer, the immunocomplexes on the membrane were visualized by chemiluminescent exposure of X-ray film (ECL Plus, Amersham). Bands were quantitated using video densitometry. The extent of ERK phosphorylation was determined by comparing the amount of phosphorylated ERK with the total ERK.

Data analysis and statistics. Comparisons within and among treatments for the rate constant data were made by ANOVA; post hoc comparisons, where appropriate, were made using Dunnett's t-test. Differences between treatments were considered significant at  < 0.05. The power of the statistical tests was calculated using the G-Power program (11). Data are reported as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Stimulation of bumetanide-sensitive ⁸⁶Rb transport in skeletal muscle. The basal unstimulated ⁸⁶Rb influx rate constant was 0.0158 ± 0.0013 and 0.0144 ± 0.0013 (g/ml)¹ · min¹ for the soleus and plantaris muscles, respectively (mean ± SE, n = 6 muscle pairs each). The basal unstimulated bumetanide-sensitive ⁸⁶Rb influx rate constant was 0.0013 ± 0.0020 and 0.0013 ± 0.0017 (g/ml)¹ · min¹ for the soleus and plantaris muscles, respectively (n = 6 muscle pairs each). Both the total and the bumetanide-sensitive ⁸⁶Rb-rate constants in the soleus and plantaris muscles were significantly increased by electrically stimulated, unloaded twitch contractions in vitro (P < 0.0001 by ANOVA, Fig. 2), with the bumetanide-sensitive transport accounting for approximately 20-25% of the total uptake. There was also a significant effect of vehicle or bumetanide treatment on the stimulated muscle (P < 0.003 by ANOVA, Fig. 2). Post hoc analysis indicated a significant increase in the bumetanide-sensitive rate constant for the soleus muscle at 0.3-, 1-, 3-,and 10-Hz stimulation and for the plantaris muscle at 10-Hz stimulation (P < 0.05). There was not a significant effect of stimulation frequency ranging between 0.3 and 30 Hz on the total or bumetanide-insensitive rate constant for either muscle. We measured muscle potassium content immediately after 1 min of 10-Hz stimulation to test whether bumetanide had an effect on muscle potassium levels during stimulation. Compared with paired vehicle-treated controls, bumetanide-treated soleus and plantaris muscle lost potassium during stimulation (23 ± 14 and 30 ± 23 meq/kg dry wt, respectively). This loss was significantly different from the change in potassium content in the unstimulated, bumetanide-treated muscles (P < 0.02 and P < 0.01, respectively, for duplicate comparisons among paired samples of 6 muscle pairs, each). In the unstimulated muscles, bumetanide caused an apparent, although not statistically significant, increase in potassium content of the soleus and plantaris muscles by 43 ± 25 and 35 ± 16 meq/kg dry wt, respectively. Stimulation of vehicle-treated muscles caused a decrease in potassium content of 21 and 12 meq/kg dry wt in the soleus and plantaris muscles, respectively, from the unstimulated values of 344 ± 30 and 340 ± 13 meq/kg dry wt, respectively. The unstimulated values are consistent with those reported previously (23).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   A: electrical stimulation had a significant effect on soleus muscle in vitro 86Rb uptake rate constant for both vehicle- and bumetanide-treated muscles (P < 0.0001, n = 6 muscles per point); there was not a significant interaction between frequency and treatment. The power of the test for detecting 0.5 SD unit differences at P < 0.05 was 0.98, 0.90, and 0.90 for the effects of treatment, frequency, and interaction, respectively. B: similarly, electrical stimulation also had a significant effect on plantaris muscle in vitro 86Rb uptake rate constant for both vehicle- and bumetanide-treated muscles (P < 0.003, n = 6 muscles per point). Again, there was not a significant interaction between frequency and treatment; the power of the test was the same as for the soleus muscle. C: compared with quiescent control muscles (0 Hz), post hoc analysis indicated the bumetanide-sensitive 86Rb uptake rate constant increased in electrically stimulated soleus and plantaris muscles (P < 0.05, n = 6 muscle pairs per point, power of 0.93 and 0.99 for soleus and plantaris muscles, respectively).

1

1

1

1

1

1

1

1

1

1

1

1

View larger version (25K): [in this window] [in a new window]   Fig. 3.   A: 1-adrenergic receptor agonist phenylephrine significantly increased the bumetanide-sensitive 86Rb uptake in both the soleus and plantaris muscles (P = 0.002 and P = 0.01, respectively, n = 6 muscle pairs per point). Propranolol (1 µM) was included in the incubation media to prevent cross-reactivity with the -adrenergic receptor. The phenylephrine-stimulated increase was blocked by the 1-adrenergic receptor antagonist prazosin (50 µM, n = 6 muscle pairs per point, power of 0.87 and 0.63 for soleus and plantaris muscles, respectively, for P = 0.05). Neither prazosin nor propranolol alone or in combination affected the bumetanide-sensitive uptake in vehicle-treated muscle. B: -adrenergic receptor agonist isoproterenol significantly increased the bumetanide-sensitive 86Rb uptake in both the soleus and plantaris muscles (P = 0.0003 and P = 0.05, respectively, n = 6 muscle pairs per point). Prazosin (50 µM) was included in the incubation media to prevent cross-reactivity with the 1-adrenergic receptor. The isoproterenol-stimulated increase was blocked by the -adrenergic receptor antagonist propranolol (1 µM, n = 6 muscle pairs per point, power of 0.99 and 0.38 for soleus and plantaris muscles, respectively, for P = 0.05). As in A, neither prazosin nor propranolol alone or in combination affected the bumetanide sensitive uptake in vehicle-treated muscle.

1

1

1

1

View larger version (38K): [in this window] [in a new window]   Fig. 4.   Epinephrine increased the bumetanide-sensitive 86Rb uptake in both the soleus and plantaris muscles in a dose-dependent manner (P = 0.007 and P = 0.02, respectively, n = 6 muscle pairs per point). The normal physiological range for circulating epinephrine is between 1 and 10 nM (stippled area). Compared with the unstimulated muscle, 10 nM epinephrine significantly increased the bumetanide-sensitive 86Rb uptake rate constant in both muscles (P = 0.002 and P = 0.005, respectively, by Dunnett's t-test, power of 0.91 and 0.86 for soleus and plantaris muscle, respectively).

1

1

1

1

1

1

1

1

1

1

1

1

MEK1/2 inhibition and NKCC activity. After treatment of muscle with PD-98059, 50% of the phenylephrine-stimulated, bumetanide-sensitive ⁸⁶Rb uptake was abolished in both slow-twitch and fast-twitch skeletal muscle (Fig. 5). When the more potent MEK1/2 inhibitor U-0126 was used, all of the phenylephrine-stimulated, bumetanide-sensitive ⁸⁶Rb uptake was abolished (Fig. 5). U-0124, the inactive analog of U-0126, did not affect the phenylephrine-stimulated, bumetanide-sensitive ⁸⁶Rb uptake in soleus or plantaris muscles. PD-98059, as well as U-0126, abolished all of the isoproterenol-stimulated, bumetanide-sensitive ⁸⁶Rb uptake in the soleus and plantaris muscles (Fig. 5). The isoproterenol-stimulated, bumetanide-sensitive ⁸⁶Rb uptake was unchanged by the addition of U-0124.

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Inhibition of mitogen-activated protein kinase (MAPK) kinase 1/2 (MEK1/2) activity decreased the phenylephrine (A)- and isoproterenol (B)-stimulated, bumetanide-sensitive 86Rb uptake in the soleus and plantaris muscles. Muscles were treated with the inhibitor for 15 min before stimulation with the appropriate agonist during a 10-min 86Rb uptake period. U-0124, the inactive analog of U-0126, did not affect the adrenergic agonist-stimulated values. (* P < 0.05 relative to the adrenergic agonist-stimulated value, n = 6 muscle pairs per point).

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Inhibition of MEK1/2 activity decreased the electrically stimulated, bumetanide-sensitive 86Rb uptake in the soleus and plantaris muscles. Muscles were treated with the inhibitor for 15 min before stimulation for 1 min at 10 Hz. 86Rb uptake was determined over the subsequent 10 min. In the plantaris muscle, the decreased activity with U-0126 treatment was significant at P = 0.055 with  = 0.49; for others significance is indicated at the P < 0.05 level. (* P < 0.05 relative to the adrenergic agonist-stimulated value, n = 6-11 muscle pairs per point).

ERK1/2 phosphorylation. As has been shown by others (32, 35), adrenergic receptor activation is a relatively weak stimulator of whole tissue or cell immunoreactive ERK1/2 phosphorylation (Fig. 7). Muscles were stimulated with isoproterenol for 5 min. SDS-PAGE protein blots were probed with an anti-doubly phosphorylated ERK antibody, stripped, and reprobed with an anti-ERK2 antibody to normalize the phosphorylation for ERK expression; data are expressed as the fractional change in phosphorylation in treated muscle compared with its contralateral control. Normalized for total ERK2, isoproterenol-treated soleus and plantaris muscle had 1.23 ± 0.21 and 1.21 ± 0.21 times more phospho-ERK, respectively, than vehicle-treated contralateral muscle. Pretreatment with the MEK1/2 inhibitor U-0126 decreased phospho-ERK in isoproterenol-treated soleus and plantaris muscle 0.87 ± 0.06 and 0.78 ± 0.20 times, respectively, compared with isoproterenol- plus U-0124-treated contralateral muscle (U-0124 is the inactive isomer of U-0126). U-0124 had no effect on the level of isoproterenol-induced phosphorylation of ERK (Fig. 7). Although stimulation of the ERK pathway is necessary for stimulation of the bumetanide-sensitive transport (Figs. 5 and 6), at the whole tissue level it is not sufficient. This was demonstrated by an increased ERK1/2 phosphorylation after 5 min of insulin stimulation. Insulin increased phospho-ERK 1.40 ± 0.15 and 1.19 ± 0.17 times in soleus and plantaris muscles, respectively, compared with vehicle-treated contralateral control muscle (Fig. 7). Nonetheless, insulin was ineffective for stimulating bumetanide-sensitive transport in the muscle.

View larger version (42K): [in this window] [in a new window]   Fig. 7.   Isoproterenol and insulin modestly increase extracellular signal-regulated kinase 1 (ERK1) and ERK2 phosphorylation. Doubly phosphorylated ERK1 and ERK2 were quantified by Western blot (typical blot shown in A for soleus muscle), and the data were normalized by total ERK2 expression, as determined by stripping and reprobing the Western blots with an anti-ERK2 antibody (not shown). Relative to its vehicle-treated control, 10 min exposure of the contralateral muscle to isoproterenol or insulin increased the ERK phosphorylation by 20-40% in the soleus and plantaris muscles (B; n = 5-6 muscle pairs per determination). Exposure of both muscles of each pair to isoproterenol, with one being pretreated with the MEK1/2 inhibitors U-0126 or PD-98059, resulted in the pretreated muscle showing less ERK phosphorylation (C; n = 5-6 muscle pairs per determination). The inactive U-0124 had no effect on ERK phosphorylation levels in response to isoproterenol. [* P < 0.05 compared with the control level of phosphorylation for either vehicle treatment (B) or isoproterenol treatment without MEK1/2 inhibitor (C)].

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ion transport by muscle fibers must be regulated to provide a compensatory response to the dynamic movement of ions during contraction (30, 42). Previously, we demonstrated that NKCC mRNA, protein, and activity are present in adult rat skeletal muscle (46). However, whether the NKCC activity could be modulated remained a question. From the data presented here, we conclude that NKCC activity can be modulated by some, but not all, of the stimuli that are also known to stimulate Na⁺-K⁺-ATPase activity (15) in a manner that is dependent on MAPK pathways. Three major findings provide the support for this conclusion. First, the skeletal muscle bumetanide-sensitive ⁸⁶Rb transport activity can be stimulated to levels accounting for as much as 30% of the potassium congener uptake after electrical stimulation of the muscle or exposure of the muscle to epinephrine and its mimetics. Second, unlike the contractile activity or catecholamines, insulin did not stimulate NKCC activity. Finally, MAPK inhibitors blocked stimulation of NKCC activity. We will discuss the significance of these findings in the following paragraphs.

The rate constants for the basal rate of transport reported here allow comparison with previously reported data. Multiplying the rate constant by the potassium concentration in the bathing solution gives an estimate of the maximum rate of potassium transport under basal conditions. These values for total uptake in the ex vivo muscle preparation here [rate constants ranging from 0.015 to 0.030 (g/ml)¹ · min¹] are comparable to values reported for perfused hindlimb with low-flow perfusion (28) and cardiac myocytes in culture (44). However, the values reported here are less than those measured under similar conditions (9, 10, 36), the preparations differing primarily in the source of the rats and the method of muscle excision (pentobarbital surgery in this study vs. decapitation). Basal NKCC activity in skeletal muscle is smaller than the activity observed when the muscle is stimulated. In fact, bumetanide treatment of resting skeletal muscle sometimes causes increased ⁸⁶Rb influx (as demonstrated by a negative value for the bumetanide-sensitive ⁸⁶Rb rate constant), although this is not statistically significant. This trend may be due to an increase in the activity of other inwardly directed potassium channels and transport mechanisms. Analogously, it has been shown by us and by others that the blockade of the Na⁺-K⁺-ATPase can stimulate NKCC or other inwardly directed transport activity (8, 38, 46). Nevertheless, the conclusion from these data is that quiescent muscle has a very low NKCC-mediated inward flux of isotope compared with total inward flux.

Electrical stimulation of the muscles was a powerful stimulator of the bumetanide-sensitive ⁸⁶Rb uptake. As few as 18 electrically induced contractions (0.3 Hz for 1 min) were sufficient stimulus to increase the bumetanide-sensitive ⁸⁶Rb uptake significantly above basal levels (Fig. 2). The relative magnitude of the increase in total transport rate is comparable to that reported previously by Everts et al. (13). In addition, we found that the stimulation of NKCC activity was not frequency dependent. It has previously been suggested that deoxygenation can stimulate NKCC activity in red blood cells (34); protons, potassium, and epinephrine are also well-documented stimuli for potassium transport in red blood cells during exercise, apparently through stimulation of both Na⁺-K⁺- ATPase and NKCC activity (29). Although oxygen in our bathing solution was maintained throughout the preparation, contracting muscle consumes oxygen, creating a possible oxygen gradient within the tissue. However, if the stimulation of NKCC activity due to electrical stimulation occurs by hypoxia, the mechanism is either independent of stimulation frequency or the stimulus threshold is very low.

Adrenergic receptor stimulation was a powerful stimulator of the total and bumetanide-sensitive ⁸⁶Rb uptake in skeletal muscle. Comparable to values reported by others, adrenergic receptor stimulation increased total uptake by ~40% (7). Stimulation with either an ₁- or -adrenergic receptor agonist significantly increased NKCC activity in skeletal muscle (Fig. 3). Although -adrenergic receptor-mediated pathways are often thought to oppose -adrenergic receptor-mediated pathways, in some cases their activities do complement each other and activate NKCC activity in some cell types (1, 18, 47). Defining the specific receptor-mediated pathways responsible for NKCC activation involved pharmacological treatment of the skeletal muscle with the appropriate agonists and antagonists. The results of these experiments indicate that NKCC activity was stimulated specifically by both ₁- and -adrenergic receptor-mediated pathways. In the case of the -adrenergic receptor-mediated pathway, the bumetanide-sensitive transport accounted for nearly one-third of the total ⁸⁶Rb uptake. Furthermore, the maximum NKCC activity stimulated by epinephrine (Fig. 4) was not significantly different from that caused by stimulation with phenylephrine or isoproterenol (P < 0.05). This may indicate that a maximal level of stimulation of NKCC activity was obtained with either the phenylephrine or isoproterenol and that the ₁- and -adrenergic receptor-mediated pathways converge to activate NKCC. Such a convergence would be consistent with the putative Ca²⁺ and cAMP-mediated stimulation of muscle potassium transport by methylxanthines and papaverine (20, 27, 28). Our data may also explain part of the reason why potassium uptake by muscle is impaired in patients taking -adrenergic receptor blockers (17). Furthermore, considering the relatively low ₁-adrenergic receptor density in skeletal muscle (33), the demonstration of a potential physiological role for this receptor is a novel result of these experiments.

Both ₁- and -adrenergic receptor activation were stimulatory for NKCC activity in the muscle. Therefore, we tested the possibility of an interaction between the two adrenergic receptor-mediated pathways. Treatment of the muscle simultaneously with agonists for both receptors resulted in a significant negative interaction for the stimulation of NKCC activity in the predominantly slow-twitch soleus muscle. It is reasonable to conclude that the ₁-adrenergic pathway inhibits the -adrenergic pathway. This also may explain why soleus muscle NKCC activity declines when treated with high doses of epinephrine (Fig. 4).

Perhaps the most significant finding presented here is the lack of NKCC stimulation by insulin. As with catecholamines, insulin is a well-known stimulator of potassium uptake by skeletal muscle (15, 21). In both adipocytes and fibroblasts, insulin stimulates Na⁺-K⁺ ATPase activity and NKCC (31). Because skeletal muscle is one of the primary targets of insulin, we tested whether insulin stimulated NKCC activity in muscle. Although physiological resting levels of insulin in the rat are 10 µU/ml, 100 µU/ml was used to ensure activation of its receptor. We confirmed that insulin indeed stimulated ⁸⁶Rb uptake in both soleus and plantaris muscles. However, the increase in uptake was not due to stimulation of NKCC activity because insulin did not significantly stimulate the bumetanide-sensitive ⁸⁶Rb uptake. We conclude that the insulin-stimulated ⁸⁶Rb uptake is primarily due to the well-documented stimulation of Na⁺-K⁺ ATPase activity (6, 15). Taken with results from catecholamine stimulation and electrical stimulation, these data show that NKCC activity is regulated and stimulated in response to some stimuli, but not by all stimuli known to stimulate potassium uptake by skeletal muscle.

Contractile activity, phenylephrine, and isoproterenol have all been shown to rapidly stimulate MAPK activity (39, 47). Therefore, we were interested in the potential involvement of MAPK in the stimulation of NKCC activity in skeletal muscle. Specifically, we tested the involvement of the ERK arm of the MAPK pathways. ERK1/2, key substrates in this signal cascade, are activated by phosphorylation by MEK1/2. In our experiments, known inhibitors of MEK1/2 abolished the isoproterenol-stimulated NKCC activity (Fig. 5) and the electrically stimulated NKCC activity (Fig. 6). Phenylephrine-stimulated NKCC activity showed the same trend, although the effect was not as striking (Fig. 5). From these data, we conclude that the ERK1/2 pathway is necessary for stimulation of NKCC activity in muscle. However, insulin is also known to stimulate the ERK1/2 pathway (Fig. 7 and Ref. 24), so why does insulin not stimulate NKCC activity in skeletal muscle? One possibility is that ERK1/2 activation alone may not be sufficient to stimulate skeletal muscle NKCC activity. Another possibility is that insulin stimulates a unique subset or compartment of ERK1/2 kinases. It is also possible that insulin activates other intracellular signaling mechanisms (5, 16) that may inhibit ERK1/2 activation of skeletal muscle NKCC activity. Unfortunately, no MAPK-specific activators are currently available that will specifically activate the ERK1/2 pathway without stimulating other mechanisms (e.g., tyrosine kinase receptor agonists). Nevertheless, the demonstration by these data of a difference within skeletal muscle in the consequences of stimulating the ERK1/2 pathway through distinctly different mechanisms is a significant finding. Furthermore, NKCC activity in skeletal muscle may prove to be a valuable model to study the interactions of various intracellular signaling pathways.

In conclusion, we have shown that electrical stimulation, ₁-adrenergic stimulation, and -adrenergic stimulation of muscle increased muscle NKCC activity. The intracellular mechanism responsible for the increased NKCC activity, regardless of stimulus, involved the activation of the ERK1/2 arm of the MAPK pathways. However, complexity in the intracellular signal pathways was evident in that stimulation of the ERK1/2 pathway with insulin did not stimulate NKCC activity. Functionally, the rapid stimulation of NKCC activity in skeletal muscle may serve as an additional potassium sequestering mechanism by muscle cells. This could provide the body with another rapid, efficient mechanism to buffer potassium and may be of particular importance for insulin-independent control of plasma potassium.

Perspectives