Exercise-induced changes in plasma composition increase erythrocyte Na⁺, K⁺-ATPase, but not Na⁺–K⁺–2Cl⁻ cotransporter, activity to stimulate net and unidirectional K⁺ transport in humans

Abstract

We tested the hypothesis that exercise-induced changes in plasma composition result in peak stimulation of erythrocyte unidirectional K⁺ (J_K,in) and net K⁺ (J_K,net) transport within the first 120 s. In experimental series 1 (7 men; 2 women), plasma [K⁺] was continuously measured in vitro (37 °C) after the addition of red blood cells (RBCs) obtained from rested subjects (resting RBCs) into an exercise-simulated plasma (ESP; increased plasma osmolality, [K⁺], [H⁺], [lactate] and [adrenaline] (epinephrine)), and J_K,net calculated. In experimental series 2 (7 men; 4 women), resting RBCs were incubated in true exercise plasma (TEP) obtained after two 30 s bouts of high intensity leg cycling exercise to determine J_K,net and J_K,in (via RBC ⁸⁶Rb accumulation). J_K,net of resting RBCs increased from 0.9 ± 28.7 in resting plasma to 285 ± 164 mmol (l RBCs)⁻¹ h⁻¹ in ESP and to 178 ± 60 mmol (l RBCs)⁻¹ h⁻¹ after 10 s in TEP. Both J_K,net and J_K,in peaked within 10 s of incubation and decreased rapidly during the initial 120 s. The use of inhibitors for the Na⁺,K⁺-ATPase (ouabain) and the Na⁺-K⁺-2Cl⁻ cotransporter (NKCC; bumetanide) indicated that rapid increases in J_K,in and J_K,net upon incubation of resting RBCs in TEP were due primarily to increased Na⁺,K⁺-ATPase activity; the NKCC appeared to be involved only when the Na⁺,K⁺-ATPase was blocked. It is concluded that RBCs rapidly increase J_K,in and J_K,net in response to exercise-induced changes in plasma composition.

During high intensity exercise, the net loss of K⁺ from contracting skeletal muscle generates pronounced alterations in plasma and whole body K⁺ balance (Lindinger et al. 1995). Contracting skeletal muscle loses K⁺ in proportion to contraction intensity, and net loss may exceed 400 μmol s⁻¹ (kg contracting muscle)⁻¹ (Lindinger, 1995). Repeated bouts of short duration, high intensity exercise produces elevations in plasma [K⁺] of as a high as 9 mmol l⁻¹ in humans (Medbo & Sejersted, 1990) and 12 mmol l⁻¹ in horses (Harris & Snow, 1992). In addition to the hyperkalaemia, high intensity exercise also produces pronounced increases in plasma osmolarity, [H⁺], [adrenaline] and [lactate] that affect extrarenal K⁺ transport processes. There occurs a rapid ‘regulatory’ K⁺ uptake by extrarenal tissues including skeletal muscle (Lindinger et al. 1992, 1995; McDonough et al. 2002), liver (Bia & DeFronzo, 1981) and red blood cells (McKelvie et al. 1991, 1992; Lindinger et al. 1999).

There has been controversy regarding the capability of red blood cells (RBCs) to regulate plasma [K⁺] during exercise, and this appears to have been related to the exercise intensity. We (McKelvie et al. 1991; Lindinger et al. 1992, 1995) have suggested that RBCs play a role in the regulation of plasma [K⁺] during supramaximal intensity exercise because increases in both RBC [K⁺] and content are greater than what could be accounted for by decreases in RBC volume. In contrast, studies using lower exercise intensities concluded that RBCs did not play such a role (Vollestad et al. 1994; Juel et al. 1999). Using in vitro methods, Maassen et al. (1998) were also unable to demonstrate an increase in K⁺ uptake by RBCs with exercise; however, it appears that the washing techniques used to prepare the RBCs, and the use of an artificial incubation medium, may have been responsible for the negative result. When we avoided such washing procedures and incubated RBCs obtained from subjects at rest in true exercise plasma (TEP) obtained from the same subjects immediately after high intensity exercise, there was a pronounced increase in RBC K⁺ uptake (Lindinger et al. 1999). We concluded that RBCs are capable of regulating plasma [K⁺] under conditions associated with very high intensity exercise, and that this response was not evident at lower exercise intensities.

A limitation of the previous study (Lindinger et al. 1999) was that the first time point of data collection did not occur until 2 min of incubation of RBCs in plasma. An absence of significant effect of TEP on unidirectional K⁺ influx (J_K,in), despite the significant net uptake of K⁺ (J_K,net), suggested that peak transport rates occurred during the initial 2 min of incubation. Furthermore, 2 min is a long time in terms of circulatory transit of blood through contracting skeletal muscle and in the whole body, and it is highly likely in vivo that responses are approaching completion in that time. We tested the hypotheses that J_K,in and J_K,net are highest immediately upon introduction of RBCs to TEP and that J_K,in and J_K,net are primarily affected by increased Na⁺,K⁺-ATPase and NKCC activities. In order to test these hypotheses, erythrocyte J_K,in and J_K,net were measured during the first 2 min of incubation using both an exercise-simulated plasma (ESP) and TEP in the presence and absence of K⁺ transport inhibitors.

METHODS

Subjects

Eighteen male and female university students between the ages of 21 and 27 years volunteered for the study. Potential risks and procedures were explained to the subjects prior to them signing a written, informed consent form for participation. The subjects were determined, by questionnaire, to be healthy, free from known cardiovascular problems, and moderately active. Subjects agreed to abstain from ingesting alcohol and caffeine for 2 days prior to their participation in the study, and reported compliance upon arrival to the laboratory. The study and its procedures were approved by University of Guelph's Human Subjects Committee and performed according to the Declaration of Helsinki.

Chemicals

NaCl and KCl were from Fisher Scientific (Fair Lawn, NJ, USA). Bumetanide, ouabain, adrenaline, ascorbic acid, l-lactic acid and DMSO were obtained from Sigma Chemical Co. (St Louis, MO, USA). A 20 mm stock solution of bumetanide was made by dissolving the chemical in pure DMSO, from which 5 μl were added per millilitre of plasma to obtain a plasma concentration of 0.1 mm. A 10 mm stock solution of ouabain was made in 0.9 % NaCl, from which 10 μl were added per millilitre of plasma to obtain a plasma concentration of 0.1 mm. A stock solution of adrenaline of 0.01 mm was made daily in 0.9 % NaCl containing 2 mg ml⁻¹ of ascorbic acid to prevent oxidation of adrenaline; from this solution, 1 μl was added to 1 ml of plasma to obtain a plasma [adrenaline] of 10 nm.

Experimental protocol

Two series of experiments were performed, with series 1 using exercise-simulated plasma (ESP) and series 2 using true exercise plasma (TEP).

Series 1: exercise-simulated plasma (ESP)

Experiments using ESP were identical to our previous study (Lindinger et al. 1999) except the responses were measured during the initial 120 s of incubation. ESP had elevated osmolality, [adrenaline], [K⁺] and [H⁺] mimicking levels seen with high intensity exercise. It has been reported that high intensity exercise increases plasma osmolality from 285 to 320 mosm l⁻¹ (Buono & Faucher, 1985), [K⁺] from 4 to 9 mmol l⁻¹ (Medbo & Sejersted, 1990), [adrenaline] from 1 to 20 nmol l⁻¹ (Brooks et al. 1990) and [H⁺] from 40 to 100 nmol l⁻¹ (Lindinger et al. 1992).

Subjects (5 male; 2 female, aged between 22 and 24 years) reported to the laboratory in the morning, 2–3 h after a light breakfast consisting of juice with cereal or toast. An antecubital vein blood sample (50 ml) was transferred into five 10 ml lithium-heparin vacutainer tubes and gently mixed, and cells were separated from plasma by centrifugation (10 min at 15 000 g). After carefully removing only plasma, the remaining plasma with entire buffy coat layer and underlying red cells were removed leaving only RBCs. This method of separation also effectively removes leukocytes that would otherwise contribute to measured K⁺ fluxes (Maassen et al. 1998).

A 1 ml aliquot of plasma was pipetted into a 1.6 ml plastic vial to be used as resting plasma (negative control). To 4 ml of the remaining plasma, the following additions were made to create ESP: (a) 4 μl of 0.01 mm adrenaline to obtain 10 nm; (b) 160 μl of 1 m NaCl, increasing plasma osmolality by 80 mosmol kg⁻¹; (c) 120 μl of 1 m lactic acid increased plasma [H⁺] and [lactate] by 50 nm and 30 mm, respectively; and (d) 20 μl of 1 m KCl to increased the K⁺ concentration was by 5 mm. After these additions, the plasma was swirled to ensure thorough mixing while avoiding foaming.

J_K,net experiments

Experiments were performed within a Plexiglas chamber maintained at 37 °C. The K⁺-selective electrode (valinomycin membrane, filled with 0.1 m KCl; Kwik-Tip, World Precision Instruments, Sarasota, FL, USA) with a separate reference electrode (Flex-Ref, World Precision Instruments), were calibrated in 4 and 20 mm KCl standards (Nova Biomedical, Waltham, MA, USA). The concentrations of the standards were verified in duplicate using a blood gas/electrolyteanalyser (Statprofile 9+ analyzer, Nova Biomedical) prior to calibration of the electrodes. Electrodes were calibrated ≈1 h prior to performing an experimental series, and the calibration was verified immediately after. Data were accepted when statistical analysis of pre-test and post-test calibrations showed that there was not a significant drift in electrode sensitivity.

RBCs and aliquots of either ESP or resting plasma (matched by subject) were warmed to 37 °C and the K⁺ electrode placed into the plasma for the collection of 10 min of baseline data. Then, 0.72 ± 0.02 ml of RBCs was pipetted into the plasma (yielding a hematocrit of ≈40 %) and the mixture rapidly swirled so as to disperse the RBCs. Within 6 s of introducing the RBCs, plasma [K⁺] was recorded at frequent intervals for 5 min; intermittent manual swirling minimized settling of RBCs.

Series 2: true exercise plasma (TEP)

Seven male and four female subjects (Table 1) between the ages of 21 and 27 participated in these experiments. The number of complete data sets used for each experiment is given in Table 1 for J_K,net data and Table 2 for J_K,in data.

Table 1.

Subject characteristics by experiment for J_{K, net} data obtained in Series 2

Experiment	1	2	3	4	5	6
Number of subjects	5	9	9	8	9	6
Age (years)	23.2 ± 0.4	23.3 ± 0.6	23.6 ± 0.5	23.6 ± 0.6	23.4 ± 0.6	23.8 ± 0.7
Height (cm)	175.7 ± 1.90	174.5 ± 2.43	174.8 ± 2.41	176.8 ± 1.58	174.7 ± 2.42	176.9 ± 1.92
Weight (kg)	75.2 ± 4.6	71.5 ± 3.6	73.3 ± 3.7	75.9 ± 3.0	72.7 ± 3.8	76.2 ± 3.9
Exercise 1 (W)	770.0 ± 33.9	757.8 ± 36.6	785.6 ± 31.8	806.3 ± 27.4	774.4 ± 37.1	791.7 ± 35.2
Exercise 2 (W)	693.0 ± 30.5	682.0 ± 32.9	707.0 ± 28.6	725.6 ± 24.7	697.0 ± 33.4	712.5 ± 31.6

Values are means ±s.e.m.

Table 2.

Subject characteristics by experiment for J_K,in data for Series 2

Experiment	1	2	3	4	5	6
Number of subjects	7	7	5	7	5	6
Male	6	7	4	5	4	5
Female	1	0	1	2	1	1
Age (year)	23.7 ± 0.7	23.8 ± 0.6	23.4 ± 1.0	23.1 ± 0.8	23.6 ± 1.0	23.7 ± 0.8
Weight (kg)	75.9 ± 3.6	76.5 ± 3.4	75.7 ± 5.6	74.8 ± 4.2	77.1 ± 4.1	72.2 ± 4.4
Exercise 1 (W)	785.7 ± 40.4	821.4 ± 26.4	820.0 ± 43.6	792.9 ± 44.2	810.0 ± 53.4	783.3 ± 40.1
Exercise 2 (W)	707.1 ± 36.4	739.3 ± 23.8	738.0 ± 39.2	713.6 ± 39.8	729.0 ± 48.04	705.0 ± 36.1

Values are means ±s.e.m.

Exercise protocol

Subjects were pre-tested on a cycle ergometer (Lode, Quinton Excalibur Sport) several days prior to the test day in order to determine the resistance required to achieve the maximal exercise intensity at which a pedalling frequency of 100–120 r.p.m. could be maintained for each 30 s exercise bout. In order to achieve maximal exercise-related blood changes, subjects performed two modified Wingate leg cycling exercise bouts while being verbally encouraged. Similarly high intensity leg cycling exercise showed that the largest increases in plasma constituents occurred within femoral venous plasma, compared to arterial plasma, at the end of the second exercise bout (Lindinger et al. 1992).

Upon reporting to the laboratory on the experiment day, the subject lay supine on an examination table. Using the Seldinger technique with aseptic procedures, a 16 gauge, 15 cm single lumen central venous catheter (Cook Inc., Bloomington, IN, USA) was inserted into the femoral vein under local anaesthetic (Xylocaine). A 20 inch extension with a 4-way stopcock was fitted to the catheter and secured with an OpSite tape bandage along the subject's iliac crest. The stopcock allowed for the removal of saline remaining in the catheter, ensuring that the blood sample was not contaminated. Subjects rested in a seated position for 15 min post catheterization in order to minimize the changes in plasma composition due to the catheterization procedure (Lindinger et al. 1992) and postural fluid shift (Mack et al. 1998). A resting blood sample was obtained via a 60 ml syringe, and immediately divided into six 10 ml vacutainer tubes (Becton Dickinson, Franklin Lakes, NJ, USA) containing lithium-heparin as anticoagulant.

The first exercise bout started at the predetermined peak power which was maintained for the first 3 s. During the next 27 s, power output decreased linearly to ≈50 % of peak power. For the next 4 min, the subjects pedalled slowly at, or below, 25 W. With 10 s of the 4 min rest period remaining, subjects increased their cycling frequency to between 100 and 120 r.p.m. to initiate the second exercise bout. Peak power output was ≈10 % less then the first exercise bout, with the same rate of power decline. At 15 s into the second exercise bout and for 15–30 s after cessation of exercise, a blood sample was continuously drawn until 50 ml was obtained through the catheter. This blood sample was immediately divided into five heparinized vacutainer tubes. The femoral venous catheter was removed and subjects recuperated, under supervision, for at least 30 min prior to leaving the testing facility.

Blood analyses

Microhaematocrit was immediately determined in duplicate by centrifugation (15 000 g for 15 min) of blood-filled capillary tubes. A 1.5 ml whole blood sample (rest and exercise) was immediately cooled in ice water and analysed in duplicate for blood gases and plasma [Na⁺], [K⁺], [Cl⁻], [glucose], [lactate], pH and P_CO2 at 37 °C within 1 h of sampling (Statprofile 9+ analyser). The remaining whole blood was immediately centrifuged (15 000 g for 15 min). Plasma was removed via transfer pipettes and dispersed into capped vials. The buffy coat and top red cell layers were completely removed and discarded. Plasma [protein] was measured with a clinical refractometer (Atago, SPR-T2, Japan).

The separated RBCs and plasma were allowed to cool to room temperature and separated into aliquots for the experiment. During the time between sample collection and performing the experiments, samples were kept in sealed microcentrifuge tubes to minimize gas exchange with air.

Preparation of TEP

Radioactive ⁸⁶Rb, a K⁺ analogue, was added to 20 ml of TEP to enable measurement of the amount and rate of K⁺ uptake by RBCs. Also, 20 μl of 1 g l Evans Blue dye (Fisher Scientific) was added to the 20 ml of TEP in order to quantify net water flux between RBCs and plasma. For two subjects, ¹⁴C-labelled inulin was also added to the plasma in order to determine the amount of plasma that remained trapped within the packed RBCs obtained at the end of the experiment; inulin is membrane impermeable and not taken up by RBCs. The TEP, with additions, was mixed for several minutes.

The TEP was divided into five aliquots. The following inhibitors were added to achieve desired concentrations. Ouabain (0.1 mm) effectively inhibits Na⁺,K⁺-ATPase activity (Sachs, 1971). Bumetanide was added to TEP to obtain a plasma total [bumetanide] of 0.1 mm. Given that albumin binds 97 % of bumetanide (Walker et al. 1989) the free (unbound) plasma [bumetanide] was 0.003 mm. This concentration of free bumetanide should achieve ≈95 % inhibition of NKCC activity (Duhm, 1987). The positive control (resting RBCs in TEP) and negative controls (exercise RBCs in TEP and resting RBCs in resting plasma) received no drug additions.

The following experiments were performed in a randomized order for each subject, and all experiments were completed within 5 h of blood sampling.

1. Resting RBCs with TEP. We hypothesized that J_K,in and J_K,net of resting RBCs incubated in plasma obtained after high intensity exercise (TEP) would be increased compared to the control series (resting RBCs incubated in resting plasma).

2. Exercise RBCs with TEP. We hypothesized that the transport of K⁺ by ‘exercised’ RBCs would not be elevated.

3. Resting RBCs with TEP + 0.1 mmol l⁻¹ ouabain. We hypothesized that increased Na⁺,K⁺-ATPase activity would contribute to increased J_K,in and J_K,net.

4. Resting RBCs with TEP + 0.1 mmol l⁻¹ bumetanide. We hypothesized that increased Na⁺-K⁺-2Cl⁻ cotransporter activity would contribute to increased J_K,in and J_K,net.

5. Resting RBCs with TEP + 0.1 mmol l⁻¹ ouabain + 0.1 mmol l⁻¹ bumetanide. We hypothesized that inhibition of both Na⁺-K⁺-2Cl⁻ cotransporter and Na⁺,K⁺-ATPase activities would be additive to inhibition of each alone, and that the residual inward K⁺ flux would be increased in RBCs that are incubated in TEP.

6. Resting RBCs with resting plasma. We hypothesized that there would not be a stimulation of K⁺ transport when plasma constituents are at resting concentrations.

In order to achieve inhibition of RBC J_K,in and J_K,net from the onset of incubation with plasma, it was necessary to preincubate RBCs with the inhibitor of choice. This procedure avoided the time-dependent delays for full inhibition when inhibitors were present only in plasma (data not shown). For experiments 3, 4 and 5 (above), RBCs were incubated in resting plasma containing twice the final plasma concentration of the inhibitors (0.2 mm ouabain; 0.2 mm bumetanide) and continuously rocked for 30 min at room temperature. This allowed for inhibitor binding to NKCC and Na⁺,K⁺-ATPase sites in the RBC plasma membrane. Upon completion of incubation, plasma and RBCs were separated by centrifugation (15 min at 15 000 g) and this plasma discarded.

J_K,net was measured using a K⁺-selective electrode as described above. Between each experiment, the K⁺ electrode was rinsed with double distilled water and the sides gently wiped to remove water before being placed in the plasma for the next series. Upon completion, the blood was centrifuged (5 min at 15 000 g) to separate the plasma and RBCs.

J_K,in

Aliquots of plasma and RBCs were always matched by subject. Plasma (2 ml, within a 10 ml glass vial) and RBCs (2 ml; haematocrit (Hct), 85 ± 1 %) within a 3 ml syringe were warmed for 15 min in a Plexiglas chamber heated to 37 °C. The experiment was initiated upon expulsion of RBCs from the syringe into the vial containing the plasma. The blood was vortex-mixed for 7 s. From this, 180 μl samples were removed by pipettor at 10, 20, 30, 45, 60, 75, 90, 105, 120, 135, 150, 165, 180, 210 and 240 s. Analysis of plasma [K⁺] was also performed using the Statprofile 9+ analyser prior to addition of RBCs and at 30 and 240 s of incubation.

Each 180 μl blood sample was pipetted into a 1.5 ml conical, polyethylene centrifuge tube containing 300 μl of cold (≈0 °C) dibutyl phthalate (DBP, Fisher Chemical) and immediately centrifuged at 15 000 g for 2 min. The DBP forms a layer between the packed RBCs and the plasma, facilitating their complete separation (Kirk et al. 1992). Some plasma was removed for subsequent measurement of Evans Blue and haemoglobin concentrations and for ¹⁴C-labelled inulin and ⁸⁶Rb activities. The remaining plasma and the top of the DBP layer were removed by aspiration and discarded.

The following procedure was performed to completely remove extracellular radioactivity adhering to the plastic vial and the surface of the DBP layer. The conical tube, now containing only packed RBCs with a protective top layer of DBP, was gently submerged into water until filled, and then the water and top of the DBP layer were removed by aspiration. This procedure was performed three times, with the final aspiration removing most of the remaining DBP but none of the packed RBCs - occasionally there was loss of some RBCs, and this contributed to the different n values used for different experiments. Periodic analysis of the DBP layer after centrifugation of the blood sample showed no ⁸⁶Rb radioactivity in this layer when using this technique. The packed cells were solubilized by the addition of 0.5 ml of 0.5 % Triton X-100 (Fisher Scientific) and vortex mixing. The lysed cells were deproteinized by the addition of 0.8 ml of 5 % trichloroacetic acid (Sigma Chemical Co.) and vortex mixing. The mixed samples were left to sit for 5 min, then centrifuged for 5 min at 15 000 g. After centrifugation, 1.0 ml of the supernatant was pipetted into a plastic 7 ml scintillation vial for detection of ⁸⁶Rb activity (Cerenkov method using a Beckman LS5000 TA liquid scintillation counter; Beckman Instrument Co., Mississauga, ON, Canada). From the two experiments using ¹⁴C-labelled inulin, an additional 50 μl RBC supernatant sample was added to 5.0 ml of Beckman Ready Protein and ¹⁴C activity measured using the scintillation counter. A 50 μl plasma sample removed prior to addition of RBCs served as the plasma reference value. The ¹⁴C data from these two experiments were very similar and used to correct for the amount of trapped plasma within the RBC pellets of all experiments. The RBC ⁸⁶Rb activity (counts min⁻¹) attributed to the trapped plasma was subtracted from the total RBC ⁸⁶Rb activity to yield RBC ⁸⁶Rb uptake.

Plasma Evans Blue absorbance (50 μl plasma diluted into 1 ml double distilled water) was measured at 610 nm and 740 nm (Foldager & Blomqvist, 1991) using the spectrophotometer. These data were used to determine changes in plasma and RBC volumes. Plasma haemoglobin absorbance was measured at a wavelength of 420 nm (Beckman DU-70, Fullerton, CA) to quantify cell lysis.

Calculations

The quality of the data obtained from each experiment was assessed statistically as follows. For each experiment, RBC ⁸⁶Rb counts were plotted over time and the data described by a best-fit, 2nd order polynomial. Data points more than two standard deviations from the mean were discarded (attributed to pipetting and other errors) and a new line of best-fit generated. When lines of best-fit were not well described by a 2nd order polynomial (r² < 0.5), the data were excluded from further analysis. This contributed to the different numbers of experiments reported for each series (Tables 1 and 2).

Calculation of J_K,net took into account the increase in plasma water content, estimated from the increase in plasma Evans Blue concentration (see below). For example, a 10 % decrease in plasma Evans Blue concentration represented a 10 % increase in plasma water content, necessitating correction of plasma [K⁺] upwards by 10 %.

The dilution of the plasma caused by the addition of packed RBCs to plasma was taken into consideration when calculating J_K,in. The amount of plasma added with the packed RBCs was based on a measured Hct of 85 ± 1 %, therefore addition of packed RBCs increased the plasma volume from 2.0 to 2.25 ml.

The RBC ⁸⁶Rb counts were used to calculate K⁺ uptake and J_K,in as follows:

(1)

where a is plasma specific activity (counts min⁻¹μmol⁻¹).

(2)

where X₁ and X₂ represent the radioactivity (counts min⁻¹) of m litres of cells and t is the incubation period (h). Calculations were based on the initial (t = 0) volume of cells.

Statistics

Values are reported as mean ±s.e.m.J_K,net, J_K,in and K⁺ uptake data were analysed using a series of two-way (with respect to time and treatment) repeated measures ANOVA to compare each series to the control. Statistical analyses of J_K,in and J_K,net were limited to data obtained during the first 2 min of incubation because of the relative lack of change during the subsequent 3 min. When a significant F ratio was obtained, the Bonferroni test was used to compare means. Comparisons of plasma [K⁺] were performed using a paired t test. Statistical significance was accepted at P < 0.05.

RESULTS

Incubation of RBCs in ESP

The addition of resting RBCs to ESP resulted in a rapid decline in plasma [K⁺] that was essentially complete within 2 min (Fig. 1A). In contrast, the addition of resting RBCs to resting plasma had no effect on plasma [K⁺]. The J_K,net at 10 s of incubation was 285 ± 164 mmol (l RBC)⁻¹ h⁻¹ for resting RBCs incubated in ESP and 0.8 ± 20 mmol (l RBC)⁻¹ h⁻¹ for resting RBCs incubated in resting plasma. The J_K,net in ESP decreased rapidly (Fig. 1B).

Figure 1. Plasma [K⁺] and RBC J_K,net before and after addition of RBCs.

A, plasma [K⁺] measured continuously during a 10 min period without red blood cells (RBCs) and for 5 min after addition of RBCs. B, the apparent net flux of K⁺ (J_K,net) between RBCs and plasma. •, resting RBCs incubated in exercise simulated plasma (plasma from resting subjects to which adrenaline, NaCl, lactic acid and KCl were added); ○, resting RBCs incubated in resting plasma without any additions. * Significantly different from plasma alone.

Incubation of RBCs in TEP

In this series of experiments, RBCs were incubated in the presence of K⁺ transport inhibitor(s) prior to the addition of RBCs to TEP containing the matching inhibitor(s). Therefore, the K⁺ transport pathway of interest should have been inhibited prior to the addition of the RBCs. The high intensity of the exercise appears to over-ride sex or other differences amongst subjects, and there were no notable differences in the female responses compared to their male counterparts.

In each of these series of experiments, a decreased plasma Evans Blue concentration indicated a net loss of water from RBCs to plasma during the period of incubation. The Evans Blue responses were highly variable amongst RBCs from each subject, with no significant difference amongst experiments (Table 3). While values up to 60 s were not significantly different from 0, the calculation of RBC J_K,net and J_K,in took the dilution of plasma [K⁺] and ⁸⁶Rb counts into consideration.

Table 3.

The percentage change in RBC volume

Incubation duration (s)	30	60	120	240
Experiment
1 (n = 11)	−6.0 ± 7.9	−8.7 ± 9.4	−18.1 ± 7.7	−23.1 ± 8.5
2 (n = 10)	−4.7 ± 13.9	−0.6 ± 5.8	−15.2 ± 7.5	−20.2 ± 8.7
3 (n = 11)	−3.6 ± 6.1	−6.8 ± 6.0	−15.4 ± 4.0	−19.6 ± 10.1
4 (n = 10)	−8.4 ± 6.5	−7.4 ± 5.1	−12.9 ± 6.1	−14.4 ± 11.8
5 (n = 10)	−17.7 ± 10.6	−19.6 ± 9.4	−15.7 ± 8.0	−30.5 ± 7.1
6 (n = 11)	−7.9 ± 8.3	−4.2 ± 10.0	−17.0 ± 9.6	−17.2 ± 13.4

Values are means ±s.e.m.n, number of subjects.

The addition of resting RBCs to true resting plasma, and of exercise RBCs to TEP, caused no change in plasma [K⁺] (Fig. 2A). In contrast, addition of resting RBCs to TEP without ouabain (including TEP with bumetanide) resulted in rapid, significant decreases in plasma [K⁺]. In contrast, in the presence of only ouabain, there occurred a rapid increase in plasma [K⁺] (Fig. 2B), while combined ouabain and bumetanide resulted in no net change in plasma [K⁺].

Figure 2. Plasma [K⁺] before and after addition of RBCs.

Plasma [K⁺] measured continuously during a 10 min period without RBCs and for 5 min after addition of RBCs. •, resting RBCs incubated in TEP (note significant decrease at 11 min; P < 0.05). ○, exercise RBCs incubated in TEP without any additions. ▪, ‘resting’ plasma with resting RBCs; ♦, resting RBCs treated with ouabain incubated in TEP with 0.1 mm ouabain (note significant increase from 10.5 min onwards). ▾, resting RBCs treated with bumetanide incubated in TEP with 0.1 mm bumetanide (note significant decrease from 11 min onwards). ▴, resting RBCs treated with ouabain and bumetanide incubated in TEP with 0.1 mm ouabain and 0.1 mm bumetanide.

Compared to the incubation of resting RBCs in TEP, K⁺ uptake by RBCs was significantly impaired in the presence of ouabain or ouabain plus bumetanide (Fig. 3B). In contrast, bumetanide alone had no effect on K⁺ uptake during the time course of these experiments. The rate constant for unidirectional K⁺ uptake by resting RBCs in TEP was 0.372 ± 0.004 μmol (l RBCs)⁻¹ s⁻¹ and was 0.341 ± 0.003 μmol (l RBCs)⁻¹ s⁻¹ in TEP with bumetanide. TEP with ouabain, or with ouabain ± bumetanide, resulted in rate constants of 0.057 ± 0.002 and 0.034 ± 0.002 μmol (l RBCs)⁻¹ s⁻¹, respectively. RBC K⁺ uptake remained lower than in resting RBCs in TEP with incubation of resting RBCs in resting plasma (rate constant, 0.221 ± 0.002 μmol (l RBCs)⁻¹ s⁻¹) and with exercise RBCs in exercise plasma (rate constant, 0.297 ± 0.008 μmol (l RBCs)⁻¹ s⁻¹).

Figure 3. RBC unidirectional K⁺ uptake during the first 240 s of incubation of resting RBCs in TEP or resting plasma.

•, resting RBCs incubated in TEP; ○, exercise RBCs incubated in TEP; ▪, ‘resting’ plasma with resting RBCs; ♦, resting RBCs treated with ouabain incubated in TEP with 0.1 mm ouabain; ▾, resting RBCs treated with bumetanide incubated in TEP with 0.1 mm bumetanide; ▴, resting RBCs treated with ouabain and bumetanide incubated in TEP with 0.1 mm ouabain and 0.1 mm bumetanide. All increased significantly with time except ♦ and ▴.

Incubation of RBCs in TEP produced an initial (10 s) J_K,net of 160 ± 54 mmol (l RBCs)⁻¹ h⁻¹ (Fig. 4), which was markedly lower than that determined in ESP. J_K,net at 60 and 240 s of incubation were 38 ± 17 and 6.2 ± 4.8 mmol (l RBCs)⁻¹ h⁻¹. The corresponding J_K,in were 30 ± 8.7, 6.4 ± 1.6 and 2.7 ± 0.6 mmol (l RBCs)⁻¹ h⁻¹ at 10, 60 and 240 s, respectively (Fig. 5). Mathematically, J_K,net should not be greater than J_K,in; the discrepancy reflects differences in methodologies in conducting the J_K,netvs.J_K,in experiments separately and at different times after blood sampling.

Figure 4. The net flux of K⁺ between RBCs and plasma.

•, resting RBCs incubated in TEP; ○, exercise RBCs incubated in TEP; ▪, ‘resting’ plasma with resting RBCs; ♦, resting RBCs treated with ouabain incubated in TEP with 0.1 mm ouabain; ▾, resting RBCs treated with bumetanide incubated in TEP with 0.1 mm bumetanide; ▴, resting RBCs treated with ouabain and bumetanide incubated in TEP with 0.1 mm ouabain and 0.1 mm bumetanide. Some error bars have been omitted for clarity.

Figure 5. The unidirectional flux of K⁺ (J_K,in) into RBCs.

•, resting RBCs incubated in true exercise plasma (TEP); ○, exercise RBCs incubated in TEP; ▪, ‘resting’ plasma with resting RBCs; ♦, resting RBCs treated with ouabain incubated in TEP with 0.1 mm ouabain; ▾, resting RBCs treated with bumetanide incubated in TEP with 0.1 mm bumetanide; ▴, resting RBCs treated with ouabain and bumetanide incubated in TEP with 0.1 mm ouabain and 0.1 mm bumetanide. All J_K,in values were significantly greater than 0. Ouabain alone and ouabain + bumetanide, resulted in a significant inhibition of J_K,in compared to resting RBCs incubated in TEP with no inhibitors. J_K,in values of resting TBCs in resting plasma, and of exercise RBCs in TEP, were significantly lower than for resting RBCs incubated in TEP.

Incubation of resting RBCs in resting plasma, or of exercise RBCs in TEP, resulted in no change in J_K,net (Fig. 4). The J_K,in of resting RBCs in resting plasma at 10 s of incubation was 14 ± 6.2 mmol (l RBCs)⁻¹ h⁻¹, while that of exercise RBCs in TEP was 22 ± 8.5 mmol (l RBCs)⁻¹ h⁻¹ (Fig. 5A). NKCC inhibition alone (bumetanide) resulted in similar J_K,net and J_K,in values as when RBCs were incubated in TEP without inhibitors. In contrast, inhibition of only the Na⁺,K⁺-ATPase (ouabain) resulted in a significant net loss of K⁺ (negative J_K,net) that increased during the incubation period. J_K,in in the presence of ouabain was similar to that seen when exercise RBCs were incubated in TEP. Addition of RBCs treated with combined bumetanide and ouabain resulted in no change in J_K,net (Fig. 4) and J_K,in was significantly reduced compared to the absence of inhibitors (Fig. 5).

DISCUSSION

The results of the present study support the hypothesis that peak rates of unidirectional K⁺ influx and net K⁺ flux into RBCs occur during the first minute of exposure of RBCs to TEP. The incubation of resting RBCs in TEP was designed to mimic the in vivo condition where, with very high intensity exercise, RBCs within the arterial capillary circulation abruptly come into contact with venous capillary plasma draining contracting myofibres. The results of the present study are in agreement with the concept that altered composition of this venous plasma rapidly stimulates RBC K⁺ transport and volume regulatory responses. These responses serve to prevent excessive decreases in erythrocyte volume (i.e. 6 % decrease vs. 19 % decrease in plasma volume; Lindinger et al. 1999), as well as to ‘buffer’ increases in plasma [K⁺], [lactate] and P_CO2 as shown in the present study and previously (McKelvie et al. 1991; Lindinger et al. 1992). It is suggested that the rapidity of these responses is important in the regulation of local tissue and whole body fluid, ion and acid-base balance.

Methodological considerations

A unique characteristic of this experiment was the ability to approximate in vivo conditions. We avoided or minimized the use of methodological protocols that impair the ability of RBCs to transport ions and regulate cellular volume. The important characteristics of the present protocol included: (1) the RBCs were not washed to avoid altering the cell surface matrix, membrane integrity and ion transporters; (2) RBCs (resting or exercised) were only incubated in plasma from the same subject; (3) the experiments were performed at 37 °C; and (4) a Hct of ≈40 % was used to provide sufficient RBC surface area and mass with the potential to alter plasma constituents. These criteria are considered essential for testing the hypothesis that exercise-induced changes in plasma composition rapidly stimulate RBC J_K,in and J_K,net.

Previously, we used (Lindinger et al. 1999) the initial (10 s) data point to ‘correct’ RBC counts for ‘trapped’ plasma within RBC pellets, an accepted and useful procedure in transport experiments using very low haematocrit values of 1 % or less (Kirk et al. 1992). However, when using a physiological haematocrit of ≈40 %, the use of 10 s data overestimated the contribution of trapped plasma. In two subjects, we therefore directly measured the trapped plasma of RBC pellets, and observed significantly less trapped plasma than estimated using the 10 s plasma ⁸⁶Rb counts.

By definition, J_K,net is the balance between J_K,in and J_K,out, therefore J_K,in should be greater than J_K,net. A concern with the present results is that peak values for J_K,net exceeded those of J_K,in by a factor of ≈5. Values for each of these variables were obtained in separate series of experiments, and this contributes to the discrepancy even though the same sources of plasma and erythrocytes were used. As noted below, the J_K,in values are consistent with the quantity of erythrocyte Na⁺,K⁺-ATPase sites, raising the possibility of methodological errors inherent in the measurements and calculations for J_K,net. It is possible that some settling of red cells occurred at the bottom of the tubes into which the K⁺ electrode was placed. Since the K⁺-selective tip of the electrode was also at the bottom of the tube, the decrease in [K⁺] at the detection site may have exceeded that of the plasma phase as a whole, thus overestimating J_K,net. While plasma [K⁺] measured by electrode or the Nova instrument were similar at the end of the incubation period, this may not have been the case at time points taken during the first 30 s of incubation.

RBC K⁺ transport in the absence of inhibitors

It is helpful to place the observed high rates of RBC K⁺ flux in the context of skeletal muscle K⁺ flux and Na⁺,K⁺-ATPase density. While it is recognized that erythrocytes have a low density of Na⁺,K⁺-ATPase per square micrometre of cell surface area compared to the sarcolemma, this is an incomplete picture. Skeletal muscle has 1600 pumping sites per μm² (Erlij & Grinstein, 1976) and, with a total surface area of ≈0.2 cm² g⁻¹ (Eisenberg, 1983), 320 billion pump sites per gram. Red blood cells average 450 pump sites per cell (Cheng et al. 1984) and with 5 million red cells per millilitre of blood, this yields 12 millions cells per gram of packed red cells with 5.4 billion pump sites. The maximum in vivo pumping rate of rat soleus muscle equates to ≈300 mmol (kg wet weight)⁻¹ h⁻¹ (Clausen et al. 1987), compared to the peak J_K,net of ≈160 mmol l⁻¹ h⁻¹ in the present study (Fig. 4). It is possible that the J_K,net values in the present study are high, as they are about 5-fold greater than measured J_K,in. The measured J_K,in values are more in line with the number of pump sites present within the red cell population when compared to skeletal muscle.

In both series of control experiments (resting RBCs incubated in resting plasma, and exercise RBCs incubated in TEP) there was no increase in J_K,net. However, J_K,in was elevated in both series immediately after the addition of RBCs to the plasma. While these increases were less than that seen when resting RBCs were incubated in TEP, the decline over time was similar to that seen in the other experiments. These results suggest that prior ‘packing’ of RBCs or dispersion of RBCs into plasma may have increased the open probability of K⁺ channels and Na⁺,K⁺-ATPase activity.

The J_K,net of resting RBCs incubated in TEP at 10 s (178 ± 60 mmol (l RBCs)⁻¹ h⁻¹) was significantly greater than at 30 s (74 ± 27 mmol (l RBCs)⁻¹ h⁻¹) and 120 s (22 ± 11 mmol (l RBCs)⁻¹ h⁻¹), as predicted. The J_K,net values at 30 and 120 s were similar to those previously reported (Lindinger et al. 1999). This increase in J_K,net was caused by a marked increase in J_K,in (30 ± 8.7 mmol (l RBCs)⁻¹ h⁻¹) compared to that seen when resting RBCs were incubated in resting plasma (14 ± 6.1 mmol (l RBCs)⁻¹ h⁻¹). Also, incubation of resting RBCs in resting plasma resulted in no net flux of K⁺, indicating a balance between J_K,in and unidirectional K⁺ efflux (J_K,out). Incubation of resting RBCs in ESP resulted in a greater stimulation of both J_K,net and J_K,in than when resting RBCs were incubated in TEP. Both J_K,net and J_K,in decreased rapidly during the initial 120 s of incubation of resting cells in TEP and ESP. The increased J_K,net when RBCs were incubated in ESP, compared to TEP, can be attributed to higher [K⁺], osmolality, [Na⁺] and [Cl⁻] in ESP versus TEP, creating greater stimuli for K⁺ uptake (Lindinger et al. 1999). It is concluded that exercise-induced increases in plasma [K⁺], osmolality, [Na⁺] and [Cl⁻] stimulated net and unidirectional K⁺ uptake by RBCs within the first seconds of incubation of RBCs in TEP.

RBC K⁺ transport in the presence of inhibitors

The Na⁺,K⁺-ATPase and the NKCC are two primary mechanisms responsible for increases in RBC J_K,in and J_K,net (O'Neill, 1999). Although not measured, it is also important to recognize that outward K⁺ flux from cells occurs, primarily through the K⁺-Cl⁻ cotransporter (KCC). It is also germane to the present study that the KCC is partially inhibited by the concentration of bumetanide used in the present study (Culliford et al. 2003).

NKCC inhibition

In the present study, NKCC inhibition had no effect on J_K,in at 10 s of incubation, and only accounted for 4 ± 2 % (not significant) of J_K,in at 240 s of incubation. The minor effect of NKCC inhibition on J_K,in in the present study and previously (Lindinger et al. 1999) suggests that the NKCC was not activated, at least initially, under these conditions. In our previous study (Lindinger et al. 1999), the apparent contribution of the NKCC was 8.1 ± 5.6 % of the increase in J_K,in; this is an artifact because we had not pre-incubated the RBCs with inhibitor prior to adding RBCs to TEP or ESP. Therefore the apparent increase in J_K,in in the presence of bumetanide seen previously (Lindinger et al. 1999) reflected the time course of bumetanide binding to the NKCC and not a bumetanide-sensitive J_K,in component.

Na⁺,K⁺-ATPase inhibition

In the present study, inhibition of ouabain-sensitive Na⁺,K⁺-ATPase activity decreased J_K,in by 20 ± 4 % at 10 s of incubation, 26 ± 5 % at 30 s, 43 ± 12 % at 120 s and 55 ± 8 % at 240 s, compared to J_K,in of resting RBCs incubated in TEP without inhibitors at the corresponding time points. In our previous study, the Na⁺,K⁺-ATPase was responsible for 71 ± 5 % of the increase in J_K,in after 15 min of incubation of resting RBCs in TEP (Lindinger et al. 1999). In the absence of Na⁺,K⁺-ATPase inhibition, it therefore appears that exercise-induced changes in plasma composition results in a time-dependent increase in the ouabain-sensitive ATPase activity. This suggests an increasing reliance on the Na⁺,K⁺-ATPase to maintain an inward flux of K⁺ into the cells and to regulate intracellular [Na⁺]. In the presence of Na⁺,K⁺-ATPase inhibition, there was an initial increase in NKCC activity, but this was probably dependent on the Na⁺,K⁺-ATPase maintaining the transmembrane driving force for Na⁺ influx (O'Neill, 1999). Thus, in the present study, it appears that there was an increasing inhibition of NKCC activity as intracellular [Na⁺] increased, leading to combined Na⁺,K⁺-ATPase and NKCC inhibition.

In the present study, the net loss of K⁺ from ouabain-treated RBCs incubated in ouabain-treated TEP was unexpected, but consistent with similar reports in skeletal muscle (Hawke et al. 2001) and cardiac myocytes (Vasalle et al. 1962). Measurements of Evans Blue and haemoglobin concentrations ruled out the possibility of large volume changes and cell lysis in these experiments, indicating that membrane integrity was maintained. Inhibition of Na⁺,K⁺-ATPase activity would have allowed intracellular [Na⁺] to increase (O'Neill, 1999), perhaps to a sufficient degree to generate an increase in intracellular [Ca²⁺]. Increased intracellular [Ca²⁺], in turn, would cause an increased open probability of Ca²⁺-activated K⁺ channels in the plasma membrane (Dunn, 1998), increasing unidirectional flow of K⁺ out of the cells. In skeletal muscle cells, the ouabain-induced loss of K⁺ was inhibited by tetrodotoxin (Hawke et al. 2001) suggestive of an outward flow of K⁺ through voltage-gated Na⁺ channels by slip-mode conductance (Santana et al. 1998; Hawke et al. 2001). This mechanism also remains a possibility in RBCs inhibited by ouabain. Alternatively, preincubation of RBCs with ouabain may have resulted in sufficient accumulation of intracellular Na⁺ and cell swelling (O'Neill, 1999). To counter the volume increase, the outwardly directed KCC, which is stimulated in swollen mature erythrocytes (Godart & Ellory, 1996), may have been activated. This would result in a net loss of K⁺ from RBCs and regulatory volume decrease.

Combined Na⁺,K⁺-ATPase and NKCC inhibition

The combined inhibition of Na⁺,K⁺-ATPase and NKCC activities resulted in significant decreases in J_K,in and J_K,net. The decreases in J_K,in (40 %, 44 %, 57 % and 67 % at 10 s, 30 s, 120 s and 240 s, respectively) were significantly greater than values seen with either ouabain or bumetanide alone. Because there appeared to be little or no activation of NKCC activity when the Na⁺,K⁺-ATPase was not inhibited, these data suggest an increased role for the NKCC under conditions where the Na⁺,K⁺-ATPase is blocked. Thus, while one transporter is blocked (Na⁺,K⁺-ATPase), the other transporter (NKCC) may be able to compensate, at least partially, for the loss of K⁺ transport function. A similar compensatory effect was described by McKelvie et al. (1997); the β-adrenoreceptor blocker propranolol did not affect RBC K⁺ uptake, a finding interpreted to mean either the Na⁺,K⁺-ATPase was not involved in K⁺ uptake, or compensation by other transporters occurred. Furthermore, in the present study, when resting RBCs were inhibited with both bumetanide and ouabain, there was no net K⁺ uptake. In general, these results are consistent with past studies that have attributed increased RBC K⁺ uptake to the Na⁺,K⁺-ATPase and NKCC (Maassen et al. 1998; Lindinger et al. 1999). Finally, combined inhibition of the ouabain-sensitive Na⁺,K⁺-ATPase and NKCC only partially inhibited J_K,in. This effect was also similar to that seen previously (Lindinger et al. 1999) and is attributed to increased passive flux of K⁺ through channels in the erythrocyte membrane (Gardos, 1958; Richter et al. 1997).

An interesting result of combined Na⁺,K⁺-ATPase and NKCC inhibition was a positive J_K,net, compared to the negative J_K,net seen with Na⁺,K⁺-ATPase inhibition alone. This result is consistent with a swelling-induced activation of the KCC with Na⁺,K⁺-ATPase inhibition (Kaji, 1986; Duhm, 1987). However, in the presence of both bumetanide and ouabain, it appears that bumetanide may have inhibited the swelling-induced activation of KCC, thereby reducing unidirectional flux and preventing net K⁺ loss.

In vivo perspectives: regulation of plasma [K⁺] by RBCs during high intensity exercise

Approximately 80 % of the exercise-induced increase in blood K⁺ content occurred within erythrocytes after 30 s of cycling at power outputs equivalent to about 350 % (McKelvie et al. 1991, 1992; Lindinger, 1995). An estimate of the potential for RBCs to clear K⁺ from plasma in vivo can be made using the mean peak J_K,net at 10 s of incubation. Blood flow through the femoral vein during dynamic knee extension exercise at 100 % was 7.1 l min⁻¹ for each leg (Sjøgaard et al. 1985) which, when using a Hct of 40 % equates to a red cell flow of 1.4 l RBCs per leg in 30 s. Applying the mean peak J_K,net of 178 mmol (l RBCs)⁻¹ h⁻¹ for resting RBCs incubated in TEP, the 2.8 l of RBCs (two legs performing high intensity cycling) would take up 4.2 mmol of K⁺ in 30 s, raising RBC [K⁺] by ≈1.5 mmol l⁻¹. These rates of net K⁺ uptake need not be sustained because, with very high intensity exercise, power outputs and proportionate rates of skeletal muscle K⁺ loss decrease by about 50 % in 30 s. Also, when RBCs perfuse non-contracting tissues during periods of high intensity exercise, they release K⁺ to these tissues (McKelvie et al. 1992; Lindinger et al. 1992, 1995). So why are such high rates of RBC K⁺ transport required? Transit time of RBCs through the working muscle is very short, taking ≈0.2 s during maximal (100 % ) one-legged dynamic knee extension exercise (Sjøgaard et al. 1985). The high rates of net K⁺ transport by the RBC reported here and by Lindinger et al. (1999) are needed for the RBCs to achieve a lowering of plasma [K⁺]. This facilitates K⁺ removal from the interstitial compartment of contracting muscle where K⁺ accumulation contributes to muscle fatigue (Sjøgaard, 1996). The present results are consistent with a model of plasma and whole body K⁺ regulation presented by Lindinger et al. (1995) that describes a role for the RBC as a K⁺ transport vehicle. The RBCs serve to remove K⁺ from the interstitium and plasma within contracting muscles, then release K⁺ back to the plasma as RBCs perfuse non-contracting tissues where K⁺ is taken up. In contrast, during prolonged, sub-maximal exercise when contracting muscle K⁺ efflux rates are low, it is unlikely that erythrocytes play an important role in plasma K⁺ regulation.

Conclusions

When plasma composition is altered by very high intensity exercise, initial (10–30 s) J_K,in into RBCs was very high, facilitating the clearance of K⁺ from plasma. During the initial 120 s of incubation of RBCs in TEP there was a rapid slowing of net and unidirectional K⁺ flux into RBCs; it is perhaps fortuitous that this occurs in concert with decreasing power output and net K⁺ loss from contracting muscle during very high intensity exercise. The increased K⁺ uptake was achieved primarily by activation of the ouabain-sensitive Na⁺,K⁺-ATPase. When the Na⁺,K⁺-ATPase was blocked, there appeared to be an increased reliance on the NKCC to cause J_K,in. It is concluded that RBCs have the ability to rapidly effect a lowering of plasma [K⁺]in vivo during very high intensity exercise in humans.