A 10-mg dose of amiloride increases time to failure during blood-flow-restricted plantar flexion in healthy adults without influencing blood pressure

Abstract

Amiloride has been shown to inhibit acid-sensing ion channels (ASICs), which contribute to ischemia-related muscle pain during exercise. The purpose of this study was to determine if a single oral dose of amiloride would improve exercise tolerance and attenuate blood pressure during blood-flow-restricted (BFR) exercise in healthy adults. Ten subjects (4 females) performed isometric plantar flexion exercise with BFR (30% maximal voluntary contraction) after ingesting either a 10-mg dose of amiloride or a volume-matched placebo (random order). Time to failure, time-tension index (TTI), and perceived pain (visual analog scale) were compared between the amiloride and placebo trials. Mean blood pressure, heart rate, blood pressure index (BPI), and BPI normalized to TTI (BPI_norm) were also compared between trials using both time-matched (TM₅₀ and TM₁₀₀) and effort-matched (T₅₀ and T₁₀₀) comparisons. Time to failure (+69.4 ± 63.2 s, P < 0.01) and TTI (+1,441 ± 633 kg·s, P = 0.02) were both significantly increased in the amiloride trial compared with placebo, despite no increase in pain (+0.4 ± 1.7 cm, P = 0.46). In contrast, amiloride had no significant influence on the mean blood pressure or heart rate responses, nor were there any significant differences in BPI or BPI_norm between trials when matched for time (all P ≥ 0.13). When matched for effort, BPI was significantly greater in the amiloride trial (+5,300 ± 1,798 mmHg·s, P = 0.01), likely owing to an increase in total exercise duration. In conclusion, a 10-mg oral dose of amiloride appears to significantly improve the tolerance to BFR exercise in healthy adults without influencing blood pressure responses.

INTRODUCTION

Throughout history, physiologists have examined the contributors to exercise tolerance and muscular work capacity in humans. Much of this work has been focused on skeletal muscle fatigue and pain (1–3), which are considered undesirable side effects of intense muscular work. Although healthy adults can generally mitigate exercise-related muscle pain by simply reducing the intensity or volume of exercise, this strategy may be ineffective in certain clinical populations. For instance, individuals with peripheral artery disease (PAD) are known to experience skeletal muscle pain during ambulation, a symptom known as intermittent claudication (4–6). For these individuals, exercise-related (a.k.a. “exertional”) muscle pain limits the total volume of muscular work, which ultimately impairs the performance of basic daily activities (i.e., walking) (7, 8). Identifying a method of attenuating skeletal muscle pain, independent of exercise intensity, would be of particular benefit to this population.

The mechanisms responsible for exercise-related muscle pain are multifaceted and include both mechanical (9, 10) and chemical (11–13) stimuli. Although mechanosensitization occurs via mechanical distortions in the receptive field of Group III (and to a lesser extent, group IV) afferents (14, 15), Group IV metabosensitive (and to an extent, group III) afferents are sensitized by a wide variety of metabolic byproducts produced during muscular contractions. Examples include thromboxane A2 (16), bradykinin (17), arachidonic acid and lactic acid (18, 19), and decreasing pH (2, 11, 12). Notably, these same mechanisms are also responsible for evoking the exercise pressor reflex, a well-documented neural feedback loop that contributes to the increases in heart rate and blood pressure observed during exercise (20–24). The receptors contributing to this sensory pathway also vary, including transient receptor potential vanilloid channels [i.e., TRPv1 (25, 26)], purinergic receptors [i.e., P2X (27, 28), and acid-sensing ion channels [ASICs; (29)]. ASICs are of particular interest to individuals suffering from ischemia-related muscle pain (such as intermittent claudication). This is because, as the name implies, ASICs are activated by increases in muscle acidity, and both muscle acidity (30) and ASIC function (31, 32) increase during muscle ischemia. Therefore, it would be logical to suspect that blocking ASICs would attenuate exercise-induced muscle pain when performed under ischemia. If true, this could have significant implications for improving exercise tolerance in individuals with PAD.

One pharmacological option for antagonizing ASICs is the use of amiloride. Amiloride is a potassium-sparing diuretic, which was traditionally prescribed for individuals with hypertension and/or congestive heart failure (33, 34). However, in addition to increasing water excretion, amiloride has also been shown to nonselectively block ASICs (35, 36). In fact, a recent investigation by Campos et al. (37) found that infusing a soluble form of amiloride into the forearm of awake humans significantly reduced the blood pressure and muscle sympathetic nerve activity (MSNA) responses to voluntary handgrip exercise and metaboreflex activation. Considering the redundant mechanisms mediating the exercise pressor reflex and exercise-related muscle pain, this evidence would support the notion that amiloride may serve as a suitable analgesic for individuals suffering from ischemia-related muscle pain during exercise.

To that end, this study aimed to determine if an acute oral dose of amiloride would be effective at promoting exercise tolerance and attenuating blood pressure responses during blood-flow-restricted (BFR) isometric plantar flexion exercise in young healthy adults. We expected that 10 mg of amiloride would increase the total duration and time-tension index (TTI) of BFR plantar flexion exercise compared with a volume-matched placebo, while not influencing the pain response recorded at end exercise. For clarity, this would indicate that a greater volume of muscular work would be achieved for the same maximal level of pain, suggesting that exercise tolerance was improved. In contrast, we expected amiloride to attenuate the blood pressure, heart rate, and MSNA responses to BFR plantar flexion when matched for time against the same exercise performed after ingestion of placebo. If supported, these hypotheses would provide strong rationale for a larger investigation into the efficacy of amiloride as an analgesic for individuals suffering from ischemia-related muscle pain during exercise.

METHODS

Subjects and Study Design

All experimental protocols used in this study were approved by the Penn State Hershey Institutional Review Board (IRB no. 15300), and all subjects provided written informed consent. Using an anticipated effects size of Cohen’s d = 1.02, an a priori power analysis indicated that 10 subjects would be required to achieve statistical significance with a desired power of 0.8 and significance accepted at P < 0.05 [all power analyses were performed using G*Power version 3.1.9.7 software (38)]. This effect size was estimated from previously reported data demonstrating reduced MSNA responses during metaboreflex activation after amiloride administration (37), and total MNSA was selected based on the expectation that it would provide a reasonably conservative power estimation compared with blood pressure or time to failure. As such, a total of 11 subjects completed this study. All participants were instructed to arrive at each visit at least 3 h postprandial; having abstained from caffeine for 8 h; and alcohol, over-the-counter medication, and intense physical activity for 24 h before their visit. It was discovered that one participant had completed a marathon race roughly 48 h before one of their experimental visits, which confounded their data. Therefore, that subject’s data were excluded, resulting in a total of 10 subjects being included in the final analysis (6 males, 4 females, 30 ± 5 yr old, 173.3 ± 14.2 cm, 75.8 ± 12.0 kg, 25.2 ± 2.0 kg/m²). All subjects were generally healthy and defined as having no known cardiovascular, metabolic, renal, pulmonary, or musculoskeletal diseases that would preclude participation. Day of menstrual cycle phase was also recorded in all female subjects. Of the four female subjects included in the final analysis, three reported a normally occurring menstrual cycle. These three subjects completed their placebo and amiloride visits during days 14 and 18, 14 and 7, and 20 and 9 of their menstrual cycles, respectively.

This study followed a within-subjects repeated-measures design and included three total visits to the Penn State Hershey Clinical Research Center (CRC) in the Clinical and Translational Science Institute (CTSI) located at the Penn State Milton S. Hershey Medical Campus. The first visit served as a familiarization trial, whereas visits 2 and 3 served as experimental trials. To control for dietary influences, subjects completed a 24-h dietary recall form during visit 1, a copy of which was returned to each subject. Subjects were then instructed to follow the same diet 24 h before their second and third visits. Each subject also completed their visits during the same time of day and in a temperature-controlled environment (20°C–22°C).

Experimental Procedures

Each visit followed the same general design, with subjects first being instrumented for data collection. After instrumentation, the force corresponding to each subject’s maximal voluntary isometric plantar flexion contraction (MVC) was determined for the nondominant leg. MVC assessments were performed three times, each separated by 60 s, with the highest force being recorded as the MVC. Next, after a 2-min baseline, subjects performed continuous isometric plantar flexion exercise of their nondominant leg at 30% of their predetermined MVC until volitional fatigue. All exercise (including evaluation of MVC) was performed in a semiseated position, as illustrated in Fig. 1. In visit 1, participants performed both free-flow and BFR plantar flexion in a randomized order. BFR was achieved using an automated pneumatic pressurized cuff (E20 Rapid Cuff Inflator, DE Hokanson, Bellevue, WA) inflated to 100% of the pressure required to completely occlude tibial artery blood flow at baseline [validated via Doppler ultrasound (iE33, Phillips, Andover, MA)]. In brief, the pressure within the occlusion cuff was slowly inflated in a stepwise manner (beginning with 10- to 15-mmHg increments and ending with ∼5-mmHg increments) with the leg secured in the exercising position. Blood flow velocity was continuously recorded from the tibial artery by an experienced investigator (T.H. and J.S.), and the same investigator recorded blood flow within each subject. The cuff pressure corresponding to the disappearance of the flow velocity signal (via Doppler) was recorded as the 100% limb occlusion pressure, and this value was reassessed at the beginning of each visit (occlusion pressures: 151 ± 21 mmHg, 165 ± 16 mmHg, and 153 ± 21 mmHg in the familiarization, placebo, and amiloride visits, respectively; F_2,18=2.107, P = 0.15). In the experimental visits (visits 2 and 3), participants again performed BFR exercise at 30% MVC (determined at the beginning of each visit) until volitional fatigue after ingesting either a 10-mg oral dose of amiloride or a volume-matched placebo (described in more detail in the Dosing Protocol). The order of experimental trials (placebo vs. amiloride) was randomized before visit 1 and blinded to the participants and primary investigators. Randomization was overseen by the investigational drug pharmacy at the Penn State Hershey Medical Center. In an abundance of caution (in the unlikely event of an adverse reaction to the study drug), two licensed nurses on the study team (A.C. and C.B.) were unblinded to the order of trials.

Figure 1.

A schematic of the study design. A: experimental setup, with the subject sitting in a semireclined position with the nondominant foot secured into a force plate. B: time-matching protocol between the amiloride and placebo visits. T₅₀ and T₁₀₀ represent the 50% and 100% effort time points for each trial, respectively, whereas TM₅₀ and TM₁₀₀ represent the time points for each trial corresponding to 50% and 100% of the placebo trial, respectively. BFR, blood flow restriction; FF, free flow; v1, visit 1.

Dosing Protocol

A commercially available supply of 5-mg amiloride tablets (Par Pharmaceuticals, NDC no. 49884-0117-01) were compounded, stored, and provided to study subjects by the investigational drug services pharmacy at the Penn State Milton S. Hershey Medical Center. Active capsules consisted of a 5-mg amiloride tablet and methylcellulose filler within a gelatin capsule, whereas the placebo dose consisted of only methylcellulose within a gelatin capsule. Two capsules were provided for each 10-mg dose of amiloride and the volume-matched placebo. The dose randomly assigned for visit 2 was dispensed to subjects at the end of visit 1, and likewise, the dose randomly assigned for visit 3 was dispensed at the end of visit 2. Subjects were instructed to take this dose ∼3 h before the scheduled start of their experimental visit. This instruction was based on the 6-h half-life of amiloride and the assumption that peak serum concentrations would be reached between 3 and 4 h (39, 40).

Instrumentation

Plantar flexion force, blood pressure, and heart rate were all continuously recorded via a multichannel data acquisition system (PowerLab, AD Instruments, Colorado Springs, CO). Plantar flexion force (kg) was recorded via a force transducer (Transducer Techniques, Temecula, CA) inserted into the footplate of the plantar flexion dynamometer, and visual force feedback was displayed to subjects on a computer screen (Fig. 1). Heart rate was recorded from a standard three-lead electrocardiogram (Cardiocap/5, GE Healthcare, Waukesha, WI), and beat-by-beat blood pressure was recorded via finger photoplethysmography (Finometer, FMS, Arnhem, The Netherlands). To improve the accuracy of the beat-by-beat blood pressure recording, the baseline Finometer reading for each trial was adjusted to a brachial blood pressure reading collected at the same time (SureSigns VS3, Phillips, Andover, MA). Blood flow velocity was also recorded from the anterior or posterior tibial artery during assessment of occlusion pressure and throughout each plantar flexion trial via Doppler ultrasound (iE33, Phillips, Andover, MA) using a linear-array transducer (L11-3) operating at a frequency of 3.6 MHz and maintaining an insonation angle ≤60°. The blood flow velocity signal was only used to confirm that BFR was maintained throughout BFR trials. Perceived pain and Rating of Perceived Exercise (RPE) were also recorded at the end of each exercise trial using a 12-cm visual analog scale (VAS) and the Borg 6–20 RPE scale, respectively. To determine if amiloride had any influence on metabolite accumulation, whole blood lactate (mmol/L; used as a surrogate marker of metabolic byproduct accumulation) was collected from the forefinger ∼5 min before and after each BFR trial during the experimental visits using a handheld lactate meter (Lactate Plus, Nova Biomedical, Waltham, MA).

MSNA was successfully recorded from the peroneal nerves of the nonexercising leg in seven subjects. First, the course of the peroneal nerve was mapped using short trains of external stimulation (10–60 V, 1–5 ms, 1 Hz) with the leg secured in an extended position. After the optimal location for signal acquisition was identified, a tungsten microelectrode (∼200 µm diameter with a tapered noninsulated tip) was inserted into the peroneal nerve by a trained microneurographer (J.C., J.S., or T.H.) and a reference electrode was inserted ∼2–3 cm away from the recording site. The recording electrode was adjusted until MSNA burst activity was clearly identified (using both audio and visual feedback) according to previously established criteria (41). The raw nerve signal was amplified, band-pass filtered (500–5,000 Hz), and integrated at 0.1 s (Iowa Bioengineering, Iowa City, IA). This integrated nerve signal was then analyzed for burst rate (bursts/min), burst incidence (bursts/100 beats), and total MSNA (burst area/HR; a.u.).

Data Analysis

In the present study, exercise tolerance was quantified as the total time to failure (seconds), the pain response recorded at the end of each exercise trial (cm), and as the total time tension index (TTI) achieved during each exercise bout. TTI was calculated as the area under the curve (AUC) for plantar flexion force across all R-R intervals and was expressed relative to seconds (kg × s). Cardiovascular responses were quantified as the mean heart rate (HR), mean arterial pressure (MAP), systolic blood pressure (SBP), and diastolic blood pressure (DBP) responses recorded during each trial, as well as the total MAP blood pressure index (BPI) recorded throughout each exercise bout. Similar to TTI, BPI was recorded as the AUC for MAP across all R-R intervals and was expressed relative to seconds (mmHg × s). Considering the underlying hypothesis that amiloride would increase the duration of BFR plantar flexion, BPI was expected to be significantly greater during the amiloride trial. To adjust for this, BPI was also expressed relative to TTI (BPI_norm; mmHg/kg) at each time point, normalizing the blood pressure response to the exercise stimulus.

To account for the presumed differences in total exercise time, exercise tolerance and cardiovascular responses were also compared between the placebo and amiloride trials using effort-matched and time-matched comparisons. Specifically, 60-s averages of heart rate and blood pressure were collected at the 50% time point of the placebo trial and 50% time point of the amiloride trial, independently (T₅₀). Likewise, BPI and BPI_norm were calculated at T₅₀ as the AUC beginning at the onset of exercise and ending at the 50% time points for the placebo trial and the amiloride trial, independently. Next, heart rate and blood pressure values were once again averaged in 60-s bins around the time points in the amiloride trial that corresponded to the 50% and 100% time points previously identified in the placebo trial (TM₅₀ and TM₁₀₀, respectively). Likewise, additional BPI and BPI_norm values were also calculated in the amiloride trial as the AUC beginning at baseline and ending at the time point corresponding to the 50% and 100% time points in the placebo trial. This time-matching approach was performed within each individual subject and is illustrated in Fig. 1.

Statistical Approach

This study examined the primary hypothesis that a 10-mg dose of amiloride would increase exercise tolerance while mitigating the blood pressure and MSNA responses to BFR plantar flexion, which was selected as a method of simulating the peripheral artery insufficiency associated with PAD. Before testing these hypotheses, we first needed to provide proof-of-concept that the application of BFR would, in fact, reduce exercise tolerance and exaggerate sympathetic responses to plantar flexion in healthy adults. Accordingly, time to failure, RPE, pain, and mean blood pressure responses were compared between the BFR and free-flow trials completed during visit 1 using two separate treatments (FF vs. BFR) by time (baseline vs. T₅₀ vs. T₁₀₀ and baseline vs. TM₅₀ vs. TM₁₀₀) repeated-measures analyses of variance (RMANOVA). Next, the influence of amiloride on preexercise baseline values were examined using a paired samples t test comparing resting blood pressure, heart rate, MSNA, MVC, and the assigned contraction intensity between the placebo and amiloride trials. The influence of amiloride on relative exercise tolerance was then tested by comparing time to failure, TTI, RPE, and pain responses between the amiloride and placebo trials using separate one-way RMANOVA. Next, the influence of amiloride on mean blood pressure and heart rate responses to BFR plantar flexion were examined using two separate treatments (placebo vs. amiloride) by time (baseline vs. T₅₀ vs. T₁₀₀ and baseline vs. TM₅₀ vs. TM₁₀₀) RMANOVA, and peak cardiovascular responses were compared using a single treatment (placebo vs. amiloride) by time (baseline, peak value) RMANOVA. The influence of amiloride on the BPI and BPI_norm responses was examined using paired-samples t tests (placebo vs. amiloride). Finally, MSNA responses were compared between the amiloride and placebo trials using separate t tests for each time-matched and effort-matched comparison. Any significant main effects of treatment or interactions were examined further using post hoc analyses with corrections for multiple comparisons (Sidak). All data are presented as means ± standard deviation (SD).

RESULTS

Proof of Concept

As indicated in Fig. 2A, the total duration of exercise was significantly attenuated during BFR plantar flexion exercise compared with free-flow exercise in visit 1 (483.4 ± 262.8 s vs. 791.7 ± 362.6 s in BFR and free-flow trials, respectively, P = 0.01). Likewise, the application of BFR significantly increased the pain response (3.2 ± 2.3 cm vs. 1.5 ± 1.3 cm in BFR vs. free-flow exercise, respectively, P < 0.01) and RPE response (17 ± 1 vs. 15 ± 2 in BFR vs. free-flow exercise, respectively, P = 0.03) compared with free-flow exercise. Results also indicated a significantly augmented mean MAP response during BFR plantar flexion when matched for time against free-flow exercise (F_1,9= 6.321, η_P² = 0.413, P = 0.03; Fig 2B), which was not observed when MAP was matched for relative effort (F_1,9= 3.272, η_P² = 0.267, P = 0.10; Fig 2C). Main effects of time were also observed for MAP in both trials when matched for time (F_1,9= 22.058, η_P² = 0.769, P < 0.01; Fig 2B) and effort (F_1,9= 21.223, η_P² = 0.766, P < 0.01; Fig 2C). These data support the proof-of-concept that the application of BFR would significantly impair exercise tolerance while exaggerating the exercise pressor reflex in young healthy adults.

Figure 2.

Time to failure (A) and mean arterial pressure (MAP) responses matched for time (B) and relative effort (C) compared between the blood flow restriction (BFR) and free-flow (FF) trials using repeated measures analyses of variance. *Significant differences between trials at the identified time points. †Significant increase from baseline for both trials. Significance was accepted at P < 0.05, and data are presented as means ± SD. n = number of subjects included in analysis. In A, white circles represent male subjects and white triangles represent female subjects.

Influence of Amiloride on Exercise Tolerance

Results indicated that amiloride had no significant influence on the MVC recorded at baseline compared with the placebo trial (Table 1) nor was there any significant effect on the assigned (30%) exercise intensity. However, when time to failure was compared between the amiloride and placebo trials, results indicated a significant increase in time to failure during the amiloride trial (Table 2, Fig. 3C). The total TTI tended to be significantly greater during the amiloride trial compared with the placebo trial when initially examined with a standard pairwise comparison (d = 0.496 [95% CI: −0.176 to 1.143]; P = 0.07). However, when the relative difference in assigned contraction intensity between trials (within subjects) was included as a covariate, results indicated a significantly greater TTI during the amiloride trial compared with placebo (Table 2, Fig. 3A). For clarity, although there were no significant differences in assigned contraction intensity between trials, there was substantial within-subject variability that could have influenced the calculation of TTI. In contrast, no significant differences were observed for the peak pain response (Fig. 3B) or RPE response between trials (Table 2).

Table 1.

Preexercise values compared between the amiloride and placebo trials

	n	Placebo	Amiloride	Mean Diff	95% CI	Cohen’s d	P
Maximal voluntary contraction, kg	10	40.5 ± 3.6	46.5 ± 20.2	6.0 ± 17.3	−4.7/16.7	−0.347	0.15
Assigned contraction intensity, kg	10	15.7 ± 4.5	16.8 ± 7.2	1.0 ± 5.8	−2.5/4.6	0.187	0.28
Mean arterial pressure, mmHg	10	83 ± 7	83 ± 7	0 ± 7	−5/4	0.039	0.45
Systolic blood pressure, mmHg	10	112 ± 9	114 ± 10	1 ± 6	−5/2	0.251	0.22
Diastolic blood pressure, mmHg	10	71 ± 7	71 ± 8	0 ± 8	−5/5	−0.028	0.46
Heart rate, beats/min	10	60 ± 6	62 ± 5	2 ± 4	−5/0	0.475	0.16
MSNA burst rate, bursts/min	7	22.5 ± 10.0	17.8 ± 10.0	−4.7 ± 7.4	−10.2/0.8	0.632	0.07
MSNA burst incidence, burst/100 beats	7	35.3 ± 15.1	27.3 ± 15.0	−8.0 ± 10.6*	−15.9/−0.1	0.758	0.04
Total MSNA, a.u.	7	408.8 ± 175.3	372.5 ± 195.7	−36.2 ± 126.4	−129.9/57.3	0.287	0.23

Data are means ± SD for placebo, amiloride, and mean diff; n, number of participants. CI, confidence intervals; mean diff, mean difference; MSNA, muscle sympathetic nerve activity. *Significant difference between conditions.

Table 2.

Markers of exercise tolerance compared between amiloride and placebo trials

	n	Placebo	Amiloride	Mean Diff	95% CI	ηp2	P
Time to failure, s	10	570.1 ± 255.0	639.6 ± 264.1	69.4 ± 63.2*	30.3/108.6	0.573	<0.01
Time tension index, kg·s/1,000a	10	6.72 ± 4.31	8.16 ± 5.68	1.44 ± 0.63*	−361.0/3244.2	0.496	0.02
Pain, cm	10	3.1 ± 2.5	3.5 ± 3.2	0.4 ± 1.7	−1.5/0.6	0.06	0.46
RPE	10	16 ± 2	16 ± 2	0 ± 0	0/0	0.18	0.19

Data are means ± SD for placebo, amiloride, and mean diff; n, number of participants.

^aDifferences in assigned contraction intensity included as covariate. Mean diff, mean difference; η_p², partial η squared; RPE, rating of perceived exertion. *Significant difference between conditions.

Figure 3.

Time-tension index (TTI; A), pain collected via visual analog scale (B), and time to failure (C) compared between the amiloride (Amil) and placebo trials using one-way repeated measures analyses of variance. *Significant difference between trials. Significance was accepted at P < 0.05, and data are presented as means with individual data points overlayed. White circles represent male subjects and white triangles represent female subjects.

Influence of Amiloride on Pressor Responses

At baseline, amiloride had no significant influence on resting blood pressure or heart rate but did significantly reduce MSNA burst incidence (Table 1). When mean blood pressure and heart rate responses were compared across time and between trials, results indicated significant main effects of time for HR, SBP, DBP, and MAP when matched for time (Table 3, Fig. 4, E–H) and when matched for effort (Table 4, Fig. 4, A–D). However, no significant main effects of treatment or interactions were observed for any comparison (Tables 3 and 4). Likewise, when peak HR, SBP, DBP, and MAP responses were compared across time and between trials, results indicated only significant main effects of time for all comparisons (Table 4), except for DBP, which demonstrated a significant treatment by time interaction (P = 0.04). This interaction can be explained by a greater increase in DBP from baseline to peak during the amiloride trial compared with placebo (Table 4). Likewise, no significant differences were observed for MSNA burst rate or burst incidence for any time-matched or effort-matched comparison (all P ≥ 0.290; Fig. 5). Total MSNA tended to be higher during the amiloride trial at the effort-matched 50% time point (T₁₀₀; d = −0.797 [95% CI: −1.635 to 0.088], P = 0.08; Fig. 5) but not at any other time point (all d ≤ −0.693, P ≥ 0.15).

Table 3.

Blood pressure and heart rate responses compared between amiloride and placebo trials matched for time

		Placebo			Amiloride			RMANOVA
	n	BL	T50	T100	BL	TM50	TM100	Time	Drug	Int.
HRmean, beats/min	9	59 ± 6	71 ± 8	76 ± 8	62 ± 5	72 ± 7	76 ± 8	F = 33.43 ηP2 = 0.80 *P < 0.01	F = 0.68 ηP2 = 0.07 P = 0.43	F = 3.73 ηP2 = 0.31 *P = 0.04
MAPmean, mmHg	9	83 ± 7	91 ± 8	93 ± 8	82 ± 7	93 ± 11	94 ± 10	F = 38.0 ηP2 = 0.82 *P < 0.01	F = 0.06 ηP2 < 0.01 P = 0.81	F = 1.54 ηP2 = 0.16 P = 0.24
SBPmean, mmHg	9	110 ± 8	121 ± 11	124 ± 11	112 ± 8	125 ± 16	128 ± 16	F = 22.21 ηP2 = 0.73 *P < 0.01	F = 1.19 ηP2 = 0.13 P = 0.30	F = 1.42 ηP2 = 0.15 P = 0.27
DBPmean, mmHg	9	71 ± 7	78 ± 7	80 ± 7	70 ± 8	78 ± 10	80 ± 10	F = 44.32 ηP2 = 0.84 *P < 0.01	F < 0.01 ηP2 < 0.01 P = 0.94	F = 1.44 ηP2 = 0.15 P = 0.26
BPI, mmHg·s/1,000	10		25.9 ± 13.2	49.4 ± 25.9		27.9 ± 13.6	41.8 ± 27.4	F = 14.11 ηP2 = 0.61 *P < 0.01	F = 2.72 ηP2 = 0.23 P = 0.13	F = 6.49 ηP2 = 0.49 *P = 0.03
BPInorm, mmHg/kg	10		8.1 ± 4.1	8.4 ± 4.3		8.2 ± 4.1	8.3 ± 4.1	F = 7.57 ηP2 = 0.45 *P = 0.02	F < 0.01 ηP2 < 0.01 P = 0.98	F = 1.06 ηP2 = 0.10 P = 0.32

Data are means ± SD for placebo and amiloride; n, number of participants. BL, baseline; BPI, blood pressure index; BPI_norm, BPI normalized to the time-tension index; DBP, diastolic blood pressure; HR, heart rate; MAP, mean arterial pressure; η_p², partial η squared; SBP, systolic blood pressure; T₅₀, 50% of total exercise time within each trial; T₁₀₀, 100% of total exercise time within each trial; TM₅₀, time matched for 50% of total exercise time in the placebo trial; TM₁₀₀, time matched for 100% of the total exercise time in the placebo trial. *Statistically significant P value.

Figure 4.

Heart rate (HR; A and E), mean arterial pressure (MAP; B and F), systolic blood pressure (SBP; C and G), and diastolic blood pressure (DBP; D and H) compared between the amiloride (Amil) and placebo trials matched for time (E–H) and relative effort (A–D) using repeated measures analyses of variance. †Significant increase from baseline for both trials. Significance was accepted at P < 0.05, and data are presented as means ± SD. n = number of subjects included in analysis.

Table 4.

Blood pressure and heart rate responses compared between amiloride and placebo trials matched for effort

		Placebo			Amiloride			RMANOVA
	n	BL	T50	T100	BL	T50	T100	Time	Drug	Int.
HRmean, beats/min	10	60 ± 6	72 ± 8	77 ± 8	62 ± 5	73 ± 6	78 ± 8	F = 47.70 ηP2 = 0.84 *P < 0.01	F = 1.12 ηP2 = 0.11 P = 0.31	F = 0.39 ηP2 = 0.04 P = 0.68
MAPmean, mmHg	10	83 ± 7	92 ± 8	94 ± 8	83 ± 7	95 ± 11	97 ± 13	F = 42.49 ηP2 = 0.82 *P < 0.01	F = 0.56 ηP2 = 0.05 P = 0.47	F = 1.32 ηP2 = 0.12 P = 0.27
SBPmean, mmHg	10	112 ± 9	123 ± 13	127 ± 14	114 ± 10	127 ± 16	133 ± 21	F = 26.46 ηP2 = 0.74 *P < 0.01	F = 1.69 ηP2 = 0.15 P = 0.22	F = 1.11 ηP2 = 0.11 P = 0.35
DBPmean, mmHg	10	71 ± 7	78 ± 7	80 ± 7	71 ± 8	80 ± 9	82 ± 11	F = 52.50 ηP2 = 0.85 *P < 0.01	F = 0.22 ηP2 = 0.02 P = 0.64	F = 1.70 ηP2 = 0.16 P = 0.20
HRpeak, beats/min	10	60 ± 6		85 ± 8	62 ± 5		87 ± 8	F = 169.77 ηP2 = 0.95 *P < 0.01	F = 1.77 ηP2 = 0.16 P = 0.21	F = 0.01 ηP2 < 0.01 P = 0.90
MAPpeak, mmHg	10	83 ± 7		102 ± 12	83 ± 7		106 ± 13	F = 70.25 ηP2 = 0.88 *P < 0.01	F = 0.85 ηP2 = 0.08 P = 0.38	F = 4.19 ηP2 = 0.31 P = 0.07
SBPpeak, mmHg	10	112 ± 9		135 ± 19	114 ± 10		144 ± 21	F = 44.59 ηP2 = 0.83 *P < 0.01	F = 2.53 ηP2 = 0.22 P = 0.14	F = 3.03 ηP2 = 0.25 P = 0.11
DBPpeak, mmHg	10	71 ± 7		87 ± 10	71 ± 8		90 ± 10	F = 88.29 ηP2 = 0.90 *P < 0.01	F = 0.31 ηP2 = 0.03 P = 0.59	F = 5.73 ηP2 = 0.38 *P = 0.04
BPI, mmHg·s/1,000	10		25.9 ± 13.2	49.4 ± 25.9		29.5 ± 13.2	56.5 ± 26.2	F = 28.0 ηP2 = 0.75 *P < 0.01	F = 8.68 ηP2 = 0.49 *P = 0.01	F = 5.22 ηP2 = 0.36 *P = 0.04
BPInorm, mmHg/kg	10		8.1 ± 4.1	8.4 ± 4.3		8.2 ± 4.1	8.6 ± 4.3	F = 10.5 ηP2 = 0.54 *P = 0.01	F < 0.01 ηP2 < 01 P = 0.95	F = 0.188 ηP2 = 0.02 P = 0.67
Lactate, mmol/L	10	1.4 ± 1.0		2.0 ± 1.0	1.5 ± 0.8		2.1 ± 1.3	F = 15.35 ηP2 = 0.63 *P < 0.01	F = 0.05 ηP2 < 0.01 P = 0.82	F < 0.01 ηP2 < 0.01 P = 0.96

Data are means ± SD for placebo and amiloride; n, number of participants. BL, baseline; BPI, blood pressure index; BPI_norm, BPI normalized to the time-tension index; DBP, diastolic blood pressure; HR, heart rate; MAP, mean arterial pressure; η_p², partial η squared; SBP, systolic blood pressure; T₅₀, 50% of total exercise time within each trial; T₁₀₀, 100% of total exercise time within each trial; TM₅₀, time matched for 50% of total exercise time in the placebo trial; TM₁₀₀, time matched for 100% of the total exercise time in the placebo trial. *Statistically significant P value.

Figure 5.

Muscle sympathetic nerve activity (MSNA) burst rate (A and D), burst incidence (B and E), and total activity (C and F) compared between the placebo and amiloride (Amil) trials matched for time (D–F) and effort (A–C) using separate paired samples t tests. Data are presented as mean responses with individual data overlayed. n = number of subjects included in analysis. White circles represent male subjects and white triangles represent female subjects. BL, baseline; T₅₀, effort-matched 50% time point; T₁₀₀, effort-matched 100% time point; TM₅₀, time-matched 50% time point; TM₁₀₀, time-matched 100% time point.

When BPI was compared between trials, results indicated a significant main effect of time and a significant time by treatment interaction for time-matched comparisons (Table 3; Fig 6C), which can be explained by a greater increase from TM₅₀ to TM₁₀₀ in the placebo trial compared with the amiloride trial. When matched for effort, BPI demonstrated significant main effects of time, treatment, and a significant time by treatment interaction (Table 4), which can be explained by a greater total BPI in the amiloride trial compared with the placebo trial (Fig 6A), likely due to the increase in the total duration of exercise. However, when BPI_norm was compared across time and between trials, only significant main effects of time were observed (Table 4, Fig. 6, B and D), indicating no effect of amiloride. Likewise, amiloride had no significant influence on whole blood lactate accumulation (Table 4).

Figure 6.

Blood pressure index (BPI; A and C) and BPI normalized to the total time under tension (BPI_norm; B and D) compared between the amiloride (Amil) and placebo trials matched for time (C and D) and relative effort (A and B) using paired samples t tests. *Significant differences between trials at the identified time points. Significance was accepted at P < 0.05, and data are presented as mean responses with individual data overlayed. n, number of subjects included in the analysis. White circles represent male subjects and white triangles represent female subjects. BPI, blood pressure index; BPI_norm, blood pressure index normalized to the time-tension index.

DISCUSSION

The data presented here support the primary hypothesis that a 10-mg oral dose of amiloride will promote exercise tolerance during BFR plantar flexion. Specifically, time to failure (Fig 3C) and total TTI (Fig 3A) were both increased in the amiloride trial compared with placebo (Table 2), despite no significant increases in perceived pain (Fig 3B) or RPE. In other words, in the amiloride trials, subjects were able to perform more work and for a longer period of time at the same relative level of discomfort. As will be discussed later, this may have significant implications for improving walking tolerance in individuals with PAD-related muscle pain. However, it should be noted that we did not observe a direct influence of amiloride on reducing pain responses either. This may be partially explained by the fact that pain was only assessed at end exercise (limiting the ability to interpret the cumulative pain response), as well as an increased exercise stimulus during the amiloride trial. It should also be noted that results demonstrated a high degree of variability in time to failure, as indicated by the large standard deviation of the mean difference (Table 2). This can largely be explained by the response of a specific subject who performed ∼58 more seconds of exercise in the placebo trial compared with the amiloride trial, representing the only diverging response within the entire sample (all other subjects exercised longer during the amiloride trial, with differences ranging from 15 to 153 s). Indeed, when the data from this subject are removed from this analysis, the mean difference in time to failure between the amiloride and placebo trials increases from 69.4 to 83.6 s, and the standard deviation of the mean difference decreases from 63.2 to 47.3 s.

In contrast to the first hypothesis, our findings did not support the secondary hypothesis that amiloride would attenuate the exaggerated blood pressure responses observed during BFR plantar flexion. This is because no significant main effects of condition were observed for any mean or peak blood pressure or heart rate response, regardless of whether comparisons were matched for time or effort (Tables 3 and 4, Fig. 4). Moreover, the total sum of blood pressure responses (BPI) was not significantly different between amiloride and placebo trials when matched for time (Fig 6C) and was actually augmented during the amiloride trial when matched for relative effort (Table 4, Fig 6A). As noted in the results, this can easily be explained by an increase in total exercise time, which would inflate the total area under the curve for blood pressure (hence, the purpose for including time-matched comparisons). Indeed, this difference in BPI was ameliorated when normalized to TTI (BPI_norm; Table 4, Fig. 6, B and D). Also, although we did not observe any statistically significant influence of amiloride on MSNA responses to BFR plantar flexion, we did observe a large net increase in the Total MSNA response during the amiloride trials (Fig. 5). In fact, this difference becomes statistically significant at the T₅₀ time point (d = −1.055 [95% CI: −2.047 to −0.003], P = 0.04) with the removal of one subject (to the authors’ knowledge, there is no physiological or experimental rationale to justify the removal of this subject in the main analysis). Therefore, the possibility that amiloride increased MSNA burst amplitude or burst area cannot be ruled out. Nevertheless, it remains highly unlikely that the 10 mg dose of amiloride used in the present study had any influence on MSNA burst occurrence (i.e., burst rate or burst incidence). In the following sections, we provide speculation regarding why blood pressure and heart rate were unaffected by amiloride.

Pain and Exercise Tolerance

As discussed in the introduction, exercise-related muscle pain is attributed to a myriad of factors, including the stimulation of nociceptive subgroups of both mechanosensitive (9, 10) and metabosensitive (11, 13) muscle afferents. ASICs have been shown to be a significant contributor to the activation of these nociceptive subgroups in various models of pain (36, 42, 43). In one particular study, Khataei et al. provided compelling evidence that ASICs contribute heavily to exercise-induced muscle pain when they found that ASIC3 knockout mice demonstrated a significantly attenuated pain response to acute high-intensity exercise (29). These authors also found that 8 wk of high-intensity interval training resulted in a downregulation of ASIC expression in C57BL/6J mice, which was significantly correlated to increases in time to exhaustion. Therefore, it would appear that ASICs play a crucial role in evoking exercise-related muscle pain, and consequently, limiting the capacity for muscular work. This was also observed by Campos et al. (37), who reported a significant decrease in perceived discomfort during metaboreflex activation [postexercise muscle ischemia (PEMI)] after infusing amiloride into the occluded forearm of human subjects but not after saline infusion. It should also be noted that the contribution of ASICs to nociception is not strictly limited to the periphery, as ASICs are also known to contribute to sensory pathways within the central nervous system (44). Therefore, the 10-mg oral dose used in the present study may also have resulted in central ASIC inhibition. This may have contributed, in part, to the lack of any significant decrease in the pain response observed in the amiloride trial.

Another interesting area of speculation relates to the influence of afferent stimulation on changes in central motor drive to the contracting skeletal muscle (45–48). Specifically, Amann et al. have demonstrated that pharmacological attenuation of groups III and IV muscle afferent feedback augments motor output (iEMG) to exercising muscles and accelerates the development of peripheral fatigue (47, 49). Hureau et al. (50) expanded on this concept by examining cycling trial performance with and without intact groups III and IV afferent feedback, while also controlling for the rate of convective oxygen delivery. The results from this study provide strong evidence that, when the rate of oxygen delivery is unchanged, muscle afferent mediated inhibition of central motor drive effectively limits exercise performance. Considering that ASICs are known to be expressed on the nerve endings of nociceptive subgroups of group IV muscle afferents (12), it may be possible that amiloride promoted motor activation of the triceps surae by limiting afferent mediated motor inhibition. If true, this could lead to an improvement in exercise tolerance that is independent of exercise-related muscle pain.

Blood Pressure and Heart Rate

The exercise pressor reflex has been a primary focus of human cardiovascular research for many years and includes two primary afferent signaling mechanisms: the mechanoreflex and the metaboreflex (15, 51). Although ASICs have been shown to contribute to both mechano- and metabotransduction (12, 52, 53), it is generally thought that ASIC activation contributes more heavily to metaboreflex activation. For instance, Kim et al. found that genetic knockout of ASIC3 resulted in a significantly attenuated blood pressure response to metabolite infusion (capsaicin, lactic acid, and diprotonated phosphate) but not passive tendon stretch after femoral artery ligation in decerebrated rats (54). Likewise, Campos et al. (37) found that blocking ASICs with a local infusion of amiloride significantly attenuated the mean blood pressure and muscle sympathetic nerve activity (MSNA) responses to isometric handgrip and metaboreflex activation (via PEMI). Based on this, we hypothesized that amiloride would also attenuate the blood pressure responses to ischemic plantar flexion exercise in the present study. However, that is not what we observed. Instead, we found that amiloride had no significant influence on blood pressure (Fig. 4, B–D, F–H) or MSNA burst occurrence (Fig. 5, A, B, D and E), regardless of whether comparisons were matched for time or relative effort. Moreover, this discrepancy could not have been explained by differences in exercise duration or intensity, as no differences were observed even when the total blood pressure responses were normalized to the exercise stimulus (BPI_norm; Fig. 6, B and D). In other words, the intensity and duration of the muscular contraction remained the only significant contributor to the total blood pressure response, as indicated by the significant increases across time in both trials, as well as the significantly greater BPI observed at T₅₀ and T₁₀₀ in the amiloride trial (during which, exercise duration was increased).

At first glance, these findings may seem to contradict those reported by Campos et al. (37). However, that is not the position of the authors. Instead, the discrepancies in the blood pressure results reported in these two studies are likely explained by differences in the administration of amiloride. In the present study, amiloride was administered in pill form roughly 3 h before each study visit, whereas, in the study by Campos et al., amiloride was infused locally after exsanguination of the limb (via Bier-block). Therefore, it is very likely that the amiloride infused in the Campos et al. study acted almost exclusively on ASICs expressed on local sensory afferents. This would support the notion that the pressor response-lowering effects of amiloride are primarily due to inhibition of ASICs at the peripheral sensory neurons, as opposed to centrally mediated ASIC inhibition. In contrast, the amiloride dose administered in our study would likely have had a more systemic influence on ASIC inhibition, including both the peripheral and central nervous systems. Although highly speculative, this may have resulted in a centrally mediated improvement in exercise tolerance, but no significant peripherally mediated antihypertensive response. Indeed, this will be discussed in the next section as a potential focus of future research.

It is also worth noting that the amiloride dose administered in the study by Campos et al. was normalized to body mass (2.5% of forearm volume, 0.75 mg/mL amiloride), whereas the 10-mg oral dose administered in the present study was not. This may also have contributed to differences in blood pressure responses observed between these two studies. Nevertheless, both studies provide support for the notion that the ASIC-blocking properties of amiloride can be exploited to promote exercise tolerance under flow-limited conditions.

Implications for Future Research

As alluded to previously, the dosing and administration protocols used in investigations of amiloride in humans differ based on experimental constraints. With evidence mounting to support a general influence of amiloride on exercise tolerance, future research should begin examining the influence of dosing and administration (i.e., intravenous vs. oral) on exercise-related analgesia and antihypertensive responses. Such research may identify a minimum dosing threshold necessary to achieve antihypertensive responses, which may explain the lack of change in blood pressure observed in the present study.

Similarly, future research should also consider examining the relative influences of centrally mediated versus peripherally mediated ASIC inhibition on exercise-related analgesia and blood pressure regulation. However, answering this question would require a unique experimental approach. For instance, the contributions of centrally versus peripherally expressed ASICs could not easily be examined by targeting specific ASIC subunits, as both ASIC1a and ASIC3 have been shown to be expressed within the central and peripheral nervous systems (55–57). Campos et al. (37) addressed this by infusing amiloride locally and under limb occlusion, which would presumably isolate ASIC inhibition to the peripheral sensory neurons. However, there is no comparable method to restrict ASIC inhibition to the central nervous system. That is, until recently, when investigators from the United States and Canada developed an aerosolized form of amiloride that can be delivered through the nasal passages (58). This could provide a valuable method of localizing amiloride administration to the central nervous system, allowing investigators to determine the relative contributions of central versus peripheral ASIC activation on exercise-related algesia and blood pressure responses.

Perhaps the most promising area of future research is the potential application of amiloride as a method of promoting exercise tolerance in patients with PAD. In the present study, BFR was applied to plantar flexion as a method of simulating the impaired exercise tolerance and exaggerated blood pressure responses often observed during weight-bearing exercise (i.e., walking) in PAD patients (59, 60). Indeed, this approach was supported by our findings (Fig. 2, A and B). The next logical step would be to determine if the improvements in exercise duration and total work capacity observed after amiloride administration in our study translate to a sample of PAD patients. If so, this would provide strong support for the use of amiloride as an acute treatment for exercise-related skeletal muscle pain. Reducing exercise-related muscle pain would have significant benefits for this population that extend far beyond general comfort. For instance, limiting muscle pain during walking exercise would likely result in significant increases in the total volume of weight-bearing exercises performed during rehabilitation, which has been shown to improve patient outcomes (61–63). Of course, the long-term efficacy of amiloride also needs to be examined to assure sustained sensitivity and patient safety (i.e., risk of hyperkalemia).

Limitations

As is the case with any investigation, this study is subject to certain experimental limitations. One of the primary limitations of this study is that we did not analyze serum concentrations of amiloride at the time of data collection. Despite this, the time course for peak serum concentration is reported to be 3–4 h, and the half-life of amiloride (taken orally) is reported to be 6 h (39, 40), both of which are within the 3-h time frame between dosing and data collection. A similar limitation was that amiloride was prescribed as a single 10-mg dose for all participants, rather than being prescribed relative to body mass (i.e., mg/kg). This approach was intended to balance the effect of the drug while limiting the risk to subjects and was selected in consultation with a team of physicians. However, the primary limitation to this approach is that it may result in some subjects taking too small of a dose, which would increase the risk of a Type II error. Considering that we observed significant changes in markers of exercise tolerance, this likely had little, if any influence on the overall findings of this study. The collection of pain at end exercise also presents a limitation, as this measure only provides a single peak value for the entire exercise period. Future studies may consider using a continuous method of pain assessment. Other limitations include the restriction to a small sample size and the use of a healthy model of arterial insufficiency rather than a sample of PAD patients. The limitation to the overall sample size (n = 10) is primarily due to logistical issues with conducting in-person research during the COVID-19 pandemic (data collection began in December of 2020). As noted above, however, the observation of significant changes to markers of exercise tolerance suggests that this had a limited influence on the interpretation of our findings. The decision to use BFR in lieu of performing these experiments in a sample of PAD patients was primarily based on balancing patient safety against the potential benefits of the study. Although BFR falls short of mimicking the systemic nature of PAD (i.e., atherosclerosis, inflammation, etc.), it does serve as a useful tool for examining ischemia-related skeletal muscle pain. As such, our findings provide strong support for future studies aimed at examining the benefits of amiloride for patients with PAD during walking exercise.

Perspectives and Significance

Our study was the first to examine the influence of an oral dose of amiloride, an ASIC-blocking agent, on exercise tolerance and pressor responses during BFR exercise in healthy adults. Our findings support the hypothesis that amiloride will increase total exercise duration and volume during BFR plantar flexion, without increasing pain or perceived effort. In contrast, our findings do not support the secondary hypothesis that amiloride would attenuate the exaggerated blood pressure responses observed during ischemic exercise. Ultimately, these data support the notion that ASICs are primary contributors to exercise intolerance under ischemic conditions, and that amiloride may serve as a valuable treatment for promoting exercise tolerance under flow-limited conditions.

DATA AVAILABILITY

The data sets generated and/or analyzed during the current study are not publicly available due to data sharing policies of the host institution. However, data may be made available by the corresponding author on reasonable request, and with the permission of the Pennsylvania State University.

GRANTS

This study was supported by National Institutes of Health (NIH) Grants P01-HL134609 (to L. I. Sinoway), UL1-TR002014 (to L. I. Sinoway), R01-HL164571 (to J.C.), the Department of Anesthesiology and Perioperative Medicine at the Penn State Milton S. Hershey Medical Center, and the University of Southern Mississippi (to J.S.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.S., J.C., V.R., P.J., and Z.C. conceived and designed the research; J.S., J.C.L., T.H., C.B., A.C., M.D.H., and J.C. performed the experiments; J.S., J.C.L., T.H., and P.D. analyzed data; J.S., J.C.L., T.H., and J.C. interpreted the results of the experiments; J.S. prepared figures; J.S. drafted the manuscript; J.S., J.C.L., T.H., and J.C. edited and revised the manuscript; J.S., J.C.L., T.H., C.B., A.C., P.D., M.D.H., V.R., P.J., Z.C., and J.C. approved the final version of the manuscript.