The muscle reflex and chemoreflex interaction: ventilatory implications for the exercising human

Abstract

We examined the interactive influence of the muscle reflex (MR) and the chemoreflex (CR) on the ventilatory response to exercise. Eleven healthy subjects (5 women/6 men) completed three bouts of constant-load single-leg knee-extension exercise in a control trial and an identical trial conducted with lumbar intrathecal fentanyl to attenuate neural feedback from lower-limb group III/IV muscle afferents. The exercise during the two trials was performed while breathing ambient air ($\text{S}{\text{a}}_{{\text{O}}_2}$ ~97%, $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$~84 mmHg, $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$ ~32 mmHg, pH ~7.39), or under normocapnic hypoxia ($\text{S}{\text{a}}_{{\text{O}}_2}$ ~79%, $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ ~43 mmHg, $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$ ~33 mmHg, pH ~7.39) or normoxic hypercapnia ($\text{S}{\text{a}}_{{\text{O}}_2}$ ~98%, $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ ~105 mmHg, $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$ ~50 mmHg, pH ~7.26). During coactivation of the MR and the hypoxia-induced CR (O₂-CR), minute ventilation (V̇e) and tidal volume (V_T) were significantly greater compared with the sum of the responses to the activation of each reflex alone; there was no difference between the observed and summated responses in terms of breathing frequency (f_B; P = 0.4). During coactivation of the MR and the hypercapnia-induced CR (CO₂-CR), the observed ventilatory responses were similar to the summated responses of the reflexes (P ≥ 0.1). Therefore, the interaction between the MR and the O₂-CR exerts a hyperadditive effect on V̇e and V_T and an additive effect on f_B, whereas the interaction between the MR and the CO₂-CR is simply additive for all ventilatory parameters. These findings reveal that the MR:CR interaction further augments the ventilatory response to exercise in hypoxia.

NEW & NOTEWORTHY Although the muscle reflex and the chemoreflex are recognized as independent feedback mechanisms regulating breathing during exercise, the ventilatory implications resulting from their interaction remain unclear. We quantified the individual and interactive effects of these reflexes during exercise and revealed differential modes of interaction. Importantly, the reflex interaction further amplifies the ventilatory response to exercise under hypoxemic conditions, highlighting a potential mechanism for optimizing arterial oxygenation in physically active humans at high altitude.

INTRODUCTION

The ventilatory response to exercise is tightly regulated to optimize arterial oxygenation and acid-base balance and ensure, in conjunction with the circulatory response, that the metabolic demand of skeletal muscle is met during physical activity (23). Although central command, a feedforward mechanism, has traditionally been considered the primary driver of exercise hyperpnea (5, 21), the importance of the muscle reflex (MR) and the chemoreflex (CR) has increasingly been recognized in recent years (19, 23).

The muscle reflex is triggered by the excitation of mechano- and metabosensitive group III and IV muscle afferents during exercise (16, 34, 42, 45, 57). Both animal (6, 30, 52) and human (1, 27) studies have demonstrated a crucial role of the neural feedback from these muscle afferents in determining adequate ventilatory responses to exercise. The chemoreflex is predominantly triggered by the stimulation of O₂-sensitive arterial chemoreceptors in the carotid and aortic bodies (i.e., the peripheral chemoreflex) and CO₂-sensitive medullary chemoreceptors (i.e., the central chemoreflex). In contrast to maximal whole body exercise, arterial blood gases and other humoral chemoreceptor stimuli (e.g., plasma K⁺, norepinephrine) seem to only play a minor role in activating the chemoreflex during submaximal effort (18). Nevertheless, despite that the magnitude of the increase in ventilation with exercise intensity may not be directly dictated by the chemoreflex, a constant afferent discharge from the carotid chemoreceptors to medullary respiratory neurons is required to maintain eupneic breathing at rest and during exercise (8, 23, 24). Although various studies have attempted to delineate the individual significance of the muscle reflex and the chemoreflex for the control of breathing, the effect of their interaction for regulating the hyperpnea of voluntary exercise remains obscure (10).

Previous studies in humans, focusing on the interaction between the muscle reflex and the chemoreflex, have altered the concentration of blood gases (to stimulate chemoreceptors) during postexercise circulatory occlusion (to stimulate metabosensitive muscle afferents; 11, 20, 40), or during passive limb stretch/movement (to stimulate mechano-sensitive muscle afferents; 11, 53). However, because the reflexes were engaged at rest, excluding both the background influence of central command (10, 37) and the interplay between mechano- and metabosensitive afferents (13), the outcome of these studies may not fully represent the scenario during exercise. This issue was considered by Fregosi and Seals (26) who had subjects perform dynamic handgrip exercise with a hypoxic inspirate (and also in normoxia) while occluding forearm circulation to activate the muscle reflex and to prevent the confounding effect of systemic hypoxemia on the metabolic milieu in the active muscles. The authors found that the ventilatory response to hypoxic exercise exceeded the summation of the responses during normoxic exercise alone and during hypoxia alone (i.e., at rest). Hence, their findings implicated a hyperadditive interaction when the muscle reflex and the chemoreflex are activated concurrently during exercise (26). However, the ischemic exercise may have, inadvertently, stimulated the metabonociceptors, a subset of metabosensitive afferent fibers not activated by typical dynamic exercise (33, 38, 48). Thus, the exact nature of the MR:CR interaction and the associated consequence for the ventilatory response to exercise remain uncertain.

Therefore, by using pharmacological blockade of muscle afferent feedback to decrease the muscle reflex during leg exercise, while triggering the chemoreflex via hypoxia (O₂-CR) and hypercapnia (CO₂-CR), this study aimed to elucidate the effect of the MR:CR interaction on ventilation during exercise. Of note, despite a diminished effect of the muscle reflex in determining the ventilatory response to exercise when engaging only a small muscle mass, we chose moderate single-leg knee-extension exercise for a short duration to minimize neural feedback from other muscles (e.g., upper body), confounding limitations of the cardiopulmonary systems, and disparities in intramuscular metabolic perturbation often evident during whole body exercise in hypoxia or hypercapnia. We hypothesized that the MR:CR interaction would potentiate the exercise-induced ventilatory responses.

METHODS

Subjects.

All experimental protocols were approved by the Institutional Review Board of the University of Utah and by the Salt Lake City Veterans Affairs Medical Center. Following written and verbal explanations of all procedures and risks, written informed consent was obtained from each subject before enrollment in the study. Eleven recreationally active healthy volunteers participated in the study (5 females and 6 males; age: 26 ± 3 yr; body mass: 75 ± 15 kg; height: 176 ± 12 cm). The participants were nonsmokers, nonmedicated, and asymptomatic for cardiovascular or respiratory disorders. All participants refrained from exercise for 24 h and from caffeine and alcohol for 12 h before each study visit. Female participants were premenopausal and studied during the early follicular phase of the menstrual cycle, which was confirmed by their circulatory sex hormone levels obtained during each experimental trial (Table 1). Data collected in these subjects during this study, focused on the exercise pressor reflex and CR interaction and the cardiovascular implications, have been published previously (58).

Table 1.

Circulating levels of sex hormones in female subjects during the study trials

	Estrogen, pg/mL	Progesterone, ng/mL	FSH, mIU/mL	LH, mIU/mL
Control	133 ± 73	0.1 ± 0.1	4.7 ± 2.5	5.3 ± 5.9
Fentanyl	107 ± 53	0.8 ± 1.2	3.0 ± 3.0	5.2 ± 6.6

Values are means ± SD. Data were assessed from arterial blood samples; n = 5. Control, trial with intact leg muscle afferent feedback; fentanyl, trial with attenuated leg muscle afferent feedback using fentanyl blockade; FSH, follicle-stimulating hormone; LH, luteinizing hormone. No differences between the trials, P ≥ 0.31.

Experimental design.

Participants initially performed an incremental single-leg knee-extensor test (10 + 10 W/min) to task failure to determine their peak work rate (W_peak, 55 ± 15 W) and thereafter were familiarized with all noninvasive study procedures during an independent practice session. Subsequently, on two separate occasions, participants completed a control trial (Ctrl) and an identical trial during which feedback from µ-opioid receptor-sensitive muscle afferents was blocked via lumbar intrathecal fentanyl (Fent). All visits were separated by ≥48 h and conducted at the same time of day. The participants performed three bouts [4 min each, 60 revolutions/min (rpm)] of dynamic knee-extension exercise with the right leg at 60% of their W_peak (33 ± 9 W). The three bouts of exercise in each of the two experimental trials (Ctrl and Fent) were conducted in ambient room air, i.e., normoxia (Norm_Ctrl and Norm_Fent), normocapnic hypoxia (Hypo_Ctrl and Hypo_Fent), and normoxic hypercapnia (Hyper_Ctrl and Hyper_Fent) conditions (Fig. 1). The exercise bouts were separated by a 6-min break breathing room air. The experimental trials and conditions were initially randomized and then counterbalanced.

Fig. 1.

Schematic illustration of the individual and the concurrent activation of the muscle reflex (MR) and the chemoreflex (CR) during exercise. Each of the two experimental trials (control and fentanyl) included 3 conditions (normoxia, normocapnic hypoxia, and normoxic hypercapnia). This design resulted in 6 bouts of single-leg knee-extension exercise, performed at a constant workload. Norm_Ctrl and Norm_Fent, control and fentanyl trials conducted in ambient room air, i.e., normoxia; Hypo_Ctrl and Hypo_Fent, control and fentanyl trials conducted under normocapnic hypoxia conditions; Hyper_Ctrl and Hyper_Fent, control and fentanyl trials conducted under normoxic hypercapnia conditions. The straight arrows denote the comparisons used to estimate the ventilatory effects of activation of the MR (black arrow), activation of the CR (light gray arrow), and coactivation of the MR and CR (dark gray arrow).

Intrathecal fentanyl injection.

Participants were seated in a flexed position, and 0.5 mL of an anesthetic solution containing 0.05 mg/mL fentanyl was delivered intrathecally at the vertebral interspace between L3 and L4 (3). To assess whether fentanyl migrated beyond the cervical level, the cardiorespiratory response to an arm-cranking test (15 and 30 W, 3 min each, 60 rpm; Monark-Crescent, Varberg, Sweden) was measured before the fentanyl administration and again after the exercise bouts were completed (1). Heart rate (HR) was monitored using a three-lead electrocardiogram (CardioCard; Nasiff Associates, Central Square, NY). The arm-cranking test was performed in both Ctrl and Fent to ensure a similar protocol across study trials. The time between fentanyl injection and the first exercise bout was between 15 and 25 min, and all participants completed the entire protocol within 75 min after the injection of fentanyl.

Ventilatory responses to exercise in normoxia, normocapnic hypoxia, and normoxic hypercapnia.

Ventilatory responses, i.e., minute ventilation (V̇e), breathing frequency (f_B) and tidal volume (V_T), and pulmonary gas exchange, i.e., end-tidal partial pressure of O₂ ($\text{P}\mathrm{E}{\mathrm{T}}_{{\text{O}}_2}$) and CO₂ ($\text{P}\mathrm{E}{\mathrm{T}}_{\text{C}{\text{O}}_2}$), O₂ consumption, and CO₂ production, were measured breath-by-breath using an open-circuit calorimetry system (Innocor; Innovision, Glamsbjerg, Denmark). Oxyhemoglobin saturation ($\text{S}{\text{p}}_{{\text{O}}_2}$) was estimated using pulse oximetry (OxiMax N-600x; Nellcor, Minneapolis, MN) with a forehead sensor. To induce normocapnic hypoxia, participants inspired from a Douglas bag containing 8% O₂, 3%–4% CO₂, and balance N₂ to achieve a decrease in $\text{S}{\text{p}}_{{\text{O}}_2}$ to ~80%. $\text{S}{\text{p}}_{{\text{O}}_2}$ and $\text{P}\mathrm{E}{\mathrm{T}}_{\text{C}{\text{O}}_2}$ were then held constant by titrating pure O₂ or CO₂ in the inspirate. For the normoxic hypercapnia conditions, an increase in $\text{P}\mathrm{E}{\mathrm{T}}_{\text{C}{\text{O}}_2}$ to ~50 mmHg was induced by inspiring a gas mixture of 6%–7% CO₂, 21% O₂, and balance N₂; $\text{P}\mathrm{E}{\mathrm{T}}_{\text{C}{\text{O}}_2}$ was maintained by titrating pure CO₂ in the inspirate when needed. After the target level for hypoxemia or hypercapnia was reached (<45 s), participants sat quietly during a 2-min lead-in phase preceding the exercise bout. During each exercise bout, arterial blood samples were collected anaerobically via a radial artery catheter and assessed in a cooximeter and blood gas analyzer (GEM 4000; Instrumentation Laboratory, Bedford, MA).

Data processing.

For all respiratory responses, averages were calculated from every minute during each exercise bout. Based on the assumption that normoxic exercise performed with fentanyl blockade is characterized by minimal MR and CR activity, the ventilatory response during Norm_Fent was considered as the baseline. The individual and interactive effects of the reflexes on the ventilatory response during the final minute of exercise were defined as follows (Fig. 1): 1) the individual effects of the MR, the O₂-CR, and the CO₂-CR were estimated by the differences between Norm_Ctrl and Norm_Fent (△Norm_Ctrl-Norm_Fent), between Hypo_Fent and Norm_Fent (△Hypo_Fent-Norm_Fent), and between Hyper_Fent and Norm_Fent (△Hyper_Fent-Norm_Fent), respectively, where a difference = 0 indicates no effect resulting from the reflex activation; 2) the interactive effects of the MR and O₂-CR and of the MR and CO₂-CR during the reflex coactivation were calculated as the differences between Hypo_Ctrl and Norm_Fent (△Hypo_Ctrl-Norm_Fent) and between Hyper_Ctrl and Norm_Fent (△Hyper_Ctrl-Norm_Fent), respectively, where a difference = 0 indicates no effect resulting from the reflex coactivation; 3) to investigate the mode of interaction between the reflexes, the observed ventilatory changes during the coactivation of the MR and the CR (O₂-CR or CO₂-CR) were compared with the arithmetic sum of the changes elicited by each reflex alone (△Hypo_Ctrl-Norm_Fent versus △Norm_Ctrl-Norm_Fent + △Hypo_Fent-Norm_Fent or △Hyper_Ctrl-Norm_Fent versus △Norm_Ctrl-Norm_Fent + △Hyper_Fent-Norm_Fent). The definition of the interaction mode has been previously described (53, 58). Briefly, the interaction was deemed hyperadditive if the response observed during the coactivation of the reflexes was greater than the summation of the responses evoked by each reflex alone, reflecting that the interactive effect was synergistic, while the interaction was deemed additive if the observed response was equal to the summated response, indicating that the interactive effect was a simple addition of the individual effects of the reflexes.

Statistical analysis.

Data were analyzed using statistical analysis software (SPSS 22; IBM, Armonk, NY). Descriptive statistics were used for the description of subject characteristics. The Mauchly’s test was used to determine sphericity, and the Geiser-Greenhouse correction was applied when sphericity was violated. The Student’s paired t-test was used to detect the difference in female sex hormone levels between the two experimental trials as well as the effect of fentanyl blockade on cardiorespiratory responses during the arm-cranking test. Two-way ANOVA (trial × time) with repeated measures was used to examine temporal changes in the gas exchange and ventilatory responses across the 4 min of exercise under each condition: normoxia, normocapnic hypoxia, and normoxic hypercapnia. Because this study aimed to activate the CR with hypoxia or hypercapnia and to investigate its interaction with the MR, data obtained from the hypoxic and hypercapnic conditions were also separately compared with those from the normoxic conditions. Thus, two-way ANOVA (trial × condition) with repeated measures was used to test alterations in arterial blood chemistry and ventilatory responses during the final minute of exercise. If the ANOVA detected a significant main effect or interaction, the Tukey’s post hoc analysis was performed to identify the differences. To directly test our hypotheses, a priori planned comparisons were made to determine the individual/interactive effects and the interaction mode of the reflexes on the ventilatory changes during exercise, using the Holm-Bonferroni method to correct for familywise error of multiple comparisons. The 95% confidence interval (95% CI) was also calculated for the difference between the observed and summated responses. All data are presented as means ± SD, and statistical significance was set at P < 0.05.

RESULTS

Cardiorespiratory responses to arm exercise.

Lumbar intrathecal fentanyl did not affect the cardiorespiratory responses to arm cranking (pre- vs. postinjection at 15 W/30 W; HR: 110 ± 20 versus 111 ± 20/121 ± 22 versus 122 ± 23 beats/min; f_B: 23 ± 5 versus 22 ± 5/25 ± 4 versus 24 ± 6 breaths/min; V_T: 1.1 ± 0.2 versus 1.1 ± 0.2/1.3 ± 0.3 versus 1.3 ± 0.2 L; ventilatory equivalent for CO₂: 35 ± 3 versus 37 ± 3/34 ± 3 versus 36 ± 4; all P ≥ 0.13). This reflects the lack of significant cephalad migration and direct binding of fentanyl to cerebral µ-opioid receptors (36).

Gas exchange and ventilatory responses during exercise.

During the 4 min of exercise in normoxia, $\text{S}{\text{p}}_{{\text{O}}_2}$ was not different between Ctrl and Fent (Fig. 2A) while $\text{P}\mathrm{E}{\mathrm{T}}_{{\text{O}}_2}$ was significantly higher and $\text{P}\mathrm{E}{\mathrm{T}}_{\text{C}{\text{O}}_2}$ significantly lower in Ctrl than in Fent (Fig. 2, B and C; significant main effects of trial). Hypoxia and hypercapnia altered blood gases during exercise under each condition, with no significant differences between Ctrl and Fent in terms of $\text{S}{\text{p}}_{{\text{O}}_2}$, $\text{P}\mathrm{E}{\mathrm{T}}_{{\text{O}}_2}$, and $\text{P}\mathrm{E}{\mathrm{T}}_{\text{C}{\text{O}}_2}$ (Fig. 2, D–I).

Fig. 2.

Oxyhemoglobin saturation and gas exchange during exercise. Exercise was performed with intact (control) and blocked (fentanyl) leg muscle afferent feedback in nomoxia, normocapnic hypoxia, and normoxic hypercapnia conditions. $\text{S}{\text{p}}_{{\text{O}}_2}$, estimated oxyhemoglobin saturation; $\text{P}\mathrm{E}{\mathrm{T}}_{{\text{O}}_2}$, end-tidal partial pressure of O₂; $\text{P}\mathrm{E}{\mathrm{T}}_{\text{C}{\text{O}}_2}$, end-tidal partial pressure of CO₂.

The ventilatory changes in response to the 4 min of exercise are illustrated in Fig. 3. During the normoxic exercise, f_B was significantly decreased in Fent compared with Ctrl (Fig. 3B; significant main effect of trial) while V̇e and V_T were similar between the trials (Fig. 3, A and C). During the hypoxic exercise, V̇e and f_B were significantly lower in Fent compared with Ctrl (Fig. 3, D and E; significant main effects of trial) while V_T was not different between the trials (Fig. 3F). Moreover, there was a significant interaction (trial × time) in terms of both V̇e and V_T during the hypoxic exercise. During the hypercapnic exercise, V̇e was significantly lower in Fent compared with Ctrl (Fig. 3G; significant main effect of trial) while f_B and V_T were not different between the trials (Fig. 3, H and I).

Fig. 3.

Ventilatory responses during exercise. Exercise was performed with intact (control) and blocked (fentanyl) leg muscle afferent feedback in nomoxia, normocapnic hypoxia, and normoxic hypercapnia conditions. V̇e, minute ventilation; f_B, breathing frequency; V_T, tidal volume. *Significantly different from the fentanyl trial, P < 0.05.

Individual and interactive ventilatory effects of the MR, the O₂-CR, and the CO₂-CR.

The arterial blood chemistry and ventilatory data collected during the final minute of exercise in normoxia, hypoxia, and hypercapnia are documented in Tables 2 and 3. Figure 4 presents the separate and combined ventilatory effects triggered by the activation of the MR and the O₂-CR during exercise. Individual activation of the MR (i.e., △Norm_Ctrl-Norm_Fent) had no significant effect on V̇e, f_B, and V_T. Individual activation of the O₂-CR (i.e., △Hypo_Fent-Norm_Fent) significantly increased V̇e, f_B, and V_T during exercise. When both reflexes were activated concurrently (i.e., △Hypo_Ctrl-Norm_Fent), V̇e, f_B, and V_T significantly increased. As for the mode of interaction between the MR and the O₂-CR (i.e., △Hypo_Ctrl-Norm_Fent versus △Norm_Ctrl-Norm_Fent + △Hypo_Fent-Norm_Fent), the V̇e and V_T responses were significantly greater during the coactivation of the two reflexes compared with the summation of the responses elicited by each reflex alone (95% CI: V̇e = 10.3 ± 6.2 L/min; V_T = 0.3 ± 0.2 L), reflecting a hyperadditive interaction. By contrast, the f_B response to the coactivation of the MR and the O₂-CR was not different from the summated response (P = 0.37; 95% CI = −0.5 ± 2.8 breaths/min), indicating an additive interaction.

Table 2.

Arterial blood chemistry and ventilatory responses during the final minute of normoxic and hypoxic exercise

	Normoxia		Normocapnic Hypoxia		P Value
	Control	Fentanyl	Control	Fentanyl	Main effect of fentanyl	Main effect of hypoxia	Interaction effect
$\text{S}{\text{a}}_{{\text{O}}_2}$, %	97 ± 0	96 ± 2	79 ± 2†	79 ± 1†	0.12	<0.01	0.23
$\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, mmHg	84 ± 4	81 ± 3*	43 ± 3†	44 ± 5†	0.03	<0.01	<0.01
$\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$, mmHg	32 ± 3	33 ± 2*	33 ± 3†	33 ± 2	0.05	0.23	0.02
Arterial pH	7.39 ± 0.03	7.37 ± 0.04	7.39 ± 0.03	7.37 ± 0.04	0.03	0.49	0.43
Arterial K+, mmol/L	3.9 ± 0.3	4.0 ± 0.3	3.9 ± 0.2	3.9 ± 0.2	0.46	0.26	0.23
Arterial lactate, mmol/L	3.0 ± 1.2	3.3 ± 1.1	3.0 ± 1.0	3.3 ± 0.8	0.04	1.00	0.63
V̇e, L/min	37 ± 6	35 ± 6	80 ± 15†	69 ± 14*†	0.02	<0.01	<0.01
fB, breaths/min	31 ± 5	29 ± 7	38 ± 6†	36 ± 7†	0.09	<0.01	0.75
VT, L	1.2 ± 0.3	1.3 ± 0.3	2.2 ± 0.6†	2.0 ± 0.6*†	0.09	<0.01	<0.01
V̇o2, L/min	1.2 ± 0.2	1.3 ± 0.2
V̇co2, L/min	1.1 ± 0.2	1.2 ± 0.2
V̇e/V̇o2	30 ± 4	27 ± 2*
V̇e/V̇co2	32 ± 3	31 ± 2*

Values are means ± SD. Exercise was performed at 60% of peak work rate (33 ± 9 W); n = 11. Control, exercise performed with intact leg muscle afferent feedback; fentanyl, exercise performed with attenuated leg muscle afferent feedback using fentanyl blockade; $\text{S}{\text{a}}_{{\text{O}}_2}$, arterial oxyhemoglobin saturation; $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, arterial partial pressure of O₂; $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$, arterial partial pressure of CO₂; V̇e, minute ventilation; f_B, breathing frequency; V_T, tidal volume; V̇o₂, O₂ consumption; V̇co₂, CO₂ production; V̇e/V̇o₂, ventilatory equivalent for O₂; V̇e/V̇co₂, ventilatory equivalent for CO₂.

^*P < 0.05 vs. control.

^†P < 0.05 vs. normoxia. Note, V̇o₂ and V̇co₂ were unable to be obtained in the hypoxic conditions because of gas titration and were analyzed using paired t-test; otherwise, all data were analyzed using 2 × 2 repeated-measure ANOVA and Tukey’s post hoc test.

Table 3.

Arterial blood chemistry and ventilatory responses during the final minute of normoxic and hypercapnic exercise

	Normoxia		Normoxic Hypercapnia		P Value
	Control	Fentanyl	Control	Fentanyl	Main effect of fentanyl	Main effect of hypercapnia	Interaction effect
$\text{S}{\text{a}}_{{\text{O}}_2}$, %	97 ± 0	96 ± 2	98 ± 1†	97 ± 2*	0.05	0.04	0.78
$\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, mmHg	84 ± 4	81 ± 3*	105 ± 4†	104 ± 7†	0.01	<0.01	0.01
$\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$, mmHg	32 ± 3	33 ± 2*	50 ± 2†	50 ± 1†	0.01	<0.01	0.02
Arterial p	7.39 ± 0.03	7.37 ± 0.04	7.26 ± 0.03†	7.26 ± 0.03†	0.37	<0.01	0.06
Arterial K+, mmol/L	3.9 ± 0.3	4.0 ± 0.3	4.0 ± 0.4	3.8 ± 0.4†	0.61	0.18	0.02
Arterial lactate, mmol/L	3.0 ± 1.2	3.3 ± 1.1	2.4 ± 0.6	2.8 ± 0.9	0.14	0.10	0.96
V̇e, L/min	37 ± 6	35 ± 6	91 ± 20†	85 ± 22*†	0.05	<0.01	0.13
fB, breaths/min	31 ± 5	29 ± 7	36 ± 7	36 ± 10†	0.55	0.03	0.57
VT, L	1.2 ± 0.3	1.3 ± 0.3	2.6 ± 0.8†	2.4 ± 0.5†	0.44	<0.01	0.26
V̇o2, L/min	1.2 ± 0.2	1.3 ± 0.2
V̇co2, L/min	1.1 ± 0.2	1.2 ± 0.2
V̇e/V̇o2	30 ± 4	27 ± 2*
V̇e/V̇co2	32 ± 3	31 ± 2*

Values are means ± SD. Exercise was performed at 60% of peak work rate (33 ± 9 W); n = 11. Control, exercise performed with intact leg muscle afferent feedback; fentanyl, exercise performed with attenuated leg muscle afferent feedback using fentanyl blockade; $\text{S}{\text{a}}_{{\text{O}}_2}$, arterial oxyhemoglobin saturation; $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, arterial partial pressure of O₂; $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$, arterial partial pressure of CO₂; V̇e, minute ventilation; f_B, breathing frequency; V_T, tidal volume; V̇o₂, O₂ consumption; V̇co₂, CO₂ production; V̇e/V̇o₂, ventilatory equivalent for O₂; V̇e/V̇co₂, ventilatory equivalent for CO₂.

^*P < 0.05 vs. control.

^†P < 0.05 vs. normoxia. Note, V̇o₂ and V̇co₂ were unable to be obtained in the hypercapnic conditions because of gas titration and were analyzed using paired t-test; otherwise, all data were analyzed using 2 × 2 repeated-measure ANOVA and Tukey’s post hoc test.

Fig. 4.

Ventilatory changes during individual and concurrent activation of the muscle reflex (MR) and the hypoxia-induced chemoreflex (O₂-CR). MR & O₂-CR, observed changes during coactivation of the MR and O₂-CR; MR + O₂-CR, sum of the changes induced by each reflex alone; V̇e, minute ventilation; f_B, breathing frequency; V_T, tidal volume. Individual subject data are shown for females (○ and broken lines) and males (△ and solid lines). *Significantly different from zero, P < 0.05; †significantly different between MR & O₂-CR and MR + O₂-CR, P < 0.05.

Figure 5 exhibits the separate and combined ventilatory effects triggered by the activation of the MR and the CO₂-CR during exercise. Individual activation of the CO₂-CR (i.e., △Hyper_Fent-Norm_Fent) significantly increased V̇e and V_T during exercise while f_B was unaffected. When the MR and the CO₂-CR were activated simultaneously (i.e., △Hyper_Ctrl-Norm_Fent), V̇e, f_B, and V_T significantly increased. The mode of interaction between the MR and the CO₂-CR (i.e., △Hyper_Ctrl-Norm_Fent versus △Norm_Ctrl-Norm_Fent + △Hyper_Fent-Norm_Fent) was additive for all variables, since the observed and the summated changes were similar (all P ≥ 0.13; 95% CI: V̇e = 5.5 ± 6.4 L/min; f_B = −1.9 ± 6.3 breaths/min; V_T = 0.3 ± 0.4 L).

Fig. 5.

Ventilatory changes during individual and concurrent activation of the muscle reflex (MR) and the hypercapnia-induced chemoreflex (CO₂-CR). MR & CO₂-CR, observed changes during coactivation of the MR and CO₂-CR; MR + CO₂-CR, sum of the changes induced by each reflex alone; V̇e, minute ventilation; f_B, breathing frequency; V_T, tidal volume. Individual subject data are shown for females (○ and broken lines) and males (△ and solid lines). *Significantly different from zero, P < 0.05. No differences between MR & CO₂-CR and MR + CO₂-CR.

DISCUSSION

The goal of this study was to examine the individual and interactive influences of the muscle reflex and the chemoreflex on the ventilatory response to exercise, and to evaluate the mode of interaction between these reflexes. Muscle afferent blockade and hypoxia/hypercapnia were employed to manipulate the reflexes individually and in parallel. The results of these experiments suggest that the MR:O₂-CR interaction is hyperadditive, meaning that, mainly through augmenting tidal volume, the ventilatory responses to exercise were larger than expected based on the summation of each reflex alone. When the chemoreflex is activated via hypercapnia, the interactive effect of the muscle reflex and CO₂-CR was simply additive, that is, equaling the sum of the effects of each reflex alone. Taken together, this study documents that the interaction effect and associated ventilatory consequences of MR:CR coactivation are distinct when chemoreflex activation is achieved via hypoxia versus hypercapnia, revealing that the MR:CR interaction further amplifies the ventilatory response to exercise in hypoxia. Because the MR:CR coactivation in hypoxia has a synergistic effect on the ventilatory response to exercise, it is likely of particular importance for attempting to maintain arterial oxygenation in the physically active human at high altitude.

Ventilatory effects of the individual activation of the muscle reflex and the chemoreflex.

To assess the effect of individual activation of the muscle reflex, the ventilatory responses to exercise in the normoxic control and fentanyl trials were compared. Although the suppressing effect of the fentanyl blockade did not attain statistical significance in terms of minute ventilation (P = 0.08, Fig. 3A), the attenuation of the muscle reflex significantly compromised the ventilatory equivalents for O₂ and CO₂ (Table 2), and this alveolar hypoventilation resulted in consistent CO₂ retention and reductions in arterial and end-tidal Po₂ during exercise (Table 2 and Fig. 2, B and C). While these findings corroborate previous studies using the same exercise modality (4) and highlight the role of the muscle reflex in regulating the ventilatory response to single-leg knee-extension exercise, the observed ventilatory effect of the fentanyl blockade is of a much smaller magnitude than that documented during cycling exercise (1–3, 27, 47). The less significant impact of the fentanyl blockade on the ventilatory response to knee-extensor, compared with cycling, exercise is likely attributed to the smaller active leg muscle mass recruited during this exercise modality and, in turn, the lower group III/IV-mediated afferent feedback from the leg (22, 32, 43, 50).

It should be noted that intrathecal fentanyl exclusively affects μ-opioid receptor-sensitive muscle afferents and only blocks ~60% of group III/IV-mediated feedback (31). Although both group III and IV muscle afferents contain μ-opioid receptors, the exact distribution remains unknown, despite some findings indirectly suggesting predominance of δ-opioid receptors on the spinal endings of thinly myelinated pain fibers (51). Regardless, δ-opioid receptor-sensitive group III/IV muscle afferent neurons were unaffected in the present study and continued to discharge in response to muscle contractions (35), which caused an underestimation of the ventilatory impact of the muscle reflex. In addition, the ventilatory effect of the fentanyl blockade might also be underrated, due in part to the influence of attenuating leg muscle afferent feedback on the blood pressure response to exercise. Specifically, decreases in arterial blood pressure can increase minute ventilation (while increases in pressure decrease ventilation), a phenomenon known as the “ventilatory baroreflex” (12, 44, 56). Because the fentanyl blockade indeed effectively curtailed arterial blood pressure during the current knee-extension exercise (i.e., from ~108 to ~99 mmHg; data shown in Ref. 58), the ventilatory component of the arterial baroreflex could have elevated breathing, and this might have partially masked the fentanyl-induced reduction in minute ventilation during exercise. However, this influence was presumably small, since the impact of the ventilatory baroreflex is limited at rest (12) and becomes negligible above a mean arterial blood pressure of ~90 mmHg (56).

Finally, to assess the individual activation of the chemoreflex, hypoxia or hypercapnia was imposed during exercise with attenuated muscle afferent activity (i.e., little muscle reflex activation). As expected (11, 20, 26, 40, 53), the ventilatory responses increased remarkably (light gray bars in Figs. 4 and 5).

Ventilatory effect of the MR:CR coactivation.

To evaluate the effect of the MR:CR coactivation, the ventilatory response during exercise performed with high muscle reflex and chemoreflex activity (i.e., hypoxic/hypercapnic exercise with intact muscle afferent feedback) was compared with the ventilatory response during exercise performed with low activity of both reflexes (i.e., normoxic exercise with attenuated muscle afferent feedback). The differences in the ventilatory response between the two conditions were profound and indicate that the MR:CR coactivation further increases minute ventilation, breathing frequency, and tidal volume during exercise (dark gray bars in Figs. 4 and 5). In effect, the observed ventilatory changes are not surprising and attributable to the combined afferent inputs from both group III/IV muscle afferents and chemoreceptors to the pontomedullary respiratory center and the resultant augmentation in ventilatory neural drive (29, 39, 49). However, accumulating evidence has pointed to a potential interaction between the neural mechanisms (10), and it has yet to be determined whether these ventilatory changes are a simple addition of the effects of the muscle reflex activation and an increased chemoreflex, or are a consequence of the interactive effect of coactivating the two reflexes. To address this issue, as discussed below, we attempted to unravel the muscle reflex and the chemoreflex interaction by comparing the ventilatory response during the coactivation of these reflexes with the mathematical sum of the discrete responses evoked by the stimulation of each reflex.

Mode of the MR:CR interaction.

The coactivation of the muscle reflex and the O₂-CR resulted in various modes of interaction with corresponding consequences for the ventilatory response to exercise. Specifically, the MR:O₂-CR coactivation evoked a hyperadditive effect upon minute ventilation and tidal volume and an additive effect upon breathing frequency (Fig. 4). These findings corroborate an earlier investigation using circulatory occlusion during handgrip exercise in normoxia and hypoxia (26), a strategy used to establish similar metabolic milieus in the working muscle and assure a comparable engagement of central command and the muscle reflex in both gas conditions. The current observations are also in agreement with a study using passive exercise, an approach that does not engage central command and the metaboreflex. These authors found that the increase in minute ventilation during passive leg movement in hypoxia was greater than the summation of the minute ventilation increases elicited by leg movement and hypoxia alone and concluded that the interaction between the muscle mechanoreflex and the O₂-CR was hyperadditive (53).

By contrast, a study using postexercise circulatory occlusion following handgrip exercise in hypoxia, an approach that does not engage central command and the mechanoreflex, demonstrated that the interaction effect resulting from the coactivation of the muscle metaboreflex and the O₂-CR on minute ventilation was simply additive (20). Because the muscle mechanoreflex and the muscle metaboreflex are two major components of the muscle reflex, it is tempting to combine the outcome of those studies and postulate that the hyperadditive effect of MR:O₂-CR coactivation in the present study is primarily driven by the activation of the muscle mechanoreflex. Nevertheless, a direct comparison between the current findings and those previous observations is rather problematic, since the current experiments, utilizing voluntary exercise, included central command, whereas the aforementioned studies were conducted in the absence of central command. This poses a critical discrepancy, since it has recently been documented that the ventilatory effect of the muscle metaboreflex could only be revealed in the presence of central command (37). Thus, it is plausible that the muscle mechanoreflex, the muscle metaboreflex, or both may have contributed to the hyperadditive effect of the MR:O₂-CR coactivation observed in the present study.

In contrast to the findings from the MR:O₂-CR trial (i.e., hyperadditive interaction), the coactivation of the muscle reflex and the CO₂-CR resulted in an additive interaction for all ventilatory variables (Fig. 5). This difference is in correspondence with prior investigations which consistently documented that the gain of the hypoxic ventilatory response increases from mild to moderate exercise, i.e., enhanced chemosensitivity (17, 41, 59), while the gain of the hypercapnic ventilatory response remains unchanged with increases in exercise intensity (14, 15, 59). The current data suggest that the distinction of the relationship between minute ventilation and exercise intensity in hypoxia and hypercapnia may, at least in part, be a consequence of the different modes of MR:CR interaction.

The specificity of the gases used to activate the chemoreceptors in the present study likely contributed to the different interaction modes in MR:O₂-CR compared with MR:CO₂-CR. In particular, whereas normocapnic hypoxia predominately activates peripheral chemoreceptors, normoxic hypercapnia stimulates both central and, albeit to a lesser extent, peripheral chemoreceptors (25). Therefore, the interaction between the peripheral chemoreflex and the central chemoreflex may have further modulated the ventilatory response during the hypercapnic exercise (9, 28, 46, 54, 55) and, thus, may have led to the difference in the interaction mode between MR:O₂-CR and MR:CO₂-CR. Second, the excitatory inputs from peripheral and central chemoreceptors to the respiratory center in the brain stem are conveyed through divergent neuronal circuits and can be modified by other influences via synaptic interconnectivities (28, 29, 39). Third, the ventilatory effect of O₂-CR can be potentiated by a muscle reflex-mediated increase in sympathetic outflow to the carotid bodies (7, 17) while the impact of an increase in sympathetic outflow to the central chemoreceptors is not clear. Consequently, chemoreflex-related neural inputs to the medulla may have been quite different in hypoxia compared with hypercapnia, potentially altering the central integration of the feedback from the muscle afferents and chemoreceptors, and, ultimately, resulting in differential modes of interaction.

Conclusion and significance.

This study documents that the interaction effect and associated ventilatory consequences of MR:CR coactivation are distinct when chemoreflex activation is achieved via hypoxia versus hypercapnia, revealing that the MR:CR interaction further amplifies the ventilatory response to exercise in hypoxia. Because the MR:CR coactivation in hypoxia has a synergistic effect on the ventilatory response to exercise, it is, therefore, likely of particular importance for attempting to maintain arterial oxygenation in the physically active human at high altitude. The ramifications of this reflex synergy are also relevant for clinical populations characterized by abnormal muscle reflexes and chemoreflexes (e.g., hypertension, heart failure) and offer a potential explanation for the patients’ aberrant ventilatory control and heightened exertional dyspnea.

GRANTS

This study was supported by the National Heart, Lung, and Blood Institute (HL-116579, HL-139451, and HL-091830) and the U.S. Department of Veterans Affairs (E6910-R, E1697-R, E1433-P, E9275-L, and E1572-P).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

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