Exaggerated sympathetic and cardiovascular responses to dynamic mechanoreflex activation in rats with heart failure: role of endoperoxide 4 and thromboxane A₂ receptors

Abstract

The primary purpose of this investigation was to determine the role played by endoperoxide 4 receptors (EP4-R) and thromboxane A₂ receptors (TxA₂-R) during isolated dynamic muscle mechanoreflex activation in rats with heart failure with reduced ejection fraction (HF-rEF) and sham-operated healthy controls. We found that injection of the EP4-R antagonist L-161,982 (1μg) into the arterial supply of the hindlimb had no effect on the peak pressor response to dynamic hindlimb muscle stretch in HF-rEF (n=6, peak ΔMAP pre: 27±7; post: 27±4 mmHg; P=0.99) or sham (n=6, peak ΔMAP pre: 15±3; post: 13±3 mmHg; P=0.67) rats. In contrast, injection of the TxA₂-R antagonist daltroban (80μg) into the arterial supply of the hindlimb reduced the pressor response to dynamic hindlimb muscle stretch in HF-rEF (n=11, peak ΔMAP pre: 28±4; post: 16±2 mmHg; P=0.02) but not sham (n=8, peak ΔMAP pre: 17±3; post: 16±3; P=0.84) rats. Our data suggest that TxA₂-Rs on thin fibre muscle afferents contribute to the exaggerated mechanoreflex in HF-rEF.

Introduction

Sympathetic nervous system activity (SNA) is augmented at rest (Leimbach et al., 1986), and during exercise (Murai et al., 2009; Notarius et al., 1999) in patients with heart failure with reduced ejection fraction (HF-rEF) compared to healthy control subjects. Exaggerated SNA in HF-rEF patients is linked to decreased exercise tolerance (Notarius et al., 2019; Piepoli et al., 1996) and increased mortality (Barretto et al., 2009). The lack of tolerance for exercise and activities of daily living reduces functional independence and quality of life, and is a significant driver of the escalated cost of care for heart failure patients (Borlaug, 2014). Thus, advancing our understanding of the exaggerated SNA in the over 26 million people worldwide with HF-rEF (Ponikowski et al., 2014) is fundamentally important towards improving quality of life for this patient population.

The exercise pressor reflex is one mechanism that contributes to the heightened SNA during exercise in HF-rEF patients (Sinoway et al., 2005). This reflex is activated when the sensory endings of group III and IV thin fibre muscle afferents are stimulated by mechanical and/or metabolic signals associated with skeletal muscle contraction (Kaufman et al., 1982; Kaufman et al., 1983; Kaufman et al., 1984; McCloskey et al., 1972). A substantial body of evidence suggests that this reflex is exaggerated in patients with HF-rEF compared to the reflex found in healthy control subjects (Murphy et al., 2011; Piepoli et al., 2014). The exaggerated exercise pressor reflex in HF-rEF contributes to an impairment in oxygen delivery to contracting skeletal muscles (Amann et al., 2014; Ives et al., 2016; Kaur et al., 2018; O’Leary et al., 2004; Smith et al., 2020b), which contributes to reductions in exercise capacity in this patient population (Smith et al., 2020b). Thus, there is pressing need to investigate the mechanisms of exaggerated exercise pressor reflex activation in HF-rEF patients.

The mechanically sensitive portion of the exercise pressor reflex (i.e., the mechanoreflex) contributes importantly to the overall reflex exaggeration in HF-rEF patients (Antunes-Correa et al., 2014; Middlekauff et al., 2001) and in animals with HF-rEF (Koba et al., 2008; Li et al., 2004; Morales et al., 2012; Smith et al., 2005). Importantly, in rats with HF-rEF compared to healthy counterparts, Wang et al. (2010) found greater activation of primarily mechanically sensitive group III muscle afferents in response to static hindlimb muscle contraction and static hindlimb muscle stretch (a model of static mechanoreflex activation isolated from contraction-induced metabolic stimuli; Stebbins et al., 1988). The augmented thin fibre muscle afferent responsiveness to the contraction and stretch protocols were likely attributable, at least in part, to a chronic sensitization of mechanically activated channels on the sensory endings of those afferents produced by cyclooxygenase (COX) products of arachidonic metabolism (Antunes-Correa et al., 2014; Middlekauff et al., 2008; Morales et al., 2012; Smith et al., 2020a). For example, Morales et al. (2012) showed that inhibition of the COX-2 isoform, but not the COX-1 isoform, within hindlimb skeletal muscles of rats with HF-rEF reduced the increase in renal SNA (RSNA) and blood pressure evoked during static hindlimb skeletal muscle stretch. Conversely, no effect of COX-1 or COX-2 inhibition on the increase in RSNA and blood pressure in response to static stretch was found in healthy control rats (Morales et al., 2012). We recently attempted to build on those findings by investigating in HF-rEF rats the role played by two major receptors for products of COX metabolism present on sensory neurons, namely endoperoxide 4 receptors (EP4-Rs) and thromboxane A₂ receptors (TxA₂-Rs), in the mechanoreflex sensitization (Butenas et al., 2020). However, we found that neither injection of an EP4-R antagonist nor a TxA₂-R antagonist into the arterial supply of the hindlimb of HF-rEF rats had an effect on the reflex increase in RSNA and blood pressure evoked during static hindlimb skeletal muscle stretch (Butenas et al., 2020).

Static stretch of rat hindlimb skeletal muscles serves as a valuable model of mechanoreflex activation isolated from the metabolic stimuli present during static contractions. However, static stretch does not reproduce the pattern of rhythmic mechanical stimuli present during dynamic contractions such as those that produce locomotion (e.g., walking, climbing stairs, etc.). Moreover, static versus dynamic mechanoreflex activation may stimulate substantially different classes of mechanically activated channels (Nakamoto et al., 2007; Sanderson et al., 2019) which may exhibit different inactivation or adaptation kinetics (Drew et al., 2002; Lewin et al., 2004; Wood et al., 2012) and which may be governed by different intracellular signaling pathways. To address these issues, our laboratory has used a 1 Hz dynamic rat hindlimb skeletal muscle stretch model of isolated mechanoreflex activation which closely replicates the rhythmic nature of skeletal muscle contraction during locomotory exercise (Butenas et al., 2019; Kempf et al., 2018; Rollins et al., 2020; Rollins et al., 2019; Sanderson et al., 2019). Whether this dynamic hindlimb skeletal muscle stretch model of mechanoreflex activation results in greater reflex sympathetic and/or cardiovascular responses in rats with HF-rEF compared to healthy control rats is unknown. Additionally, whether the reflex sympathetic and cardiovascular responses to isolated dynamic mechanoreflex activation in HF-rEF rats are governed by the intracellular pathways linked to EP4-Rs and/or TxA₂-Rs, is unknown.

Based on the information above, the purpose of this investigation was twofold. First, we sought to determine whether dynamic hindlimb skeletal muscle stretch results in greater reflex sympathetic and cardiovascular responses in HF-rEF rats compared to healthy (SHAM operated) control rats. Second, we sought to determine if the dynamic stretch modality would unmask an important role for EP4-Rs and TxA₂-Rs on thin fibre muscle afferents in the regulation of mechanoreflex activation in HF-rEF. Specifically, we tested the hypotheses that: 1) 30 seconds of 1 Hz dynamic hindlimb skeletal muscle stretch would result in greater reflex increases in RSNA, mean arterial pressure (MAP), and heart rate (HR) in HF-rEF rats compared to SHAM rats, and 2) the injection of the EP4-R antagonist L-161,982 (1 μg) or the TxA₂-R antagonist daltroban (80 μg) into the arterial supply of the hindlimb would reduce the reflex increases in RSNA, MAP, and HR evoked in response to dynamic hindlimb skeletal muscle stretch in HF-rEF rats, but not SHAM rats.

2. Methods

2.1. Ethical Approval

All experimental procedures were approved by the Institutional Animal Care and Use Committee of Kansas State University (Protocol #4076) and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (2011).

2.2. Animals

Experiments were performed on ~14–19 week old male Sprague-Dawley rats (n=37; Charles River Laboratories). Rats were housed two per cage in temperature (maintained at ~22°C) and light (12–12 hour light-dark cycle running from 7 AM to 7 PM)-controlled accredited facilities with standard rat chow and water provided ad libitum. Data from 25 of the 37 rats in this manuscript were collected as part of a set of experiments in which we investigated the role played by EP4-Rs and TxA₂-Rs during both static and dynamic mechanoreflex activation in HF-rEF rats. The data related to static mechanoreflex activation in those 25 rats have been reported previously (Butenas et al., 2020) whereas data related to dynamic mechanoreflex activation has not been reported previously. The additional 12 experiments in which only dynamic mechanoreflex activation was performed were necessary to increase statistical power of the primary study end points (n=7) or because control experiments were required (n=5).

2.3. Surgical Procedure

Myocardial infarction (MI) was induced in 23 of the 37 rats by surgically ligating the left main coronary artery (Musch et al., 1992). Briefly, rats were anesthetized initially with a 5% isoflurane-O₂ mixture (Butler Animal Health Supply, Elk Grove Village, IL, and Linweld, Dallas, TX) and maintained subsequently on 2.5% isoflurane-O₂ and then intubated and mechanically ventilated with a rodent respirator (Harvard model 680, Harvard Instruments, Holliston, MA) for the duration of the surgical procedure. A left thoracotomy was performed to expose the heart through the fifth intercostal space, and the left main coronary artery was ligated 1–2 mm distal to the edge of the left atrium with a 6–0 braided polyester suture. The thorax was then closed with 2–0 gut, and the skin was closed with 2–0 silk. Prior to termination of anesthesia, bupivacaine (1.5 mg/kg sc) and buprenorphine (~0.03 mg/kg im) were administered to reduce pain associated with the surgery, along with ampicillin (50 mg/kg im) to reduce the risk of infection. After rats were removed from mechanical ventilation and anesthesia, they were monitored closely for ~6 h post-surgery. In the remaining 14 of 37 rats, a sham ligation of the coronary artery was performed in which 6–0 braided polyester suture was passed under the left main coronary artery, but not tied. These rats are referred to as “SHAM” rats from this point forward. Following completion of either MI or SHAM procedures, rats were housed one per cage for ten days to minimize risk of infection of the surgical site. During these ten days, the antibiotic baytril (100 mg/mL) was administered in the drinking water. Following completion of the baytril treatment, rats were housed two per cage as described above. All animals were monitored daily for 14 days following MI or SHAM procedure for changes in behavior, gait/posture, breathing, appetite and body weight.

2.4. Echocardiograph Measurements

Transthoracic echocardiograph measurements were performed with a commercially available system (Logiq S8; GE Health Care, Milwaukee, WI) ~24–96 hours before the final experimental protocol. Briefly, the rats were anaesthetized as described above. Once the rat was fully anesthetized, the isoflurane mixture was reduced to 2% isoflurane-O₂. Following 5 minutes at 2% isoflurane, echocardiograph measurements began. The transducer was positioned on the left anterior chest, and left ventricular dimensions were measured. The left ventricular fractional shortening (FS), ejection fraction (EF), end diastolic (LVEDV), end systolic volume (LVESV), and stroke volume (SV) were determined by echocardiographic measurements as previously described (Craig et al., 2019a).

2.5. Surgical Procedures for Experimental Protocols

In vivo experiments were performed between four and six weeks following the MI or SHAM procedure and experiments incorporating the dynamic stretch maneuver were successfully performed on 37 rats (SHAM n=14, HF-rEF n=23). On the day of the experiment, rats were anesthetized as described above. Adequate depth of anesthesia was confirmed by the absence of toe-pinch and blink reflexes. The trachea was cannulated, and the lungs were mechanically ventilated (Harvard Apparatus, Holliston, MA) with a 2% isoflurane-balance O₂ gaseous mixture until the decerebration was completed (see below). The right jugular vein and both carotid arteries were cannulated with PE-50 catheters which were used for the injection of fluids, measurement of arterial blood pressure (physiological pressure transducer, AD Instruments), and sampling of arterial blood gasses (Radiometer). HR was measured by electrocardiogram (AD Instruments). The left superficial epigastric artery was cannulated with a PE-8 catheter whose tip was placed near the junction of the superficial epigastric and femoral arteries and a reversible snare (2–0 silk suture) was placed around the left iliac artery and vein as described previously (Rollins et al., 2019). The left calcaneal bone was severed and linked by string to a force transducer (Grass FT03), which, in turn, was attached to a rack and pinion. A retroperitoneal approach was used to expose bundles of the left renal sympathetic nerve, which were then glued (Kwik-Sil, World Precision Instruments) onto a pair of thin stainless-steel recording electrodes connected to a high impedance probe (Grass Model HZP) and amplifier (Grass P511). Multiunit signals from the renal sympathetic nerve fibers were filtered at high and low frequencies (1 KHz and 100 Hz, respectively) for the measurement of RSNA. Successful RSNA recordings were made in 11 SHAM rats and 14 HF-rEF rats.

Upon completion of the initial surgical procedures, rats were placed in a Kopf stereotaxic frame. After administering dexamethasone (0.2 mg i.v.) to minimize swelling of the brainstem, a pre-collicular decerebration was performed in which all brain tissue rostral to the superior colliculi was removed (Smith et al., 2001). Following decerebration, anesthesia was terminated, and the rats’ lungs were ventilated with room air. Rats were allowed a minimum of 60 minutes to recover from anesthesia before initiation of any experimental protocol. Experiments were performed on decerebrate, unanesthetized rats because anesthesia has been shown to markedly blunt the mechanoreflex and exercise pressor reflex in the rat (Smith et al., 2001). Body core temperature was measured via a rectal probe and maintained at ~37–38°C by an automated heating system (Harvard Apparatus) and heat lamp. Arterial pH and blood gases were analyzed periodically throughout the experiment from arterial blood samples (~75 μL) and maintained within physiological ranges (pH: 7.35–7.45, PCO₂: ~38–40 mmHg, PO₂: ~100 mmHg) by administration of sodium bicarbonate and/or adjusting ventilation as necessary.

2.6. Primary Experimental Protocols

We compared the RSNA, pressor, and cardioaccelerator responses evoked in response to 30 seconds of 1 Hz dynamic stretch of the triceps surae muscle before and after the injection of either the EP4-R antagonist L-161,982 (SHAM n=6, HF-rEF n=6) or the TxA₂-R antagonist daltroban (SHAM n=10, HF-rEF n=12) into the arterial supply of the left hindlimb. In detail, at least 60 minutes following the decerebration procedure and termination of anesthesia and ~5 minutes following a 30 second static hindlimb muscle stretch in the 25 rats for which static mechanoreflex responses were reported recently (Butenas et al., 2020), baseline muscle tension was set at ~80–100 g by manually turning the rack and pinion. Baseline blood pressure and heart rate were measured for ~30 seconds. An experienced investigator then initiated the stretch protocol in which the rack and pinion was turned back and forth manually at a 1 Hz frequency with the aid of a metronome for 30 seconds. The investigator aimed for a developed tension of ~0.6–0.8 kg on each stretch which is the tension generation typically developed during 1 Hz intermittent contraction protocols in decerebrate rats (Copp et al., 2016b; Kempf et al., 2018). This dynamic stretch protocol was adapted from protocols described originally by Stebbins et al. (1988) and Daniels et al. (2000) in the cat. Approximately five minutes following the completion of the control dynamic stretch protocol, and after ensuring that blood pressure had returned to its pre-stretch baseline value, the iliac snare was tightened and either the EP4-R antagonist L-161,982 (Tocris Bioscience; 1 μg dissolved in 0.2 ml of 0.15% DMSO in saline solution; Yamauchi et al., 2013) or the TxA₂-R antagonist daltroban (Santa Cruz Biotechnology, Inc.; 80 μg dissolved in 0.4 mL of 1% DMSO in saline solution; Leal et al., 2011) was injected into the arterial supply of the hindlimb through the superficial epigastric artery catheter. The receptor antagonist remained snared in the hindlimb circulation for 5 minutes, at which time the iliac snare was released. In EP4-R blockade experiments, the hindlimb was reperfused for 25 min. This dose of L-161,982 and timing of the injection protocol have been shown to reduce the pressor response to static hindlimb muscle contraction in rats with simulated peripheral artery disease (Rollins et al., 2020; Yamauchi et al., 2013). In TxA₂-R blockade experiments, the hindlimb was reperfused for 10 minutes. This dose of daltroban and the timing of the injection protocol have been shown to reduce the pressor response to hindlimb muscle stretch and contraction in rats with simulated peripheral artery disease (Leal et al., 2011; Rollins et al., 2020). Following the 25 or 10 min hindlimb reperfusion period and ~5 minutes following a 30 second static hindlimb muscle stretch in the 25 rats for which static mechanoreflex responses were reported recently (Butenas et al., 2020), the isolated dynamic stretch protocol was repeated exactly as described above. The tension generated during the post antagonist maneuvers was matched as closely as possible to that produced during the control stretch. Evans blue dye was then injected into the arterial supply of the hindlimb in the same manner as the receptor antagonists to ensure that the antagonists had access to the triceps surae muscle circulation. The triceps surae muscles were observed to stain blue in all but one experiment. Specifically, in one HF-rEF rat in which daltroban was injected into the hindlimb, the triceps surae muscles were not observed to stain blue and, therefore, only the data from the control stretch maneuver were used (for the SHAM vs. HF-rEF comparison).

At the end of all experiments in which RSNA was measured, postganglionic sympathetic nerve activity was abolished with administration of hexamethonium bromide (20 mg/kg iv) to allow for the quantification of background noise as described previously (Kempf et al., 2018). The decerebrate rats were then euthanized with an injection of saturated (>3 mg/kg) potassium chloride. The lungs were then excised and weighed. The heart was excised and the atria and right ventricle (RV) were separated from the left ventricle (LV) and septum, and the RV, LV, and atria were weighed. In rats with HF-rEF, the LV infarction surface area was measured using planimetry and expressed as percent of LV endocardial surface area as described previously (Craig et al., 2019b).

2.7. Control experiments

A total of five HF-rEF rats were used to investigate the potential effects of either 1% DMSO (i.e., the vehicle for daltroban) and/or whether systemic circulation of daltroban may have accounted for the attenuating effects in the main experimental group in which daltroban was injected into the hindlimb arterial circulation. In one rat, the DMSO protocol described below was the only protocol performed. In a different rat, the intravenous injection protocol described below was the only protocol performed. In three rats, the DMSO and the intravenous injection protocols were performed in series with ~5 minutes recovery time between protocols. The 1% DMSO (n=4 total) protocols were performed exactly as described above for the daltroban experiments except 0.4 ml of 1% DMSO alone was injected into the arterial supply of the hindlimb via the left superficial epigastric artery catheter. The systemic circulation control protocols (n=4 total) were performed as described above except 80 μg of daltroban was injected into the catheter placed in the jugular vein. The dynamic hindlimb muscle stretch maneuver was then performed 15 minutes after injection to match the timing of the daltroban injection into the arterial supply of the hindlimb described above.

2.8. Data analysis

Muscle tension, blood pressure, HR and RSNA were measured and recorded in real time with a PowerLab and LabChart data acquisition system (AD Instruments). The original RSNA data were rectified and corrected for the background noise determined after the administration of hexamethonium bromide. Baseline mean arterial pressure (MAP), RSNA and HR were determined from the 30-second baseline periods that preceded each maneuver. The peak increase in MAP (peak ΔMAP), RSNA (peak ΔRSNA), and HR (peak ΔHR) during dynamic stretch were calculated as the difference between the peak values wherever they occurred during the maneuvers and their corresponding baseline value. The tension-time indexes (TTIs), integrated RSNA for the first 5 sec of the stretch maneuver (∫ΔRSNA_{5 sec}) and blood pressure indexes (BPIs) were calculated by integration of the area under curve during the stretch maneuver and subtracting the integrated area under the curve during the baseline period. We calculated the peak ΔRSNA and the ∫ΔRSNA_5sec because the renal sympathetic nerve response to dynamic stretch is often characterized by large bursts early (i.e., the first ~5 secs) that may occur in sync with tension development which then subside over the duration of the stretch maneuver. Time courses of the increase in RSNA, MAP and HR were plotted as their change from baseline. Data were first assessed for the presence of a normal distribution, equal variance, and/or effectiveness of pairing as appropriate. Baseline MAP, baseline HR, Peak ΔMAP, Peak ΔHR, Peak ΔRSNA, BPI, ∫ΔRSNA_{5 sec}, and TTIs were compared with Sidak multiple comparisons tests. Time course data were analyzed with two-way ANOVAs and Sidak multiple comparisons tests. Data for echocardiograph measurements, body, heart, and lung weights were analyzed with unpaired Student’s t-tests or Mann-Whitney U tests as appropriate. Data are expressed as mean±SEM. Statistical significance was defined as P ≤ 0.05.

3. Results

3.1. Body weight and heart morphometrics

Body weight was not different between SHAM and HF-rEF rats (Table 1). The ratios of the lung weight to body weight and atria weight to body weight were significantly higher in HF-rEF rats compared to SHAM rats. There was no difference in the ratios of either the LV or RV weight to body weight between groups. Additionally, LVEDV and LVESV were significantly higher, while ejection fraction and fractional shortening were significantly lower, in HF-rEF rats compared to SHAM rats. Stroke volume was not different between groups.

Table 1.

Body and tissue weights and heart morphometrics in SHAM and HF-rEF rats

	SHAM (n=14)	HF-rEF (n=23)	P-Value
Body weight (g)	502 ± 16	503 ± 8	0.39
Lung/body weight (mg/g)	3.19 ± 0.12	3.51 ± 0.13	0.05*
RV/body weight (mg/g)	0.55 ± 0.02	0.59 ± 0.02	0.07
LV/body weight (mg/g)	1.92 ± 0.10	1.90 ± 0.03	0.41
Atria/body weight (mg/g)	0.20 ± 0.01	0.26 ± 0.02	0.03*
LV EDV (mL)	1.28 ± 0.08	2.12 ± 0.14	<0.01*
LV ESV (mL)	0.23 ± 0.04	1.13 ± 0.11	<0.01*
Stroke volume (mL)	1.05 ± 0.05	0.96 ± 0.05	0.12
Fractional shortening (%)	48 ± 2	22 ± 1	<0.01*
Ejection fraction (%)	83 ± 2	49 ± 2	<0.01*
Infarct size (%)	-	28 ± 2	-

Values are means ± SEM. LV, left ventricle; RV, right ventricle; EDV, end diastolic volume; ESV, end systolic volume. Data were compared using student’s t-test or Mann-Whitney tests as appropriate. Asterisks indicate a statistically significant difference between groups.

3.2. Effect of HF-rEF on responses to dynamic mechanoreflex activation

Pooled data from control conditions across all experiments in the present investigation (i.e., 14 SHAM and 23 HF-rEF rats) indicate that, compared to SHAM rats, HF-rEF rats had significantly larger peak ΔMAP, BPI, and peak ΔHR (Fig. 1) responses to 30 seconds of dynamic hindlimb muscle stretch. In experiments in which RSNA was successfully recorded (11 SHAM and 15 HF-rEF rats), the ∫ΔRSNA for the first 5 sec (i.e., ∫ΔRSNA_5sec) of the stretch maneuver was significantly larger (P=0.04) in HF-rEF compared to SHAM rats, with a trend (P=0.07, Cohen’s D=0.77, see Fig. 1D) towards a larger peak ΔRSNA in HF-rEF rats compared to SHAM rats. Furthermore, analysis of the time course of the responses to dynamic stretch revealed significant differences between SHAM and HF-rEF rats at one time point for the ΔRSNA, and multiple time points for both the ΔMAP and ΔHR (Fig. 2). The TTIs of the stretch maneuver were not different between groups (Fig. 1F). Examples of original tracings of the increase in RSNA, blood pressure and HR in response to 30 seconds of 1 Hz dynamic hindlimb skeletal muscle stretch from one SHAM rat and one HF-rEF rat are shown in Figure 3.

Figure 1: Effect of HF-rEF on the sympathetic and cardiovascular responses to dynamic mechanoreflex activation.

The peak Δ mean arterial pressure (MAP; A), peak Δ renal sympathetic nerve activity (RSNA; B), peak Δ heart rate (HR; C), blood pressure index (BPI, D), and the first 5 seconds of the integrated change in RSNA (∫ΔRSNA_5sec, E) in response to 30 seconds of dynamic hindlimb muscle stretch in SHAM (n=14) and HF-rEF (n=23) rats. TTI, tension-time index. Data were analyzed with Student’s t-tests or Mann-Whitney tests as appropriate and are expressed as mean±SEM with individual data points. Asterisks indicates statistically significant differences between groups.

Figure 2: Effect of HF-rEF on the time course of the reflex sympathetic and cardiovascular responses to dynamic mechanoreflex activation.

The Δ renal sympathetic nerve activity (RSNA; A), Δ mean arterial pressure (MAP; B), and Δ heart rate (HR; C) response to 30 seconds of dynamic hindlimb muscle stretch in SHAM (n=14) and HF-rEF (n=23) rats. Data were analyzed with two-way ANOVAs with Sidak multiple comparisons tests and are expressed as mean±SEM. Asterisks and/or black lines indicate time points where comparisons were statistically significant (P < 0.05).

Figure 3:

Examples of original tracings of the renal sympathetic nerve activity (RSNA), blood pressure (BP) and heart rate (HR) response to 30 seconds of dynamic hindlimb muscle stretch in a SHAM rat (left) and HF-rEF rat (right).

3.3. Effect of an EP4-R antagonist on the mechanoreflex

In SHAM (n=6) and HF-rEF (n=6) rats, we found that injection of the EP4-R antagonist L-161,982 into the arterial supply of the hindlimb had no effect on peak ΔMAP, BPI, peak ΔHR, peak ΔRSNA, or ∫ΔRSNA_5sec response to dynamic stretch (Fig. 4). Likewise, analysis of the time course of the responses to dynamic stretch revealed no effect of the EP4-R antagonist on the increase in RSNA, MAP, or HR in SHAM and HF-rEF rats (Fig. 5). The TTIs of the stretch maneuvers were not different between control and EP4-R blockade conditions in SHAM or HF-rEF rats (Fig. 4F). Baseline MAP and HR were not different between conditions in SHAM or HF-rEF rats (Table 2; P > 0.23 for all).

Figure 4: Effect of EP4-R blockade on the sympathetic and cardiovascular responses to dynamic mechanoreflex activation.

The peak Δ mean arterial pressure (MAP; A), peak Δ renal sympathetic nerve activity (RSNA; B), peak Δ heart rate (HR; C), blood pressure index (BPI, D), and the first 5 seconds of the integrated change in RSNA (∫ΔRSNA_5sec, E) in response to 30 seconds of dynamic hindlimb muscle stretch before (Control) and after injection of the EP4-R antagonist L-161,982 (1 μg) into the arterial supply of the hindlimb in SHAM (n=6) and HF-rEF (n=6) rats. TTI, tension-time index. Data were analyzed with Sidak multiple comparisons tests and are expressed as mean±SEM overlaid with individual responses.

Figure 5: Effect of EP4-R blockade on the time courses of the sympathetic and cardiovascular responses to dynamic mechanoreflex activation.

The Δ renal sympathetic nerve activity (RSNA; A & B), Δ mean arterial pressure (MAP; C & D), and Δ heart rate (HR; E & F) response to 30 seconds of dynamic hindlimb muscle stretch before (Control) and after injection of 1 μg of the EP4-R antagonist L-161,982 into the arterial supply of the hindlimb in SHAM (n=6) and HF-rEF (n=6) rats. Data were analyzed with two-way ANOVAs and Sidak multiple comparisons tests and are expressed as mean±SEM.

Table 2.

Baseline MAP and HR in SHAM and HF-rEF rats.

	MAP (mmHg)	HR (bpm)
SHAM (n=6)
Control	103 ± 4	492 ± 12
Post L-161,982	92 ± 3	477 ± 9
HF-rEF (n=6)
Control	90 ± 7	495 ± 11
Post L-161,982	94 ± 6	495 ± 15
SHAM (n=9)
Control	89 ± 5	493 ± 17
Post daltroban	87 ± 4	496 ± 19
HF-rEF (n=11)
Control	92 ± 5	486 ± 9
Post daltroban	90 ± 5	490 ± 9

Values are means ± SEM. MAP, mean arterial pressure; HR, heart rate. Data were compared with Sidak multiple comparisons tests. There was no significant difference for any comparison.

3.4. Effect of a TxA₂-R antagonist on the mechanoreflex

In SHAM rats (n=8), we found that the injection of the TxA₂-R antagonist daltroban into the arterial supply of the hindlimb had no effect on the peak ΔMAP, BPI, peak ΔHR, peak ΔRSNA, or ∫ΔRSNA_5sec response to dynamic stretch (Fig 6). Conversely, in HF-rEF rats (n=11), we found that the injection of the TxA₂-R antagonist into the arterial supply of the hindlimb significantly reduced the peak ΔMAP, BPI, ∫ΔRSNA_5sec, and peak ΔHR response to dynamic stretch. There was no effect of TxA₂-R blockade on the peak ΔRSNA response to stretch in HF-rEF rats (Fig. 6). Analysis of the time course of the responses to stretch revealed that TxA₂-R blockade had no effect in SHAM rats, but significantly reduced the increase in RSNA, MAP, and HR at various time points throughout the maneuver in HF-rEF rats (Fig. 7). The TTIs of the stretch maneuvers were not different between control and TxA₂-R blockade conditions in either SHAM rats or HF-rEF rats (Fig. 6F). Baseline MAP and HR were not different between conditions in SHAM or HF-rEF rats (Table 2; P > 0.59 for all).

Figure 6: Effect of TxA₂-R blockade on the sympathetic and cardiovascular responses to dynamic mechanoreflex activation.

The peak Δ mean arterial pressure (MAP; A), peak Δ renal sympathetic nerve activity (RSNA; B), peak Δ heart rate (HR; C), blood pressure index (BPI, D), and the first 5 seconds of the integrated change in RSNA (∫ΔRSNA_5sec, E) in response to 30 seconds of dynamic hindlimb muscle stretch before (Control) and after injection of the TxA₂-R antagonist daltroban (80 μg) into the arterial supply of the hindlimb in SHAM (n=8) and HF-rEF (n=11) rats. TTI, tension-time index. Data were analyzed with Sidak multiple comparisons tests and are expressed as mean±SEM overlaid with individual responses. Asterisks indicate statistically significant differences between groups.

Figure 7: Effect of TxA₂-R blockade on the time courses of the sympathetic and cardiovascular responses to dynamic mechanoreflex activation.

Δ renal sympathetic nerve activity (RSNA; A & B), Δ mean arterial pressure (MAP; C & D), and Δ heart rate (HR; E & F) during 30 seconds of dynamic hindlimb muscle stretch before (Control) and after injection of 80 μg of the TxA₂-R antagonist (daltroban) into the arterial supply of the hindlimb in SHAM (n=8) and HF-rEF (n=11) rats. Data were analyzed with two-way ANOVAs and Sidak multiple comparisons tests and are expressed as mean±SEM. Asterisks and/or black lines indicate time points where comparisons were statistically significant (P < 0.05).

3.5. Control experiments

In four HF-rEF rats, we investigated the effect of injection of 1% DMSO in 0.4 mL saline (i.e., the vehicle for daltroban) into the arterial supply of the hindlimb on the pressor and cardioaccelerator response to dynamic stretch. We found that 1% DMSO had no effect on the peak ΔMAP (control: 22 ±6, 1% DMSO: 28±9 mmHg; P=0.16), BPI (control: 390±129, 1% DMSO: 536±186 mmHg·s; P=0.22), or peak ΔHR (control: 13±5, 1% DMSO: 18±6 bpm; P=0.16) response to dynamic stretch. The TTIs of the stretch maneuvers were not different between control (17±2 kg·s) and 1% DMSO (18±2 kg·s; P=0.25) conditions. Baseline MAP (Control: 90±6; 1% DMSO: 91±9 mmHg, P=0.96) and HR (Control: 447±17; 1% DMSO: 446±17 bpm, P=0.81) were not different between conditions. These findings suggest that the attenuating effect of daltroban produced when it was injected into the arterial supply of the hindlimb of HF-rEF rats was not attributable to effects produced by its vehicle 1% DMSO.

In four HF-rEF rats, we investigated the effect of intravenous (i.v.) injection of daltroban on the pressor and cardioaccelerator response to dynamic stretch. We found that i.v. injection of daltroban had no effect on the peak ΔMAP (control: 30±8, daltroban i.v.: 34±10 mmHg; P=0.36), BPI (control: 597±154, daltroban i.v.: 617±128 mmHg·s; P=0.81), or peak ΔHR (control: 19±5, daltroban i.v.: 19±3 bpm; P=0.77) response to dynamic stretch. The TTIs of the stretch maneuvers were not different between control (16±2 kg·s) and daltroban i.v. (16±2 kg·s; P=0.65) conditions. Baseline MAP (control: 96±10; daltroban i.v.: 92±7 mmHg, P=0.77) and HR (control: 447±18; daltroban i.v.: 458±16 bpm, P=0.31) were not different between conditions. These findings suggest that the attenuating effect of daltroban produced when it was injected into the arterial supply of the hindlimb of HF-rEF rats is most likely attributable to blockade of TxA₂-R on the sensory endings of thin fiber muscle afferents and not systemic effects elsewhere in the mechanoreflex arc such as the brainstem and/or the spinal cord.

4. Discussion

We investigated the effect of HF-rEF on the reflex sympathetic and cardiovascular responses to dynamic mechanoreflex activation isolated from acute contraction-induced metabolite production, and the possible roles played by EP-4s and TxA₂-Rs in producing those responses. Consistent with our hypothesis, we found that, in HF-rEF rats, the increase in RSNA, blood pressure, and HR in response to 1 Hz dynamic hindlimb skeletal muscle stretch was greater than the increases observed in healthy SHAM rats. Moreover, we found that hindlimb arterial injection of an antagonist for TxA₂-Rs, but not EP4-Rs, reduced the exaggerated responses to dynamic hindlimb muscle stretch in HF-rEF rats, whereas there was no effect of either antagonist in SHAM rats. The present findings suggest that, in HF-rEF rats, TxA₂-Rs on the sensory endings of thin fibre muscle afferents contribute to a chronic sensitization of the mechanically activated channels that underlie dynamic mechanoreflex activation. The findings enhance our understanding of the altered reflex control of the sympathetic nervous system and cardiovascular system during exercise in HF-rEF patients.

The isolated mechanoreflex protocol used in the present investigation produces a rhythmic mechanical stimulus similar in pattern and magnitude to the mechanical stimulus present during electrically-induced dynamic skeletal muscle contractions with this same decerebrate rat model (Kempf et al., 2018). The protocol resulted in substantial reflex increases in RSNA and MAP. The reflex increases in HR were relatively modest when considered in relation to the high resting HR in the rat. Our present finding that the sympathetic and cardiovascular responses to 1 Hz repetitive/dynamic skeletal muscle stretch were exaggerated in HF-rEF rats compared to SHAM rats builds upon several previous findings in which various models of isolated intermittent/dynamic mechanoreflex activation have been used in HF-rEF subjects. Koba et al. (2008) found that the increase in RSNA and lumbar SNA averaged across 12 individual, one-second duration hindlimb muscle stretches was larger in HF-rEF rats compared to healthy rats. Middlekauff et al. (2004) found that three minutes of passive wrist flexion (30 times/min) produced larger increases in muscle SNA (MSNA) in HF-rEF patients than in aged-matched healthy counterparts, an effect later attributed to a sensitization of mechanoreceptors by COX products of arachidonic metabolism (Middlekauff et al., 2008). More recently, Antunes-Correa et al. (2014) found that passive knee joint flexion/extension (30 times/min) produced a larger increase in MSNA in HF-rEF patients before exercise training compared to after exercise training. Those prior studies provided valuable information regarding the augmented mechanoreflex control of SNA in HF-rEF. However, the various isolated intermittent/dynamic mechanoreflex models produced little to no change in cardiovascular variables. In contrast, Witman et al. (2015) found that continuous 1 Hz passive knee joint flexion/extension produced larger increases in HR in healthy control subjects than in HF-rEF patients, whereas the increase in MAP was not different between groups. The reason for the differences between the present investigation and that of Witman et al. (2015) is unclear but may include the pharmacological therapies of the HF-rEF patients.

The present investigation employed established and validated experimental protocols (see Methods) to examine the role played by EP4-Rs and TxA₂-Rs on the sensory endings of thin fibre muscle afferents during isolated dynamic mechanoreflex reflex activation in SHAM and HF-rEF rats. In SHAM rats, we found that hindlimb arterial injection of the EP4-R antagonist had no effect on the sympathetic, pressor, or cardioaccelerator response to dynamic stretch, which is consistent with our recent findings (Rollins et al., 2020). Those findings are similar to previous findings in SHAM/healthy rats in which EP4-R blockade had no effect on the sympathetic and/or pressor responses to static hindlimb muscle stretch (Butenas et al., 2020; Yamauchi et al., 2013), or static hindlimb muscle contraction (Stone et al., 2015a; Yamauchi et al., 2013). Moreover, our present finding in SHAM rats that the hindlimb arterial injection of the TxA₂-R antagonist had no effect on the pressor response to dynamic hindlimb muscle stretch is consistent with our recent investigations in which dynamic (Rollins et al., 2020) and static (Butenas et al., 2020) hindlimb stretch was performed. Collectively, the present results, and the bulk of the available literature (Butenas et al., 2019; Butenas et al., 2020; Rollins et al., 2020; Rollins et al., 2019; Yamauchi et al., 2013), suggest that EP4-Rs and TxA₂-Rs on the sensory endings of thin fibre muscle afferents in SHAM/healthy rats do not produce a chronic sensitization of the mechanically activated channels that underlie isolated static or dynamic mechanoreflex activation.

The present investigation of the role played by EP4-Rs and TxA₂-Rs during dynamic mechanoreflex activation in HF-rEF rats was undertaken to shed further mechanistic light on previous findings suggesting a role for COX metabolism in the exaggerated sympathetic responses to dynamic mechanoreflex/exercise pressor reflex activation in HF-rEF subjects (Antunes-Correa et al., 2014; Middlekauff et al., 2008; Smith et al., 2020a). Our finding in HF-rEF rats that the EP4-R antagonist had no effect on the sympathetic, pressor, or cardioaccelerator response to dynamic hindlimb muscle stretch extends our recent findings in HF-rEF rats that the same EP4-R blockade protocol had no effect on the reflex responses to static hindlimb muscle stretch (Butenas et al., 2020). Moreover, the present findings are consistent with our recent investigation in which EP4-R blockade had no effect on the pressor response to dynamic stretch in a rats with simulated peripheral artery disease (PAD, Rollins et al., 2020). The possibility that redundancy among various receptors on the sensory endings of thin fibre muscle afferents may have masked a role for EP4-Rs in the exaggerated mechanoreflex in HF-rEF rats should also be considered (Stone et al., 2015b). Furthermore, a role for EP4-Rs in the exaggerated exercise pressor reflex in HF-rEF may yet exist, wherein contraction-induced increases in metabolite production (e.g., PGE₂) results in EP4-R stimulation. In support of this notion, Scott et al. (2002) showed that, compared with age-matched control subjects, HF-rEF patients had a greater concentration of PGE₂ and PGF_1α within venous effluent from rhythmically contracting forearm muscles, a maneuver that resulted in larger increases in HR and ventilation in HF-rEF patients compared to age-matched healthy control subject. Moreover, the PGE₂ and PGF_1α concentrations correlated positively with the contraction-induced increase in ventilation in HF-rEF patients (Scott et al., 2002).

Our present finding in HF-rEF rats that hindlimb arterial injection of the TxA₂-R antagonist daltroban reduced the sympathetic, pressor, and cardioaccelerator response to dynamic stretch is consistent with our recent report that daltroban reduced the pressor and cardioaccelerator response to dynamic stretch in rats with simulated PAD (Rollins et al., 2020). Our present finding stands in contrast, however, to our recent finding that daltroban had no effect on the reflex responses to static hindlimb muscle stretch in HF-rEF rats (Butenas et al., 2020). The contrasting findings related to the role played by TxA₂-Rs during different mechanoreflex modalities may be best explained by the fact that the different modalities may activate substantially different classes of mechanically activated channels, and that, in HF-rEF, TxA₂-R signaling may modulate the function of the channels that are stimulated to a greater extent during dynamic compared to static mechanoreflex activation. There is at least indirect support for this explanation in the literature. For example, Nakamoto et al. (2007) suggested that some mechanically activated channels associated with the mechanoreflex are located within the belly of skeletal muscles and respond primarily to changes in muscle length whereas others are located near myotendinous junctions and respond primarily to muscle tension generation. More recently, Copp et al. (2016a) found that, in healthy rats, the tarantula peptide GsMTx4 (a modulator of mechanically gated channel function that is partially selective for specific classes of channels including piezo channels, Bae et al., 2011) reduced the pressor response only during the very early phase of static hindlimb muscle stretch when muscle length changed rapidly. In a follow-up study, Sanderson et al. (2019) found that GsMTx4 reduced the pressor response throughout the duration of dynamic rat hindlimb muscle stretch wherein muscle length changed repetitively. Based on the findings above, it is reasonable to speculate that, in HF-rEF rats, TxA₂-Rs on the sensory endings of thin fibre muscle afferents mediate a chronic sensitization of the mechanically activated channels that respond primarily to changes in muscle length; the identity of which may include piezo channels. It is also possible that TxA₂-Rs are expressed differently on sensory neurons innervating different regions of the muscle (i.e., muscle belly vs. myotendinous junction). These possibilities remain to be investigated.

The specific mechanism that results in the TxA₂-R mediated mechanoreflex sensitization in HF-rEF rats is unknown. An increase in the expression of TxA₂-Rs on sensory endings of muscle afferents does not appear to contribute, however, as we previously reported no difference between SHAM and HF-rEF rats in TxA₂-R protein or mRNA expression in lumbar dorsal root ganglia tissue (Butenas et al., 2020). We should note, however, that in heart failure patients, exercise training reduced mechanoreflex control of MSNA and TxA₂-R gene expression in skeletal muscle homogenates (Antunes-Correa et al., 2014). A possible explanation for the effect of daltroban in HF-rEF rats in this study is that the development of HF-rEF resulted in increased basal levels of COX products of arachidonic metabolism. There is at least indirect evidence supporting this notion. For example, we (Butenas et al., 2020) and others (Morales et al., 2012; Smith et al., 2020a) have reported greater COX-2 protein expression within skeletal muscle homogenates of HF-rEF patients/animals compared to healthy controls. Another possible explanation is that an amplification of inositol trisphosphate (IP₃) and diacylglycerol signaling (the G protein-coupled signaling pathways associated with TxA₂-Rs) within sensory neuron endings may exist in HF-rEF rats compared to SHAM rats. Possible HF-rEF-induced alterations in IP₃ signaling within the sensory ending of thin fibre muscle afferents, for example, may result in an increased cytosolic calcium concentration which, in turn, may sensitize mechanically activated channels (Zhuang et al., 2015). Whether the effect of TxA₂-R inhibition in HF-rEF rats presented herein reflects an alteration in IP₃ second messaging remains unclear.

A few experimental considerations should be noted. First, the rat hindlimb stretch model of isolated mechanoreflex activation is that it divorces mechanical stimuli from the influence of acute contraction-induced metabolite production. Thus, hindlimb muscle stretch allows for investigation of the presence and mechanisms of chronic sensitization of mechanically activated channels. A limitation of the model is that it stimulates mechanically activated channels differently than they are stimulated when skeletal muscles contract and intramuscular pressure increases (Gallagher et al., 2001). Importantly, however, Stone et al. (2015b) found that the vast majority (i.e., >86%) of the muscle afferents that respond to rat hindlimb muscle stretch also respond to hindlimb muscle contraction. Moreover, investigations that have identified a mechanism that modulates the reflex responses to rat hindlimb muscle stretch have consistently found that mechanism also modulates the reflex responses to hindlimb muscle contraction (Copp et al., 2016a; Copp et al., 2016b; Downey et al., 2017; Kim et al., 2019; Leal et al., 2011; McCord et al., 2011; Nakamoto et al., 2008; Schiller et al., 2019; Smith et al., 2003; Smith et al., 2005; Tsuchimochi et al., 2011; Wang et al., 2013; Yamauchi et al., 2012). Second, RSNA was not measured in the vehicle and systemic control experiments. We elected to perform the vehicle and systemic control protocols in experiments in which technical factors precluded successful RSNA recording in order to maximize the statistical power of the RSNA comparisons in the main experimental protocols. Finally, HF-rEF patients constitute roughly half of all patients with heart failure (Owan et al., 2006). Whether our present findings extend to patients with heart failure with preserved ejection fraction and/or non-ischemic heart failure remains unknown.

5. Conclusions

In the present investigation, we found that the sympathetic, pressor, and cardioaccelerator responses to 1 Hz dynamic hindlimb muscle stretch were larger in rats with HF-rEF compared to the responses in SHAM operated healthy control rats. Furthermore, we found that the hindlimb arterial injection of a TxA₂-R antagonist, but not an EP4-R antagonist, reduced the exaggerated responses to dynamic hindlimb muscle stretch in HF-rEF rats. The data suggest that TxA₂-Rs on the sensory endings of thin fiber muscle afferents contribute to a chronic sensitization of the mechanically activated channels that underlie dynamic/rhythmic mechanoreflex activation. The data enhance our understanding of the reflex control of the sympathetic nervous system during dynamic exercise and the mechanistic underpinnings of exercise intolerance and cardiovascular risk for the over 26 million people worldwide with HF-rEF (Ponikowski et al., 2014).

Highlights.

• Sympathetic responses to dynamic muscle stretch are augmented in heart failure

• Endoperoxide 4 receptors do not contribute to the mechanoreflex in heart failure

• Thromboxane A₂ receptors contribute to mechanoreflex exaggeration in heart failure

• Neither of these receptors contribute to the mechanoreflex in healthy rats