Sympathoexcitation in response to cardiac and pulmonary afferent stimulation of TRPA1 channels is attenuated in rats with chronic heart failure

Abstract

Excessive sympathoexcitation characterizes the chronic heart failure (CHF) state. An exaggerated cardiac sympathetic afferent reflex (CSAR) contributes to this sympathoexcitation. Prior studies have demonstrated that the CSAR to capsaicin [transient receptor potential (TRP) vanilloid 1 agonist] is exaggerated in CHF animal models. We recently discovered that capsaicin application to the lung visceral pleura in anesthetized, vagotomized, open-chested rats increases mean arterial pressure (MAP), heart rate (HR), and renal sympathetic nerve activity (RSNA). We named this response the pulmonary spinal afferent reflex (PSAR). Due to the similarities between TRP vanilloid 1 and TRP ankyrin 1 (TRPA1) channels as well as the excessive sympathoexcitation of CHF, we hypothesized that stimulation of the CSAR and PSAR with a specific TRPA1 agonist would result in an augmented response in CHF rats (coronary ligation model) compared with sham control rats. In response to a TRPA1 agonist, both CSAR and PSAR in sham rats resulted in biphasic changes in MAP and increases in HR and RSNA 10–12 wk postmyocardial infarction (post-MI). These effects were blunted in CHF rats. Assessment of TRPA1 expression levels in cardiopulmonary spinal afferents by immunofluorescence, quantitative RT-PCR, and Western blot analysis 10–12 wk post-MI all indicates reduced expression in CHF rats but no reduction at earlier time points. TRPA1 protein was reduced in a dorsal root ganglia cell culture model of inflammation and simulated tissue ischemia, raising the possibility that the in vivo reduction of TRPA1 expression was, in part, caused by CHF-related tissue ischemia and inflammation. These data provide evidence that reflex responses to cardiopulmonary spinal afferent TRPA1 stimulation may be attenuated in CHF rather than enhanced.

NEW & NOTEWORTHY Excessive sympathoexcitation characterizes chronic heart failure (CHF). The contribution of transient receptor potential ankyrin 1 (TRPA1) channel-mediated reflexes to this sympathoexcitation is unknown. We found that application of TRPA1 agonist to the heart and lung surface resulted in increased heart rate and sympathetic output and a biphasic change in mean arterial pressure in control rats. These effects were attenuated in CHF rats, decreasing the likelihood that TRPA1 channels contribute to cardiopulmonary afferent sensitization in CHF.

INTRODUCTION

Chronic heart failure (CHF) is a common and life-threatening health condition that is growing in prevalence. A significant contributor to CHF disease progression is excessive sympathoexcitation. For example, CHF is associated with elevated basal sympathetic tone (2, 6, 8), an intensified exercise pressor reflex (47), and an augmented cardiac sympathetic afferent reflex (CSAR) (51, 52). The CSAR originates in cardiac spinal sensory afferents and results in increases in mean arterial pressure (MAP), heart rate (HR), and sympathetic outflow as measured by renal sympathetic nerve activity (RSNA). Previous work from our laboratory has indicated the CSAR is tonically activated in animal models of CHF and that stimulation of the CSAR with both bradykinin (a B₂-receptor agonist) and capsaicin [a transient receptor potential (TRP) cation channel subfamily vanilloid member 1 (TRPV1) receptor agonist] produces an amplified sympathoexcitatory response in animals with CHF compared with sham control animals (49, 52). This suggests a potential role of cardiac afferent-containing B₂ receptors and TRPV1 receptors in mediating cardiac afferent sensitization in CHF. However, whether related receptor channels such as the TRP ankyrin 1 (TRPA1) receptor are also involved in the enhanced CSAR in CHF remains unknown. Given the known role of TRPV1 agonists in the CSAR, the close structural/functional relationship between TRPV1 and TRPA1 (21, 27, 43), and the colocalization of TRPV1 and TRPA1 within dorsal root ganglia (DRG) (36), we hypothesized that TRPA1 channels may also contribute to the exaggerated cardiac sympathetic neural circuitry endemic to CHF. Moreover, TRPA1, a promiscuous channel, is sensitive to a number of endogenous substances associated with myocardial ischemia and CHF-related inflammation, including bradykinin, prostaglandins, and protons (14, 39). Thus, in the present study, we hypothesized that stimulation of epicardial sensory afferents by the specific TRPA1 agonist ally isothiocyanate (AITC) (1, 3, 40) would result in a sympathetic reflex that would be exaggerated within the setting of CHF.

Using an experimental design similar to CSAR activation, we recently demonstrated that topical application of bradykinin and capsaicin directly to the lung visceral pleura results in a sympathetic reflex in anesthetized, vagotomized, open-chested rats (33). This reflex, to the best of our knowledge previously undescribed, includes increases in MAP, HR, and RSNA (33). We named this reflex the pulmonary spinal afferent reflex (PSAR). We know that both the PSAR and CSAR originate at the level of T1−T4 DRG, which provide sensory afferent innervation to the heart and lung, because chemical denervation of T1−T4 DRG sensory afferents with the potent neurotoxin resiniferatoxin abolishes both reflexes (33, 49). However, whether pulmonary spinal sensory afferents are also sensitive to TRPA1 receptor agonists remains unknown. Furthermore, it is not known if reflex responses to cardiopulmonary, spinal, sensory afferent TRPA1 stimulation is altered in the CHF state. Based on these considerations, we further hypothesized that activation of pulmonary spinal sensory afferents by topical application of AITC to the lung visceral pleura would result in a sympathoexcitatory response that would also be exaggerated within the setting of CHF.

We tested these hypotheses by 1) assessing hemodynamics and sympathetic output in response to direct application of AITC to the epicardium and lung visceral pleura, respectively, in sham and CHF rats; 2) measuring mRNA and protein expression of TRPA1 in sham and CHF rat T1−T4 DRG; and 3) assessing TRPA1 protein levels in DRG cell culture models of tissue ischemia and inflammation, respectively.

METHODS

Ethical approval and animals.

All animal experimentation was approved by the Institutional Animal Care and Use Committee of the University of Nebraska Medical Center and were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Experiments were performed on adult, male, 300- to 400-g Sprague-Dawley rats purchased from the Charles River. Animals were housed on site and given a 1-wk acclimation period before experimentation. Food and water were supplied ad libitum, and rats were on 12:12-h light-dark cycles.

We were unable to perform all animal experiments in a single cohort of animals; thus, multiple animal cohorts were used. Here, we describe how many animals were used for each experiment. Table 1 shows the total number of animals for cardiopulmonary spinal afferent reflex stimulation experiments. We studied a total of 21 sham rats and 23 CHF rats in both cardiac and pulmonary spinal afferent experiments combined. We performed echocardiography on these animals. These data are shown in Table 2. Some but not all of the animals used in the in vivo CSAR and PSAR assays received both the 30 and 5 mM AITC dose, whereas others received only a single dose. As shown in Table 3, we studied seven CHF rats in the 30 mM group and five of seven rats also received 5 mM AITC stimulation. As shown in Table 3, we studied seven sham rats; six rats received both doses. As shown in Table 4, we studied eight CHF rats at 30 mM and five of eight rats also received 5-mM AITC stimulation. As shown in Table 4, we studied six sham rats and all received both doses. Application of 30 and 5 mM AITC was separated by 20 min. Saline was used to wash away residual AITC in between doses. Concerning rats treated with vehicle, we carried out both epicardial and pulmonary application of saline on several common CHF and sham rats. In total, we used eight CHF rats and eight sham rats to complete the vehicle experiments shown in Tables 3 and 4. Tissue samples from sham and CHF rats that underwent the CSAR and PSAR experiments were used for the 10- to 12-wk postmyocardial infarction (post-MI)/sham surgery Western blot analysis and quantitative RT-PCR. Additional cohorts of sham and CHF rats were used for the immunofluorescence analysis (n = 4 for sham and CHF rats). Additional rat cohorts were also used to complete the Western blot analysis at 1, 3–4, and 6–8 wk post-MI/sham surgery time points (n = 6) for CHF and sham rats at each of these three time points.

Table 1.

Hemodynamic and morphological data in sham and CHF rats

	Sham	CHF
Number of rats/group	21	23
Body weight, g	412 ± 7	434 ± 10
Heart weight, mg	1167 ± 13	1757 ± 44*
Heart weight/body weight, mg/g	2.8 ± 0.1	4.1 ± 0.1*
Wet lung weight/body weight, mg/g	4.5 ± 0.1	7.6 ± 0.4*
Left ventricular end-systolic pressure, mmHg	127.4 ± 1.7	105.9 ± 1.7*
Left ventricular end-diastolic pressure, mmHg	4.3 ± 0.3	16.2 ± 1.5*
Heart rate, beats/min	373 ± 6	331 ± 3.9
dP/dtmax, mmHg/s	9,018 ± 318	4,840 ± 178*
dP/dtmin, mmHg/s	−8,081 ± 269	−3,520 ± 172*
Infarct size, %	0	41.3 ± 1.4*

Values are means ± SE. CHF, chronic heart failure. Data were obtained 10–12 wk postmyocardial infarction/sham surgery.

^*P < 0.05 by and unpaired two-tailed t-test.

Table 2.

Echocardiography data in sham and CHF rats

	Sham	CHF
Number of rats/group	21	23
Left ventricular end-diastolic diameter, mm	7.3 ± 0.1	10.2 ± 0.2*
Left ventricular end-systolic diameter, mm	4.1 ± 0.1	8.2 ± 0.2*
Left ventricular end-systolic volume, µl	74 ± 5	380 ± 24*
Left ventricular end-diastolic volume, µl	268 ± 10	592 ± 27*
Ejection fraction, %	72.6 ± 1.2	36.8 ± 1.5*
Fractional shortening, %	44.2 ± 1.1	19.0 ± 0.9*

Values are means ± SE. CHF, chronic heart failure. Data were obtained 6 wk postmyocardial infarction/sham surgery.

^*P < 0.05 by an unpaired two-tailed t-test.

Table 3.

Baseline mean arterial pressure and heart rate in sham and CHF rats that received epicardial AITC

	Vehicle		5 mM AITC Dose		30 mM AITC Dose
	Sham	CHF	Sham	CHF	Sham	CHF
Number of rats/group	7	5	6	5	7	7
Mean arterial pressure, mmHg	91 ± 4	83 ± 7	90 ± 3	80 ± 6	90 ± 4	85 ± 4
Heart rate, beats/min	365 ± 10	332 ± 9*	364 ± 11	329 ± 9	365 ± 11	329 ± 6*

Values are means ± SE. CHF, chronic heart failure; AITC, ally isothiocyanate Data were obtained from anesthetized, open-chested, vagotomized rats 10–12 wk postmyocardial infarction/sham surgery.

^*P < 0.05, CHF vs. sham rats by a Mann-Whitney rank-sum test.

Table 4.

Baseline mean arterial pressure and heart rate in sham and CHF rats that received lung visceral pleura AITC

	Vehicle		5 mM AITC Dose		30 mM AITC Dose
	Sham	CHF	Sham	CHF	Sham	CHF
Number of rats/group	6	5	6	5	6	8
Mean arterial pressure, mmHg	84 ± 6	89 ± 5	93 ± 5	90 ± 6	89 ± 7	94 ± 5
Heart rate, beats/min	372 ± 12	330 ± 7*	359 ± 22	330 ± 7	385 ± 8	327 ± 6*

Values are means ± SE. CHF, chronic heart failure; AITC; ally isothiocyanate. Data were obtained from anesthetized, open-chested, vagotomized rats 10–12 wk postmyocardial infarction/sham surgery.

^*P < 0.05, CHF vs. sham rats by a Mann-Whitney rank-sum test.

Rat model of CHF.

CHF was produced by coronary ligation as previously described (46, 47, 49). In brief, rats were anesthetized with 3% isoflurane and mechanically ventilated at 60 breaths/min. A thoracotomy was performed through the left fifth intercostal space, and the pericardium was opened, exposing the epicardium. The left anterior descending coronary artery was ligated followed by thorax closure and manual reestablishment of intrapleural pressure. Sham animals underwent the same procedure, including the thoracotomy, except no coronary ligation was performed. Buprenorphine (0.05 mg/kg sc) was given once immediately after surgery and on postoperative days 1 and 2 for alleviation of pain.

Hemodynamics were assessed with echocardiography (Vevo 3100, Visual Sonics, Toronto, ON, Canada) in sham and CHF rats 6 wk postsurgery, as previously described (45–47, 49). After the acute terminal experiments (described below), rats were euthanized with a rapid intravenous injection of saturated KCl. The hearts and lungs were removed and weighed. The ratio of the infarct area to whole left ventricle (LV) minus septum was measured.

In vivo measurement of arterial blood pressure, HR, and RSNA before and after PSAR and CSAR stimulation.

Surgical preparation was performed as previously described (33, 47, 49) in rats 10–12 wk post-MI/sham operation. For the acute, terminal experiments, rats were anesthetized with urethane (800 mg/kg ip) and α-chloralose (40 mg/kg ip). Supplemental doses of α-chloralose (20 mg/kg iv) were administered every 1.5–2 h to maintain an appropriate level of anesthesia. The anesthetic plane was monitored by establishing that rats were unresponsive to pedal withdrawal and corneal reflexes. The trachea was cannulated, and rats were mechanically ventilated with room air supplemented with oxygen. At the beginning of the acute experiments, a Millar catheter (SPR 524, size: 3.5-Fr, Millar Instruments, Houston, TX) was advanced through the right carotid artery into the LV to determine LV end-diastolic pressure, LV systolic pressure, dP/dt_max, and dP/dt_min. The transducer was then pulled back into the aorta and left in place to record arterial pressure. MAP and HR were derived from the arterial pressure pulse using LabChart 7.1 software and an analog-to-digital converter (PowerLab model 16S, AD Instruments, Colorado Springs, CO). The right jugular vein was cannulated for intravenous injections and administration of saline at a rate of 3 ml/h. Body temperature was maintained between 37 and 38°C by a heating pad. In all experiments, the cervical vagal nerves were cut bilaterally to prevent any reflex responses originating from vagal afferent activation, so as to observe only responses related to cardiac or pulmonary spinal afferent activation.

In the acute, terminal surgical preparations described above, RSNA was recorded as previously described (33, 49). In brief, the left kidney, renal artery, and nerves were exposed through a left retroperitoneal flank incision. Sympathetic nerves running on or beside the renal artery were identified. The renal sympathetic nerves were placed on a pair of platinum-iridium recording electrodes and cut distally to avoid recording afferent activity. Nerve activity was amplified (×10,000) and filtered (bandwidth: 100–3,000 Hz) using a Grass P55C preamplifier. The nerve signal was displayed on a computer through which it was rectified, integrated, sampled (1 kHz), and converted to a digital signal by the PowerLab data acquisition system. At the end of the experiment, the rat was euthanized with an overdose of saturated KCl. Respective noise levels were subtracted from the nerve recording data before percent changes from baseline were calculated. Integrated RSNA was normalized as 100% of mean baseline (1 min) during the control period.

To activate the PSAR, a 3 × 3-mm square of filter paper was soaked in the TRPA1 agonist AITC (Sigma-Aldrich, Atlanta, GA) and then applied to the ventral surface of the left superior lung lobe as previously described (33). The CSAR was similarly activated except that AITC was applied to the LV epicardial surface, in a region adjacent to but not overlapping with the LV scar area, as previously described (49). PSAR and CSAR experiments were performed in different animals to avoid any potential interaction between the two reflexes. Baseline and post-AITC levels of arterial blood pressure, MAP, HR, and RSNA were continuously recorded. Two concentrations of AITC were used to soak the filter paper: 5 and 30 mM. We repeated CSAR and PSAR activation using vehicle only. The vehicle for AITC was DMSO.

Immunofluorescence.

The immunofluorescence protocol was as described in our previous publications (33, 44, 49). Animals were anesthetized with pentobarbital sodium (40 mg/kg ip) and then perfused through the aorta with 200 ml of heparinized saline followed by perfusion with 200 ml of 4% paraformaldehyde. T1−T4 DRG were resected and submersion fixed in 4% paraformaldehyde for 24 h and then transferred to 30% sucrose and PBS solution for 72 h. DRG samples were embedded and sectioned (14-µm thickness) using a Leica cryostat (−20°C). Slices were thaw mounted onto glass slides in preparation for immunofluorescence. Before incubation with the primary antibodies, slides were washed in PBS and Triton X-100 (0.2%) for 10 min and then rinsed twice for 5 min each in PBS. After the rinse, slides were blocked for 1 h with 10% donkey serum albumin (Jackson ImmunoResearch Laboratories, West Grove, PA) and then incubated with primary antibodies prepared in 10% donkey serum albumin overnight at 4°C. The following primary antibodies were used: rabbit anti-TRPA1 (1:200 dilution, product no. NB100-91319, Novus Biologicals, Littleton, CO) and mouse anti-NF200 (A-fiber neuronal marker, 1:200 dilution, product no. N0142, Sigma-Aldrich). After an overnight incubation, slides were rinsed three times with PBS for 5 min each and incubated with secondary antibodies (prepared in PBS) for 1 h at room temperature. Secondary antibodies were Pacific blue-conjugated goat anti-mouse IgG (1:100 dilution, product no. p31582, ThermoFisher, Waltham, MA) and Alexa 594-conjugated goat anti-rabbit IgG (1:400 dilution, product no. A-11037, ThermoFisher). Alexa Fluor 488-conjugated isolectin-B4 (1:100 dilution, product no. 121411, ThermoFisher) was used to identify neuronal C-fibers. Slides were rinsed twice (5 min each) with PBS and Triton X-100 and then once with sterile ultrapure water, sealed with a glass coverslip with mounting medium, and then imaged with a Leica confocal microscope. No staining was seen when a negative control was performed with PBS instead of the primary antibody (data not shown).

Western blot analysis.

Western blots were performed to measure the relative expression of TRPA1 protein in resected T1−T4 DRG and in our DRG cell culture models. Hypoxia-inducible factor (HIF)-1α was also assessed in a cell culture model of tissue ischemia. In preparation for Western blot analysis, tissues were processed using a lysis buffer made from a 100:1 dilution of RIPA buffer and protease inhibitor cocktail (Sigma-Aldrich). This mixture was then homogenized, centrifuged, and stored at −80°C until use.

Protein concentration was measured using the BCA protein assay using known concentrations of BSA as the standard. Protein samples were loaded onto a 10% SDS-PAGE gel with protein standards (Bio-Rad, Hercules, CA) and then transferred to a polyvinylidene difluoride membrane. The membrane was probed for TRPA1 (1:1,000 dilution, product no. NB100-91319, Novus Biologicals) or HIF-1α (1:1,000 dilution, product no. NB100-479, Novus Biologicals) and then normalized against GAPDH (1:1,000 dilution, product no. sc-32233, Santa Cruz Biotechnology, Dallas, TX). Bands were detected by chemiluminescence (SuperSignal West Femto and SuperSignal West Dura, ThermoFisher Scientific) with a UVP BioImaging System (AnalytikJena, Upland, CA). The specificity of this TRPA1 antibody has been prevalidated by the antigen-retrieve experiment (data not shown). The Western blot protocol was similar to those previously described (44, 49).

Real-time quantitative RT-PCR for TRPA1 mRNA in T1−T4 DRG.

Real-time quantitative RT-PCR was performed to determine the expression of TRPA1 mRNA in resected T1−T4 DRG samples. Tissues were lysed, homogenized, resuspended in PBS, and then incubated with 1 µl TRIzol reagent (Invitrogen, Carlsbad, CA). RNA concentration and quality were measured using a Nanodrop (ThermoFisher Scientific).

Reverse transcription was performed using the QuantiTect RT kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Then, 1 µl of each reverse transcribed cDNA product was used for the quantitative PCR. PCR master mix was prepared using 5 µl SYBR green (Life Technologies), 3.8 µl RNase-free water, 0.1 µl forward primer (TRPA1 or β-actin, Integrated DNA Technologies, Coralville, IA), and 0.1 µl reverse primer (TRPA1 or β-actin) per 10-µl reaction. The relative expression of TRPA1 mRNA was normalized against β-actin mRNA using the ΔΔC_t method (where Ct is threshold cycle) and expressed as fold changes ($2^{-\Delta \Delta {\text{C}}_{\text{t}}}$). The following primer sequences were used for TRPA1: forward 5′-CTGTGAAGCGCTGAATGTAATG-3′ and reverse 5′-GCTCCTTGGCTGAGAAGAAA-3′. The following primer sequences were used for β-actin: forward 5′-ACAGGATGCAGAAGGAGATTAC-3′ and reverse 5′-ACAGTGAGGCCAGGATAGA-3′.

Cell culture.

To assess potential molecular mechanisms underlying the alteration in T1−T4 DRG TRPA1 protein expression in CHF rats, we performed several DRG cell culture experiments with the 50B11 DRG cell line. This is a rat cell line and was provided by Ahmet Hoke of Johns Hopkins University. The 50B11 cell culture protocol was based on a previous publication (5). 50B11 DRG cells were maintained in GIBCO Neurobasal Medium (500 ml, Life Technologies) supplemented with 20% glucose (5.5 ml, Life Technologies), 0.2 M l-glutamine (1.37 ml, Life Technologies), FBS (50 ml, Sigma-Aldrich), B-27 supplement (10 ml, Life Technologies), and penicillin-streptomycin (5 ml, Life Technologies) in a 37°C incubator with 5% CO₂. Cells were differentiated with 75 µM forskolin (Sigma-Aldrich) 1 day before starting treatment and then maintained in 100 µM for the duration of the treatment. In the IL-6, IL-1β, and TNF-α treatment experiments, cells were treated with medium containing IL-6 (10 ng/ml, R&D Systems, Minneapolis, MN), IL-1β (10 ng/ml, R&D Systems), or TNF-α (10 ng/ml, R&D Systems), respectively, or a combination of all three cytokines (10 ng/ml each) for 3 days. For each of these cell culture experiments, IL-6, IL-1β, and TNF-α medium was changed daily. In the simulated tissue ischemia experiment, cells were maintained in complete neurobasal medium in a hypoxia chamber (HypOxystation H135, HypOxygen, Frederick, MD) at 1% O₂ for 3 days without any reexposure to atmospheric O₂. After the 3-day treatment, cells were processed using lysis buffer made from a 1:100 dilution of RIPA buffer and protease inhibitor cocktail (Sigma-Aldrich), homogenized, centrifuged at 14,000 g at 4°C for 20 min, and then stored until use at −80°C.

Statistical analysis.

All values are expressed as means ± SE. Statistically significant differences between groups were determined by two-tailed unpaired t-test for parametric data (Tables 1 and 2) and by the Mann-Whitney rank sum test for nonparametric data. The Wilcoxon matched-pair signed-rank test was used to determine significant changes within an animal cohort before and after PSAR and CSAR activation (see Figs. 2 and 3, C–E). P < 0.05 was considered statistically significant. All plots and statistics were performed using SigmaPlot11.0 software.

Fig. 2.

Epicardial application of ally isothiocyanate (AITC). A and B: representative tracings of a sham (A) and chronic heart failure (CHF; B) rat’s reflex to application of 30 mM AITC to the epicardium. C−E: corresponding changes in mean arterial pressure (ΔMAP; C), heart rate (ΔHR; D), and renal sympathetic nerve activity (ΔRSNA; E) for 50 and 300 mM doses. Data are means ± SE. Veh, vehicle; P, pressor response; DP, depressor response; BPM, beats/min; BL, baseline. Animals were 10–12 wk postmyocardial infarction (MI)/sham surgery. *P < 0.05 comparing sham and CHF groups by a Mann-Whitney rank-sum test; #P < 0.05 compared with BL by a Wilcoxon matched-pair signed-rank test.

Fig. 3.

Lung visceral pleura application of ally isothiocyanate (AITC). A and B: representative tracings of a sham (A) and chronic heart failure (CHF; B) rat’s reflex to application of 30 mM AITC to the lung visceral pleura. C−E: corresponding changes in mean arterial pressure (ΔMAP; C), heart rate (ΔHR; D), and renal sympathetic nerve activity (ΔRSNA; E) for 50 and 300 mM doses. Data are means ± SE. Veh, vehicle; P, pressor response; DP, depressor response; BPM, beats/min; BL, baseline. Animals were 10–12 wk postmyocardial infarction (MI)/sham surgery. *P < 0.05 comparing sham and CHF groups by a Mann-Whitney rank-sum test; #P < 0.05 compared with BL by a Wilcoxon matched-pair signed-rank test.

RESULTS

Evaluation of body weight, organ weight, and baseline hemodynamics.

Hemodynamics and organ weights of sham and CHF rats are shown in Table 1. Echocardiographic data are shown in Table 2. MI-induced cardiac dilatation in CHF rats was indicated by increased LV systolic and diastolic diameters and volumes 6 wk post-MI. Furthermore, CHF rats exhibited reduced ejection fraction and fractional shortening compared with sham rats, indicating decreased cardiac systolic function (Table 2 and Fig. 1A). Hemodynamic data collected at the time of the terminal experiments (10–12 wk post-MI) further demonstrated that there was a significant increase in LV end-diastolic pressure in CHF rats compared with sham rats. LV dP/dt_max and dP/dt_min were also significantly lower in CHF rats. Heart weight and lung weight to-body weight ratios were elevated in CHF rats compared with sham rats, suggesting cardiac hypertrophy and substantial pulmonary congestion in the CHF state. Moreover, in rats with CHF, a gross examination revealed a dense scar in the anterior ventricular wall. The mean infarct area was 41.3 ± 1.4% of the LV area (Table 1 and Fig. 1B). No infarcts were identified in sham rats. Pleural fluid and ascites were also found in some CHF rats, whereas none was observed in sham rats.

Fig. 1.

Echocardiography and necropsy evaluation. A: M-mode, long-axis (left), and two-dimensional (right) echocardiography images of the left ventricle (LV) from a sham rat and chronic heart failure (CHF) rat. B: photo of the dissected LV at necropsy from one sham rat and one CHF rat. The outline marks the scar area in CHF.

Stimulation of cardiac spinal afferents with AITC in sham and CHF rats.

We stimulated cardiac spinal afferents by applying AITC directly to the LV epicardium in anesthetized, vagotomized, open-chested sham and CHF rats that were 10–12 wk postsurgery. At baseline (before AITC application), there was no statistically significant difference in MAP between sham and CHF rats in the three experimental dose groups (vehicle, 5 mM AITC, and 30 mM AITC), but HR was significantly lower in CHF rats in the vehicle and 30 mM AITC groups (Table 3). Figure 2 shows the changes in MAP, HR, and RSNA we observed after epicardial application of AITC. At the 5 mM AITC dose, both sham and CHF rats experienced biphasic blood pressure responses that included a pressor phase and a depressor phase. On average, at the 5 mM AITC dose, the pressor response in CHF rats was reduced compared with sham rats, although this reduction did not reach statistical significance, whereas the CHF depressor response was significantly reduced compared with sham rats. At the 30 mM dose, the pressor response was significantly blunted in CHF rats, whereas there was no statistically significant difference between groups in the depressor response (Fig. 2, A–C). HR increased in response to epicardial application of AITC for both sham and CHF rats at both doses, but these increases in HR were significantly blunted in CHF rats (Fig. 2D). Similarly, RSNA increased in both sham and CHF rats in response to epicardial applied AITC; however, this effect was attenuated in CHF rats (Fig. 2E). Application of vehicle alone (DMSO) resulted in near-zero changes in MAP, HR, and RSNA in both sham and CHF rats, although the increase in HR for sham rats, despite a small absolute change, did reach statistical significance (Fig. 2D). Collectively, these data indicate that both sham and CHF rats are sensitive to epicardial AITC application. However, counter to our hypothesis, the measured response to cardiac afferent stimulation with AITC was reduced in CHF rats.

Stimulation of pulmonary spinal afferents with AITC in sham and CHF rats.

Table 4 shows MAP and HR at baseline for the cohorts of sham and CHF rats that underwent lung visceral pleura application of AITC. There was no statistically significant difference in baseline MAP between sham and CHF rats for any dosing group (vehicle, 5 mM AITC, and 30 mM AITC), but HR was significantly reduced in the vehicle and 30 mM CHF group. Figure 3 shows the changes in MAP, HR, and RSNA that we observed in response to AITC application. Rats from both groups exhibited a biphasic change in blood pressure after AITC application. The pressor phase was significantly blunted in CHF rats, whereas the depressor phase was not (Fig. 3, A–C). HR increased in sham rats in response to both doses, an effect that was blunted in CHF rats (Fig. 3D). RSNA increased in sham and CHF rats in response to topical application of AITC to the lung. This increase was significantly blunted in CHF rats (Fig. 3E). Treatment with vehicle alone (DMSO) resulted in near-zero changes in MAP, HR, and RSNA in both sham and CHF animals (Fig. 3, C–E). Taken together, these data suggest, first, that pulmonary spinal afferents are sensitive to the TRPA1 agonist AITC, and, second, that the measured response to lung visceral pleura application of AITC in CHF rats was attenuated compared with sham rats 10–12 wk post-MI/Sham surgery.

TRPA1 expression in T1−T4 DRG of sham and CHF rats.

The physiological response to cardiac and pulmonary afferent stimulation with AITC in CHF rats was blunted compared with sham rats 10–12 wk post-MI/sham surgery. Based on this observation, we hypothesized that CHF rats would have reduced TRPA1 expression in their T1−T4 DRG, which provide sensory afferent innervation for both the heart and lung. We tested this hypothesis by assessing TRPA1 expression in T1−T4 DRG using immunofluorescence, Western blot analysis, and quantitative RT-PCR.

We first performed Western blots to measure TRPA1 protein in T1−T4 DRG in Sham and CHF rats over a time course post-MI/sham surgery (Fig. 4B). There was no statistically significant difference at 1, 3–4, and 6–8 wk post-MI. However, there was a statistically significant reduction of TRPA1 protein in CHF rats 10–12 wk post-MI surgery. Furthermore, TRPA1 mRNA levels in T1−T4 DRG, measured with quantitative RT-PCR, were significantly reduced in CHF rats 10–12 wk postsurgery (Fig. 4C). Finally, we acquired immunofluorescence images of sham and CHF T1−T4 DRG expression of TRPA1, neuronal A-fibers (NF200), and neuronal C fibers (IB-4) (Fig. 4A). Our data demonstrated that 1) TRPA1 colocalized with both A- and C-fiber DRG neurons in sham rats and 2) TRPA1 staining was moderately reduced in CHF rat T1−T4 DRG compared with sham rats.

Fig. 4.

T1−T4 dorsal root ganglia (DRG) transient receptor potential ankyrin 1 (TRPA1) expression in sham and chronic heart failure (CHF) rats. A: immunofluorescence images of T1−T4 DRG in a sham and CHF rat 10 wk postmyocardial infarction (MI)/sham surgery. Arrows in the merged image mark colocalization of TRPA1 and IB4 in two fibers. Triangles mark colocalization of TRPA1 and NF200 in two fibers. B: sham-normalized Western blots for TRPA1 in T1−T4 DRG in sham and MI rats from 1, 3–4, 6–8, and 10–12 wk post-MI. MI size is listed below as a percentage of left ventricular surface ± SE. C: sham-normalized TRPA1/β-actin mRNA levels in T1−T4 DRG. *P < 0.05 by a Mann-Whitney rank-sum test.

DRG cell culture model of inflammation and tissue ischemia.

Cardiac injury, including ischemic injury in CHF, increases circulating inflammatory cytokines including IL-6, IL-1β, and TNF-α (15, 20, 23, 32, 41, 42). We hypothesized that reduction of TRPA1 protein in CHF rats is mediated by inflammation. To provide evidence in support of this hypothesis, we cultured cells from the 50B11 rat DRG cell line and treated them with IL-6, IL-1β, and TNF-α individually and then a cocktail of all three combined. After treatment with the respective cytokines, expression of TRPA1 protein was assessed by Western blot analysis. Treatment with IL-6, IL-1β, and TNF-α individually did not cause a significant reduction in TRPA1 expression (Fig. 5, A–C). However, the combination of all three cytokines did significantly reduce TRPA1 expression (Fig. 5D).

Fig. 5.

Transient receptor potential ankyrin 1 (TRPA1) protein in dorsal root ganglia (DRG) cell culture models of inflammation. A–D: Western blot analysis for TRPA1 from a cell culture of DRG cells (50B11 rat line) under normal conditions (control) and with exposure to the proinflammatory cytokines IL-6 (A), IL-1β (B), and TNF-α (C) or a cocktail of all three combined (D). *P > 0.05 between control and the respective cytokine group for all four cell culture experiments by a Mann-Whitney rank-sum test.

Impaired heart function can lead to inadequate organ tissue perfusion and oxygenation. Thus, we also hypothesized that the reduction of TRPA1 protein we observed in CHF rats could be caused, at least in part, by ischemic tissue conditions. To produce a model of ischemic-like tissue conditions, we exposed 50B11 DRG cells to 1% O₂ for 3 days. Under these cell culture conditions, we observed a significant increase in HIF-1α protein, a marker of cellular hypoxia. We also observed a significant reduction in TRPA1 protein (Fig. 6).

Fig. 6.

Transient receptor potential ankyrin 1 (TRPA1) protein in a dorsal root ganglia (DRG) cell culture model of tissue ischemia. A and B: hypoxia-inducible factor (HIF)-1α (A) and TRPA1 (B) protein expression in a DRG cell culture (50B11 DRG rat cell line) grown under normoxic and simulated tissue ischemia (hypoxic) conditions. *P < 0.05 by a Mann-Whitney rank-sum test.

DISCUSSION

The present study illustrates that pulmonary and cardiac spinal sensory afferents, both previously demonstrated to be sensitive to bradykinin and capsaicin (33, 49, 52), are also sensitive to AITC, a TRPA1 channel agonist. In sham rats, stimulation of cardiac and pulmonary spinal sensory afferents with topical application of AITC resulted in a biphasic arterial blood pressure response, increased HR, and increased RSNA. We hypothesized that the measured physiological response evoked by stimulation of these sensory afferents with AITC would be augmented in CHF rats, because excessive sympathoexcitation characterizes CHF and because previous work from our laboratory indicates an augmented reflex response to capsaicin and bradykinin application in CHF animal models. However, in contrast to our hypothesis, the measured physiological response to topical application of AITC to both the lung and heart surface was blunted in CHF rats. These attenuated responses were associated with reductions in TRPA1 mRNA and protein in T1−T4 DRG.

Persistent and exaggerated sympathoexcitation contributes to CHF disease progression and is mediated through a variety of physiological mechanisms. CHF is associated with an exaggerated exercise pressor reflex (16, 34, 47), enhanced peripheral and central chemoreflex sensitivity (37, 38, 54), depressed baroreflex sensitivity (26, 54), as well as an exaggerated CSAR (49, 52). Over the past two decades, studies from our laboratory (10, 22, 49, 51, 52) have focused on both peripheral and central mechanisms by which the CSAR contributes to the enhanced global sympathetic outflow of CHF. These cardiac spinal afferents are silent in the normal state and were originally considered to be essential pathways only for transmission of cardiac nociception to the central nervous system during myocardial ischemia. However, studies from our laboratory and others (10, 22, 25, 49, 51, 52, 55) demonstrated that stimulation of these afferents by endogenous and exogenous chemicals such as bradykinin and capsaicin results in increased sympathetic outflow, arterial blood pressure, and HR and decreased baroreflex sensitivity. We have also demonstrated that the discharge of cardiac sympathetic afferents is increased and cardiac reflex responses of arterial blood pressure, HR, and RSNA are exaggerated in CHF animals (50–52), suggesting that these afferents become sensitized and are tonically active in CHF. Our study confirmed that the enhanced CSAR contributes to autonomic dysfunction including increased global sympathetic outflow and impaired baroreflex function in the CHF state (10, 22, 47, 49, 51). More recently, we have published new evidence (48, 49) showing that chronic and selective ablation of cardiac TRPV1-positive afferents with resiniferatoxin markedly attenuated cardiac fibrosis, apoptosis, hypertrophy, and cardiac diastolic dysfunction in post-MI rats. These observations suggest a critical role for enhanced input from cardiac spinal afferents in mediating the deleterious cardiac remodeling and cardiac dysfunction in heart failure. Given the fact that the sensitivity of cardiac spinal afferents in response to epicardial application of bradykinin and capsaicin was enhanced in CHF (49, 52), we were surprised to find that sympathetic outflow in response to epicardial application of AITC was depressed in CHF rats. The response to lung visceral pleura application of AITC was also blunted in CHF rats. While the present study data do add to the large body of evidence that illustrates a dysregulated sensory/autonomic environment in CHF, they are unusual in demonstrating a depressed (rather than enhanced) cardiopulmonary afferent activity in response to a specific nociceptor agonist (i.e., TRPA1). These data imply that cardiopulmonary spinal nociceptor channels do not respond to post-MI CHF in a uniform manner. TRPA1 is most widely known as the receptor of mustard oil and other exogenous stimuli. However, being a promiscuous channel, TRPA1 is sensitive to a host of endogenous substances including bradykinin, prostaglandins, and protons (14, 39), all of which are associated with myocardial ischemia and inflammation post-MI. It is also possible that reduction of TRPA1 expression is a protective mechanism initiated in response to an otherwise exaggerated cardiopulmonary spinal afferent activity in the CHF state.

We also found that sham rats exhibit a sympathetic response to TRPA1 agonist stimulation of the lung visceral pleura. This observation is new and has a number of clinical implications. TRPA1 is sensitive to acrolein (3), a chemical found in common pollutants including cigarette smoke and vehicular exhaust. Exposure to such pollutants has been demonstrated to be capable of causing acute changes in blood pressure, HR, and autonomic balance (9, 11, 12, 30). Shanks et al. (33) recently demonstrated that the effects of topical lung pleura application of bradykinin in vagotomized, anesthetized rats are largely recapitulated upon bradykinin inhalation directly into the bronchi. This observation may suggest that topical application of agonist to the lung surface may act as a surrogate assay for lung inhalation-initiated reflexes. Thus, perhaps the acute cardiovascular effects experienced by people exposed to air pollutants are mediated, at least in part, by TRPA1 expressing pulmonary T1−T4 DRG sensory afferents. In addition, our data suggest that these acute cardiovascular events may occur in the absence of preexisting disease. Interestingly, CHF rats had a blunted response to visceral lung pleura application of AITC. These data, in conjunction with our sham rat data, raise the possibility that stimulation of lung spinal afferent TRPA1 channels in healthy (non-CHF) conditions may contribute to autonomic imbalance with a shift toward enhanced sympathetic output but in CHF could possibly lead to reductions in MAP with only modest increases in HR and sympathetic tone.

Whereas the present lung reflex assay emphasized sympathoexcitation, most studies that have examined the physiological response to pulmonary afferent TRPA1 stimulation (13, 18, 19) have reported predominantly parasympathetic responses while implicating the effects of vagal C-fiber stimulation. Acute physiological responses reported in these studies included bronchoconstriction, bradypnea, bradycardia, and hypotension. Methodological differences between the aforementioned studies and the present study likely account for their disparate results. These methodological differences include animal model (mice vs. rats), TRPA1 agonist (acrolein vs. AITC), and route of application (smoked box or intratracheal versus topical visceral pleural application). Yet, perhaps the largest methodological difference is that our in vivo experiments followed bilateral vagotomy, whereas the aforementioned studies largely used animals with intact vagi. However, Lee et al. (19) did perform a bilateral vagotomy as an experimental control, which largely abolished the parasympathetic effects in response to intratracheally administered acrolein. Although interesting, careful examination of their postvagotomy data does reveal modest, transient tachypneic and hypertensive intervals after stimulation with acrolein, possibly representing the effects of sympathetic spinal afferents. This observation may suggest that in vagal-intact animals the parasympathetic effects mask and dominate the sympathetic effects. Alternatively, the recent work of Shanks et al. (33) revealed the sympathetic response to topical lung application of bradykinin and capsaicin in vagotomized rats to be largely recapitulated in rats with intact vagi. This observation may suggest that reflexive responses to lung visceral pleura applied chemical agonists may be predominantly mediated by spinal afferents. Perhaps there are few chemically sensitive vagal sensory afferents at (or near) the lung visceral pleura. A study designed to carefully map spinal and vagal sensory afferents throughout the lung, including the visceral pleura, would facilitate assessment of these possibilities. Most lung sensory afferent mapping studies to date have focused on the vagus. For example, Coleridge and colleagues carried out many studies (7, 28, 31), most in anesthetized, mechanically ventilated, open-chested dogs, that identified the location of the sensory afferents within the lung often by mechanical probing of the lung surface. More recently, Chang et al. (4) used reporter viruses to map two distinct vagal sensory neuron subtypes (P2ry1 and Npy2r) within the mouse lung and found the majority of them to be located beneath the airway smooth muscle layer. Lung spinal sensory afferents have been identified in the lung by a number of studies using retrograde labeling (17, 29, 35). However, in these studies, the labeling solution was usually applied to the lung parenchyma or airways and not specifically to the lung surface.

Cardiac and pulmonary spinal reflexes are initiated at the level of the T1−T4 DRG sensory afferents. Thus, upon discovering an attenuated physiological response to both epicardial and lung surface application of AITC in CHF rats, we hypothesized there to be decreased TRPA1 expression in CHF rat T1−T4 DRG. We measured TRPA1 protein levels using Western blots at 1, 3–4, 6–8, and 10–12 wk post-MI/sham surgery and observed a significant reduction at the 10- to 12-wk time point. Our TRPA1 immunofluorescence data and TRPA1 mRNA data are consistent with the Western blot data, confirming downregulated TRPA1 receptors in T1−T4 DRG 10–12 wk post-MI. Our immunofluorescence data suggested that both NF200-positive A-fiber and IB4-positive C-fiber DRG express TRPA1 receptors. Since the TRPA1 receptor is also sensitive to mechanical stimuli, it is possible that the TRPA1 receptor on cardiopulmonary spinal afferents may mediate some mechanical sensation from lung and heart. This hypothesis needs to be further confirmed by additional studies. We believe that the reduction in TRPA1 protein level in T1−T4 DRG contributed to the markedly attenuated physiological response to lung and heart surface application of AITC in CHF rats 10–12 wk post-MI/sham surgery. It is possible that the reduction in TRPA1 protein in T1-T4 DRG in 10–12 wk post-MI rats was underestimated by Western blot analysis. This is because the cardiopulmonary DRG neurons represent only a fraction of the T1−T4 DRG soma. Therefore, if the noncardiopulmonary portion had normal (sham) levels of TRPA1 protein, the signal within the cardiopulmonary portion would, in effect, be diluted. If this is true, then, by extension, we may infer that the statistically significant reduction in TRPA1 protein 10–12 wk post-MI could be more pronounced if we were able to assess cardiac and pulmonary DRG neurons in isolation. It is also possible that part of the attenuated response in CHF rats was caused by posttranslation modifications: the cysteine residues of TRPA1 NH₂-terminal ankyrin repeat region are known targets of covalent modification (24).

To explore possible molecular mechanisms underlying the decreased TRPA1 protein expression in CHF rat T1−T4 DRG, we performed DRG cell culture experiments investigating the respective effects of proinflammatory cytokines and simulated tissue ischemia on TRPA1 protein expression. We did not observe a reduction in TRPA1 protein after individual treatment with IL-6, IL-1β, or TNF-α. However, we did observe a significant reduction when cells were treated with a cocktail of all three cytokines. This observation increases the likelihood that the CHF proinflammatory environment accounted for the reduced TRPA1 expression in vivo. In the simulated ischemia cell culture model, we observed reduced TRPA1 protein compared with normoxic controls. This finding lends support to the possibility that tissue ischemia in CHF may have contributed to reduced TRPA1 protein expression in cardiopulmonary sensory neurons.

Given the above discussion, there are still some potential limitations in the present study. One is that all experiments were conducted under anesthesia. We acknowledge that anesthetics could have affected the results. However, the experimental data we obtained required anesthesia, and animals from both experimental groups received equal anesthetic treatment. Second, since our study focused on spinal sensory afferents, we were unable to assess the potential contribution of vagal afferents and potential supraspinal effectors including, for example, medullary signal integration and output. Third, we used simple cell culture models of inflammation and tissue ischemia to provide evidence in support of possible explanations behind the reduction of TRPA1 protein we observed in CHF rat T1−T4 DRG. Despite this, we recognize the in vivo events that caused the reduction in TRPA1 expression are likely complex and multifactorial. Fourth, because we used only one agonist (AITC), we are unable to rule out the possibility that the physiologic responses we observed were unique to AITC and could differ for other agonists, even agonists specific to TRPA1 such as cinnamaldehyde. Fifth, the elevated wet lung weight-to-body weight ratio in CHF rats (Table 1) indicates the presence of pulmonary edema. The effect of pulmonary edema, and any CHF-related pulmonary pathology, on the physiological response to topical application of AITC to the lung and heart is unknown and requires additional studies to elucidate. Sixth, we acknowledge that TRPA1 expression in the DRG may not be reflective of protein expression at the afferent nerve terminals in the heart and lung. We measured TRPA1 expression in the soma because direct quantitative measurement of TRPA1 protein expression in afferent terminals is technically unavailable and because all proteins are synthesized in the neuronal cell body (soma) and then delivered to the afferent terminal. Seventh, while bilateral vagotomy was performed in our study, carotid sinus baroreceptors are innervated by the glossopharyngeal nerve and thus, despite vagotomy, remain intact. Hence, intrinsic differences between sham and CHF rat baroreflex may have influenced the results of our in vivo experiments. We know that CHF rats have a blunted baroreflex (26, 54), which engenders an exaggerated sympathoexcitatory response. Given that we observed an attenuated reflex response to lung and heart surface application of AITC in CHF rats, we believe it is likely that even with denervated carotid baroreceptor afferents, the differences we observed between Sham and CHF rats would be maintained. Eighth, similar to most published animal studies to date, the present study was completed using only male rats. However, growing evidence indicates that CHF pathophysiology significantly differs between the sexes, including within the neural-hormonal axis. For example, Yu et al. (53) recently demonstrate that both male and female rats have elevated plasma norepinephrine and plasma arginine vasopressin levels 4 wk post-MI compared with non-MI control rats; however, both are reduced in female MI rats compared with male MI rats. The results presented in the present study may differ for female CHF rats. Finally, the application of agonist was limited to the epicardium and lung visceral pleura. We chose this route of administration because it gave us specificity in afferent targeting. For instance, we know the stimulated epicardial and lung visceral pleura sensory afferents are located at (or near) the surface of their respective organs, because those were the specific locations of agonist application. Specificity of afferent targeting is important, as Shanks et al. (33) demonstrated the response to topical application of capsaicin and bradykinin to the lung visceral pleura to be location dependent. Specificity in sensory afferent targeting is limited in other experimental designs including, for example, the “smoked box” where the precise locations of the transducing afferents are ambiguous (e.g., sinuses, lower airways, skin etc.).

In conclusion, these data illustrate that epicardial and visceral lung pleural spinal sensory afferents are sensitive to TRPA1 agonists and that sympathetic outflow in response to such stimuli is blunted in CHF rats.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants R01-HL-126796-A1 (to H. J. Wang and I. H. Zucker) and R01-HL-121012-A1 (to H. J. Wang).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

I.H.Z. and H.-J.W. conceived and designed research; R.J.A., Z.X., K.P., J.H., A.J.C., I.H.Z., and H.-J.W. performed experiments; R.J.A., Z.X., K.P., J.H., A.J.C., H.D.S., I.H.Z., and H.-J.W. analyzed data; R.J.A., Z.X., K.P., J.H., A.J.C., H.D.S., S.L., I.H.Z., and H.-J.W. interpreted results of experiments; R.J.A., Z.X., K.P., I.H.Z., and H.-J.W. prepared figures; R.J.A., Z.X., K.P., A.J.C., H.D.S., S.L., I.H.Z., and H.-J.W. drafted manuscript; R.J.A., Z.X., K.P., A.J.C., H.D.S., S.L., I.H.Z., and H.-J.W. edited and revised manuscript; R.J.A., Z.X., K.P., J.H., A.J.C., H.D.S., S.L., I.H.Z., and H.-J.W. approved final version of manuscript.