Chronic activation of hypothalamic oxytocin neurons improves cardiac function during left ventricular hypertrophy-induced heart failure

Abstract

Aims

A distinctive hallmark of heart failure (HF) is autonomic imbalance, consisting of increased sympathetic activity, and decreased parasympathetic tone. Recent work suggests that activation of hypothalamic oxytocin (OXT) neurons could improve autonomic balance during HF. We hypothesized that a novel method of chronic selective activation of hypothalamic OXT neurons will improve cardiac function and reduce inflammation and fibrosis in a rat model of HF.

Methods and results

Two groups of male Sprague–Dawley rats underwent trans-ascending aortic constriction (TAC) to induce left ventricular (LV) hypertrophy that progresses to HF. In one TAC group, OXT neurons in the paraventricular nucleus of the hypothalamus were chronically activated by selective expression and activation of excitatory DREADDs receptors with daily injections of clozapine N-oxide (CNO) (TAC + OXT). Two additional age-matched groups received either saline injections (Control) or CNO injections for excitatory DREADDs activation (OXT NORM). Heart rate (HR), LV developed pressure (LVDP), and coronary flow rate were measured in isolated heart experiments. Isoproterenol (0.01 nM–1.0 µM) was administered to evaluate β-adrenergic sensitivity. We found that increases in cellular hypertrophy and myocardial collagen density in TAC were blunted in TAC + OXT animals. Inflammatory cytokine IL-1β expression was more than twice higher in TAC than all other hearts. LVDP, rate pressure product (RPP), contractility, and relaxation were depressed in TAC compared with all other groups. The response of TAC and TAC + OXT hearts to isoproterenol was blunted, with no significant increase in RPP, contractility, or relaxation. However, HR in TAC + OXT animals increased to match Control at higher doses of isoproterenol.

Conclusions

Activation of hypothalamic OXT neurons to elevate parasympathetic tone reduced cellular hypertrophy, levels of IL-1β, and fibrosis during TAC-induced HF in rats. Cardiac contractility parameters were significantly higher in TAC + OXT compared with TAC animals. HR sensitivity, but not contractile sensitivity, to β-adrenergic stimulation was improved in TAC + OXT hearts.

1. Introduction

Heart failure (HF) affects 5.7 million adults in the United States and prevalence is projected to increase 46% in the next 15 years.¹ Approximately 50% of patients diagnosed with HF die within 5 years, necessitating the development of new treatments.¹ A hallmark of HF is elevated cardiac sympathetic activity and parasympathetic withdrawal,^2,3 an imbalance that contributes to ventricular dysfunction, structural remodelling, and electrical instability.⁴ In the initial stages of HF, parasympathetic tone decreases as early as 3 days after the development of cardiac dysfunction, typically preceding increases in sympathetic activity.^5,6 Elevated sympathetic activity is often managed with β-blockers, which alleviate HF symptoms; however, β-blockade does not address the functionally important reduction of cardiac parasympathetic tone that occurs with HF.

Previous studies have demonstrated the benefit of vagal nerve stimulation (VNS) to elevate parasympathetic tone during HF.⁷ Stimulation of the right vagus nerve during HF in rats improved LV function, prevented contraction deficits, reduced ventricular weight, and increased survival.^8,9 In humans, clinical trials with left or right VNS demonstrated improved heart rate (HR) variability and 6-min walk distance in HF patients, but also indicated adverse effects that include dysphonia, cough, and throat pain—likely due to the non-specificity of electrical VNS.^10,11 A disadvantage of using electrical current to activate the vagus nerve is that although efferent cardiac fibres are activated, efferent fibres that innervate non-cardiac visceral organs, as well as sensory afferent fibres, are likely also activated. The efficacy of VNS is also dependent upon proper tuning of the stimulating current amplitude and frequency¹¹ and maximum efficacy might require implantation of cuff electrodes around both the right and left vagus nerve. Approaches that selectively activate only cardiac parasympathetic neurons could be effective without the associated confounding variables and side effects that occur with VNS. One such approach would be to selectively activate cardiac vagal neurons (CVNs) in the dorsal motor nucleus of the vagus to achieve the benefits of left and right VNS without the side effects of electrical stimulation. Demonstrating the efficacy of such an approach in an animal model of HF could form the basis of a new clinical therapy.

CVNs within the brainstem regulate parasympathetic activity to the heart to maintain normal HR and coronary flow.¹² CVNs receive powerful excitation from a population of oxytocin (OXT) neurons within the paraventricular nucleus (PVN) of the hypothalamus^12,13 that co-release OXT and glutamate to excite CVNs. This higher brain centre is responsible for regulating both autonomic function in normal situations and cardiac responses in high-stress conditions.¹⁴ We have recently demonstrated that in rats with LV hypertrophy that progresses to HF, CVNs have diminished excitation due to both an increase in spontaneous inhibitory GABAergic neurotransmission frequency and a decrease in amplitude and frequency of excitatory glutamatergic neurotransmission to CVNs.¹⁵ This finding indicates that augmentation of the excitatory PVN OXT/glutamate pathway to CVNs could be a promising approach to maintain cardiac parasympathetic activity during HF.

OXT is important for maintaining cardiovascular homeostasis and parasympathetic cardiac activity, particularly during anxiety and stress.¹⁶ For example, OXT administration prevented increased HR and diminished HR variability that occurs with social isolation.^17,18 In other studies, rats subjected to daily restraint stress had increased cardiac infarct size and increased incidence of severe arrhythmias during myocardial ischemia–reperfusion,¹⁹ while intra-cerebroventricular administration of OXT, that did not increase plasma OXT levels, reduced the cardiac injury that occurred following episodes of ischemia–reperfusion.²⁰ We have recently demonstrated that chronic activation of PVN OXT neurons in rats restores the release of OXT from PVN fibres, decreases HR and blood pressure (BP), and more importantly, prevents the hypertension that occurs during chronic intermittent hypoxia/hypercapnia.²¹ Overall, this previous work supports our hypothesis that chronic activation of PVN OXT neurons could increase cardiac parasympathetic activity and blunt the progression of cardiac dysfunction during HF.

Our objective was to test whether chronic activation of hypothalamic PVN OXT neurons in a rat model of HF [trans-ascending aortic constriction (TAC)] prevents loss of cardiac contractile function and reduces cardiac inflammation and fibrosis compared with age-matched untreated HF rats (Control). To test this hypothesis, Designer Receptors Specifically Activated by Designer Drugs (DREADDs) were expressed in hypothalamic PVN OXT neurons of all animals. DREADDs were activated with daily injections of clozapine N-oxide (CNO) in the treatment HF group (TAC + OXT) and in a control group (OXT NORM). LV function, LV fibrosis, and the expression level of the inflammatory cytokine interleukin-1β (IL-1β) were assessed via excised perfused heart experiments, histology, and western blot assays. Our results indicate that chronic activation of hypothalamic OXT neurons could be an effective approach to slow the development of cardiac damage and dysfunction that occurs during pressure overload HF.

2. Methods

2.1 Ethical approval

All animal procedures were completed in agreement with the George Washington University institutional guidelines and in compliance with suggestions from the panel of Euthanasia of the American Veterinary Medical Association and the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals.

2.2 Surgical procedure for TAC

Pressure overload induced LV hypertrophy was initiated in male Sprague–Dawley rats using a minimally invasive TAC procedure, similar to our previous studies.¹⁵ Rats at one week of age were aestheticized by hypothermia and underwent TAC surgery. A 0.5 cm incision was made at the level of the chest, the chest was opened and the thymus was retracted to reveal the aorta. A 4-0 silk suture was passed around the ascending aorta, and with a 25-gauge needle temporarily placed adjacent to the aorta, the suture was tied around both the aorta and needle (TAC and TAC + OXT groups). The needle was then removed, leaving the constricting suture around the aorta. Buprenorphine was applied as an analgesic. Successful constriction was confirmed upon examination of the aorta after each animal was sacrificed.

2.3 Selective expression and activation of DREADDs in PVN OXT neurons

Selective activation of PVN OXT neurons was achieved using DREADDs with a highly selective OXT promoter to drive expression of DREADDs in PVN OXT neurons, as previously described.²¹ Our previous work has demonstrated that injections of clozapine-N-oxide (CNO) increases the firing of PVN OXT neurons for at least 1 h.²¹ Our prior work has also shown that selective activation of DREADDs in PVN OXT neurons decreases both mean arterial pressure and HR in telemetry instrumented conscious unrestrained animals.²¹

To ensure robust and highly selective expression in PVN OXT neurons, two viral vectors were used in combination with the Cre-Lox recombination system. In this system one viral vector expresses Cre recombinase under an OXT promotor. The second vector expresses the excitatory hM3Dq DREADDs.²² This is a Cre-dependent vector that has silencing double-floxed inverse open reading frames, which insures expression is only in OXT neurons that selectively express Cre. In each animal, 30–50 nl containing both viral vectors was selectively microinjected into the PVN over a 20 min period at 1 week of age.

2.4 Activation of PVN OXT neurons in vivo

PVN OXT neurons expressing DREADDs were exclusively activated by clozapine-N-oxide (CNO), a molecule that is otherwise biologically inert and does not cross the blood–brain-barrier. TAC + OXT and OXT NORM animals received intraperitoneal (IP) injections of CNO (1 mg/kg) daily, beginning at 5 weeks of age until the animal was sacrificed.

2.5 In vivo assessments of changes in autonomic tone upon PVN OXT neuron activation

Sprague–Dawley rats at 4 weeks of age that had DREADDS expression in PVN OXT neurons were anesthetized (isoflurane) and implanted with a telemetry device (HDX-11, DSI) with the pressure catheter inserted into the descending abdominal aorta to measure BP. The ECG leads of the device were inserted subcutaneously to measure HR. Rats were allowed to recover from this surgery for a week, followed by IP saline injections for 3 days to acclimate the animals to IP injections. BP & HR signals were then recorded 15 min before and 1 h after IP injections of either CNO (1 mg/kg), CNO (1 mg/kg) + atropine (1 mg/kg), or CNO (1 mg/kg) + atropine (1 mg/kg) + atenolol (10 mg/kg). One injection of one of the three solutions was administered each day on consecutive days.

2.6 Ex vivo assessments of cardiac function

At nine weeks of age rats were anesthetized with an IP injection of 0.2 mL Telazol and subject to isoflurane inhalation. Following cessation of pain reflexes, hearts were rapidly excised (n = 26) and Langendorff perfused via the aorta at constant pressure (65 mmHg) and temperature (37 °C) with a Krebs-Henseleit solution containing (in mM) 115 NaCl, 3.3 KCl, 2.0 CaCl₂, 1.4 MgSO₄, 25.0 NaHCO₃, 1.0 KH₂PO₄, 5.0 glucose, and 1.0 lactate. Perfusate was oxygenated with 95% O₂–5% CO₂. After a stabilization period of 10 min, a size five balloon (Harvard Apparatus) was inserted into the LV to measure isovolumic LV developed pressure (LVDP), as we have previously described.^23,24 Diastolic pressure was set to 10 mmHg and LVDP was computed as the difference between systolic and diastolic pressures. HR, LVDP, and coronary flow rate (CFR) were measured for at least 15 min during sinus rhythm. An isoproterenol (I6504, Sigma–Aldrich) dose response protocol was then administered using concentrations of 0.01, 0.1, 1, 10, 100, 1000 nM, and HR, LVDP, and CFR were measured at each concentration once function stabilized. After each study, rate pressure product (RPP), an indirect measure of LV work,²⁵ was calculated as the product of HR and LVDP. Contractility and relaxation were measured as the maximum and minimum values of the first derivative of the LVP waveform, respectively, to assess inotropy and lusitropy.

2.7 Anatomic measurements and histology

In a subset of animals, hearts were excised and cannulated via the aorta, flushed of blood, and weighed after clearing fluid from the chambers. The hearts were cut transversely, halfway between the apex and base, and the thickness of the LV free wall and septum were measured. The tissue was then preserved in 10% formalin for histology. Other hearts from each group were cannulated, flushed with high potassium solution, and then formalin-fixed via retrograde perfusion at constant pressure (60 mmHg). The hearts were cut longitudinally and preserved in 10% formalin for histology. All heart specimens were stained (H&E and Trichrome) to quantify cell size and collagen deposition. Average myocyte transverse cross-sectional (CS) area (µm²) and the amount of fibrosis (µm²) observed in a histological section were measured using ImageJ for at least three transverse slices from each heart.

2.8 Measurement of plasma OXT

At least 2 mL of blood was collected from each rat before heart excision. Blood was centrifuged (2000 g) at 4 °C for 15 min. Supernatant (plasma) was then decanted and stored at -80 °C. Plasma OXT level was then measured for plasma samples from each group (n = 5-6/group) by Cayman Chemicals using the Oxytocin ELISA Kit (Cayman Chemical Company, Ann Arbor, MI).

2.9 Western blotting

Samples of LV myocardium (n = 5–6/group) were flash frozen in liquid nitrogen and stored at -80 °C. For western blot analysis of protein expression, samples were thawed and homogenized using the Qproteome Mammalian Protein Prep Kit (Qiagen) in tubes containing metallic beads. Samples were then centrifuged at 14 000 g for 20 min and protein concentrations were determined by the Pierce BCA Protein Assay (Thermo Scientific). Laemmli buffer (Bio Rad) with 10% 2-mercaptoethanol (Sigma–Aldrich) was added to the samples and samples were heated to 98 °C. Equal protein concentration was loaded into wells containing 4–20% Mini-PROTEAN^® TGX™ Gels (Bio Rad). The samples were then run at 100 V for 1–1.5 h to separate proteins by electrophoresis.

After electrophoresis, the samples were transferred to PVDF membranes for 10 min at 25 V using the Trans-Blot^® Turbo™ Transfer System (Bio Rad). The membranes were blocked with 5% milk for 18 h. The membrane was incubated for 1 h at room temperature with one of three primary antibodies: collagen III (Abcam) 1:2000; interleukin-1β (IL-1β) (Cell Signaling Technology) 1:1000; or GAPDH (Sigma–Aldrich) 1:6000. The membranes were then washed and incubated for 1 h with the HRP-conjugated secondary antibodies anti-rabbit 1:6000 and anti-mouse 1:5000 (Santa Cruz). The membranes were washed again and gels were imaged using the Azure cSeries c600 (VWR). Band intensities for collagen III and IL-1β were measured using ImageJ and normalized to the corresponding band intensity for GAPDH.

2.10 Statistical measurements

Measurements in each group were compared using analysis of variance with Tukey post hoc comparisons (Minitab 17 Statistical Software) to identify significant differences between groups. Myocyte area measurements were analysed using fully nested ANOVA to account for both number of cells and number of animals. The isoproterenol dose response data were analysed using a general linear model (GLM) to evaluate the group effect and the isoproterenol dose effect. Tukey post hoc tests identified pairwise changes between groups at each isoproterenol concentration. Differences were considered significant if P < 0.05.

3. Results

At 8 weeks post-TAC, ex vivo LV function and sensitivity to β-adrenergic stimulation was measured in isovolumic working heart experiments for animals in each group (control n = 5; TAC n = 7; TAC + OXT n = 7; OXT NORM n = 6). Additional hearts were prepared for histological analysis, as described above (Control n = 5; TAC n = 5; TAC + OXT n = 5; OXT NORM n = 3). Average myocyte transverse CS area, fibrosis, and anatomic differences between groups are listed in Table 1.

Table 1.

Average measurements of myocardial tissue structure for each group of animals. Myocyte crossection area and the area of fibrosis were measured from histological sections, such as those shown in Figure 1A. Anatomical measurements (LV and septal thickness) were measured from the heart of each animal. Values are listed as mean ± SE.

	Control	TAC	TAC+OXT	OXT NORM
Myocyte area (µm2)	356 ± 98	713 ± 88a	584 ± 51b	365 ± 90
Fibrosis (µm2)	980 ± 309	7749 ± 1273a	2407 ± 723b	918 ± 112
Body weight (g)	350 ± 27	275 ± 17	298 ± 17	357 ± 24
Heart weight (g)	1.17 ± 0.10	1.62 ± 0.05a	1.75 ± 0.21	2.19 ± 0.17a
LV wall thickness (mm)	3.78 ± 0.25	5.87 ± 0.19a,c	5.53 ± 0.34a,c	3.69 ± 0.10
Septum thickness (mm)	3.75 ± 0.18	4.95 ± 0.23a,b,c	3.99 ± 0.14	3.18 ± 0.12

^aDifferent from Control.

^bDifferent from TAC + OXT.

^cDifferent from OXT NORM.

3.1 Myocyte hypertrophy and myocardial fibrosis

Images of typical H&E and Trichrome histological sections are shown in Figure 1A. Myocyte CS area was 389 ± 20 µm² in Control hearts and was no different in OXT NORM hearts (368 ± 20 µm²; Figure 1B). Myocyte CS area was significantly higher in TAC hearts (720 ± 35 µm²) compared with Control and TAC + OXT (601 ± 23 µm²) hearts (Figure 1B). Hearts from TAC rats also developed significantly more fibrosis (7749 ± 1273 µm²) than hearts from Control, TAC + OXT, and OXT NORM rats (980 ± 309 µm², 2407 ± 723 µm², and 918 ± 112 µm², respectively; Figure 1C).

Figure 1.

Histological analysis of myocyte hypertrophy and fibrosis (n = 6/group, two slides each). (A) Representative H&E stained longitudinal slices of Control, TAC, and TAC + OXT hearts are shown above images (20× magnification) of H&E and Trichrome-stained transverse sections (LV free wall) from each group. Higher collagen content (blue) is evident in the TAC heart. (B) LV myocyte CS area was significantly greater in TAC compared with Control, TAC + OXT, and OXT NORM hearts (P < 0.05). While still smaller than TAC, TAC + OXT myocyte CS area was greater than Control and OXT NORM (P < 0.05). (C) LV collagen content was significantly higher in TAC compared with Control, TAC + OXT, and OXT NORM hearts (P < 0.05).

3.2 Anatomic differences

Typical transverse slices of hearts from animals of each group are shown in Figure 2A. There was no significant difference in body weight between groups at the time of sacrifice (Figure 2B). Heart weight was not significantly different between Control (1.17 ± 0.1 g), TAC (1.62 ± 0.05 g), and TAC + OXT (1.74 ± 0.2 g); however, OXT NORM (2.19 ± 0.2 g) hearts weighed more than Control (P < 0.05; Figure 2C). LV free wall thickness (Figure 2D) was greater than Control and OXT NORM (3.78 ± 0.3 mm, 3.69 ± 0.1 mm, respectively) in both TAC and TAC + OXT animals (5.87 ± 0.2 mm, 5.22 ± 0.3 mm, respectively). The wall thickness of the ventricular septum was significantly greater in TAC (4.95 ± 0.2 mm) than all other groups (Control: 3.75 ± 0.2 mm; TAC + OXT: 3.9 ± 0.2 mm; OXT NORM: 3.18 ± 0.1 mm; Figure 2E).

Figure 2.

PVN OXT neuron activation reduces morphological changes during TAC-induced pressure overload. (A) Representative hearts for Control, TAC, and TAC + OXT hearts and transverse sections for Control, TAC, TAC + OXT, and OXT NORM hearts. The region where the transverse section (‘slice’) was cut is shown. (B) Body weight 8 weeks after TAC was the same between groups. (C) Heart weight was not different between Control (n = 7), TAC (n = 8), and TAC + OXT (n = 9) hearts (ns). (D) The LV free wall was thicker in TAC (n = 9) and TAC + OXT (n = 9) hearts than in Control (n = 7) and OXT NORM (n = 6) hearts. (E) The septal wall was thicker in TAC hearts (n = 9) than Control (n = 7), TAC + OXT (n = 9), and OXT NORM (n = 6) (P < 0.05).

3.3 Blood OXT levels and myocardial levels of collagen III and IL-1β

ELISA measurements of plasma OXT levels indicated no significant difference in plasma OXT between the four groups (P = 0.827; Figure 3A). Expression of proteins integral to inflammation (IL-1β) and fibrosis (collagen III) was measured by western blot assays and compared between groups to reveal a significant increase in IL-1β expression in TAC hearts (Figure 3B). IL-1β expression in TAC hearts was significantly elevated, 1.84 ± 0.4 times higher than Control expression (P = 0.0001). There was no difference in IL-1β expression between Control, TAC + OXT, and OXT NORM hearts (0.50 ± 0.2, and 0.21 ± 0.14 change from Control, respectively). There was no difference in collagen III expression between the groups (Figure 3C); however, TAC and TAC + OXT (2.6 ± 0.6 and 1.95 ± 0.7 times Control) collagen III expression trended greater than that of Control and OXT NORM (1.31 ± 0.5 change from Control).

Figure 3.

OXT neuron activation did not elevate plasma oxytocin levels. OXT neuron activation blunted cardiac levels of IL-1B, thereby reducing inflammation and blunting increased fibrosis. (A) Plasma OXT was not different between any group (ns, Control, TAC, OXT NORM: n = 5; TAC + OXT: n = 6), indicating that beneficial effects observed are due to specific activation of PVN OXT neurons and not global increases in blood OXT. (B) Western blot assays revealed significant elevation of IL-1β in TAC hearts compared with Control, TAC + OXT, and OXT NORM hearts (P = 0.0001, n = 5/group). Representative blots for IL-1β and GAPDH for each group are shown on the right. Band intensities for collagen III and IL-1β were measured using ImageJ and normalized to the corresponding band intensity for GAPDH. (C) Collagen III protein expression was not different between groups; however, collagen III expression in TAC and TAC + OXT hearts trended higher (ns, n = 6/group). Representative blots for collagen III and GAPDH for each group are shown on the right.

3.4 In vivo assessments of autonomic tone

Acute activation of PVN OXT neurons by CNO in awake and conscience animals significantly reduced BP an average of 13 mmHg (from 105.8 ± 1.9 mmHg to 93.4 ± 1.5 mmHg; n = 4; P < 0.001) and significantly reduced HR an average of 56 bpm (from 475.4 ± 15.9 bpm to 419 ± 12.4 bpm; n = 4; P < 0.001; Figure 4). The muscarinic receptor antagonist atropine prevented these responses. Blocking β1 receptors with atenolol had no significant additional effects on the responses to PVN OXT neuron activation (Figure 4C).

Figure 4.

Acute activation of PVN OXT neurons by CNO in awake and conscience animals significantly reduced blood pressure and heart rate. (A) Representative in vivo arterial pressure in Control and DREADDs activated animals. (B) Representative ECG measured from implanted telemetry device in Control and DREADDs activated animals. (C) Acute activation of PVN OXT neurons by CNO significantly reduced both BP and HR (n = 4, P < 0.001). The muscarinic receptor antagonist atropine prevented these responses. Blocking β1 receptors with atenolol had no significant additional effects on the responses to PVN OXT neuron activation.

3.5 Contractile function of excised hearts

The isovolumic contractile function of hearts excised from animals of each group was assessed during normal sinus rhythm, which was the same between groups (Figure 5A). Of note, the LV function of TAC + OXT hearts closely matched that of Control hearts. CFR for TAC + OXT (13.8 ± 2.9 mL/min) and OXT NORM hearts (16 ± 1.9 mL/min) was similar to that of Control hearts (13.3 ± 2.4 mL/min) while CFR trended lower for TAC hearts (8.4 ± 1.1 mL/min), although the difference was not significant (Figure 5B). The LVDP of TAC hearts was significantly lower (52 ± 7 mmHg) than that of Control, TAC + OXT, and OXT NORM hearts (Figure 5C), which maintained average LVDPs of 98 ± 3 mmHg, 126 ± 14 mmHg, and 127 ± 16 mmHg, respectively. RPP, an indirect measure of work, was also significantly lower for TAC hearts (Figure 5D) during sinus rhythm (11 081 ± 1612 mmHg*bpm) compared with Control (27 473 ± 1894 mmHg*bpm), TAC + OXT (26 874 ± 4036 mmHg*bpm), and OXT NORM (26 978 ± 2274 mmHg*bpm) hearts. Average contractility and relaxation (Figure 5F) for TAC hearts was also significantly less (1165 ± 121 mmHg/s and -917 ± 144 mmHg/s) than Control (3383 ± 313 mmHg/s and -2557 ± 416 mmHg/s), TAC + OXT (3124 ± 383 mmHg/s and -2203 ± 231 mmHg/s), and OXT NORM (4631 ± 931 mmHg/s and -2926 ± 397 mmHg/s) hearts. These data clearly reveal improved LV function and potentially unimpaired coronary flow in animals treated with PVN OXT neuron activation.

Figure 5.

Hearts from TAC animals with PVN OXT treatment had significantly improved LV function above that of hearts from untreated TAC animals. (A) Sinus rate in ex vivo hearts was not different among groups (P > 0.05). (B) CFR trended lower in TAC hearts (n = 5) in comparison with Control (n = 5), TAC + OXT (n = 6), and OXT NORM (n = 6) hearts (ns). (C) LVDP was significantly higher in Control (n = 5), TAC + OXT (n = 7), and OXT NORM (n = 6) than TAC (n = 6) hearts (P < 0.05). (D) Rate pressure product (RPP) was significantly depressed in TAC hearts compared with all other groups (P < 0.05). (E) Representative LVDP signals for each group demonstrate the reduced function of untreated TAC hearts. (F) LV contractility and relaxation were significantly greater in Control (n = 5), TAC + OXT (n = 7), and OXT NORM (n = 6) compared with untreated TAC hearts (n = 7) (P < 0.05 for both contractility and relaxation).

3.6 HR sensitivity to β-adrenergic stimulation

Results from the isoproterenol dose response studies are shown in Figures 6 and 7. Significant differences were not detected between Control and OXT NORM in any metric at the highest isoproterenol concentrations; therefore, for clarify of presentation, OXT NORM data are not shown in Figures 6 and 7. Average HR for each group trended higher with increasing isoproterenol concentration and TAC + OXT hearts responded with the greatest increase in HR (P < 0.05; Figure 6A). Of note, baseline (no isoproterenol) HR for TAC + OXT hearts matched that of TAC hearts yet HR increased in TAC + OXT hearts with increasing isoproterenol concentration to match that of Control hearts at the highest isoproterenol concentration (1000 nM).

Figure 6.

Isoproterenol dose–response curves reveal that hearts from TAC + OXT animals had improved heart rate response to β-adrenergic sensitivity. Asterisks indicate significant differences between all groups at specific isoproterenol dose (GLM and Tukey pairwise analysis). (A) HR was significantly different between all groups and between isoproterenol doses (P < 0.05, GLM and Tukey pairwise analysis). On average, HR increased in all hearts with increased isoproterenol concentration. HR was greater in control (n = 5) and TAC + OXT (n = 7) hearts and was lower in untreated TAC (n = 7) hearts for all concentrations. (B) CFR was not different between control (n = 5) and TAC + OXT (n = 5) hearts, but was significantly lower in untreated TAC hearts (n = 5) (GLM and Tukey pairwise comparisons, P < 0.05). Control hearts exhibited substantial vasodilation with increasing isoproterenol concentration. CFR for TAC + OXT hearts was consistently higher than that of TAC hearts, indicating improved coronary perfusion. (C) On average, RPP increased in all hearts with increased isoproterenol concentration and was significantly different between all groups (P < 0.05, GLM and Tukey pairwise analysis). RPP for TAC + OXT (n = 7) and Control (n = 5) hearts was similar at low isoproterenol concentrations but was lower for TAC + OXT hearts at high concentration (P < 0.05).

Figure 7.

Hearts from TAC + OXT animals had higher contractility and relaxation; however, contractility and relaxation did not significantly increase with increasing concentrations of isoproterenol. Contractility and relaxation dose–response curves were significantly different between the three groups (P < 0.05, GLM and Tukey pairwise analysis). Asterisks indicate significant differences between all groups at specific isoproterenol dose (GLM and Tukey pairwise analysis). Top: Control (n = 5) and TAC + OXT (n = 7) contractility were similar at isoproterenol concentrations of 1 nM and less. Above 1 nM, the contractility of TAC + OXT hearts did not increase with increasing concentration. The contractility of TAC hearts (n = 7) remained low for all concentrations. Bottom: Control (n = 5) and TAC + OXT (n = 7) relaxation were similar at isoproterenol concentrations of 1nM and less. Above 1 nM, the relaxation of TAC + OXT hearts did not increase with increasing concentration while that of control hearts increased dramatically. The relaxation of TAC hearts (n = 7) remained low for all concentrations.

3.7 LV contractile sensitivity to β-adrenergic stimulation

CFR remained low (P < 0.05) in TAC hearts as isoproterenol concentration increased (Figure 6B), maintaining an average of 8.4 ± 0.8 mL/min for all concentrations. CFR for TAC + OXT hearts was higher than that of TAC hearts at baseline (14 ± 2.9 mL/min) and increased to (15 ± 1.9 mL/min) at the highest isoproterenol concentration. The CFR of Control hearts increased from 13.3 ± 2.4 mL/min at baseline to 25.3 ± 3.7 mL/min at the highest concentration, exhibiting significant CFR reserve. The mean CFR for Control and TAC + OXT hearts was not different, and CFR in both Control and TAC + OXT hearts was significantly higher than TAC hearts (GLM, P < 0.05). The RPP response to increased isoproterenol concentration for Control and TAC + OXT hearts was similar for isoproterenol concentrations from baseline until a concentration of 1 nM (Figure 6C). Control RPP increased to 52 741 ± 14 328 mmHg*bpm and TAC + OXT RPP increased to 38 017 ± 4303 mmHg*bpm at the highest isoproterenol concentration. In contrast, TAC hearts only increased to 21 787 ± 3998 mmHg*bpm at the highest concentration.

Changes in LV contractility and relaxation with increasing isoproterenol concentration are shown in Figure 7. Contractility and relaxation were similar in Control and TAC + OXT hearts for isoproterenol concentrations between baseline and 1 nM. At the higher concentrations, the contractility of Control hearts increased to 7164 ± 788 mmHg/s at the highest isoproterenol concentration. Relaxation of control hearts dropped to -5402 ± 484 mmHg/s at the highest concentration. Although TAC + OXT hearts initially matched controls, correspondence was lost at concentrations above 1 nM. At the highest concentration, TAC + OXT hearts only reached contractility values of 3936 ± 589 mmHg/s and relaxation values of -3258 ± 495 mmHg/s. However, throughout the protocol TAC + OXT contractility and relaxation remained elevated over TAC hearts, which only reached a maximum contractility rate of 2322 ± 364 mmHg/s and minimum relaxation rate of -2023 ± 361 mmHg/s (Figure 7).

4. Discussion

These are the first studies to evaluate the effect of chronic activation of PVN OXT neurons in an animal model of pressure overload induced hypertrophy that progresses to HF. These studies also demonstrate the first use of DREADDs to modulate autonomic activity with the goal of mitigating the deleterious effects of cardiac pressure overload. It is generally accepted that parasympathetic tone is cardioprotective²⁶ and our studies further demonstrate that chronic activation of PVN OXT neurons confers significant cardioprotection during TAC in rats. Taken together, our results support the claim that PVN OXT neuron activation could be a novel target for elevating cardiac parasympathetic tone to alleviate the damaging effects of pressure overload induced hypertrophy.

In ex vivo isovolumic contracting heart studies, we found that, compared with diseased (TAC) animals, the hearts from animals treated with chronic PVN OXT neuron activation (TAC + OXT) developed higher pressures and had greater contractility and relaxation kinetics. Myocyte CS area was lower and the amount of fibrosis observed in LV tissue histology sections was also lower. These results are consistent with a study of HF following myocardial infarction in which administration of an acetylcholinesterase inhibitor, to increase the duration of muscarinic receptor activation by acetylcholine, improved contractility and reduced myocyte hypertrophy and collagen deposition.²⁷ In our studies, we also observed that PVN OXT neuron activation during TAC additionally improved cardiac chronotropic response to isoproterenol (Figure 6).

OXT has been shown to buffer cardiovascular responses to stress and promote cardiac healing by increasing cardiac parasympathetic tone and reducing cardiac sympathetic activation. However, the specific autonomic mechanism behind the cardioprotective effect of PVN OXT activation in HF is unknown. To better understand this, in vivo BP and HR responses were recorded in DREADDs expressing rats. In those studies we found that acute activation of DREADDs in PVN OXT neurons reduced both HR and BP, in accordance with our previously published results.²¹ We have extended these findings to demonstrate that the beneficial effects observed with PVN OXT neuron activation are likely the result of cardiac-specific increases in cholinergic activity. This is supported by the observations that plasma OXT was not different between groups and reductions in BP and HR following DREADDs activation with CNO were completely blocked by atropine (Figures 3A and 4C).

Our finding that LV function was significantly improved in TAC + OXT animals after 8 weeks of TAC indicates a sustained beneficial shift in cardiac autonomic balance. In addition to the down-stream effects of directly increasing parasympathetic activity, PVN OXT neuron activation may have initiated a parasympathetic mediated reduction in sympathetic activity. Interactions between adrenergic and cholinergic pathways are complex, with multiple factors responsible for activation and antagonism. In the sinus node, cholinergic activation dominates the control of HR over that of adrenergic activation.²⁸ In ventricular myocytes, it is generally accepted that M2 muscarinic receptor activation attenuates the production of cyclic AMP to reduce the inotropic effects of β-adrenergic receptor activation.^29,30 During chronic sympathetic stimulation, M2 activation reduces myocardial stress by lowering cyclic AMP to reduce the cellular hypercontractile state and increase relaxation during diastole, thereby improving myocyte viability³¹ and slowing the progression of hypertrophy.^32,33 Vagal tone and circulating acetylcholine also maintain beneficial dilation of coronary arteries,^34,35 an effect that has been shown to be independent of left ventricular (LV) pre-load, afterload, and HR.³⁶ Other studies have shown that parasympathetic nerve stimulation releases endogenous vasoactive intestinal peptide to cause vasodilation and increased coronary flow.³³ Each of these cholinergic-induced outcomes could be cardioprotective during chronic pressure overload and mediated by the activation of PVN OXT neurons.

LV function was dramatically impaired in untreated TAC animals. The high level of fibrosis and elevated collagen III expression we observed in untreated hearts increases wall stiffness, adversely affects contractile function, and negatively impacts vasodilation due to increased myocardial stiffness and increased perivascular collagen.^37,38 Reduced vasodilation within the context of pressure overload, which increases myocardial oxygen consumption, creates conditions of ischemia, causing further myocardial damage and inflammation. Indeed, we found elevated levels of the inflammatory cytokine IL-1β in TAC animals. As myocytes die, they are replaced by collagen to maintain structural integrity in the absence of cells,³⁷ which reduces working myocardial mass. Overall, the result of this detrimental cascade was observed in our TAC animals as reduced contractile function, low coronary flow, and a high level of fibrosis.

Our results are consistent with the conclusion that improved autonomic balance attenuated the loss of cardiac function in the TAC + OXT animals. Although LV hypertrophy was not significantly lower in TAC + OXT animals, likely to compensate for TAC-induced pressure overload, myocyte CS area was less, fibrosis was less, and there was a lower level of IL-1β expression. The ratio of working myocardium to wall thickness was therefore higher in TAC + OXT than in TAC animals, which likely contributed to the impressive maintenance of LV function that we observed. Improved coronary flow likely augmented the increased myocardial oxygen demand of pressure overload, thereby reducing the incidence of ischemia, preventing myocardial necrosis, preventing the loss of working myocardium, and blunting the progression of myocyte hypertrophy. The interesting finding in TAC + OXT hearts of reduced fibrosis measured via Trichrome staining, yet no significant reduction in collagen III expression, is likely due to increased in perivascular collagen. Increased perivascular collagen would not have been detected in the myocardial fibrosis assessments we conducted using Trichrome-stained myocardial slices. However, increased perivascular collagen would elevate the level of collagen III measured in the western blot assays. Overall, the beneficial results of PVN OXT neuron activation are consistent with the results reported for VNS during HF in rats⁹ as well as systemic treatment of rats that had infarction-induced HF with an acetylcholinesterase inhibitor.²⁷

We found that PVN OXT neuron activation in TAC animals partially blunts sinus node desensitization to β-adrenergic stimulation. The result that TAC + OXT animals had improved HR sensitivity to β-adrenergic stimulation, but no improvement in contractile sensitivity, provides new insight into how the impact of heterogeneous cardiac nerve density may affect the outcomes of parasympathetic nerve activation. The dense innervation of cholinergic nerve fibres in the atria and sinus node, as well as greater expression of M2 receptors in the atria compared with the ventricles,^39,40 at least partially explains the observed differences in HR and LV contractile response of TAC + OXT animals to adrenergic stimulation (Figures 6 and 7).

Although the contractility and relaxation kinetics of hearts from TAC + OXT animals, for all isoproterenol concentrations, was greater than that of hearts from TAC animals, it is intriguing that there was almost no increase in these values as isoproterenol concentration increased. Indeed, cholinergic nerve fibres and muscarinic receptors are found in the ventricles of many species, including rodents.^41–43 One explanation for the observed result could be a lower ratio of ventricular parasympathetic to sympathetic innervation and a lower expression of M2 receptors in the ventricles compared with β-receptors.⁴² The ratio of cholinergic to adrenergic innervation is close to 2:1 in the atria and 1:2 in the ventricles.⁴⁴ Within the context of this intrinsic anatomic imbalance between cholinergic and adrenergic activation of the LV, and the elevation of sympathetic tone that occurs during chronic pressure overload, our results indicate that PVN OXT neuron activation is not enough to halt ventricular adrenergic desensitization.

4.1 Limitations

The unavoidable shortcomings of ex vivo perfused heart experiments apply, which include perfusion with a crystalloid perfusate instead of blood and the non-physiologic working condition of isovolumic contraction. Additional studies are required to fully characterize the initial promising results presented here, which would include longitudinal in vivo assessments of improved cardiac function and experiments to identify the optimal time during the initial stages of myocardial hypertrophy to begin treatment via PVN OXT neuron activation.

5. Conclusions

Using DREADDs for the first time in a rat TAC model of ventricular hypertrophy that progresses to HF, we found that chronic activation of PVN OXT neurons, beginning 4 weeks after TAC, significantly improved LV function, including inotropy and lusitropy, and reduced cellular hypertrophy, fibrosis, and inflammation at 8 weeks after TAC. PVN OXT neuron activation also improved HR sensitivity to β-adrenergic stimulation but did not improve contractile sensitivity to β-adrenergic stimulation. Our results indicate that the selective activation of hypothalamic PVN OXT neurons could be an effective approach to counteract the loss of cardiac function and mitigate myocardial damage during pressure overload hypertrophy.