Interference with PPARγ in Vascular Smooth Muscle Causes Baroreflex Impairment and Autonomic Dysfunction

Abstract

S-P467L mice expressing dominant negative Peroxisome Proliferator Activated Receptor-γ selectively in vascular smooth muscle, exhibit impaired vasodilation, augmented vasoconstriction, hypertension and tachycardia. We hypothesized that tachycardia in S-P467L mice is due to baroreflex dysfunction. S-P467L mice displayed increased sympathetic traffic to the heart and decreased baroreflex gain and effectiveness. Carotid arteries exhibited inward remodeling but no changes in distensibility or stress/strain. Aortic depressor nerve activity in response to increased arterial pressure was blunted in S-P467L mice. However, the arterial pressure and heart rate responses to aortic depressor nerve stimulation were unaltered in S-P467L mice suggesting the central and efferent limbs of the baroreflex arc remain intact. There was no transgene expression in nodose ganglion, and no change in expression of the acid-sensing ion channel-2 or -3 in nodose ganglion. There was a trend for decreased expression of transient receptor potential vanilloid-1 receptor mRNA in nodose ganglion, but no difference in the immunochemical staining of transient receptor potential vanilloid-1 receptor in the termination area of the left aortic depressor nerve in S-P467L mice. Although there was no difference in the maximal calcium response to capsaicin in cultured nodose neurons from S-P467L mice, there was decreased desensitization of transient receptor potential vanilloid-1 receptor channels. In conclusion, S-P467L mice exhibit baroreflex dysfunction due to a defect in the afferent limb of the baroreflex arc caused by impaired vascular function, altered vascular structure, or compromised neurovascular coupling. These findings implicate vascular smooth muscle Peroxisome Proliferator Activated Receptor-γ as a critical determinant of neurovascular signaling.

Introduction

Peroxisome proliferator-activated receptor-gamma (PPARγ) is a ligand-activated transcription factor belonging to the nuclear receptor superfamily. Classically, PPARγ plays an important role in adipogenesis and metabolic processes but has recently emerged as a crucial element in cardiovascular diseases such as hypertension. Mutations of PPARγ lead to hypertension in humans and in experimental animal models.^1,2 Moreover, thiazolidinediones, high affinity PPARγ agonists, which increase insulin sensitivity, have been shown to be beneficial for lowering arterial pressure (AP) and improving vascular function.³

We have previously shown that mice carrying a dominant-negative form of PPARγ specifically in smooth muscle (S-P467L mice) exhibit hypertension, severe aortic dysfunction and hypertrophy of cerebral arterioles.⁴ Vascular dysfunction occurs because of impaired expression of the PPARγ target gene Regulator of G-protein signaling 5 (RGS5) leading to enhanced myogenic tone and angiotensin-II-mediated contraction in the mesenteric circulation, and impaired function of the Cullin-3, Cullin Ring Ligase (CRL) pathway leading to increased RhoA and Rho kinase activity in aorta.^5,6 Interestingly, these transgenic mice also exhibited robust tachycardia despite increased systemic AP, suggesting that interference with vascular smooth muscle PPARγ may impair the baroreflex.⁴

The baroreflex is an important beat-to-beat regulator of AP and heart rate (HR). Baroreceptors, located in the adventitia of the aortic arch and carotid sinus, provide a highly sensitive system that detects changes in AP. These receptors send afferent signals to the central nervous system, and the efferent innervation of blood vessels, heart, and kidneys make adjustments to maintain normal AP.⁷ Clinical and experimental hypertension commonly involves cardiovascular autonomic impairment associated with high sympathetic activity and blunted baroreflex sensitivity with baroreceptor resetting.⁸ In this study, we hypothesize that the tachycardia observed in S-P467L mice is attributable to cardiovascular autonomic impairment and baroreflex dysfunction. Our data suggest that impairment in PPARγ function in smooth muscle can impair neurovascular coupling which can attenuate afferent signaling to the central nervous system.

Materials and Methods

A detailed description of the transgenic mouse model, baseline cardiovascular measures, assessment of baroreflex sensitivity and efficacy, spectral analysis, mechanical properties of carotid artery, primary cultures of mouse nodose ganglia (NG), neurovascular coupling including calcium imaging as a measure of transient receptor potential vanilloid 1 receptor channels (Trpv1) functional properties, molecular measures, and data and statistical analysis are provided in the online-only data supplement. All studies were approved by the Institutional Animal Care and Use Committee in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Results

We first measured AP and HR with radiotelemetry. S-P467L mice exhibited elevated AP (Figure 1A) and were markedly tachycardic (Figure 1B). Intrinsic HR, measured after simultaneous sympathetic and parasympathetic blockade, was elevated in S-P467L mice (Figure 1B). Consistent with increased sympathetic outflow to the heart, we observed a larger bradycardic response in S-P467L mice after administering the β-adrenergic antagonist propranolol (Figure 1C). Similarly, the methyl-atropine induced tachycardic response was significantly smaller in S-P467L mice. Power spectral analysis revealed comparable relative low frequency and high frequency components between groups (Figure S1A). The low frequency component of HR (Figure S1B) and AP (Figure S1C) predominated in S-P467L mice. S-P467L mice also exhibited increased systolic AP variability (Figure S1D). Applying the sequence method to examine spontaneous baroreflex fluctuations of AP revealed no changes in the number of sequences (Figure S1E), but S-P467L mice exhibited decreases in both baroreflex gain (Figure S1F) and effectiveness (Figure S1G).

Figure 1. Arterial Pressure and Heart Rate.

Measurement of systolic arterial pressure (SAP), mean arterial pressure (MAP), diastolic arterial pressure (DAP), heart rate (HR) and intrinsic heart rate (iHR) using radiotelemetry. A) Average 1-hour recording of SAP, MAP and DAP. B) Average HR over a 1-hour period calculated from the AP recordings and iHR calculated from the AP recordings after simultaneous blockage of sympathetic and vagal tone to the heart with propranolol and methyl-atropine, respectively. C) HR response to intraperitoneal injection of methyl-atropine (ATR) and propranolol (PROP). S-P467L (black) and non-transgenic (gray) mice. *, P<0.05 vs NT (n=8).

To assess whether the impairment in baroreflex function is due to alterations in vessel structure and/or function, the mechanical properties of carotid arteries were determined in the absence of vascular tone. The lumen diameter was significantly reduced in S-P467L carotid arteries across all pressures tested (Figure 2A). However, there were no differences in vascular distensibility (Figure 2B), and although the stress-strain curve was shifted to the right in S-P467L carotid arteries, the slope was not statistically different between the groups (Figure 2C, NT: 2.61±0.03, S-P467L: 2.38±0.11, p>0.05), suggesting a similar stiffness.

Figure 2. Vascular Mechanics.

A) The passive diameter of the carotid artery was assessed under calcium-free conditions in response to increases in intraluminal pressure. B) Distensibility was calculated from the pressure diameter relationship. S-P467L (N=6); NT (N=5); *P<0.05 vs NT. C) A stress/strain curve was calculated from the internal and external diameter of the vessels that were exposed to a series of pressure steps from 0, 10, 20, 30, 50, 75, 100, 125 to 150 mmHg. S-P467L (N=4); NT (N=3).

Since baroreflex dysfunction in S-P467L mice could originate from either the afferent or central (efferent) limb, we analyzed AP and HR in conjunction with aortic depressor nerve (ADN) activity at baseline and in response to various stimuli. We first examined the HR response to phenylephrine (PE) and sodium nitroprusside (SNP) in NT and S-P467L mice under conscious conditions. S-P467L mice exhibited tachycardia at baseline (621±18 vs 527±26 bpm, P<0.01), and moderately elevated systolic AP (149.3±10.1 vs 139.5±8.4 mmHg). S-P467L and NT mice exhibited equivalent pressor and depressor responses to PE and SNP, respectively (Figure 3A). Whereas both S-P467L and NT mice exhibited similar tachycardic responses to SNP, the bradycardic response to PE was significantly blunted in S-P467L mice (Figure 3B).

Figure 3. BP and HR Response to PE and SNP.

The change in systolic BP (A) and HR (B) in conscious S-P467L and NT mice is shown in response to slow intravenous infusion of phenylephrine (PE, 5µg) followed by sodium nitroprusside (SNP, 5µg).

We next measured ADN activity in NT and S-P467L mice under anesthesia. ADN activity occurred in bursts and was synchronous with the arterial pressure pulse in both NT and S-P467L mice (Figure S2, Figure 4A–B).⁹ There was no difference in baseline ADN activity comparing S-P467L (95±13 spikes/sec) with NT (97±9 spikes/sec) mice. The ADN response to SNP was comparable between NT and S-P467L mice (Figure 4C–D). ADN activity rose abruptly in response to the pressor effects of PE in both groups, but the magnitude of the response was markedly blunted in S-P467L mice (Figure 4C–D).

Figure 4. Relationship Between Arterial Pressure, HR and Aortic Depressor Nerve Activity.

Representative tracings of blood pressure and ADN activity in anesthetized NT (A) and S-P467L (B) mice. The change in ADN activity (C–D) is shown in response to changes in arterial pressure induced by infusion of sodium nitroprusside (SNP) followed by phenylephrine (PE). S-P467L (black) and non-transgenic (gray) mice. NT (N=10); S-P467L (N=8); *, P<0.05 vs NT.

The sigmoidal fitting ADN-AP curve is shown in Figure 5. The baroreflex range, gain, and saturation pressure (MAPsat) were all significantly decreased in S-P467L mice, whereas the MAP at 50% of ADN activity (MAP50) and threshold pressure (MAPth) were not different between groups (Table 1). To examine the central integration of the afferent signal, we directly stimulated the ADN and monitored changes in AP and HR. Unlike the impaired response in S-P467L induced by acute increase in AP, the systolic AP and HR responses to ADN stimulation exhibited no differences between groups (Figure 6).

Figure 5. Baroreflex Curves.

Analysis of baroreceptor afferent function curves derived from responses to intravenous injections of sodium nitroprusside and phenylephrine in S-P467L (filled circles) and NT (open circles) mice. ADN, aortic depressor nerve; MAP, mean arterial pressure.

Table 1.

Baroreflex Parameters

Parameter	NT	S-P467L
Range (mmHg)	114 ± 4	85 ± 4*
Gain (%max)	1.69 ± 0.08	1.26 ± 0.05*
MAP50 (mmHg)	101 ± 2.4	95 ± 2
MAPth (mmHg)	45 ± 2	46 ± 4.9
MAPsat (mmHg)	172 ± 5.3	156 ± 3.8*

Tabulated data derived from individual curves. NT (N=10); S-P467L (N=8);

^*, P<0.05 vs NT.

Figure 6. HR and AP Responses to ADN Stimulation.

Reflex decreases in systolic arterial pressure (ΔSAP, A) and heart rate (ΔHR, B) in response to electric stimulation of the aortic depressor nerve (ADN) in anesthetized S-P467L (filled circles) and NT (open circles) mice. N=9 per group.

We next determined if defects in the nodose ganglia (NG) contribute to the impaired baroreflex sensitivity in S-P467L mice. Since PPARγ expression was recently identified in nodose neurons, we first assessed whether the smooth muscle promoter-driven transgene was expressed in nodose ganglia (NG) in S-P467L mice.¹⁰ No signal was detected in either aorta or NG of NT mice. Although a signal (Ct value ~28) was detected in NG from S-P467L mice, it was expressed approximately 500-fold less than in aorta (Figure S3A). There was no significant change in expression of the PPARγ target genes CD36 or aP2 in NG from P467L compared with NT (data not shown). Previous studies reported that the acid-sensing ion channel 2 (ASIC2) is expressed in aortic baroreceptor neurons and that Asic2-deficient mice exhibit baroreflex dysfunction.¹¹ However, there was no difference in the level of expression of either Asic2 (Figure S3B) or Asic3 (Figure S3C) mRNA in the NG of S-P467L mice. Transient receptor potential vanilloid-1 (Trpv1) receptor channels have been implicated in baroreflex function,¹² but there was no detectable expression of Trpv1 mRNA in the aorta. Initial investigations suggested a decrease in mRNA encoding Trpv1 in NG from S-P467L mice, but ultimately, this trend was not significant (Figure S3D). Based on our early finding, we assessed if there was histological evidence to support a defect in neural vascular coupling at the aorta arch. No obvious differences were observed between NT and S-P467L mice in the immunochemical staining pattern of Trpv1 in the termination area of the left ADN (Figure 7A–F). In all specimens, both thick and thin fibers could be seen arising from the ADN in the aortic wall and bouton-like swellings and heavily branched terminal arrays similar to those reported elsewhere.¹³ The mean numbers of grid intersections were 18.9±15.5 and 18.6±10.7 (NT vs S-P467L, t=0.237, p=1.0) for large fibers and 6.57±7.9 and 9.2±7.81 (NT vs S-P467L, t=0.851, p=0.4) indicating there were no difference between the mean number of large fiber intersections and only a modest difference for fine fiber intersections.

Figure 7. Neurovascular Coupling and Function of Trpv1 Channels in Nodose Ganglia Neurons.

Fibers and terminals of the aortic depressor nerve (ADN) on the aortic arch (aa) immunochemically stained with an antibody against Trpv1 in non-transgenic (NT) and S-P467 transgenic mice. A) Low power micrograph from a whole mount preparation from a NT animal showing the ADN innervating the region of the aortic arch at the junction with the left common carotid artery (cc). B) Similar preparation as in A from a S-P467L mouse. Note the similar appearance of the ADN and its arborizations in both mice. C–D) Complex endings of Trpv1 positive ADN axons in the vicinity of the aortic arch-left carotid junction. E–F) Arborizations of thin Trpv1 positive axons on the anterior wall of the aortic arch. Scale bars in A and B=250μm; scale bars in C–F=25μm. G) Representative traces of Ca²⁺ imaging in response to two successive capsaicin applications (100 nM, 15 sec), followed by KCl application (50 mM, 30 sec) in NG neurons. H) Maximal capsaicin-induced [Ca²⁺]_i responses normalized to KCl. I) Magnitude of Trpv1 desensitization, expressed as the ratio of 2^nd vs 1^st capsaicin-induced peak [Ca²⁺]_i in the same neuron. N=38 for NT and N=58 for S-P467L. *, P<0.05 vs NT.

Finally, we examined if there was any change in the number of functional Trpv1 channels in the NG neurons by monitoring Trpv1-mediated increase in intracellular Ca²⁺ ([Ca²⁺]_i) responses (Figure 7G).¹⁴ No significant change was observed in the magnitude of capsaicin-induced [Ca²⁺]_i rise, normalized to the KCl-induced [Ca²⁺]_i rise in NG neurons of either genotype (Figure 7H), indicating no change in the levels of functional Trpv1 channel in these neurons. Interestingly, there was a significant decrease in the magnitude of Trpv1 channel desensitization in the NG neurons from S-P467L mice (Figure 7I).

Discussion

There is emerging evidence that the ligand activated transcription factor PPARγ, best known for its regulation of adipogenesis, lipid metabolism and insulin sensitivity, plays autocrine roles in tissues outside of its classical sites of action (e.g. liver, skeletal muscle and adipose tissue). Endothelial PPARγ-deficiency (through gene ablation) or PPARγ-interference (employing a dominant negative mutant) results in impaired NO-mediated vasodilatation and increased AP in response to a high fat diet.^15–17 Smooth muscle PPAR-γdeficiency blunts the protective effects of thiazolidinediones on atherosclerosis and augments angiotensin II-induced oxidative stress, inflammation and vascular remodeling.^18,19 Selective vascular smooth muscle interference with PPARγ causes a loss of NO-responsiveness, an increase in vasoconstriction and hypertension.⁴

Despite mild hypertension, S-P467L mice exhibit robust tachycardia suggesting impaired autonomic function.⁴ S-P467L mice exhibited a blunted tachycardic response to vagal blockade and a markedly augmented bradycardic response to sympathetic blockade. The direct measurements of HR in response to pharmacological blockade suggest there is increased sympathetic activity to the heart. This conclusion was further supported by spectral analysis, although it should be noted that the application and usefulness of spectral analysis remains controversial.

HR can be controlled by baroreceptors and chemoreceptors on free nerve endings in the aortic arch and carotid sinus, which transmit afferent signals to the brain in response to the local stretch of the vessel or changes in oxygen or pH (reviewed in^20,21). Theoretically, a lesion in any portion of the baroreflex arc can lead to baroreflex impairment. That the ADN response to an increase in AP was severely impaired in S-P467L mice suggests impairment in the afferent limb of the baroreflex arc. Because the HR response to direct ADN stimulation was normal, suggests that the central and efferent limbs of the baroreflex are functioning normally. This stands in contrast with other models of hypertension including the spontaneously hypertensive rat (SHR), DOCA-salt-treated rat and obese Zucker rat, where the depressor response elicited by direct stimulation of ADN was attenuated suggesting a defect in the central component of the baroreflex arc.^22–24 Unlike the SHR, which exhibits marked hypertension, the S-P467L mouse model displays only a small increase in arterial blood pressure but profound baroreflex dysfunction. Therefore, it is unlikely that the baroreflex dysfunction observed in S-P467L mice is due to increased arterial pressure.

The mechanical properties of the vascular wall and substances released by vascular cells have major influence on baroreceptor activity. For example, impaired release of prostacyclin, increased generation of reactive oxygen species and platelet aggregation have been shown to reduce baroreceptor activity during hypertension and atherosclerosis.²⁵ ApoE-deficient mice were reported to exhibit both hypertension and tachycardia suggesting that hypercholesterolemia, atherosclerosis and/or vascular structure and function may affect baroreflex regulation.²⁶ Vascular hypertrophy has been previously associated with the degree of baroreflex impairment.²⁷ We showed here that there was a decrease in the diameter (without a change in wall thickness) of the carotid artery, and our previous observations indicate that cerebral arterioles, but not mesenteric arteries in S-P467L mice undergo hypertrophy, increased distensibility and remodeling.^4,5 Thus, we must consider the possibility that the baroreceptor afferents in S-P467L mice may be unresponsive to increases in AP due to structural changes in the vessel. Alternatively, functional defects in the vessel may be playing an important role. Previous ex vivo studies revealed that the aorta from S-P467L mice exhibits markedly impaired responses to both endogenous and exogenous nitric oxide.^4,6,28 Interestingly, mesenteric vessels from S-P467L mice display BK_Ca channel dysfunction,⁵ and polymorphisms in the KCNMB1 gene encoding the 1 subunit is associated with HR variability and baroreflex function in humans.²⁹

PPARγ is expressed in primary cultured microglia and cortical neurons where its protective effects may be due to its anti-oxidant and anti-inflammatory actions.^30,31 PPARγ is also expressed in the arcuate nucleus of the hypothalamus where it regulates the activity of a subset of neurons to control food intake.^32,33 Nevertheless, it remains unclear if PPARγ is expressed in the afferent nerves communicating baroreflex signals from the peripheral vasculature to the brain. Recently, PPARγ expression in the vagal nerve was reported to play a role in high fat diet-induced thermogenesis.¹⁰ We found no evidence for the expression of the transgene encoding dominant negative PPARγ, nor a change in expression of two PPARγ target genes in peripheral NG nerves. Thus, PPARγ in nerves derived from the aortic arch and carotid sinus baroreceptors was unlikely to be affected. Asic2, Asic3 and Trpv1 have been mechanistically implicated to modulate the baroreflex.^11,12 There was no change in Asic channel expression, but initial studies suggested that expression of Trpv1 might be decreased. This was attractive because Trpv1 channels are localized on the nerve fibers and terminals in the aortic arch and in NG neurons, and because ablation or inhibition of Trpv1 containing neurons blunts the baroreflex.¹² Although there were no detectable differences in the number of Trpv1-expressing fibers at the termination of the left ADN, and no change in the maximal response to the Trpv1 ligand capsaicin, there was a significant decrease in the extent of channel desensitization in cultured NG neurons from S-P467L mice. This decrease in the magnitude of Trpv1 desensitization could lead to prolonged channel opening duration, which could promote action potential firing, that is, increased excitation of these neurons upon Trpv1 activation. Physiologically, the Trpv1 channel could be activated by acidic pH (≤6.0) or noxious temperatures (≥43°C).³⁴ However, pathological conditions (e.g. inflammation) could lead to the activation of Trpv1 at normal body temperature, as well as by relatively weak acidic pH (~6.8 to 6.4).^14,35 Under such circumstances, if the Trpv1 channel desensitization is decreased, it could lead to prolonged excitation of NG neurons. Of course, we recognize that this would stand in apparent contradiction to data showing that loss of Trpv1 causes baroreceptor dysfunction.¹² It is also possible that the decrease in channel desensitization constitutes a compensatory effect elicited by the decrease in Trpv1 expression, which could ultimately lead to a normalization of channel function.

Perspectives

The most important observations from this study are that mice with impaired PPARγ activity in vascular smooth muscle exhibit tachycardia mechanistically caused by any combination of: a) increased cardiac sympathetic nervous system activity, b) impaired baroreflex, and c) a blunted ADN response to increased AP localized to the afferent limb of the baroreflex arc. Therefore, there may be some degree of baroreflex failure in S-P467L mice, which drives the increased HR.

Alteration in baroreflex function is a common problem in diabetes. Baroreflex sensitivity is decreased in patients with impaired glucose tolerance,³⁶ and is blunted in diabetic patients.³⁷ Likewise, animal models of diabetes display baroreflex alterations in a similar fashion to human diabetes.^38,39 Depressed baroreflex sensitivity was reported to be predictive of other cardiovascular events in diabetes.⁴⁰ Of greatest relevance to the current study, pioglitazone, a high affinity synthetic PPARγ agonist increased baroreflex sensitivity in type II diabetic patients.⁴¹ Similarly, rosiglitazone, another member of the thiazolidinedione family of PPARγ agonists improves baroreflex gain in rats with diet-induced obesity.⁴² Given that diabetes is often associated with PPARγ impairment, these findings combined with our current study raises the possibility that diabetes-associated baroreflex alteration is due to defective vascular PPARγ. Indeed our findings implicate vascular smooth muscle PPARγ as a critical determinant of neurovascular signaling. Thus, S-P467L mice may be a useful model to understand the decreased baroreflex sensitivity in diabetes.

Novelty and Significance.

What is New?

• Mice expressing dominant negative PPARγ specifically in vascular smooth muscle (S-P467L) exhibit increased sympathetic activity to the heart and decreased effectiveness of the baroreflex.

• S-P467L mice exhibit decreased aortic depressor nerve (ADN) activity in response to changes in arterial pressure, whereas arterial pressure and heart rate responses to direct ADN stimulation were similar in S-P467L and control mice

• These data suggest that the afferent limb of the baroreflex arc is impaired whereas the central and efferent limbs of the baroreflex arc remain intact.

What is Relevant?

• PPARγ is a ligand-activated transcription factor classically playing an important role in adipogenesis and metabolic processes but more recently cardiovascular diseases such as hypertension.

• The baroreflex is an important beat-to-beat regulator of AP and heart rate (HR), controlled by baroreceptors, located in the adventitia of the aortic arch and carotid sinus.

Summary

The most important observations from this study relate to the finding that mice with impaired PPARγ activity in vascular smooth muscle exhibit tachycardia mechanistically caused by any combination of a) increased cardiac sympathetic nervous system activity, b) impaired baroreflex regulation, c) a blunted ADN response to increased AP localized to the afferent limb of the baroreflex arc. Therefore there may be some degree of baroreflex failure in S-P467L mice, which drives the increased HR. These findings implicate vascular smooth muscle PPARγ as a critical determinant of neurovascular signaling.