Renal Denervation Prevents Heart Failure Progression Via Inhibition of the Renin-Angiotensin System

Abstract

BACKGROUND

Previously, we have shown that radiofrequency (RF) renal denervation (RDN) reduces myocardial infarct size in a rat model of acute myocardial infarction (MI) and improves left ventricular (LV) function and vascular reactivity in the setting of heart failure following MI.

OBJECTIVES

The authors investigated the therapeutic efficacy of RF-RDN in a clinically relevant normotensive swine model of heart failure with reduced ejection fraction (HFrEF).

METHODS

Yucatan miniswine underwent 75 min of left anterior descending coronary artery balloon occlusion to induce MI followed by reperfusion (R) for 18 weeks. Cardiac function was assessed pre- and post-MI/R by transthoracic echocardiography and every 3 weeks for 18 weeks. HFrEF was classified by an LV ejection fraction <40%. Animals who met inclusion criteria were randomized to receive bilateral RF-RDN (n = 10) treatment or sham-RDN (n = 11) at 6 weeks post-MI/R using an RF-RDN catheter.

RESULTS

RF-RDN therapy resulted in significant reductions in renal norepinephrine content and circulating angiotensin I and II. RF-RDN significantly increased circulating B-type natriuretic peptide levels. Following RF-RDN, LV end-systolic volume was significantly reduced when compared with sham-treated animals, leading to a marked and sustained improvement in LV ejection fraction. Furthermore, RF-RDN improved LV longitudinal strain. Simultaneously, RF-RDN reduced LV fibrosis and improved coronary artery responses to vasodilators.

CONCLUSIONS

RF-RDN provides a novel therapeutic strategy to reduce renal sympathetic activity, inhibit the renin-angiotensin system, increase circulating B-type natriuretic peptide levels, attenuate LV fibrosis, and improve left ventricular performance and coronary vascular function. These cardioprotective mechanisms synergize to halt the progression of HFrEF following MI/R in a clinically relevant model system.

CENTRAL ILLUSTRATION

Renal Denervation Protects the Failing Heart via Reduced Neprilysin Activity and RAS Inhibition

Sympathetic nervous system activity in the kidney and heart is significantly elevated in heart failure and contributes to increased sodium and water absorption, renal vasoconstriction, neprilysin enzyme activation, and increased renin release with subsequent activation of the renin-angiotensin system (RAS). This pathological kidney response further exacerbates myocardial and vascular dysfunction. Our current work in a clinically relevant model of heart failure demonstrates that renal denervation using the St. Jude Medical EnligHTN radiofrequency (RF) catheter significantly attenuates renal-mediated cardiac and vascular pathology in severe heart failure when this therapy is applied after the onset of heart failure. Our data reveal significant reductions in renal cortex norepinephrine (NE) resulting in attenuated circulating levels of both angiotensin I and II. Additionally, renal denervation also inhibited renal neprilysin activity to increase circulating levels of B-type natriuretic peptide (BNP). Increased BNP bioavailability combined with attenuated RAS signaling resulted in amelioration of cardiac fibrosis, improved left ventricular (LV) function, and improved coronary artery vascular reactivity. Our data strongly suggest that renal denervation represents a viable therapeutic approach to treat heart failure with reduced ejection fraction.

With improved care and management of patients who experience acute myocardial infarction (MI), the incidence of heart failure (HF) continues to increase dramatically, afflicting >6.5 million people in the United States with >950,000 new cases annually (1). The prevalence of HF is expected to increase by 46% and have a total cost of ~$70 billion in the year 2030 (1). A primary pathological component in the pathogenesis of HF is overactivation of the sympathetic nervous system (SNS) to compensate for reduced cardiac pump function (2,3). Down-regulation of pathological sympathetic signaling represents a key therapeutic strategy, which to date, has occurred primarily through lifelong pharmacotherapy (i.e., beta-blockers, ACE inhibitors, angiotensin receptor blockade, and so on). In the absence of a cure for HF, widespread use of these drugs reduces mortality, despite a myriad of side effects resulting in very low patient adherence (4). Novel and potent strategies are warranted, devoid of significant side effects, that attenuate the progression of HF pathology while significantly reducing mortality.

Renal denervation (RDN) is an endovascular procedure that results in the ablation of renal afferent and efferent sympathetic nerve activity in the kidneys. Initial clinical trial results failed to reach a consensus on the efficacy of RDN in the context of lowering blood pressure (5-8). However, recently, it has been shown that RDN significantly reduces blood pressure compared with sham control regardless of the use of antihypertensive pharmacotherapy (9-11). Furthermore, it has been postulated that RDN is beneficial in the setting of HF, because the SNS plays a prominent role in the pathological remodeling of the heart and vasculature, independent of blood pressure. To date there have been no large, randomized, sham controlled clinical trials to test RDN in the setting of HF, and the efficacy of RDN remains uncertain. Our group recently performed proof-of-concept work in a rodent model of HF (12), in which radiofrequency (RF)-RDN was efficacious and provided insight into novel mechanisms by which attenuated renal sympathetic nerve activity preserves left ventricular (LV) function independent of blood pressure-lowering effects. In this present study, we tested the hypothesis that RF-RDN therapy, following the onset of myocardial infarction and HF, would lead to reduced renal sympathetic activity, producing long-term improvements in cardiovascular function in a clinically relevant animal model of heart failure with reduced ejection fraction (HFrEF).

METHODS

ANIMALS AND STUDY DESIGN.

A total of 25 female Yucatan miniswine (9 to 10 months of age) weighing 30 to 35 kg were subjected to myocardial ischemia and reperfusion injury (MI/R) and underwent follow-up for a period of 18 weeks (Figure 1). At 6 weeks post-MI/R, animals underwent an RF-RDN or sham-RDN procedure. All experimental protocols were approved by the Institute for Animal Care and Use Committee at Louisiana State University Health Sciences Center and handled in compliance with the National Institutes of Health “Guide for the Care and Use of Laboratory Animals.” The Online Appendix contains additional Methods and Materials.

FIGURE 1. MI/R-Induced HF Experimental Protocol.

Yucatan miniswine were subject to 75 min of ischemia followed by 18 weeks of reperfusion. Six weeks following myocardial ischemia/reperfusion (MI/R), swine were treated with sham-renal denervation (RDN) or radiofrequency (RF)-RDN. Transthoracic echocardiography (TTE) was performed at baseline and at 3, 6, 9, 12, 15, and 18 weeks. Peripheral blood was collected pre- and post-MI/R, pre- and post-RDN, and 18 weeks post-MI/R. Cardiac and renal tissues were collected at 18 weeks for histological, biochemical, and molecular analysis. 2D = 2-dimensional; LV = left ventricular.

RESULTS

RF-RDN ATTENUATES RENAL NERVE TYROSINE HYDROXYLASE STAINING AND LEVELS OF NOREPINEPHRINE WITHIN THE KIDNEY CORTEX WHILE PRESERVING KIDNEY FUNCTION.

Tyrosine hydroxylase (TH) is the rate-limiting enzyme of catecholamine biosynthesis. We performed TH staining on the renal nerves at 12 weeks post-RDN as an assessment of catecholamine production and nerve viability following sham- and RF-RDN treatment (Figure 2). Representative photomicrographs from each group (Figure 2A [sham-RDN] and Figure 2B [RF-RDN]) illustrate reduced TH staining with RF-RDN compared with sham-RDN. The quantification of stain intensity is summarized in Figure 2C. The RF-RDN–treated animals had a >3-fold increase in the percentage of reduced TH intensity per total nerves when compared with sham-RDN (Figure 2D). Kidney dopamine and norepinephrine (NE) were measured (Figures 2E and 2F) as markers of renal sympathetic activity 12 weeks post-treatment. There was a 72% reduction in kidney dopamine and a 77% reduction in kidney NE following RF-RDN (p < 0.01). Peripheral blood samples were obtained at baseline, 6 weeks post-MI, and at 18 weeks post-MI to measure blood urea nitrogen and creatinine levels to assess kidney function (Table 1). These data clearly demonstrate that the RF-RDN procedure used did not impair renal function in the setting of heart failure. In fact, circulating creatinine levels were significantly lower in the RF-RDN–treated group at 18 weeks following MI compared with the sham-RDN group.

FIGURE 2. Renal Artery Tyrosine Hydroxylase Staining and Kidney Catecholamine Production Following Delayed RF-RDN.

Tyrosine hydroxylase (TH) staining at the 18-week endpoint following sham- or RF-RDN. (A) 4× representative renal artery TH-stained photomicrographs from 4 sham-RDN–treated swine. Squares indicate renal artery nerve and magnified photomicrographs (20×) of TH staining. (B) 4× representative renal artery TH stain from 4 RF-RDN–treated swine. Squares indicate renal artery nerve and magnified photomicrographs (20×) of TH staining. (C) Quantitative TH stain intensity. (D) Percentage of reduced TH stain intensity from total nerves counted. (E) Dopamine and (F) norepinephrine concentration per gram of kidney cortex tissue. Scale bars = 200 um (4×) and 50um (20×). “a” = renal artery; “v” = renal vein in 4× images.

TABLE 1. Circulating BUN and Creatinine.

	HF + Sham-RDN (n = 10)			HF + RF-RDN (n = 9)
	Baseline	6 Weeks	18 Weeks	Baseline	6 Weeks	18 Weeks
BUN, mg/dl	9.5 ± 0.52	11 ± 0.74	12.5 ± 0.97	9.1 ± 0.45	10.6 ± 0.84	11 ± 0.55
Creatinine, mg/dl	0.69 ± 0.03	0.83 ± 0.04	1 ± 0.06	0.69 ± 0.02	0.76 ± 0.04	0.82 ± 0.05*
BUN/creatinine ratio	14 ± 0.93	13.4 ± 1.22	12.8 ± 1.07	13.3 ± 0.85	14.6 ± 1.72	13.9 ± 1.09

Values are mean ± SEM. Blood urea nitrogen (BUN), creatinine, and the ratio were calculated from peripheral blood draws at baseline and at 6 and 18 weeks post-MI/R.

^*p < 0.05 between groups at 18 weeks post-MI/R.

HF = heart failure; MI/R = myocardial ischemia and reperfusion injury; RF-RDN = radiofrequency renal denervation.

DELAYED RF-RDN INHIBITS PROGRESSIVE LV REMODELING AND IMPROVES FUNCTION IN HF.

To assess cardiac remodeling in a temporal manner, we measured LV dimensions and volumes using transthoracic echocardiography. LV diastolic internal dimension (Figure 3A) was not statistically different between groups; however, the LV systolic internal dimension (Figure 3B) was significantly reduced starting at 12 weeks post-MI/R and was preserved throughout the study. Volumetric measurements corroborate the LV internal dimension data: we observed no differences in the LV end-diastolic volume and a significant improvement in the left ventricular end-systolic volume (LVESV) as early as 3 weeks post–RF-RDN (Online Figures 1A and 1B). The significant preservation of LVESV was maintained throughout 12 weeks post–RF-RDN when compared with sham-RDN–treated swine (p < 0.05) (Online Figure 1B). This preservation of LV systolic internal dimension and LVESV is reflected in the sustained improvement in LVEF throughout the 12 weeks post–RF-RDN (Figure 3C). Most importantly, at 18 weeks post-MI/R, in the RF-RDN group, LVEF was significantly improved compared with 6-week post-MI/R levels prior to treatment (p < 0.05) (Figure 3C). We did not observe any significant difference in intraventricular septum wall thickness in diastole between groups (Figure 3D); however, the intraventricular septum in systole (Figures 3E) was significantly preserved in the RF-RDN compared with the sham-RDN group 3 weeks post-RDN and 12 weeks post-RDN (p < 0.05). No significant differences were observed in the LV posterior wall thickness in either diastole or systole (Figures 3F and 3G).

FIGURE 3. LV Structure After MI/R Followed by Delayed Sham or RF-RDN Treatment.

Left ventricular internal dimension (LVID) was measured in (A) diastole (d) and (B) systole (s). (C) Left ventricular ejection fraction (LVEF). LV intraventricular septal wall thickness (IVS) in diastole (D) and systole (E). Left ventricular posterior wall thickness (LVPW) measurements in diastole (F) and systole (G). *p < 0.05, **p value <0.01 between groups. Abbreviations as in Figure 1.

At 18 weeks post-MI/R, 2-dimensional speckle tracking longitudinal strain analysis was performed to further assess LV function. Figure 4A illustrates representative strain analysis using transthoracic echocardiography, showing both global and regional abnormalities in longitudinal strain from baseline to 18 weeks post-MI/R and between treatment groups. Global longitudinal strain was significantly preserved at 18 weeks post-MI/R in the RF-RDN group when compared with sham-RDN–treated swine (Figure 4B). Regionally, significant improvements were only observed in the medial lateral and basal lateral segments of the heart in the RF-RDN group from this single parasternal long-axis view (Figure 4C).

FIGURE 4. LV Longitudinal Strain After MI/R Followed by Delayed Sham or RF-RDN Treatment.

(A) Representative longitudinal strain analysis of sham- and RF-RDN–treated swine at baseline and 18 weeks post-MI/R. (B) Global and (C) regional longitudinal strain analysis at 12 weeks post-RDN. AAS = anterior apical septum; AL = apical lateral; BAS = basal anterior septum; BL = basal lateral; MAS = mid anterior septum; ML = mid lateral; other abbreviations as in Figure 1.

Furthermore, at 18 weeks post-MI/R, the systemic invasive hemodynamics were measured in all animals (Table 2). There were no significant differences in heart rate or in systolic, diastolic, or mean arterial blood pressure between the sham- and RF-RDN study groups.

TABLE 2. Terminal Systemic Invasive Hemodynamics.

	HF +	HF +
	Sham-RDN	RF-RDN	p Value
Heart rate, beats/min	109 ± 8	126 ± 6	0.109
Systolic blood pressure, mm Hg	98 ± 5	96± 5	0.773
Diastolic blood pressure, mm Hg	71 ± 4	71± 4	0.932
Mean arterial blood pressure, mm Hg	83 ± 5	83± 4	0.978

Values are mean ± SEM. The p value is between groups at 18 weeks post-MI/R.

RF-RDN INHIBITS LV FIBROSIS AND PRO-FIBROTIC GENE EXPRESSION IN HF.

LV fibrosis was assessed in the infarct border (Online Figure 2) and nonischemic zones (Figure 5) at the 18-week timepoint in the sham-RDN– and RF-RDN–treated animals. The 4× representative photomicrographs of Masson’s trichrome staining are shown in Online Figures 2A and 2B in the infarct border and in Figures 5A and 5B in the nonischemic zone. Within the infarct border zone, LV % area of fibrosis and gene expression of Col1A1, Col3A1, TGFb, and IL-6 were not significantly reduced (Online Figures 2C to 2G). However, LV % area of fibrosis and gene expression of Col1A1, Col3A1, TGFb, and IL-6 were significantly reduced with RF-RDN treatment at 18 weeks post-MI/R in the nonischemic zone (p < 0.01) (Figures 5C to 5G).

FIGURE 5. LV Fibrosis and Fibrotic Gene Profile in Left Ventricle Nonischemic Zone in HF Following Delayed Sham or RF-RDN.

Representative photomicrographs (4×) of LV Masson’s trichrome staining of the nonischemic zone of (A) sham-RDN– and (B) RF-RDN–treated animals. Percent area of fibrosis (C) is quantified in each group. Fibrotic gene profile consisting a mRNA expression of (D) collagen 1 (COL1A1), (E) collagen 3 (COL3A1), (F) transforming growth factor (TGF)-b, and (G) interlukin (IL)-6. Scale bars = 200 um (A and B). NS = not significant; other abbreviations as in Figure 1.

RF-RDN inhibits the renin-angiotensin system.

Frozen kidney cortex and serum samples from sham-RDN and RF-RDN treated animals were sent for Core RAS-Fingerprint analysis (Attoquant Diagnostics, GmbH, Vienna, Austria) using liquid chromatography-mass spectrometry. There was no significant reduction in renal tissue angiotensin I or II (Figure 6A and 6B). However, the data trended toward lower levels in the RF-RDN treated animals. Furthermore, at 18 weeks post-MI/R, there was a >12-fold reduction in circulating serum angiotensin I in the RF-RDN versus sham-RDN group (p < 0.02) (Figure 6C) and >6-fold reduction in angiotensin II levels in the RF-RDN–treated animals (p < 0.03) (Figure 6D).

FIGURE 6. Kidney Cortex and Circulating Angiotensin I and II Levels.

Renal levels of (A) angiotensin I and (B) angiotensin II at 18 we ks post-MI/R. Circulating serum angiotensin I (C) and II (D) at baseline and 18 weeks post-MI/R in sham-RDN and RF-RDN treated animals. Abbreviations as in Figures 1 and 5.

RF-RDN INCREASES CIRCULATING B-TYPE NATRIURETIC PEPTIDE.

At 18 weeks post-MI/R, circulating B-type natriuretic peptide (BNP) levels in the animals treated with RF-RDN at 6 weeks following MI/R were significantly increased when compared with sham-RDN (p < 0.05) (Figure 7A). The increase in circulating BNP is not a result of increased myocardial BNP production, because LV BNP mRNA levels between groups were not statistically different (Figure 7B).

FIGURE 7. Circulating BNP Through 18 Weeks Post MI/R, BNP Gene Expression and Neprilysin Activity at 18 Weeks Following Sham or Delayed RF-RDN.

(A) Circulating B-type natriuretic peptide (BNP) at baseline, 6 weeks pre-treatment, and 18 weeks post-MI/R. (B) Left ventricular myocardial BNP mRNA expression. (C) Kidney cortex neprilysin activity at 12 weeks post-RDN. NS = not significant.

RENAL NEPRILYSIN ACTIVITY FOLLOWING RF-RDN.

Neprilysin (NEP) is a membrane-bound and soluble endopeptidase that is the primary mechanism for clearance of BNP. NEP is present in multiple organs within the body and throughout the circulation; the kidney is 1 organ where NEP is particularly abundant. Here we demonstrate a significant reduction (p < 0.01) in kidney NEP activity at 18 weeks post-MI/R in RF-RDN versus sham-RDN treated swine (Figure 7C).

IMPROVED CORONARY VASCULAR REACTIVITY FOLLOWING RF-RDN TREATMENT.

Given the importance of coronary artery vascular function on myocardial blood flow as well as cardiac myocyte function and viability, coronary vascular vasorelaxation responses were evaluated at 18 weeks post-M/R. For these experiments, we isolated the previously ischemic/reperfused LAD and left circumflex coronary arteries, evaluating their vasorelaxation responses in an ex vivo organ bath system. Vessels were challenged with increasing concentrations of endothelial-dependent (bradykinin, and substance P) and independent (sodium nitroprusside) substances. In swine treated with RF-RDN, coronary vasodilatory responses to bradykinin and EC₅₀ values were significantly improved in both the LAD and LCX (p < 0.01) (Figures 8A and 8B). Similarly, significant improvements in vasorelaxation responses and EC50 to substance P were observed in the LAD and LCX (Figures 8C and 8D). Finally, we observed significant improvements in the vasorelaxation responses, in an endothelial independent manner, in the ischemia-reperfused LAD (Figure 8E), but not the LCX (Figure 8F).

FIGURE 8. Coronary Vascular Relaxation at 18 Weeks Post-MI/R With Delayed Sham or RF-RDN Treatment.

Vascular relaxation curves and EC50 values of sham- or RF-RDN swine coronary arteries to bradykinin (A and B), substance P (C and D), and sodium nitroprusside (SNP) (E and F) in a dose-dependent manner. (A, C, E) LAD, and (B, D, F) LCX coronary arteries. *p < 0.05, **p < 0.01 between sham- and RF-RDN groups 18 weeks post-MI/R. nM = nanomolar; other abbreviations as in Figure 1.

DISCUSSION

RDN devices were originally developed to treat hypertensive patients in which medications failed to achieve clinically important blood pressure reductions. The initial hypothesis was that ablation of the renal efferent sympathetic nerves would alter sodium and water retention, inhibit the renin-angiotensin system (RAS), and ultimately reduce blood pressure in a sustained and meaningful manner. Following initial reports of remarkable blood pressure–lowering effects (13-15), the failure to achieve significant reductions in blood pressure in the Symplicity HTN-3 (Renal Denervation in Patients With Uncontrolled Hypertension) trial dampened enthusiasm and resulted in the reassessment of the promise of RDN therapies to treat hypertension (16). As a result of this important clinical trial, RDN catheters were reimagined and clinical trial designs for new trials were significantly modified. More recently, off- and on-medication sham controlled clinical trials of hypertensive patients (SPRYAL HTN OFF [Global Clinical Study of Renal Denervation With the Symplicity Spyral Multi-electrode Renal Denervation System in Patients With Uncontrolled Hypertension in the Absence of Antihypertensive Medications] and HTN ON-MED [Global Clinical Study of Renal Denervation With the Symplicity Spyral Multi-electrode Renal Denervation System in Patients With Uncontrolled Hypertension on Standard Medical Therapy] and RADIANCE HTN) have reported systolic blood pressure reductions of approximately 7- to 9-mm Hg and 5-mm Hg reductions in diastolic pressure (9-11). These reductions in blood pressure are clinically meaningful and have been shown to reduce the risk of cardiovascular diseases (17,18). Although the SNS is involved in the pathogenesis of hypertension, it is well appreciated that the involvement of the SNS is even greater in HF. However, there are many hypertensive patients lacking a sympathetic component (2,19-21). Given the importance of the SNS in HF, we investigated the potential therapeutic benefit of RDN in the setting of severe HFrEF. A number of pharmacological agents currently used to treat hypertension act by dampening sympathetic signaling, which is at the crux of HF pathology. Sympatholytics and RAS inhibitors have successfully improved outcomes, and the newest generation of drugs, the angiotensin receptor NEP inhibitors, further improve mortality and reduce hospitalization readmission rates (22). Despite these treatments, HF is a leading cause of death and has an abysmal 5-year mortality rate of 50% after diagnosis (1). Renal denervation has the potential to quell 3 major limitations of currently available drugs used to treat HF: 1) patient nonadherence; 2) untoward side effects; and 3) the lack of proximal inhibition of renal SNS and RAS pathway pathological signaling.

Previously, our laboratory investigated delayed RF-RDN for the treatment of HF in a rodent model following acute MI in hypertensive and normotensive rats (12). Similar improvements in LV function and vascular reactivity were observed in both hypertensive and normotensive rats as well as reduced cardiac fibrosis (12). Interestingly, we also observed substantial increases in circulating levels of natriuretic peptides (NPs) following RF-RDN in the setting of HF. We determined that RF-RDN resulted in the inhibition of renal NEP activity, and we believe this novel effect is responsible for the increase in circulating NPs (12).

The results of the present study confirm and extend our previous rodent study findings that delayed RF-RDN therapy maintains systolic LV function and vascular function compared with the sham-RDN in the setting of HF (Central Illustration). Also, significant advancements have been made from our previous work. We investigated, for the first time, the effects of RF-RDN in a large animal, post-MI, HF model using a device that is currently under investigation in man as an antihypertensive endovascular therapy. Confirmation of renal nerve denervation at 12 weeks post-RDN was observed by a significant reduction in TH stain within the renal arteries of the RF-RDN group. Furthermore, we measured kidney dopamine and norepinephrine levels in a blinded manner, which were reduced by 72% and 77% in the RF-RDN group, respectively. These data confirm the near total ablation of renal sympathetic nerves. We also demonstrate significant reductions in LVESV at 3 weeks post–RF-RDN and sustained throughout 12 weeks; this result was confirmed by preservation of the LV internal diameter at end-systole. These structural cardiac changes led to a significant improvement in LVEF through 12 weeks post–RF-RDN. Cardiac function, as measured by global speckle-tracking strain analysis, was significantly improved at 12 weeks post–RF-RDN when compared with sham-RDN. This work confirmed our previous findings that improvements in cardiac and vascular function following RF-RDN are blood pressure independent (12).

This study presents convincing data that RF-RDN improves coronary artery vascular reactivity in a failing heart following MI/R. Normal vascular endothelial and smooth muscle function are critical for regulating perfusion of myocardial tissue and ultimately improved outcomes in these patients (23,24). Delayed RF-RDN therapy provided a lasting improvement on vascular function, as evidenced by ex vivo vascular reactivity experiments. At 12 weeks post-treatment, LAD and LCX coronary artery vascular reactivity was significantly improved when compared with the coronary vessels of the sham-RDN group. The improvements in coronary vascular function were in part related to improved endothelial function, because relaxations to bradykinin and substance P were significantly improved. Concomitantly, we observed improvements in response to the nitro-vasodilator sodium nitroprusside, suggesting that the vascular protective effects of sympathetic modulation, through RF-RDN, also extended to coronary smooth muscle cells.

It remains unknown whether radiofrequency ablation of the renal sympathetic nerves has a lasting effect and if this therapy would be effective in chronic diseases such as essential hypertension. In the present study, a 1-time treatment resulted in a lasting reduction in renal nerve sympathetic activity without permanent adverse effects on kidney function due to the RDN procedure. Renal retention parameter in both groups increased from baseline to 18 weeks post-MI/R regardless of treatment. However, this occurred to a lesser extent in the RF-RDN–treated group when compared with the sham-RDN treated group. These values are still within normal ranges and do not suggest any substantial detriment to renal function. Both TH staining and kidney NE content were reduced at least 12 weeks following RF-RDN in the setting of severe HF. Moreover, these reductions in sympathetic nerve signaling to the kidney are accompanied by reductions in renal and systemic RAS that are notorious for having detrimental effects on the heart, vascular, and renal systems. The reductions in angiotensin I and subsequently angiotensin II, as a result of RDN therapy, played a direct role in the improvements observed in cardiac and vascular function.

A novel finding from our previous work was the identification of the relationship between renal sympathetic nerve activity and NP metabolism (12). Our study demonstrates that reductions of sympathetic nerve activity, via RF-RDN, inhibited renal activity of the endopeptidase NEP, which is critically responsible for the degradation of various bioactive peptides such as ANP, BNP, bradykinin, and angiotensin II (25). As a result, improvements in cardiac function in HF with RF-RDN treatment were associated with elevations in circulating ANP and BNP. The present study confirms these findings. RF-RDN inhibited renal NEP activity and significantly increased circulating BNP levels in a sustained fashion compared with sham-RDN. Interestingly, although angiotensin II is also a substrate for NEP cleavage, RDN did not result in elevations of angiotensin II. In fact, circulating angiotensin II levels were significantly decreased compared with sham-RDN levels. We believe this is a result of the upstream inhibition of the SNS and NE signaling, which triggers RAS activation. This is a critical finding, because not only is RDN acting as an NEP inhibitor, but it also prevents sympathetic overactivation, which leads to reduced NE content, eliminating detrimental elevation in angiotensin I and II levels that typically are accompanied by NEP inhibition. Future studies will evaluate the effects of RDN on renin activity as well as aldosterone levels.

Circulating BNP can act via several mechanisms to improve cardiovascular function. BNP increases natriuresis and diuresis within the kidney (26), inhibits hypertrophic signaling in the myocardium (27), and improves cardiac contractility (28). Furthermore, BNP has been shown to reduce the expression of pro-fibrotic gene profiles, attenuate myofibroblast transformation, and inhibit fibroblast proliferation within the heart (29,30). In the present study, we confirmed the reduction in fibrosis within the nonischemic myocardium, which can be attributed both to an increase in circulating BNP and also the reduction in SNS overactivation and inhibition in RAS signaling through RF-RDN treatment. The sustained activation of the SNS, in the setting of HF, can further drive adverse cardiac remodeling through modulation of vasculature function (31) and activation of cardiac fibroblasts (32), and can drive a persistent underlying immune response to continued cardiovascular stress (33). Using RF-RDN to attenuate renal sympathetic nerve activity may not only quell the deleterious effects towards vasculature function, as shown in our present study, but also modulate resident cardiac fibroblast activity and immune cell infiltrate within the myocardium of the failing heart.

The present study is important for several reasons. It suggests that the EnligHTN multielectrode system (St. Jude Medical, St. Paul, Minnesota) provided adequate energy to produce physiologically relevant renal sympathetic denervation and reductions in kidney NE content in the setting of severe HF, which is associated with maximum renal sympathetic nervous system activation. Furthermore, RDN inhibited RAS, leading to significantly reduced level of angiotensin II, which is a key mediator of cardiovascular pathogenesis. Importantly, we were able to confirm reports that RF-RDN may be efficacious in the treatment of HFrEF. This study marks a critical step in the translation of a therapy that was possibly rushed to clinic too quickly without a thorough understanding of its beneficial mechanisms. Return to the “bench” and the use of preclinical studies have revealed that RF-RDN therapy is not dead; in fact, its therapeutic potential may extend far beyond lowering blood pressure and should be examined closely for the treatment of HFrEF.

STUDY LIMITATIONS.

This study provides strong evidence that RF-RDN mitigates the progression of HF. However, there remain several inherent limitations. The MI-induced HF model used does not completely mimic clinical HF. Aside from myocardial ischemic injury resulting in significant infarction, the swine model that was investigated did not have typical comorbidities commonly observed in HF patients such as hypertension, dyslipidemia, or diabetes. Last, the 18-week study protocol suggested sustained functional improvement due to RF-RDN therapy. Nevertheless, approximately 50% of HF patients live over 5 years beyond initial diagnosis. Additional studies are currently underway to address these limitations and shortcomings.

CONCLUSIONS

RF-RDN halts the progression of HF in a preclinical large animal model. The cardioprotective actions found in this model confirm our previous findings in rodent models of HF (12). RF-RDN in the context of HF reduces renal catecholamine levels, inhibits the RAS, inhibits NEP activity, increases circulating BNP, improves coronary vascular vasomotion, and reduces cardiac fibrosis culminating in improved LV structure and function. Rigorous clinical investigation of RF-RDN as a strategy to treat HF is warranted. In conclusion, RF-RDN is safe, efficacious, and halts the progression of HF in a translational preclinical large animal model.

PERSPECTIVES.

COMPENTENCY IN MEDICAL KNOWLEDGE:

In an animal model of HF following MI, RF-RDN slows adverse LV remodeling and preserves systolic ventricular and coronary vascular function through reduction of renal sympathetic activity, renin-angiotensin inhibition, increased circulating levels of BNP, and antifibrotic effects.

TRANSLATIONAL OUTLOOK:

Clinical studies are needed to assess whether these observations can be translated to patients with HF in the presence and absence of hypertension and assess the durability of whatever benefit occurs.

ABBREVIATIONS AND ACRONYMS

BNP

B-type natriuretic peptide

HF

heart failure

LV

left ventricle/ventricular

MI/R

myocardial ischemia/ reperfusion

NE

norepinephrine

NP

natriuretic peptide

RAS

renin-angiotensin system

RF

radiofrequency

RDN

renal denervation

SNS

sympathetic nervous system

APPENDIX

For an expanded Methods section as well as supplemental figures and a table, please see the online version of this paper.