Renal denervation improves cardiac function and exercise duration in a miniswine model of heart failure with preserved ejection fraction

Introduction

Heart failure with preserved ejection fraction (HFpEF) affects an estimated 32 million people worldwide and its prevalence continues to rise due to an ageing population with increasing incidence of associated comorbidities including hypertension, obesity and diabetes. ¹ The complex multimorbidity, multiorgan involvement associated with HFpEF disease progression is also reflected in the disappointing results from multiple therapeutic approaches in which all but two (i.e. sodium–glucose cotransporter 2 inhibitors [SGLT2i] and glucagon‐like peptide‐1 agonists) were found ineffective in reducing HFpEF severity. ¹ , ² , ³ It is well‐known that sympathetic nerve overactivity drives pathophysiological processes of HFpEF contributory comorbidities and evidence of elevated systemic and cardiac sympathetic drive present in HFpEF patients suggests contribution to disease pathophysiology. ⁴

Renal denervation (RDN) has been shown to enhance multi‐organ function, including the heart, skeletal muscle, vasculature, and lungs by reducing renal afferent and efferent sympathetic tone. ⁵ , ⁶ Using a clinically‐relevant swine model, ⁷ we tested potential use of RDN as a minimally invasive therapy for the treatment of HFpEF.

Methods

The experimental protocol is illustrated in Figure  1A . Experiments were approved by the Institutional Animal Care and Use Committee of Louisiana State University Health Sciences Center of New Orleans (Approval no. 3538). Eight‐month‐old female Gottingen minipigs (n = 9) were obtained (Marshall BioResources, Rose, NY, USA) and handled in accordance with guidelines of Public Health Service Policy on Humane Care and Use of Laboratory Animals, the Animal Welfare Act and the Guide for the Care and Use of Laboratory Animals (NIH publication no. 85‐23, revised 2011). Telemetry transmitters (TSE Systems, Chesterfield, MO, USA) were implanted for monitoring ambulatory systolic, diastolic, mean arterial blood pressures and heart rate. HFpEF was induced by subcutaneous implantation of deoxycorticosterone acetate pellets (50 mg/kg; 90‐day release) in combination with a custom Western diet containing high levels of fat (12%), fructose (9%), and salt (2% NaCl) as previously described. ⁵ Exercise capacity testing using a motor‐driven treadmill was performed at baseline, 8 and 18 weeks. Transthoracic echocardiography was performed as previously described. ⁷ , ⁸ Left ventricular ejection fraction (LVEF) was measured using two‐dimensional (2D) speckle tracking (Simpson's method), left atrial fractional area change (LAFAC) calculated from subcostal 2D B‐mode views, pulsed‐wave Doppler for peak early (E) transmitral inflow velocities during diastole, pulsed‐wave tissue Doppler imaging for early ventricular filling velocities (e') at the mitral valve annulus were performed and the E/e' ratio calculated. Invasive left ventricular end‐diastolic pressure and pulmonary capillary wedge pressure were measured as previously described. ⁷ At 10 weeks, minipigs were randomized to receive radiofrequency (RF)‐RDN (n = 5) using the EnligHTN™ catheter (St. Jude Medical, St. Paul, MN, USA) or sham‐RDN (n = 4), as previously described. ⁷ Renal artery tyrosine hydroxylase (TH) immunohistochemistry was performed to evaluate the viability of the renal nerves, and kidney norepinephrine levels were measured in both kidneys. ⁸ , ⁹ , ¹⁰ Data are expressed as mean ± standard error of the mean. Statistical analysis was performed using repeated‐measures two‐way ANOVA with Sidak post‐hoc analysis where p < 0.05 was considered statistically significant.

Figure 1.

(A) Experimental protocol: female Göttingen minipigs were implanted with radiotelemeters and subjected to 22 weeks of excess mineralocorticoid exposure and dietary insults to drive development of prominent comorbidities associated with heart failure with preserved ejection fraction (HFpEF). At 10 weeks, HFpEF miniswine were randomized to sham‐renal denervation (RDN) (n = 4) or radiofrequency (RF)‐RDN (n = 5) treatment. Transthoracic echocardiographic assessment of cardiac function and structure was performed at baseline and at 4, 8, 10, 14, 18, and 22 weeks. Voluntary treadmill exercise testing was performed in sham‐RDN (n = 4) and RF‐RDN (n = 4) HFpEF minipigs at baseline and at 8 and 18 weeks. At 22 weeks, renal arteries were collected for tyrosine hydroxylase immunostaining and kidneys collected for renal cortex norepinephrine (NE) levels. (B) Representative micrographs of tyrosine hydroxylase immunostaining of renal arteries following sham‐ or RF‐RDN at the 22‐week endpoint at low and high magnification (scale bars: 200 μm and 50 μm, respectively). (C) Renal cortex NE content per gram of kidney tissue was measured at the 22‐week endpoint. (D) Echocardiography was performed at baseline and 4, 8, 10, 14, 18, and 22 weeks and left ventricular ejection fraction (LVEF) measured using two‐dimensional auto‐ejection fraction. (E) Transthoracic pulsed‐wave Doppler echocardiography and pulsed‐wave tissue Doppler imaging were performed at baseline and 4, 8, 10, 14, 18, and 22 weeks. Peak early atrial diastolic transmitral flow velocity (E) and early atrial diastolic medial mitral annular velocity (e′) were measured and E/e′ ratio calculated. (F) Left atrial fractional area change (LAFAC) was measured from subcostal two‐dimensional B‐mode echocardiographic images acquired at baseline and 4, 8, 10, 14, 18, and 22 weeks. (G) Invasive haemodynamic measurements of left ventricular end‐diastolic pressure (LVEDP) and (H) pulmonary capillary wedge pressure (PCWP) were obtained at baseline and 22 weeks. (I) Circulating N‐terminal pro‐B‐type natriuretic peptide (NT‐proBNP) was measured at baseline, 10, 18, and 22 weeks. (J) A minipig in the RF‐RDN treatment group was uncooperative during treadmill acclimation sessions resulting in voluntary treadmill testing performed in 8 HFpEF minipigs (sham‐RDN, n = 4; RF‐RDN, n = 4). Exercise duration was measured at baseline, 8 and 18 weeks, and percent change in duration from baseline was calculated. (K–M) Ambulatory arterial blood pressure waveforms were acquired prior to sham‐RDN and RF‐RDN treatment. Delta change in systolic blood pressure (SBP) (K), diastolic blood pressure (DBP) (L), mean arterial pressure (MAP) (M), and heart rate (HR) (N) was calculated. Data are presented as mean ± standard error of the mean, numbers in circles represent the number of animals analysed.

Results

Representative photomicrographs of renal artery nerve TH staining (Figure  1B ) illustrate reduced sympathetic nerve viability with RF‐RDN compared to sham‐RDN. Kidney norepinephrine, a biomarker of renal artery nerve activity, measured 12 weeks after treatment, was significantly reduced following RF‐RDN (Figure  1C ). LVEF (%) was preserved in both groups throughout the 22‐week protocol (Figure  1D ). Left ventricular diastolic function, was significantly impaired, with increased E/e′ as early as 4 weeks in all HFpEF minipigs prior to treatment (Figure  1E ). There was a significant reduction (p < 0.05) in E/e′ 4 weeks post‐RDN and remained significantly lower at all subsequent timepoints compared to sham‐RDN (p < 0.05). Left atrial function, as measured by the percent change in left atrial area, was significantly reduced (p < 0.001) as early as 4 weeks in all HFpEF minipigs before treatment (Figure  1F ). Following RF‐RDN, there was a significant improvement in left atrial function at all timepoints compared to sham treatment (p < 0.005). There were significant elevations in resting left ventricular filling (p < 0.01) (Figure  1G ), and pulmonary pressures (p < 0.005) (Figure  1H ) in HFpEF minipigs were unaffected by RF‐RDN treatment. A clinical marker of heart failure, N‐terminal pro‐B‐type natriuretic peptide (NT‐proBNP), was significantly elevated after 10 weeks of western diet and hypertension (p < 0.0001) (Figure  1I ). Following RDN, the significant (p = 0.007) and progressive increase in NT‐proBNP in the HFpEF group was significantly (p = 0.05) halted in the RDN‐treated cohort 8 weeks later (Figure  1I ). Exercise intolerance figures prominently in HFpEF pathophysiology and consistent with the clinical phenotype, HFpEF minipigs exhibited marked fatigue illustrated by a reduction in treadmill exercise duration (Figure  1J ). RDN therapy resulted in a significant improvement in exercise capacity as evidenced by increased treadmill running duration in the RF‐RDN‐treated minipigs compared to sham‐RDN (p = 0.05). Data calculated from conscious arterial blood pressure waveforms acquired from radiotelemeters revealed a lowering of systolic (Figure  1K ), diastolic (Figure  1L ), and mean (Figure  1M ) arterial blood pressure following RF‐RDN treatment compared to sham‐RDN. The blood pressure‐lowering effects observed in RDN‐treated HFpEF minipigs were transient lasting only 4 weeks as pressures returned to pre‐treatment levels by 15 weeks. RF‐RDN did not significantly affect heart rate when compared to sham‐RDN HFpEF minipigs (Figure  1N ).

Discussion

Our findings demonstrate that RDN significantly attenuates the pathogenesis of key markers of HFpEF with observed substantial and sustained improvement in E/e′, LAFAC and treadmill exercise. Interventional ablation of efferent and afferent sympathetic nerves along renal arteries has been deemed effective in treating refractory hypertension, a prevalent comorbidity in HFpEF patients. Clinical data of RDN use in HFpEF patients are limited to retrospective studies where RDN was shown to reduce E/e′ in a subgroup of hypertensive patients with HFpEF ¹¹ , ¹² and a small underpowered randomized trial (RDT‐PEF) of 25 HFpEF patients showing early beneficial effects following RDN but not sustained. ¹³ Whether efficacy of RDN for HFpEF depends on its ability to lower blood pressure remains controversial. ¹³ Our data may further confound the issue in that RDN significantly decreased blood pressure and improved diastolic function; however, RDN only transiently lowered blood pressure through 1 month while its improvement in diastolic function (i.e. E/e′) was sustained. Our findings are consistent with clinical data demonstrating improvement in diastolic function without significant association with systolic blood pressure in patients with pathological cardiac remodelling. ¹⁴ These data corroborate blood pressure‐independent mechanism(s) in which RDN may improve cardiac function. ⁸ Nevertheless, future studies will need to confirm these findings in a larger validation cohort and provide additional clarity on the physiological observations. The ongoing UNLOAD‐HFpEF (Renal Denervation to Treat Heart Failure With Preserved Ejection Fraction) trial will be the first randomized, sham‐controlled double‐blind trial that will address the role of RDN in HFpEF. ¹⁵ Only recently has pharmacological‐based sodium–glucose cotransporter 2 inhibition yielded beneficial effects in HFpEF patients. ³ Originally developed as an antidiabetic drug directly impacting renal function, an underappreciated aspect of SGLT2i may be its ability to reduce sympathetic nerve activity independent of glycaemic control. ¹⁶ As both SGLT2i and RDN target the kidney, further investigation into treating extracardiac organ dysfunction, perhaps with known sympathetic nerve overactivity involvement, may lead to robust treatments for HFpEF patients.

Funding

This study was supported by National Institutes of Health (NIH) grants (HL146098, HL146514, and HL151398) to Dr. Lefer; National Institutes of Health (NIH) grant (AA029984) to Dr. Sharp; and National Institutes of Health (NIH) grant (HL159428) to Dr. Goodchild.

Conflict of interest: none declared.