Pulmonary Hypertension Increases Left Ventricular Chamber Stiffness Through Ventricular Interdependence and Is Attenuated by Pulmonary-Targeted SERCA2a Gene Therapy

Abstract

Background:

Prognosis in patients with left ventricular (LV) failure is significantly worse when accompanied by pulmonary hypertension (PH). We hypothesized that right ventricular (RV) pressure overload negatively influences LV function through ventricular interdependence, and that attenuation of PH would attenuate this adverse interaction.

Methods:

Chronic post-capillary PH was induced by partial pulmonary vein banding in Yorkshire pigs. Animals underwent echocardiography and invasive LV pressure-volume relationship studies 2 months after intra-airway delivery of adeno-associated virus (AAV)-1 carrying the Sarco/endoplasmic reticulum Ca2+ ATPase (SERCA2a) gene or saline. LV function and morphology were compared among PH treated with saline (PH+Saline, n=5), AAV1.SERCA2a (PH+SERCA group: n=7), and naïve (n=8) animals.

Results:

PH+Saline pigs had elevated mean pulmonary artery pressure, lower cardiac output, and increased RV mass, but these changes were attenuate in the PH+SERCA group. The RV-to-LV diameter ratio was significantly larger in the PH+Saline group than in both the naïve and PH+SERCA groups (p=0.006), with smaller LV eccentricity index (p=0.01). LV systolic parameters as well as relaxation parameters (dP/dt minimum and tau) were not different among the groups. Meanwhile, LV pressure-volume relationship during the preload reduction revealed significantly increased LV chamber stiffness in the PH+Saline group compared to sham group, whereas the change in relationship was attenuated in the PH+SERCA group (β coefficient: PH+Saline 0.094±0.027, PH+SERCA2a 0.058±0.018, naïve 0.043±0.017; P=0.002).

Conclusion:

RV pressure overload associated with PH increases LV chamber stiffness. Attenuation of PH with SERCA2a gene therapy attenuated the increase in LV chamber stiffness in a pig model of chronic post-capillary PH.

New & Noteworthy

We found that right ventricular pressure overload from pulmonary hypertension significantly impairs left ventricular end-diastolic chamber stiffness. Gene therapy that attenuates progression of pulmonary hypertension also attenuated left ventricular chamber stiffening in a swine model. These findings suggest that adverse interventricular interactions may contribute to the worsening prognosis observed in patients with heart failure who develop pulmonary hypertension.

Introduction

Pulmonary hypertension (PH) due to left heart disease (Group 2 PH) is the most common form of PH, accounting for 65–80% of all cases¹. The development of PH and subsequent right ventricular (RV) failure markedly worsens prognosis ², yet effective therapies remain limited. The pathophysiology linking left heart disease to PH is well recognized: chronically elevated left ventricular (LV) filling pressure leads to pulmonary congestion and pulmonary vascular remodeling that result in RV pressure overload. In contrast, the effects of RV pressure overload on LV function through ventricular interdependence remain incompletely understood. This interaction is particularly important in patients with heart failure with preserved ejection fraction (HFpEF), in whom impaired LV diastolic function is the primary pathophysiological abnormality and the development of PH is associated with worse symptoms and poor prognosis³.

We previously demonstrated that intra-airway delivery of adeno-associated virus (AAV)1-mediated sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA2a) reversed pulmonary vascular remodeling and improved RV function in a porcine model of severe PH⁴. In this sub-analysis study, we investigated the effects of PH-induced RV dysfunction on LV geometry and mechanics using echocardiography and invasive pressure-volume analysis.

Methods

Study Design

This is a sub-analysis of the previously reported intra-airway SERCA2a gene therapy study in female Yorkshire pigs⁴. The original study examined the impact of intra-airway AAV1.SERCA2a gene therapy in a swine model of post-capillary PH. Animals were randomized to saline treatment (PH+Saline group) or AAV1.SERCA2a treatment (PH+SERCA group).

The focus of current study is to understand the impact of severe PH and therapeutic SERCA2a gene therapy on LV physiology and morphology.

Among the animals that completed the 4-month protocol, invasive LV pressure-volume measurement was performed in 5 PH+Saline animals and 7 PH+SERCA animals. Three animals in the PH+Saline group had severely decompensated hemodynamics after RV studies that precluded LV pressure-volume study. One animal in the PH+SERCA group did not undergo LV pressure-volume study due to time constraint. Eight size-matched naïve animals that underwent the same echocardiographic and LV pressure-volume measurements were included for reference.

All procedures were conducted in accordance with approval from the Institutional Animal Care and Use Committee (IACUC) at the Icahn School of Medicine at Mount Sinai, and animal care and management were maintained in compliance with the Guide for the Care and Use of Laboratory Animals. All animals underwent a minimum 72-h acclimatization period before participation in the study.

PH induction, intra-airway gene delivery, and terminal procedure

Detailed methods for the model induction and gene delivery have been reported previously^4-6. Briefly, left superior and lower common pulmonary veins were surgically banded to induce stenosis. Two months later, when the animals exhibited mild elevation in pulmonary artery pressure, saline or AAV1.SERCA2a with cytomegalovirus promoter (1x10¹³ viral genomes) were delivered into the airway using a microsprayer through the intubation tube. Animals were followed for 2 more months post-gene delivery and cardiac function as well as hemodynamics were evaluated prior to euthanasia. RV function and hemodynamic parameters of the complete study cohort have been reported previously⁴.

Echocardiography

Left and right ventricular functions were evaluated by two-dimensional (2D), three-dimensional (3D) and Doppler echocardiography using Philips iE33 and QLab software. Short axis views from the right intercostal space were used for assessing RV and LV diameters as well as LV eccentricity index, while 3D images from the sub-sternum were used for LV volume assessment. Mitral annulus flow was obtained through substernal access. Mitral inflow image was not available in one of the animals in the AAV1.SERCA2a group.

Pressure and volume assessments

A pressure-volume catheter (Ventricath, Millar Inc.) was inserted into the LV through the carotid arterial access established by a percutaneous puncture. Its placement was optimized using fluoroscopy and real-time monitoring of the pressure-volume loops. A 7.5Fr Swan-Ganz catheter (Edwards) was advanced to the pulmonary artery from the femoral vein access, and right heart pressures were obtained. Cardiac output was measured by the thermodilution method. Additionally, a 4-Fr catheter was inserted into the left atrium through the atrial septum to measure left atrial pressure. Trans-pulmonary resistance that incorporates the stenosis resistance was calculated as [(mean PA pressure) – (mean left atrial pressure)]/(cardiac output). Steady-state LV pressure-volume data were acquired during a brief breath-hold, followed by preload reduction studies by a brief inferior vena cava ballooning using an 11Fr Fogarty balloon (Edwards). Volume calibration was performed against stroke volume measured by thermodilution method (alpha) and 3D echocardiographic end-diastolic volume (parallel conductance). Pressure-volume relationships were calculated using iox2 software (emka TECHNOLOGIES)⁷.

Tissue analysis

To examine if altered LV function was due to AAV1 mediated SERCA2a gene transfer to the heart, we isolated DNA from the RV free wall, septum, and LV for AAV genome detection. Additionally, mRNA was isolated to examine presence of human SERCA2a. Briefly, genomic DNA was isolated from LV free wall, septum, and RV tissues using the DNeasy Blood & Tissue Kit (Qiagen). DNA input was standardized to 500 ng per reaction, with samples lacking detectable amplification reanalyzed using the maximum available DNA input (up to 2,930.4 ng). qPCR was performed using Maxima SYBR Green/ROX Master Mix (Thermo Fisher Scientific) on a QuantStudio^™ 3 instrument (Applied Biosystems). CMV primers were used to assess vector DNA. An AAV2 Reference Standard (ATCC VR-1616) was used to establish assay sensitivity and determine the limit of detection. Total RNA was isolated using the mirVana^™ miRNA Isolation Kit (Thermo Fisher Scientific) and complementary DNA (cDNA) was synthesized using the Verso cDNA Synthesis Kit (Thermo Fisher Scientific). Quantitative real-time PCR was performed using Maxima SYBR Green/ROX qPCR Master Mix (Thermo Fisher Scientific) on a QuantStudio^™ 3 Real-Time PCR Instrument (Applied Biosystems). Human SERCA2a expression was quantified using gene-specific primers (f: 5’GGACTTTGAAGGCGTGGATT3’, r: 5’GGAAGCGGTTACTCCAGTATTG3’) and normalized to 18SrRNA (f: 5’TAACGAACGAGACTCTGGCAT3’, r: 5’CGGACATCTAAGGGCATCACAG3’).

Statistical analysis

Data are presented as the mean ± standard deviation. Statistical tests were conducted in GraphPad Prism version 6.07 (GraphPad Software Inc., San Diego, CA). Normality testing with the Shapiro–Wilk test (n < 50) was performed for each group. One-way ANOVA was used to compare among the 3 groups, followed by Tukey’s post-hoc test. A Kruskal-Wallis test followed by Dunn’s post-hoc test was used for data with non-Gaussian distribution. Spearman's rank correlation analysis was performed to assess the association between the end-diastolic pressure-volume relationship (EDPVR) β coefficient and selected hemodynamic and structural parameters. A p value of < 0.05 was considered statistically significant.

Results

PH and RV parameters show similar results to complete study cohort

Because some of the animals in the original study cohort⁴ did not undergo LV pressure-volume assessment, we analyzed major PH-related parameters of the animals included in this study. There were no statistically significant differences in the mean PA pressure at the time of randomization (PH+Saline 28.2±8.2 mmHg vs PH+SERCA 24±7.6 mmHg, p=0.36), as well as pulmonary vein flow velocity at the final time point (Table 1), suggesting a comparable degree of banding between the PH groups in this study. Consistent with the prior report, PH+Saline animals exhibited significantly elevated mean PA pressure and pulmonary vascular resistance, which were attenuated by AAV1.SERCA2a treatment (Table 1). Cardiac index was significantly lower in the PH+Saline group compared to the naïve and PH+SERCA groups, and post-mortem RV to LV weight ratio was significantly higher (p=0.01).

Table 1.

PH-related endpoints

	Naïve (n=8)	PH+Saline (n=5)	PH+SERCA (n=7)	p-value (ANOVA)
Body weight (Kg)	32.4±8.2	34.2±5.8	34.9±2.8	0.73
Mean PA pressure (mmHg)	17.9±5.3	50.8±14.1*	30.9±11.1¶	<0.001
RV systolic pressure (mmHg)	31.1±3.2	69.6±21.4*	44.0±15.1¶	<0.001
RV end-diastolic pressure (mmHg)	2.8±2.1	6.8±2.0*	4.6±3.0	0.04
Mean left atrial pressure (mmHg)	5.3±3.3	6.6±3.8	5.4±1.7	0.72
TPR (dynes·sec/cm5)	310±151	1361±424*	586±425	<0.001
Cardiac index (L/min/m2)	4.9±0.7	3.6±0.6*	5.2±1.1¶	0.01
RV weight (g)	33.2±6.2	57.8±25.0*	46.0±5.6	0.01
LV weight (g)	85.3±18.0	83.4±17.8	91.7±8.8	0.61
RV/LV weight ratio	0.39±0.03	0.67±0.17*	0.51±0.11	0.001
Flow velocity at the banding site	N/A	203±18	198±26	0.73

Post-hoc test results: * P<0.05 against Naïve and ^¶ P<0.05 against PH+Saline. ANOVA followed by Tukey's test was used for normally distributed data. TPR, PVR, and RV end-diastolic pressure were analyzed by Kruskal-Wallis test followed by Dunn’s post-hoc test.

Abbreviations: LV = left ventricle, PA = pulmonary artery, RV = right ventricle, TPR = transpulmonary resistance

Pulmonary hypertension increases LV stiffness

Under echocardiographic examination, PH+Saline animals presented with enlarged RV and septal flattening that resulted in significantly higher RV to LV diameter ratio at end-diastole (Figure 1). There was no statistical difference in end-diastolic volume index (p=0.83) or end-systolic volume index (p=0.76) in PH+Saline animals compared to either the naïve or PH+SERCA groups (Figure 1). However, PH+Saline animals exhibited significantly smaller eccentricity index in both end-diastole (p=0.01; post-hoc test p=0.02 against naïve and p=0.02 against PH+SERCA) and end-systole (p=0.003; post-hoc test p=0.002 against naïve and p=0.06 against PH+SERCA), indicating the deformation of the LV. Pressure-volume analysis of the LV did not find a significant difference in contractility and relaxation parameters (Table 2), but EDPVR was steeper in the PH+Saline group compared to the naïve and PH+SERCA groups (p<0.001) indicating a stiffer LV. Using curvilinear relationship, animals in the PH+Saline group showed steeper EDPVR (Figure 2A, B). While the β value of EDPVR did not show any correlation to LV weight or LVEF, EDPVR β was significantly correlated with mean PA pressure and RV to LV weight ratio (Figure 2C), suggesting that RV-LV interdependence is the primary cause of impaired LV chamber stiffness.

Figure 1: LV deformation in the pulmonary hypertension (PH) animals treated by saline.

Echocardiographic analysis showed no significant differences in left ventricular (LV) volumes and ejection fraction, but significant LV deformation in PH+Saline group due to septal flattening. Top: Representative short-axis views of LV and right ventricle (RV) at end-diastole and end-systole. Bottom: LV volume indices and ejection fraction were not different among the three groups. However, RV to LV ratio was significantly larger with smaller eccentricity index in PH+Saline animals. These changes were attenuated in animals treated with AAV1.SERCA2a (naïve n=8, PH+Saline n=5, PH+SERCA n=7). Results of ANOVA are shown for end-diastolic volume index (EDVI), end-systolic volume index (ESVI), and ejection fraction. Tukey’s post-hoc analysis results are shown for end-diastolic LV/RV diameter ratio (ANOVA p=0.006), end-diastolic eccentricity index (ANOVA p=0.01), and end-systolic eccentricity index (ANOVA p=0.003).

Table 2.

LV parameters

	Naïve (n=8)	PH+Saline (n=5)	PH+SERCA (n=7)	p-value (ANOVA)
Echocardiography
3D end-diastolic volume (ml)	59.8±11.7	58.0±11.2	61.6±11.6	0.87
3D end-systolic volume (ml)	11.2±4.4	10.1±1.7	11.9±5.6	0.78
Mitral E wave (cm/s)	62.3±12.2	85.3±21.8	73.6±15.1	0.07
Mitral A wave (cm/s)	55.5±13.8	80.0±29.2	68.2±22.9	0.18
Mitral E/A	1.15±0.25	1.20±0.52	1.16±0.35	0.97
Mitral deceleration time (ms)	92.3±30.7	102.3±45.2	102±33.7	0.83
Pressure volume catheter
Maximum dP/dt (mmHg/s)	2384±399	2552±725	2917±831	0.31
Minimum dP/dt (mmHg/s)	−2482±540	−2086±808	−2637±810	0.42
Tau (ms)	48.8±6.8	50.2±11.4	47.8±6.9	0.86
LV EDP (mmHg)	11.6±4.0	14.7±7.8	15.9±3.8	0.28
ESPVR (mmHg/ml)	1.39±0.39	2.18±0.74	2.12±1.11	0.14
EDPVR (linear, mmHg/ml)	0.36±0.08	0.76±0.21*	0.53±0.15¶	<0.001
EDPVR β (curvilinear)	0.043±0.017	0.094±0.027*	0.058±0.018	0.003

Post-hoc test results: * P<0.05 against Naïve and ¶P<0.05 against PH+Saline. ANOVA followed by Tukey's test was used for normally distributed data. Tau and EDPVR β were analyzed by Kruskal-Wallis test followed by Dunn’s post-hoc test.

Abbreviations: EDP = end-diastolic pressure, EDPVR = end-diastolic pressure volume relationship, ESPVR = end-systolic pressure volume relationship, LV = left ventricle

Figure 2. Pressure-volume loop analysis exhibited increased left ventricular (LV) chamber stiffness in PH+Saline animals.

(A) Representative LV pressure-volume loops during preload reduction study demonstrate steeper end-diastolic pressure-volume relationship (EDPVR) in PH+Saline animals. (B) EDPVRs of individual animals based on calculated equation are shown for respective groups (naïve n=8, PH+Saline n=5, PH+SERCA n=7). (C) Scatterplots of EDPVR β against LV weight, LV ejection fraction (EF), mean pulmonary artery (PA) pressure, and LV to right ventricle (RV) weight ratio. Spearman’s correlation coefficients are shown as r with p values (n=15).

Improved LV chamber stiffness is not due to SERCA2a gene transduction

To examine if AAV-derived SERCA2a gene was expressed in the heart and led to improved LV diastolic performance, we examined AAV genomic DNA and human SERCA2a mRNA in the RV, septum, and the LV. None of the tissues examined showed positive expression, excluding the direct cardiac SERCA2a gene transduction as a cause of attenuated LV stiffness.

Discussion

We demonstrate that RV pressure overload from PH is associated with increased LV chamber stiffness in a pig model and that it can be attenuated by reducing the RV pressure overload with intra-airway AAV1.SERCA2a gene therapy. Improved LV chamber stiffness was not due to direct transduction of SERCA2a, but likely secondary to improved RV-LV interaction from reduced PA pressure.

Using a combination of echocardiography and pressure-volume evaluations of the LV in a clinically relevant post-capillary PH model, we characterized the impact of PH-induced RV overload on the LV. The observed increase in LV chamber stiffness in the PH+Saline animals can be explained by ventricular interdependence, although a contribution from intrinsic LV myocardial remodeling cannot be ruled out. Severe RV pressure overload produces RV dilation and septal displacement toward the LV cavity, altering myocardial geometry. Together with pericardial constraint that limits the total cardiac volume⁸, RV enlargement impairs LV filling. These mechanical chamber interactions increase the LV diastolic chamber stiffness despite preserved systolic function. Consistent with this mechanism, LV contractility and relaxation indices remained unchanged, whereas LV EDPVR was significantly altered and correlated strongly with pulmonary artery pressure and RV remodeling rather than LV parameters. These results indicate ventricular interaction as the dominant mechanism leading to significantly impaired LV filling in severe PH.

Since PH frequently complicates heart failure, ventricular interdependence may represent an additional mechanism contributing to impaired LV filling in these patients. However, because of diversity in etiology and variable severity of left heart disease, RV-LV interaction is difficult to examine in heart failure patients with PH. We used a pig model of PH induced by pulmonary vein banding, which mimics the pathophysiology of PH due to left heart disease. In this model, lack of LV and left atrial dysfunction offers unique opportunities to study RV-LV interaction in the absence of LV physiological abnormality. Our data suggest that development of PH in patients with LV dysfunction might pose additional negative impact on LV stiffness through interventricular dependence. This is particularly important in HFpEF patients since impaired LV filling is the major pathophysiology leading to HF that can further deteriorate.

Because SERCA2a is a central calcium handling protein in the heart, its delivery to the myocardium can potentially enhance intracellular calcium uptake during diastole, that leads to improved contractile and lusitropic function. Although stiffness is not the primary parameter that is expected to improve, we examined myocardial tissues for evidence of AAV-mediated SERCA2a gene uptake. Unlike the lung tissues that showed positive expression⁴, we found no viral genome or human SERCA2a mRNA indicating intra-airway delivery did not result in off-target expression to the heart. Meanwhile, our data suggest that intra-airway SERCA2a gene therapy could be a viable option for treating PH due to left heart disease.

Limitations of the study include small sample size and the use of only female animals. Some of the parameters examined were likely underpowered because of the small sample size. However, significantly altered LV EDPVR, which was associated with LV deformation, seems to be strongly related to the RV-LV interaction in PH.

In conclusion, we demonstrated that LV chamber stiffness is worsened by PH-induced RV pressure overload, and attenuation of PH would attenuate such negative impact on LV in a pig model. Restoration of pulmonary vascular remodeling with inhaled AAV1.SERCA2a attenuated LV stiffness without detectable LV gene transfer, highlighting the importance of RV unloading as a therapeutic strategy for preserving LV diastolic function in PH.