Left ventricular and aortic dysfunction in cystic fibrosis mice

Abstract

Background

Left ventricular (LV) abnormalities have been reported in cystic fibrosis (CF); however, it remains unclear if loss of cystic fibrosis transmembrane conductance regulator (CFTR) function causes heart defects independent of lung disease.

Methods

Using gut-corrected F508del CFTR mutant mice (ΔF508), which do not develop human lung disease, we examined in vivo heart and aortic function via 2D transthoracic echocardiography and LV catheterization.

Results

ΔF508 mouse hearts showed LV concentric remodeling along with enhanced inotropy (increased +dP/dt, fractional shortening, decreased isovolumetric contraction time) and greater lusitropy (−dP/dt, Tau). Aortas displayed increased stiffness and altered diastolic flow. β-adrenergic stimulation revealed diminished cardiac reserve (attenuated +dP/dt,−dP/dt, LV pressure).

Conclusions

In a mouse model of CF, CFTR mutation leads to LV remodeling with alteration of cardiac and aortic functions in the absence of lung disease. As CF patients live longer, more active lives, their risk for cardiovascular disease should be considered.

INTRODUCTION

Cystic fibrosis (CF) has long been characterized by its effect on epithelial tissues, however, since the discovery of the cystic fibrosis transmembrane conductance regulator (CFTR) gene, extensive research efforts have sought to characterize the effect of CFTR loss on various tissues throughout the body. CFTR expression and Cl⁻ currents were identified in cardiac myocytes 20 years ago [1]. Despite these early findings, a defined physiologic role for CFTR in the heart has remained largely elusive. With its Cl⁻ conductive properties, it has been proposed that CFTR contributes to the maintenance of resting membrane potential and minimization of action potential prolongation associated with stimulation of inward Ca²⁺ current [2, 3]. Additionally, simulation studies by Kuzumoto et al predicted that increases or decreases in CFTR channel density would positively or negatively alter action potential duration, respectively, due to β-adrenergic stimulation [4]. Our prior studies examining the effects of increasing or decreasing CFTR activity on isolated cardiomyocyte contraction support these predictions [5, 6]. Beyond maintaining normal physiologic function, CFTR is up-regulated during ischemia [7], involved in the protection of ischemic pre- and post-conditioning [8, 9], and down-regulated in patients with heart failure [10].

Numerous investigators have examined heart function in CF patients, however, with the prominent pulmonary disturbances in CF patients, the majority of these studies have focused on the right ventricle. Increased pulmonary pressures in individuals with lung disease can cause cor pulmonale and right ventricular overload [11]. In contrast, evidence for primary left ventricular dysfunction in CF has been scarce. Recent evidence for increased aortic stiffness in CF children [12] and adults [13-15] has added to the question whether left ventricular dysfunction exists in CF. These studies, coupled with our prior experiments showing CFTR involvement in murine cardiomyocyte contraction [5, 6] led us to hypothesize that loss of CFTR function would lead to primary cardiovascular disturbances independent of lung disease. To bypass the effects of lung disease, we utilized ΔF508 CFTR mutant mice, which do not display the lung pathology found in CF patients [16, 17]. Through left ventricular catheterization and 2D echocardiography, we examined cardiac and vascular function in vivo. We found that loss of CFTR function leads to left ventricular remodeling, increased inotropy and lusitropy, and diminished responsiveness to β-adrenergic stimulation, coupled with changes in aortic stiffness and flow.

METHODS

Animals

Gut-corrected ΔF508 CFTR mutant mice (strain C57/BL6-Cftr^tm1Kth-TgN(FABPCFTR)#Jaw/Cwr), hereto referred to as “ΔF508”, originated from Case Western Reserve University (Cleveland, OH), but were obtained from Dr. Deborah Nelson at the University of Chicago (Chicago, IL) and bred in-house. These mice are homozygous for the F508 deletion in the endogenous mouse CFTR gene [16], but express human CFTR from the rat fatty acid binding protein 1 promoter to prevent intestinal obstruction. These mice were developed in a similar manner as the gut-corrected CFTR null mouse by Zhou et al [17]. To date there have been no phenotypical differences between gut-corrected CFTR null and gut-corrected ΔF508 mice (communication with Dr. Craig Hodges, Case Western Reserve University). In the gut-corrected null mouse, human CFTR mRNA was detected throughout the intestinal tract, pancreas, kidney, and brain, but not in the heart [17]. In our prior experiments with gut-corrected ΔF508 mice, we did not observe any CFTR activity in isolated cardiac myocytes, as evidenced by the inability of CFTR_inh-172 to affect contraction or Ca²⁺ signaling [5].

ΔF508 CFTR breeders were used to generate progeny for experiments, while strain-matched controls (WT), C57/BL6 mice, were obtained from Jackson Laboratory (Bar Harbor, ME). Adult mice (16-20 weeks) were used. There was no difference (P > 0.05) between the ages, body weight, or numbers of males vs. female mice in each group of animals. In all sets of experiment, equal numbers of each sex were used. Use of animals and protocols were approved by University of Illinois at Urbana-Champaign and Washington University, St Louis IACUCs.

Left ventricular catheterization

The protocol for in vivo hemodynamic studies was derived from that of Rockman et al [18] and carried out by the Mouse Cardiovascular Phenotyping Core Facility at Washington University, St. Louis. Mice were anesthetized with thiopental sodium (60 mg/kg, intraperitoneally). This anesthesia produces a normal heart rate of 500 beats/min while still providing a surgical plane of anesthesia. There were no appreciable differences in the amount of anesthesia required for WT vs. ΔF508 mice. Hemodynamic measurements were recorded at baseline and 60 seconds after a constant infusion of dobutamine (Barnes-Jewish Hospital Clinical Pharmacy, St. Louis, MO). Continuous left ventricular (LV) systolic and diastolic pressures and their derivatives (dP/dt) were recorded with Millar analysis software.

Echocardiography

Non-invasive ultrasound examination of the heart and aorta of anesthetized mice was performed using a Vevo 770 Ultrasound System. Complete 2-D, M-mode, and Doppler ultrasound utilized short- and long-axis views. The Tei index was calculated by the time from mitral closing to the onset of mitral opening (isovolumetric contraction time (IVCT) + isovolumetric relaxation time (IVRT) + ejection time (ET), divided by the time from onset to the end of left ventricular outflow (ET). For pulse wave velocity measurements, the ascending aorta, aortic arch, and proximal portion of the descending aorta were imaged in one 2-D imaging plane. Following euthanization, hearts were removed and weighed for comparison to body weight.

Statistics

All data are expressed as means ± SEM. Unpaired Student’s t-test and one-way ANOVA with repeated measures and Bonferonni post-test were performed by GraphPad Prism (La Jolla, CA). Significant values are represented as *, P<0.05; **, P<0.01; ***, P<0.001.

RESULTS

CFTR mutation causes left ventricular concentric remodeling

To determine if ΔF508 CFTR mutation causes morphological changes, we performed echocardiography of the left ventricle. In diastole, the left ventricular posterior wall (LVPW) and interventricular septum (IVS) were significantly thickened in ΔF508 mice, with a concomitant decrease in left ventricular internal diameter (LVID) (Table 1). As a consequence, the LV relative wall thickness ((LVPWd + IVSd)/LVIDd) was significantly larger in comparison to WT mice (ΔF508: 0.54±0.02 vs. WT: 0.42±0.02, n=10, P<0.001). Gross examination of hearts following experiments revealed no difference in the heart:body weight ratio of ΔF508 hearts. Despite no global enlargement of the heart, ΔF508 mouse hearts undergo specific concentric remodeling of the left ventricle.

Table 1. Echocardiographic characterization of left ventricular structure.

Left ventricles of WT and ΔF508 mice (n=10) were examined using M-mode echocardiography. LVPWd, LVPWs: left ventricular posterior wall during diastole or systole, respectively; IVSd, IVSs: interventricular septum during diastole or systole, respectively; LVIDd, LVIDs: left ventricular internal diameter during diastole or systole, respectively. Values are expressed as means ± SEM.

Strain	Body Weight (g)	Heart:Body (mg/g)	LVPWd (mm)	IVSd (mm)	LVIDd (mm)	LVPWs (mm)	IVSs (mm)	LVIDs (mm)
WT	24.61±1.35	4.57±0.10	0.73±0.03	0.73±0.02	3.46±0.09	1.17±0.04	1.22±0.05	2.04±0.07
ΔF508	23.28±0.80	4.72±0.07	0.82±0.02*	0.86±0.03**	3.11±0.05**	1.21±0.03	1.25±0.04	1.66±0.07**

^*P < 0.05;

^**P < 0.01 vs. WT by Student’s t-test.

Characterization of ΔF508 hearts at rest

Subsequently, we examined if ΔF508 mice exhibit altered in vivo heart function at rest. Echocardiography showed a significant increase in the contraction of ΔF508 hearts, as indicated by increased fractional shortening (FS) and decreased isovolumetric contraction time (IVCT) (Table 2). The Tei index, an inverse measurement of overall heart performance (RV and LV), was significantly decreased in ΔF508 mice, indicating greater performance of ΔF508 hearts. These findings were supported by left ventricular catheterization experiments, which showed increases in indices of contraction (+dP/dt) and relaxation (−dP/dt, Tau) (Figure 1). Additionally, ΔF508 hearts had elevated systolic pressures, but no change in end-diastolic pressure or heart rate.

Table 2. Echocardiographic measurement of resting left ventricular function.

Echocardiographic measurement of heart function in WT and ΔF508 mice (n=10). FS: fractional shortening. E: early filling peak velocity; A: late filling or atrial peak velocity; DT: deceleration time; E’: early diastole; A’: late diastole; S’: peak systolic annular velocity; IVCT, IVRT: isovolumetric contraction or relaxation time, respectively. Values are expressed as means ± SEM.

Strain	FS (%)	E (cm/s)	A (cm/s)	E/A	DT (ms)	S’ (cm/s)	E’ (cm/s)	A’ (cm/s)	E’/A’	E/E’	IVCT (ms)	IVRT (ms)	Tei Index
WT	41.0± 1.2	777± 28	591± 30	1.33± 0.07	16.6± 1.0	22.7± 0.8	25.6± 1.4	20.2± 1.4	1.31± 0.09	31.2± 2.0	11.6± 1.1	11.5± 0.7	0.52± 0.03
ΔF508	46.9± 1.4*	715± 17	645± 17	1.12± 0.04*	16.2± 0.7	21.0± 0.7	26.3± 1.7	29.3± 1.7**	0.95± 0.11*	29.2± 2.3	7.4± 0.4**	11.0± 0.6	0.42± 0.02*

^*P < 0.05;

^**P < 0.01 vs. WT by Student’s t-test.

Figure 1. ΔF508 CFTR mice show increased baseline cardiac activity.

LV hemodynamic measurements were performed in WT (n=9) and ΔF508 (n=12) mice via catheterization. Each circle represents a single mouse, while the line represents the mean.

Trans-mitral velocities and tissue Doppler (TDI) data are presented in Table 2. These studies revealed an 8% decrease in early filling peak velocity (E) together with a 9% increase in late filling or atrial peak velocity (A), resulting in a significantly decreased E/A ratio, but no change in deceleration time (DT). TDI indicated no change in early diastole (E’), but did reveal an increase in late diastole (A’). As a result, the E’/A’ of ΔF508 hearts was significantly decreased from their WT counterparts; however, no change in E/E’ was observed.

Altered aortic compliance and increased afterload in ΔF508 mice

Based on our findings of left ventricular concentric hypertrophy, increased contraction, and increased LV systolic pressures in ΔF508 mice, we examined the aortic function of ΔF508 mice. ΔF508 mice aortas had significantly decreased internal diameters during both systole and diastole compared to WT mice (Figure 2A), and were significantly less distensible than WT aortas (Figure 2B). Furthermore, pulse wave velocity measurements showed that ΔF508 mice had a significant increase in velocities compared to WT mice, indicating an increase in aortic stiffness (Figure 2C). In ΔF508 mice, examination of the flow pattern within the aorta showed decreased diastolic forward velocity and increased diastolic retrograde velocity (Figure 2D). There were no differences in aortic valve function between either group of animals (data not shown). Thus, in vivo, ΔF508 aortas are stiffer and less distensible, leading to alterations in aortic blood flow.

Figure 2. CFTR mutation causes increases aortic stiffness.

In vivo, aortic compliance was measured using echocardiography in WT and ΔF508 mice (n=10). A. Representative images of WT and ΔF508 aortic arches using M-mode images presented with values for mean internal diameters (ID) of aorta (Ao) during diastole (d) and systole (s). B. Distensibility was calculated as the relative change in diameter between systole and diastole. C. Pulse wave velocity-measured aortic stiffness. D. Representative Doppler flow images during 2 cycles of systole and diastole through the aortic arches of WT and ΔF508 mice. The straight line represents zero blood velocity with below the line showing velocity of forward flow as blood moves away from probe and above the line showing velocity of retrograde flow as blood moves towards probe. Values for the mean forward and retrograde velocities are presented. All values are presented as means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. WT by Student’s t-test.

β-adrenergic stimulation reveals altered cardiac reserve

We further explored the functional characteristics of ΔF508 hearts via catheterization studies; wherein WT and ΔF508 mouse hearts were stimulated with incremental increases of the β-adrenergic agonist dobutamine. In both WT and ΔF508 mice, dobutamine caused significant changes in all measurements (P<0.05). However, while ΔF508 mice retained the ability to respond to dobutamine with increased contraction and relaxation, compared to WT mice, they had significantly diminished changes from baseline with a 49% reduction in +dP/dt, 51% decrease in −dP/dt, and 80% relative decrease in left ventricular systolic pressure (Figure 3). Additionally, while similar at lower concentrations, at maximal stimulation, ΔF508 mice had a significantly larger decrease in LVEDP than WT mice. There were no differences in HR or Tau between WT and ΔF508 mice at any dobutamine dose. Thus, in addition to changes at rest, ΔF508 mice display altered responses when subjected to exogenous β-adrenergic stimulation.

Figure 3. Decreased cardiac reserve in ΔF508 mice during β-adrenergic stimulation.

LV catheterization-measured hemodynamic changes following stimulation with serial concentrations of the β₁-adrenergic agonist dobutamine in WT (n=9) and ΔF508 (n=12) mice. Data are shown as percent change from the respective baseline (Figure 1) for each concentration. Values are expressed as means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. WT by Student’s t-test.

DISCUSSION

Despite extensive electrophysiological characterization of CFTR in cardiac myocytes, its role in cardiac physiology remains unclear. We previously showed that CFTR is involved in maintaining un-stimulated [5] and β-adrenergic stimulated [6] contraction of neonatal ventricular myocytes, while others have shown CFTR’s involvement in the heart’s response to ischemia [7, 9]. These, coupled with reports of ventricular arrhythmias, cardiomyopathy, and post-mortem fibrosis and necrosis in CF patients [19-21], prompts investigation into the cardiac health of individuals with CF. For the first time, we have examined, in vivo, the physiological effect of chronic CFTR loss on the heart in a manner that allows for evaluation of primary cardiovascular function.

ΔF508 mouse hearts showed specific concentric remodeling of the left ventricle. Prior histological studies comparing ventricular areas in ΔF508 and FV/B mice showed specific enlargement of the left ventricle in ΔF508 mice, with no change in right ventricular size [22]. These findings are consistent with the fact that ΔF508 mice, unlike CF patients, do not develop lung disease. Thus, for our purpose of examining primary left ventricular function, ΔF508 mice proved to be a useful model. Left ventricular hypertrophy can occur secondary to pressure or volume overload. Duan et al found that aortic banding resulted in faster progression of LV hypertrophy in CFTR null mice, indicating greater susceptibility to heart failure [23]. On a cellular level, calcineurin [24] and CaMKII [25] have been heavily implicated in mediating cardiac myocytes hypertophic response, with the latter being important for transition to dilated cardiomyopathy [26]. The potential role of CaMKII in ΔF508 hearts is intriguing, as we have previously shown that ΔF508 neonatal cardiomyocytes have an increased reliance on CaMKII to maintain normal contraction [5]. It remains to be seen if Ca²⁺ transients and/or sarcoplasmic reticulum (SERCA) Ca²⁺ regulation are altered; however, this would support our observed increases in inotropy and lusitropy [27].

Concentric hypertrophy can lead to diastolic dysfunction due to abnormal relaxation or restrictive filling. The E, A, and E/A ratio are indicators of diastolic function and used to indicate normal function (E > A), abnormal relaxation (E < A), pseudonormal filling (E > A), or restrictive filling (E >> A) [28]. In our study the E/A ratio was decreased with E > A, which could represent psuedonormal filling, however, the E/E’ ratio, which is used clinically to diagnose diastolic dysfunction was not changed in ΔF508 mice. Abnormal relaxation, associated with hypertension, ischemic cardiomyopathy, and left ventricular hypertrophy, is usually indicated by E < A, prolonged IVRT, and prolonged DT; none of which we observed in our studies. However, as indicated above, E > A values can represent normal cardiac function or pseudonormal function, which is the transition between abnormal relaxation and restrictive filling, and is characterized by normal diastolic filling values. With pseudonormal function, active myocardial relaxation, as occurs during early diastole (E’), can be mildly decreased, while late diastolic passive myocardial distention caused by atrial contraction (A’) can increase. Using these measurements, normal function is represented by E’/A’ > 1. In our tissue Doppler studies, ΔF508 hearts displayed no change in early diastole, but had significantly increased atrial contraction velocities, resulting in a E’/A’ ratio < 1, which was significantly less than WT mice. Thus, it is possible that ΔF508 mouse hearts display signs of pseudonormal or early mild diastolic dysfunction with alterations in passive left ventricular relaxation, however, targeted investigation of this in future studies is necessary.

Responsiveness to exercise is an important prognostic indicator for cardiovascular disease [29]. In lieu of exercise, we exposed ΔF508 mice to increasing concentrations of the β-adrenergic agonist, dobutamine. While ΔF508 mice retained the ability to respond to stimulation, they had a significantly reduced cardiac reserve, with attenuated changes in +dP/dt, −dP/dt, and LV pressure. A previous study found that 51% of CF patients had decreased ejection fraction in response to exercise [30]. While this study contained patients with severe lung dysfunction, the effect of which cannot be discounted, we propose that limitations of myocardial recruitment, in part due to increased basal demands, may play a role in this attenuated response.

In addition to imaging the heart, we examined, for the first time, the aortic structure and function of CF mice in vivo. In vitro, CFTR null mice display increased vascular tone and attenuated vasoactive peptide-induced relaxation [31]. Our findings that ΔF508 aortas have decreased internal diameters, are less distensible, and are stiffer, agree with in vitro data. These, coupled with our observations of altered aortic flow, are congruent with altered large vessel hemodynamics in CF patients at rest [12, 14] and during exercise [13]. Despite normal blood pressures, Hull et al observed large vessel abnormalities in CF patients that indicated enhanced vascular aging [14]. Recently, Hull et al have shown that decreasing systemic inflammation through short-course IV antibiotics can improve central hemodynamics [15]. It is unclear if systemic inflammation contributes to any of our results, but it is clear that in addition to cardiac function, it is important to elucidate the role of large vessel reactivity and hemodynamics in CF to determine the cardiovascular risk for these patients.

While once a childhood disease, advances in research and clinical care have resulted in a CF population where at least 50% of patients are now adults (www.cff.org). Our studies were done in adult mice 16-20 weeks old, which would correspond to a young adult in humans given an average life expectancy of 2 years for C57/BL6 mice. With aortic stiffness identified in CF children [12], it is possible that left ventricular changes may be seen in CF at a young age. Additionally, since age is a known risk factor for cardiovascular disease, we postulate that as individuals with CF age they may be at increased risk for cardiovascular disease compared to their non-CF counterparts. While it is clear that decreased pulmonary function in CF affects right ventricular function, we have shown left ventricular and aortic function are also affected by CFTR loss. These findings provide compelling reasons to examine the cardiovascular function of CF patients in more detail.