The authors have declared that no competing interests exist.
Principal investigator: KN. Conceived and designed the experiments: BVN HVA EVD LN DD GS KN. Performed the experiments: BVN EVD LN. Analyzed the data: BVN HVA. Contributed reagents/materials/analysis tools: BVN HVA EVD LN. Wrote the paper: BVN HVA EVD LN DD GS KN.
Left ventricular (LV) and right ventricular (RV) function have an important impact on symptom occurrence, disease progression and exercise tolerance in pressure overload-induced heart failure, but particularly RV functional changes are not well described in the relevant aortic banding mouse model. Therefore, we quantified time-dependent alterations in the ventricular morphology and function in two models of hypertrophy and heart failure and we studied the relationship between RV and LV function during the transition from hypertrophy to heart failure.
MRI was used to quantify RV and LV function and morphology in healthy (n = 4) and sham operated (n = 3) C57BL/6 mice, and animals with a mild (n = 5) and a severe aortic constriction (n = 10).
Mice subjected to a mild constriction showed increased LV mass (
Relevant changes in mouse RV and LV function following an aortic constriction could be quantified using MRI. The well-controlled models described here open opportunities to assess the added value of new MRI techniques for the diagnosis of heart failure and to study the impact of new therapeutic strategies on disease progression and symptom occurrence.
Heart failure (HF) is a progressive syndrome in which the heart is no longer capable of pumping blood at a rate commensurate with the peripheral needs
Despite considerable progress, the mechanisms responsible for the transition from compensated hypertrophy to HF are still not completely understood
The goal of this study was therefore twofold. First, to quantify longitudinal changes in LV and RV morphology as well as function during the transition from a healthy to a compensated or decompensated state of LV hypertrophy. Second, to investigate the interplay between LV and RV function during this process.
Here, the well-defined, reproducible mouse model of transverse aorta constriction (TAC) can play a prominent role
All animal experiments were performed according to the Directive 2010/63/EU of the European Parliament and approved by the Animal Care and Use Committee of Maastricht University (protocol: 2009-019).
In this study 11 weeks old male C57BL/6 mice weighing between 23 and 25 g were used. Animals were housed under standard laboratory conditions with a 12 h light/dark cycle and were maintained on a standard diet and had access to water
For MRI a total of 22 animals were randomly separated in a control group (n = 4), a group that was sham-operated (n = 3) and in two groups which underwent a surgically induced mild (n = 5) or severe (n = 10) transverse aortic constriction (TAC), resulting in LV pressure overload
Measurements were performed with a 9.4 T small animal MRI scanner (Bruker BioSpec, Ettlingen, Germany) equipped with a 740 mT/m gradient coil. A 72-mm-diameter quadrature transmit coil was used in combination with a 4 element phased-array receive coil (Bruker). Mice were anesthetized with isoflurane (4.0% for induction, 1.5–2.0% for maintenance) in medical air (0.4 L/min). The front paws were placed on ECG electrodes and a balloon pressure sensor was placed on the abdomen. Body temperature was maintained at 36–37°C with a heating pad and monitored with a rectal temperature sensor.
Cinematographic (cine) MR images were acquired using an ECG-triggered and respiratory-gated FLASH sequence, with the following parameters: pulse repetition time/echo time = 7/1.8 ms, number of signal averages = 6, α = 15, field of view = 3×3 cm2, matrix = 192×192, slice thickness = 1 mm, number of cardiac frames = 15–20. Measurements were performed in 2 long-axis and 5 short-axis planes, covering the LV from apex to base with interslice distance optimized for heart size.
Local cardiac function was measured from mid-ventricular short-axis tagged images. Tagging MRI was done using the FLASH sequence with a spatial modulation of magnetization (SPAMM) preparation module resulting in a sinusoidal modulation of magnetization that moves along with the cardiac tissue during the heart cycle. The preparation consisted of two Gaussian RF pulses (α = 45°, pulse width = 200 µs), separated by a gradient (duration = 200 µs) defining tag wavelength (0.5 mm) and orientation. Total duration of the preparation module was 2.7 ms. Tagged images were recorded with a reduced matrix of 192×96 (frequency×phase encoding) and reconstructed on a 384×384 matrix for data analysis. Tags were applied in horizontal and vertical directions and with 180° phase shift for complementary SPAMM (CSPAMM) reconstruction
MRI measurements were performed at 1, 2, 4, 7, 10 and 13 weeks after surgery. Cine MR images were obtained at all time points. During the first MRI experiment cine MR images of the aortic arch were acquired to confirm correct positioning of the TAC (
Examples of MRI scans through the aortic arch in (left) a control, (middle) mild TAC and (right) severe TAC mouse. Indicated are (IA) the innominate artery, (LCCA) left common carotid artery, (LSA) left subclavian artery, and the transverse aortic constriction (TAC).
Tagged MR images were obtained at week 2 until week 10, at which time points all animals were in experiment, to determine the relationship between local strain changes and global cardiac morphology and function. Mice with a severe TAC were euthanized 10 weeks after surgery for animal welfare reasons. Immediately after the last measurements the anesthetized animals were killed by means of perfusion of the vascular bed with phosphate buffered saline (pH 7.4) infused via a needle penetrating the apex and exsanguination from the vena cava inferior. Next, the integrity of the aortic band was visually verified and lung wet weight (LuW) and tibia length (TL) were measured.
The myocardial wall was segmented semi-automatically in the cine MR images using CAAS MRV FARM (Pie Medical Imaging, The Netherlands) to obtain LV and RV volumes, and the LV and RV ejection fractions (EF)
Local tissue motion was quantified from the tagged images using a method based on optical flow theory implemented in Mathematica 7.0 (Wolfram Research Inc., Champaign, IL)
Data are expressed as mean ± standard deviation (SD). Changes in LV and RV volumes and EF, LV mass/TL, heart rate, respiratory rate, bodyweight (BW), WT and strains were tested for statistical significance with an ANOVA for repeated measures with time and group as factor, followed by the Bonferroni post-hoc test when appropriate. In case of interaction between time and group, the effect of time was tested separately per group. Changes in LuW/TL and heart weight/TL were tested for statistical significance with a 1-way ANOVA, followed by the Bonferroni post-hoc test. For survival analysis additional data from healthy (n = 48), mild TAC (n = 2) and severe TAC mice (n = 89) available from our laboratory was included. Differences between Kaplan-Meier survival curves were tested for statistical significance by means of Log Rank analysis. Calculations were performed using SPSS 19.0 (SPSS Inc., Chicago). For all tests the level of significance was set at α = 0.05.
Survival analysis performed on a large cohort of mice (
Kaplan-Meier analysis was performed based on survival data of a large cohort of healthy mice (n = 55), mild TAC (n = 5) and severe TAC mice (n = 99) available from our laboratory. Log Rank analysis showed a significant difference in survival between the severe TAC mice as compared to the control and mild TAC mice (
Weeks | 1 | 2 | 4 | 7 | 10 | 13 | |
|
Control | 528±29 | 532±30 | 524±33 | 532±28 | 534±34 | 536±21 |
Mild TAC | - | 520±30 | 544±30 | 547±35 | 560±48 | 556±41 | |
Severe TAC | 600±31 | 592±51 | 529±61 | 530±31 | 561±33 | - | |
|
Control | 74±6 | 86±23 | 89±8 | 90±10 | 90±12 | 89±9 |
Mild TAC | - | 82±17 | 77±9 | 81±17 | 74±16 | 95±14 | |
Severe TAC | 110±15 | 88±9 | 81±18 | 89±12 | 87±12 | - | |
|
Control | 24.4±0.8 | 25.4±1.3 | 26.0±1.2 | 27.0±1.3 | 27.8±1.6 | 28.9±1.8 |
Mild TAC | - | 26.9±2.2 | 27.6±2.4 | 28.0±2.0 | 29.6±1.7 | 28.4±0.9 | |
Severe TAC | 24.5±0.9 | 25.2±1.1 | 26.5±1.2 | 27.3±1.2 | 25.8±0.7 | - |
General characteristics of the control animals and mice with a mild and severe constriction. Indicated are the time points relative to the time of surgery [weeks], the heart rate (HR) [min−1] and respiratory rate (Resp) [min−1] during the MR examination, and the body weight (BW) [g].
Representative end diastolic short-axis and long-axis images from control mice and mice subjected to a mild and severe aortic constriction 10 weeks after surgery (A). Indicated are the left ventricle (LV), right ventricle (RV), the papillary muscles (PM) and decreased apical wall thickness in the mouse with a severe TAC (↓). Corresponding movies can be found in the supplementary material. Wall thickening (WT) in the experimental groups at 2 and 10 weeks after surgery (B). At 2 weeks, WT had decreased in all sections of the heart in mice with a severe TAC as compared to controls (
Mice with a mild TAC revealed a small increase of LV mass normalized to TL (5.4±0.7 mg/mm) as compared to controls (3.9±0.4 mg/mm,
LV mass normalized to tibia length (TL) in control, mild and severe TAC mice as a function of time. Cardiac mass slightly increased in response to a mild constriction as compared to controls (*,
End diastolic volume in control, mild and severe TAC mice (left column), end systolic volume (middle column) and ejection fraction (right column) as a function of time for both the left ventricle (LV) (top row) and right ventricle (RV) (bottom row). End diastolic and end systolic volumes clearly show LV and RV dilation in the severe TAC mice, but not in the mild TAC mice as compared to the control animals. LV ejection fraction was slightly depressed in response to a mild constriction as compared to controls, and showed a progressive decline in time in the group with a severe constriction. RV ejection fraction remained unchanged in mice with a mild constriction of the aorta as compared to control mice, but showed a progressive decline in case of a severe aortic constriction. Mean and SD per time point are denoted by the corresponding symbol and error bars. Statistical differences as compared to the control group are indicated by * (
In contrast, severe TAC resulted in a progressive increase of LV mass normalized to TL from 5.8±0.6 mg/mm in week 2 to 9.1±0.5 mg/mm in week 10 (
Impaired LV function may induce lung remodeling and/or edema and subsequently RV failure
Relationship between left (LV) and right ventricular (RV) ejection fraction (EF) for all mice at all time points. The right ventricular ejection fraction progressively decreased (black arrow) in the severe TAC group only and that the changes in RVEF were preceded by a decline in LVEF apparent from the shift of the majority of the measurement points to the left.
Group | LuW/TL [mg/mm] | ww/dw [-] | HW/TL [mg/mm] |
Control | 9.4±1.5 | 6.1±0.6 | 7.9±0.5 |
Mild TAC | 10.4±2.0 | 6.8±1.8 | 10.3±2.1 |
Severe TAC | 17.5±4.8† | 5.6±0.7 | 13.6±2.3‡ |
An increased lung weight-to-tibia length (LuW/TL) ratio [mg/mm] indicated the presence of pulmonary remodeling in the mice with a severe constriction (†,
In this study we investigated the evolution of RV and LV function in well-controlled mouse models of compensated hypertrophy and decompensated HF as induced by two different degrees of transverse aortic constriction (TAC), using cinematographic and tagging MRI
There are a number of important findings to this study. First, mice with a mild TAC revealed myocardial hypertrophy and only a slightly depressed LVEF consistent with a state of compensated hypertrophy. Second, mice subjected to a severe TAC showed progressive LV hypertrophy, increased LV volumes and a drastic decline in LV function in accordance with a condition of HF. Third, myocardial principal strains were significantly reduced in severe TAC mice as compared to controls and progressively decreased over time. Fourth, changes in RV volumes and EF could be quantified in TAC mice using cardiac MRI. Fifth, the progressive deterioration of LV function in severe TAC mice was followed in time by worsening of RV function and severe pulmonary remodeling, two important hallmarks of congestive left ventricular pump failure
LV mass, end diastolic and end systolic volumes were slightly elevated at 2 weeks after mild TAC, after which these variables remained essentially constant. RV volumes and EF in mild TAC mice, however, were unchanged as compared to control mice. The severe TAC mice showed a progressive increase in both end diastolic and end systolic volumes accompanied by a decline in LVEF. In these mice also a marked increase was found in RV end systolic volume as compared to the control mice resulting in a deteriorating RVEF. The absence of RV dilation, RV end diastolic volumes remained unchanged, may point to impaired contractile properties rather than dilation as a cause for the impaired RV function. These mice likely also developed profound pulmonary remodeling, as indicated by an increased LuW/TL ratio (17.5±4.8 mg/mm) as compared to controls (9.4±1.5 mg/mm) (
Awareness is growing that RV function has an important impact on symptom occurrence, disease progression as well as exercise tolerance in various cardiac pathologies
Cardiac strains were quantified in terms of the 2D principal strains from tagged MR images using a method based on optical flow theory. While myocardial principal strain E1 is mainly oriented in the radial direction, E2 coincides with the circumferential direction
The effects of both passive and active cardiac tissue mechanics on the transition from hypertrophy to heart failure gain increased interest
The septum is believed to contribute to both LV and RV function in the normal and diseased heart, although the precise mechanisms are not fully understood
The disease progression reported in this study compares well with previous data, despite the fact that the phenotype resulting from TAC surgery may vary depending on mouse strain and the surgical technique used
Cardiac MRI is an important clinical tool for HF diagnosis
There are some limitations to this study. A valuable comparison of LV and RV volumes and function with, for example, conductance catheters measurements was not made. However, it was anticipated that this would require a large cohort of mice, since catheter measurements in mice are terminal. Instead, the number of mice required was minimized by choosing a longitudinal study design with readouts from non-invasive imaging. The experimental variation in for example LV mass, LVEDV and LVESV increased during the course of the experiment, in particular in the severe TAC group, but was comparatively small at the start of the experiment. This suggests that within group differences in systolic pressure gradient immediately after TAC were small and that the observed variation resulted from inter animal differences, but we cannot fully exclude some variation due to small differences in pressure gradient. Although the number of mice in the sham and mild TAC groups was limited, the longitudinal study design generally is more efficient and results in increased statistical power as changes over time are assessed within the same animals.
In this study longitudinal MRI measurements were performed in mice subjected to a mild or severe TAC. The mice with a mild TAC developed compensated hypertrophy, whereas the mice with a severe TAC developed congestive HF. A decline in RV function was observed following the progressive deterioration of LV function, relevant for many cases where RV failure develops secondary to LV pathologies. The well-controlled aortic banding model of HF described here therefore opens opportunities to assess the added value of various new MR imaging techniques for the diagnosis of HF, to study the impact of new therapeutic strategies on disease progression and symptom occurrence in the RV and LV, and to assess the effects of pharmacological or mechanical LV unloading on the RV. Such studies might eventually lead to improvements in care for patients suffering from pressure overload-induced HF.
(PPT)
(PPT)
(PPT)
We gratefully acknowledge D. Veraart and J. Debets for biotechnical assistance, and W.M. Blankesteijn for discussions.